You are on page 1of 203

Abstract

Across the nation, the debate over metropolitan sprawl and its impacts continues. A decade ago, Smart
Growth America (SGA) and the U.S. Environmental Protection Agency (EPA) sought to raise the level of
the debate by sponsoring groundbreaking research to quantitatively measure sprawl and its quality-oflife impacts. The resulting measures have been widely used in urban research.
In this study for the National Cancer Institute, the Brookings Institution, and Smart Growth America, we
begin in Chapter 1 by reviewing the history of sprawl-related research and writing. Before 2000,
attempts to characterize sprawl were mostly qualitative, subjective, and often polemical. Sprawl was
either very bad or very good, depending on the philosophical leaning of the author. After 2000,
researchers began to measure sprawl objectively and rank places as more or less sprawling, first mostly
in terms of density variables and later more completely as a multidimensional phenomenon. They also
began to relate quantitative measures of sprawl to quality-of-life outcomes.
In Chapter 2, we update the original county indices to 2010. As one would expect, the degree of county
sprawl does not change dramatically over a 10-year period. Also, given their fixed boundaries, most
counties become more compact (denser and with smaller blocks) over the 10-year period. Sprawl occurs
mainly as previously rural counties (in 2000) outside metropolitan areas become low density suburbs
and exurbs of metropolitan areas (in 2010).
In Chapter 3, we develop refined versions of the indices that incorporate more measures of the built
environment. The refined indices capture four distinct dimensions of sprawl, thereby characterizing
county sprawl in all its complexity. The four are development density, land use mix, population and
employment centering, and street accessibility. The dimensions of the new county indices parallel the
metropolitan indices developed by Ewing et al. (2002), basically representing the relative accessibility
provided by the county. The simple structure of the original county sprawl index has become more
complex, but also more nuanced and comprehensive, in line with definitions of sprawl in the technical
literature.
In Chapters 4, 5, and 6, we validate our county indices against data on journey to work, traffic fatalities,
and obesity, physical activity, and chronic diseases. Validation of the 2010 sprawl indices, both the
original and refined indices and four individual factors, takes the same form as the original validation in
2003. As expected based on earlier research, we find that county sprawl is associated with lower walk
and transit shares of commute trips, longer drive times, higher fatal crash rates, and higher incidence of
obesity and certain chronic diseases. Other outcomes are not as clearly related to sprawl.

DRAFT

In Chapter 7, we develop metropolitan sprawl indices that, like the refined county indices, have four
distinct dimensions-- development density, land use mix, population and employment centering, and
street accessibility. Compared to metropolitan sprawl indices from the early 2000s, these new indices
incorporate more variables and hence have more construct validity. For example, the earlier effort
defined density strictly in terms of population concentrations, while this effort considers employment
concentrations as well. The reason for developing metropolitan sprawl indices, rather than limiting
ourselves to counties, is that metropolitan areas are natural units of aggregation for certain analyses.
In Chapter 8, we relate sprawl indices to three topics of recent popular interest, for which datasets have
recently been released. The three are life expectancy, housing plus transportation costs, and upward
economic mobility. We find that life expectancy is greater in compact areas, in part due to lower
prevalence of obesity and lower vehicle miles traveled (which relates to traffic fatalities). Housing costs
are higher in compact areas, but these higher costs are more than offset by lower transportation costs,
and the net effect is a just barely significant relationship between compactness and overall housing
affordability.
In Chapter 9, we conduct one of the first longitudinal analysis of sprawl to see which areas are sprawling
more over time, and which are sprawling less or actually becoming more compact. To conduct such as
analysis, we need to employ a new level of geography, the census urbanized area. In contrast of
counties and metropolitan areas, urbanized areas expand incrementally as areas grow and rural tracts
are converted to urban and suburban uses. The analysis shows that, on average, urban sprawl in the U.S.
increased between 2000 and 2010, but that there are many exceptions to this generalization.
The appendices provide values of compactness/sprawl indices for counties, metropolitan areas, and
urbanized areas. Data are available in electronic form at

Table of Contents
Abstract ......................................................................................................................................................... 1
Chapter 1. Introduction ................................................................................................................................ 6
Urban Sprawl Generally ............................................................................................................................ 6
Early Attempts to Measure Sprawl ........................................................................................................... 7
Recent Use of Satellite Imagery ....................................................................................................... 8
Multiple Dimensions ................................................................................................................................. 9
Recap of Past Studies .............................................................................................................................. 10
Chapter 2. Updated County Sprawl Index .................................................................................................. 12
Update to 2010 ....................................................................................................................................... 14
Chapter 3. Refined County Sprawl Measures ............................................................................................. 18
Density .................................................................................................................................................... 18
Mixed Use ............................................................................................................................................... 19
Centering................................................................................................................................................. 21
Street Accessibility .................................................................................................................................. 24
Relationship Among Compactness Factors............................................................................................. 25
Composite Index ..................................................................................................................................... 25
Greater Validity of New Index................................................................................................................. 27
Chapter 4. Validation of County Sprawl Indices ......................................................................................... 32
Methods .................................................................................................................................................. 32
Data and Variables ......................................................................................................................... 32
Analytical Method .......................................................................................................................... 37
Results ..................................................................................................................................................... 37
Household Vehicle Ownership ....................................................................................................... 37
Walk Mode Shares ......................................................................................................................... 38
Transit Mode Shares ...................................................................................................................... 39
Drive Times .................................................................................................................................... 41
Discussion................................................................................................................................................ 42
Chapter 5. Traffic Safety ............................................................................................................................. 43
Methods .................................................................................................................................................. 44
3

Data ................................................................................................................................................ 44
Variables ........................................................................................................................................ 45
Method of Analysis ........................................................................................................................ 46
Results ..................................................................................................................................................... 47
Limitations .............................................................................................................................................. 53
Chapter 6. Public Health ............................................................................................................................. 54
Methods .................................................................................................................................................. 55
Geographic Scale............................................................................................................................ 55
Data and Variables ......................................................................................................................... 56
BMI and Obesity...................................................................................................................................... 58
Physical Activity ...................................................................................................................................... 60
Chronic Diseases ..................................................................................................................................... 62
Limitations .............................................................................................................................................. 63
Chapter 7. Derivation and Validation of Metropolitan Sprawl Indices....................................................... 65
Methods .................................................................................................................................................. 65
Sample ........................................................................................................................................... 65
Variables ........................................................................................................................................ 66
Individual Compactness/Sprawl Factors ................................................................................................. 69
Overall Compactness/Sprawl Index for 2010 ......................................................................................... 69
Relationship to Transportation Outcomes ............................................................................................. 73
Discussion................................................................................................................................................ 75
Chapter 8. Relationship of Sprawl to Topical Outcomes ............................................................................ 79
Life Expectancy ....................................................................................................................................... 79
Conceptual Framework .................................................................................................................. 80
Air Pollution ................................................................................................................................... 81
Crime .............................................................................................................................................. 81
Methods ......................................................................................................................................... 82
Results ............................................................................................................................................ 83
Housing Affordability (H+T) .................................................................................................................... 86
Conceptual Framework .................................................................................................................. 88
Method .......................................................................................................................................... 89
Results ............................................................................................................................................ 90
4

Chapter 9. Urbanized Areas: A Longitudinal Analysis ............................................................................... 102


Methods ................................................................................................................................................ 103
Sample ......................................................................................................................................... 103
Variables ...................................................................................................................................... 103
Statistical Methods ...................................................................................................................... 106
Results ................................................................................................................................................... 106
Individual Compactness/Sprawl Factors ...................................................................................... 106
Overall Compactness/Sprawl Index for 2010 .............................................................................. 108
Overall Compactness/Sprawl Index for 2000 .............................................................................. 109
Validation Against Transportation Outcomes ............................................................................. 110
Attempted Validation Using Change Variables............................................................................ 113
Discussion.............................................................................................................................................. 113
Chapter 10. Conclusion ............................................................................................................................. 115
References ................................................................................................................................................ 116
Appendix A. County Compactness Indices for 2010, 2000, and Changes................................................ 125
Appendix B. County Compactness Factors and Composite Indices for 2010 .......................................... 151
Appendix C. Detailed Model Results for Public Health Outcomes .......................................................... 178
Appendix D. 2010 Metropolitan Indices ................................................................................................... 185
Appendix E. Urbanized Areas Compactness Indices 2010 ........................................................................ 193
Appendix F. Urbanized Areas Compactness Indices 2000 ........................................................................ 198

Chapter 1. Introduction
Across the nation, the debate over metropolitan sprawl and its impacts continues decade after decade.
There is little agreement on the definition of sprawl or its alternatives: compact development,
pedestrian-friendly design, transit-oriented development, and the catch-all term smart growth. There
is also little consensus about how sprawl impacts everything from housing affordability and traffic
congestion to open space preservation and air quality (Kunstler, 1993; Burchell,1998 and 2002,
Bruegmann, 2006; Duany et al., 2001; Ewing, 1997 and 2002; Gordon and Richardson, 1997; Kahn, 2006;
Hayden, 2004; Hirschhorn, 2005). Duany et al. (2001) use cultural, aesthetic and ecological reasons to
reject suburban sprawl as human habitat. At the other end of the spectrum, Bruegmann (2006)
describes suburban sprawl as a natural manifestation of the American Dream of a big house in the
suburbs.
A decade ago, Smart Growth America (SGA) and the U.S. Environmental Protection Agency (EPA) sought
to raise the level of this debate by sponsoring groundbreaking research on sprawl and its quality-of-life
consequences (Ewing et al. 2002; Ewing et al. 2003a, 2003b, 2003c). The original sprawl indices were
made available to researchers who wished to explore the various costs and benefits of sprawl. They
have been widely used in outcome-related research, particularly in connection with public health.
Sprawl has been linked to physical inactivity, obesity, traffic fatalities, air quality, residential energy use,
emergency response times, teenage driving, social capital, and private-vehicle commute distances and
times (Ewing et al. 2003a; Ewing et al. 2003b; Ewing et al. 2003c; Kelly-Schwartz et al. 2004; Sturm and
Cohen 2004; Cho et al. 2006; Doyle et al. 2006; Ewing et al. 2006; Kahn 2006; Kim et al.2006; Plantinga
and Bernell 2007; Ewing and Rong 2008; Joshu et al. 2008; Stone 2008; Trowbridge and McDonald 2008;
Fan and Song 2009; McDonald and Trowbridge 2009; Trowbridge 2009; Trowbridge et al. 2009; Lee et al.
2009; Nguyen 2010; Stone et al. 2010; Schweitzer and Zhou 2010; Kostova 2011; Zolnik 2011; Holcombe
and Williams 2012; Griffin et al. 2013; Bereitschaft and Debbage 2013).
In this study for the National Cancer Institute, the Ford Foundation, and Smart Growth America, we
update the original sprawl indices to 2010. Additionally, we develop refined versions of the indices that
incorporate more measures of the built environment. Finally, we validate these indices against U.S.
American Community Survey (ACS) data for journey to work, Fatality Analysis Reporting System (FARS)
data on traffic fatality and pedestrian fatality rates, and Behavioral Risk Factor Surveillance System
(BRFSS) data on overweight and obesity. It is our hope that the updated and refined sprawl measures,
developed and validated in this report, will find as widespread use in the fields of urban planning and
public health as have the original indices.

Urban Sprawl Generally


Urban sprawl (also referred to as suburban sprawl) has become the dominant metropolitan
development pattern in the United States. Examples of compact development, the antithesis of sprawl,
are so few and far between that they seem almost quaint these days. For every New York metropolitan
area, there are dozens of Atlantas and Detroits. For every Manhattan, there are hundreds of Walton
and Lapeer counties (counties located on the periphery of the Atlanta and Detroit metropolitan areas).
6

As the costs of sprawl have become more apparent, the term urban sprawl has gone from urban
planning construct to public policy concern. But what exactly is urban sprawl? In the early 1990s, the
lead author of this report worked on a definition of sprawl for purposes of growth management in
Florida. The definition ultimately adopted by the State encompassed the following urban forms: (1)
leapfrog or scattered development, (2) commercial strip development, (3) expanses of low-density
development or (4) expanses of single-use development (as in sprawling bedroom communities).
Because these forms are prototypical, and a matter of degree, the Florida definition was supplemented
with primary indicators of sprawl that could be measured and made subject to regulation. The most
important indicator, which became part of the law, was any development pattern characterized by poor
accessibility among related land uses.
All four prototypical patterns (leapfrog, etc.) are characterized by poor accessibility (Ewing 1997). In
scattered or leapfrog development, residents and service providers must pass vacant land on their way
from one developed use to another. In classic strip development, the consumer must pass other uses
on the way from one store to the next; it is the antithesis of multipurpose travel to an activity center. Of
course, in low-density, single-use development, everything is far apart due to large private land holdings
and segregation of land uses. The potential link to public policy is clear. In sprawl, poor accessibility of
land uses to one another may leave residents with no alternative to miles and miles of automobile
travel.

Early Attempts to Measure Sprawl


The first attempts to measure the extent of urban sprawl were crude. Several researchers created
measures of urban sprawl that focused on density (Pendall 1999; Fulton et al. 2001; Malpezzi and Guo
2001; Nasser and Overberg 2001; Lopez and Hynes 2003; Burchfield et al. 2005; Anthony, 2004; Lang,
2003; Pendall and Carruthers, 2003). Pendall (1999) sought to explain the incidence of sprawl for large
metropolitan areas in terms of land values, metropolitan political organization, local government
spending, traffic congestion, and various local land use policies. Pendalls measure of sprawl was strictly
related to density. Population was divided by urban acreage to obtain density estimates for 1982 and
1992, with urban acreage taken from the U.S. Department of Agricultures National Resources
Inventory. To measure increases in sprawl over time, estimates of population change between 1982
and 1992 were divided by estimates of change in urban land during the same period. By this measure,
Los Angeles, San Francisco, and San Diego were the most compact metros in 1992, while Milwaukee,
Atlanta, Cleveland, and Denver were the most sprawling. Baltimore, Los Angeles, and San Diego grew in
the most compact manner over the decade, while Boston, Cincinnati, and Minneapolis-St. Paul grew in
the least.
Building on Pendalls earlier work, Fulton et al. (2001) studied urban land consumption relative to
population change for every U.S. metropolitan area. If land is consumed at a faster rate than population
is growing, sprawl is said to be increasing. As with Pendalls earlier work, this concept of sprawl is
strictly density-related. By this criterion, the West is home to some of the least sprawling metropolitan
areas in the nation. By contrast, the Northeast and Midwest are in some ways the nation's biggest
sprawl problems since they add few new residents, yet consume large amounts of land. In this study,
7

Honolulu and Los Angeles were rated most compact in 1997, and Las Vegas and Phoenix (often
characterized as sprawling badly) were both in the top 20 in compactness. Las Vegas and Phoenix were
first and third in density gain over the 15 years studied, 1982 to 1997.
Their report also examined the causes of sprawl. Metropolitan areas tend to consume less land for
urbanizationrelative to population growthwhen they are growing rapidly in population, rely heavily
on public water and sewer systems, and have large immigrant populations. Metropolitan areas tend to
consume more land for urbanizationagain, relative to population growthif they are compact to
begin with and have fragmented local governance.
Taking a different tack, Glaeser et al. (2001) related sprawl to the degree of decentralization of
employment using data from the U.S. Department of Commerces Zip Code Business Patterns for 1996.
Zip code business patterns data were extracted from the Standard Statistical Establishments List, a file of
all single and multi-establishment companies listed by zip code and firm size.
For the 100 largest U.S. metropolitan areas, the share of overall metropolitan employment within a
three-mile ring of the Central Business District (CBD) was computed, as were the shares inside and
outside a 10-mile ring. The share within three miles reflects the presence or absence of a well-defined
employment core, while the share beyond 10 miles captures the extent of job sprawl. Metros were
then divided into four categories, based on values of these indices. Dense employment metros like New
York have at least one quarter of their employment within three miles of the city center. Centralized
employment metros like Minneapolis-St. Paul have between 10 and 25 percent of employment within
three miles of the city center, and more than 60 percent within 10 miles. Decentralized employment
metros like Washington D.C. have 10 to 25 percent of employment within the three-mile ring, and less
than 60 percent within 10 miles. Finally, extremely decentralized employment metros like Los Angeles
have less than 10 percent of their employment within the three-mile ring. A follow-up study, also for the
Brookings Institution, used the same metrics to rate the degree of sprawl in metropolitan areas as of
2006 (Kneebone 2009).
The ease of measurement associated with the early sprawl indices came with a lack of precision that led
to wildly different sprawl ratings given to different metropolitan areas by different analysts. In one
study, Portland was listed as the most compact region and Los Angeles was ranked among the most
sprawling. In another, their rankings were essentially reversed (Glaeser et al. 2001; Nasser and
Overberg 2001). More importantly, such measures disregarded other qualities commonly associated
with sprawl such as segregated land uses, commercial strips, and leapfrog development (Ewing, 1997).
Recent Use of Satellite Imagery
The same mistakes made in early quantitative studies of sprawl, neglect of land use interactions and
quality-of-life outcomes, have been made in recent studies using satellite imagery (Huang, Lu, & Sellers,
2007; Martellozzo & Clarke, 2011). Huang et al. (2007), for instance, calculated seven spatial metrics
that capture five distinct dimensions of urban form (complexity, centrality, compactness, porosity and
density) in order to compare different cities and countries throughout the world. Such methods are
useful for comparing one metropolitan or urbanized area with another. However, these methods are
8

limited in their ability to distinguish patterns of high accessibility from patterns of low accessibility
because they ignore detailed land use and street patterns.

Multiple Dimensions
Most scholars now agree that sprawl is a multidimensional phenomenon (Galster et al, 2001; Malpezzi
and Guo, 2001; Ewing et al, 2002; Cutsinger et al, 2005; Frenkel and Ashkenazi, 2008; Torrens, 2008;
Jaeger et al., 2010; Mubareka et al., 2011). But this can lead to more questions than answers. What are
the dimensions of sprawl? How are they best measured? Should these dimensions be combined into an
overall sprawl index and, if so, how? Is sprawl in all dimensions necessary to call an urban area
sprawling? Are tradeoffs allowed?
The first multidimensional measures of sprawl were developed by Galster et al. (2001). They
disaggregated land use patterns into eight dimensions: density, continuity, concentration, clustering,
centrality, nuclearity, heterogeneity (mixing), and proximity. Sprawl was defined as a pattern of land use
that has low levels in one or more of these dimensions. The researchers operationally defined each
dimension and successfully quantified six of the eight measures for 13 urbanized areas. New York and
Philadelphia ranked as the least sprawling of the 13, and Atlanta and Miami as the most sprawling.
Since then, Galster and his colleagues have extended their sprawl measures to more than 50
metropolitan areas, confirming the multidimensional nature of sprawl. In one study, they ranked
metropolitan areas using 14 different dimensions, some related to population, others to employment,
and still others to both (Cutsinger et al. 2005). The 14 dimensions, which were reduced to seven factors
through principal component analysis, however, tended to cancel out each other. Metropolitan areas
ranking near the top on one factor were likely to rank near the bottom on another. Los Angeles, for
example, ranked second on both mixed use and housing centrality, but 48th on proximity and
49th on nuclearity. With so many overlapping variables, the analysis became confused.
Ewing, Pendall, and Chen (2002) also developed sprawl indices that like Galsters were
multidimensional, but were more focused and demonstrated wider degrees of variability among
metropolitan areas. They defined sprawl as any environment with (1) a population widely dispersed in
low-density residential development; (2) a rigid separation of homes, shops, and workplaces; (3) a lack
of major employment and population concentrations downtown and in suburban town centers and
other activity centers; and (4) a network of roads marked by very large block sizes and poor access from
one place to another. The authors used these indices to measure sprawl for 83 of the nations largest
metropolitan areas, standardizing the indices with mean values of 100 and standard deviations of 25.
The indices were constructed so that the more compact a metropolitan area was, the larger its index
value. More sprawling metropolitan areas had smaller index values. Thus, in the year 2000, the relatively
compact Portland, Oregon, metropolitan area had an index value of 126, while the slightly smaller
Raleigh-Durham metropolitan area had an index value of 54 (Figure 1.1). Los Angeles ended up near the
middle of the pack, with an index of 102.
Figure 1.1. Aerial Images of Portland and Raleigh at the Same Scale. www.maps.google.com
9

Recap of Past Studies


The most notable feature of past studies (with few exceptions like that of Galster et al.) is the failure to
define sprawl in all its complexity. Density is relatively easy to measure, and hence serves as the sole
indicator of sprawl in several studies. This flies in the face of both the technical literature and popular
conceptions of sprawl.
Another notable feature, related to the first, is the wildly different sprawl ratings given to different
metros by different analysts. With the exception of Atlanta, which always seems to rank as one of the
worst, the different variables used to operationalize sprawl lead to very different results. In one study,
Portland is ranked as most compact and Los Angeles is way down the list. In another their rankings are
essentially reversed. In a third study, certain Northeastern metros are characterized as sprawling, in a
fourth they are relatively compact.
A third notable feature was the failure to validate sprawl measures vis--vis impacts such as automobile
dependence. Related to this, earlier studies paid little attention to the impacts of sprawl generally. It is
the impacts of sprawl that concern planners and public officials, not the patterns themselves. With the
exception of a few studies focusing on individual impacts, the literature was nearly devoid of
empiricism. Sprawl was presumed to have negative consequences, or presumed to be free of them,
depending on the ideological bent of the author.

Small Area Measures


The concept of sprawl naturally brings to mind large geographic areas. When we say Atlanta sprawls
badly, we are referring to the Atlanta Metropolitan Area, or perhaps if we are a transportation planner,
to the Atlanta Urbanized Area. From the earliest writings on sprawl, sprawl was said to occur primarily
at the periphery of urbanized areas moving outward. An individual street or block may contribute to
sprawl, but we would not say it is sprawl. This distinction seems particularly poignant when we talk
about population and employment centering, which is defined by interrelationships among block
10

groups. If one block group or a group of them has a significantly higher density than those surrounding
it, we can say the former serves as a center for the block groups surrounding it.
Yet, we know from the travel and public health literatures that there is a demand in the research
community for built environmental metrics at the sub-county level, what might be described as the
community or neighborhood scale. Most of the built environment-travel studies, and most of the built
environment-obesity studies have related individual outcomes to such smaller areas. Therefore, we
have derived sprawl-like metrics for census tracts within metropolitan areas, and posted them along
with metropolitan area, urbanized area, and county sprawl metrics on the NIH website
((http://gis.cancer.gov). We have used the same type of variables as in larger area analyses, extracted
principal components from multiple variables using principal component analysis, and once again,
transformed the first principal component to an index with the mean of 100 and a standard deviation of
25. The component variables are:
Table 1.1: Variable Loadings on the Census Tract Compactness Index for 2010
Component Matrix
Data Sources
Density Factor
popden
gross population density
empden
gross employment density
urbden
net population density of urban lands
jobpop
job-population balance
jobmix
degree of job mixing (entropy)
walkscore
weighted average Walk Score
smlblk
percentage of small urban blocks
avgblksze
average block size
intden
intersection density
4way
percentage of 4-or-more-way intersections
Eigenvalue
Explained variance

Census 2010
LED 2010
NLCD
LED 2010
LED 2010
Walk Score Inc.
Census 2010
Census 2010
TomTom 2007
TomTom 2007

Factor
Loadings
0.666
0.298
0.467
0.486
0.664
0.821
0.677
-0.787
0.841
0.621
4.27
42.7%

11

Chapter 2. Updated County Sprawl Index


Ewing et al. (2003b; 2003c) originally estimated a single county sprawl index for each of 448
metropolitan counties or statistically equivalent entities (e.g., independent towns and cities). These
counties comprised the 101 most populous metropolitan statistical areas, consolidated metropolitan
statistical areas, and New England county metropolitan areas in the United States as of the 1990 census,
the latest year for which metropolitan boundaries were defined as that study began. Nonmetropolitan
counties, and metropolitan counties in smaller metropolitan areas, were excluded from the sample.
More than 183 million Americans, nearly two-thirds of the United States population, lived in these 448
counties in 2000.
Six variables were part of the original county sprawl index (as shown in Table 1). U.S. Census data were
used to derive three population density measures for each county:

gross population density in persons per square mile (popden)


percentage of the county population living at low suburban densities, specifically, densities
between 100 and 1,500 persons per square mile, corresponding to less than one housing unit
per acre (lt1500)
percentage of the county population living at medium to high urban densities, specifically, more
than 12,500 persons per square mile, corresponding to about 8 housing units per acre, the lower
limit of density needed to support mass transit (gt12500)

In deriving population density measures, census tracts were excluded if they had fewer than 100
residents per square mile (corresponding to rural areas, desert tracts, and other undeveloped lands).
Ewing et al. were only concerned with sprawl in developed areas where the vast majority of residents
live.
A fourth density variable was derived from estimated urban land area for each county from the National
Resources Inventory of the U.S. Department of Agriculture.

net population density of urban places within the county (urbden)

Data reflecting street accessibility for each county were also obtained from the U.S. Census. Street
accessibility is related to block size since smaller blocks translate into shorter and more direct routes. A
census block is defined as a statistical area bounded on all sides by streets, roads, streams, railroad
tracks, or geopolitical boundary lines, in most cases. A traditional urban neighborhood is composed of
intersecting bounding streets that form a grid, with houses built on the four sides of the block, facing
these streets. The length of each side of that block, and therefore its block size, is relatively small. By
contrast, a contemporary suburban neighborhood does not make connections between adjacent cul-desacs or loop roads. Instead, local streets only connect with the street at the subdivision entrance, which
is on one side of the block boundary. Thus, the length of a side of this block is quite large, and the block
itself often encloses multiple subdivisions to form a superblock, a half mile or more on a side. Large
block sizes indicate a relative paucity of street connections and alternate routes.
12

Two street accessibility variables were computed for each county:

average block size (avgblk)


percentage of blocks with areas less than 1/100 square mile, the size of a typical traditional
urban block bounded by sides just over 500 feet in length (smlblk).

Blocks larger than one square mile were excluded from these calculations, since they were likely to be in
rural or other undeveloped areas.
The six variables were combined into one factor representing the degree of sprawl within the county.
This was accomplished via principal component analysis, an analytical technique that takes a large
number of correlated variables and extracts a small number of factors that embody the common
variance in the original data set. The extracted factors, or principal components, are weighted
combinations of the original variables. When a variable is given a great deal of weight in constructing a
principal component, we say that the variable loads heavily on that component. The greater the
correlation between an original variable and a principal component, the greater the loading and the
more weight the original variable is given in the overall principal component score. The more highly
correlated the original variables, the more variance is captured by a single principal component.
The principal component selected to represent sprawl was the one capturing the largest share of
common variance among the six variables, that is, the one upon which the observed variables loaded
most heavily. This one component accounted for almost two-thirds of the variance in the dataset.
Because this component captured the majority of the combined variance of these variables, no
subsequent components were considered.
To arrive at a final index, Ewing et al. transformed the principal component, which had a mean of 0 and
standard deviation of 1, to a scale with a mean of 100 and standard deviation of 25. This transformation
produced a more familiar metric (like an IQ scale) and ensured that all values would be positive, thereby
allowing us to take natural logarithms and estimate elasticities.
The bigger the value of the index, the more compact the county. The smaller the value, the more
sprawling the county. Scores ranged from a high of 352 to a low of 63. At the most compact end of the
scale were four New York City boroughs, Manhattan, Brooklyn, Bronx, and Queens; San Francisco
County; Hudson County (Jersey City); Philadelphia County; and Suffolk County (Boston). At the most
sprawling end of the scale were outlying counties of metropolitan areas in the Southeast and Midwest
such as Goochland County in the Richmond, VA metropolitan area and Geauga County in the Cleveland,
OH metropolitan area. The county sprawl index was positively skewed. Most counties clustered around
intermediate levels of sprawl. In the U.S., few counties approach the densities of New York or San
Francisco.
For these counties, the original sprawl index was validated against journey to work, adult obesity, and
traffic fatality data (Ewing et al. 2003a; Ewing et al. 2003b; Ewing et al. 2003c). Later, the same county
sprawl index was used to model the built environment in a study of youth obesity (Ewing et al. 2006).
For this study, the index was computed for additional counties or county equivalents in order to have
13

sprawl data for more National Longitudinal Survey of Youth (NLSY97) respondents. The 954 counties or
county equivalents in the expanded sample represented the vast majority of counties lying within U.S.
metropolitan areas, as defined by the U.S. Census Bureau in December 2003. Almost 82% of the U.S.
population lived in metropolitan counties for which county sprawl indices were now available. Most
recent research on sprawl and its impacts has made use of this expanded dataset.

Update to 2010
In updating the original county sprawl index to 2010, five of the six variables were derived in the exact
same way as for 1990 and 2000. U.S. Census files for summary levels 140 (census tracts) and 101
(census blocks) were downloaded from American FactFinder. Population data were extracted for all
census tracts in all metropolitan counties. Land area data were extracted for all census blocks in all
metropolitan counties. Ninety-nine metropolitan counties were lost to the sample because they had no
census tracts averaging 100 persons per square mile or more. They were deemed to be rural.
The sixth variable, net density of urban areas within the county, was originally computed using data on
urban and built up uses from the National Resources Inventory of the U.S. Department of Agriculture.
The most recent NRI (2007) does not provide data at the county level. Therefore the U.S. Geological
Surveys National Land Cover Database (NLCD) was used instead. NLCD serves as the definitive Landsatbased, 30-meter resolution, land cover database for the Nation. It is a raster dataset providing spatial
reference for land surface classification (for example, urban, agriculture, forest). It can be geoprocessed to any geographic unit.
For the current work, the urban land area was generated at the county level using NLCD 2006 (the latest
product) and county geography (2010) for the entire U.S. Using the Tabulate Area spatial analyst tool
within ArcGIS, urban land areas within each county were calculated. The noncontiguous areas in the
same county were aggregated resulting in total urban area in square miles. The value codes treated as
urban were:
21. Developed, Open Space - Areas with a mixture of some constructed materials, but mostly
vegetation in the form of lawn grasses. Impervious surfaces account for less than 20% of total
cover. These areas most commonly include large-lot single-family housing units, parks, golf
courses, and vegetation planted in developed settings for recreation, erosion control, or
aesthetic purposes.
22. Developed, Low Intensity - Areas with a mixture of constructed materials and vegetation.
Impervious surfaces account for 20% to 49% percent of total cover. These areas most commonly
include single-family housing units.
23. Developed, Medium Intensity Areas with a mixture of constructed materials and
vegetation. Impervious surfaces account for 50% to 79% of the total cover. These areas most
commonly include single-family housing units.

14

24. High Intensity - Highly developed areas where people reside or work in high numbers.
Examples include apartment complexes, row houses and commercial/industrial. Impervious
surfaces account for 80% to 100% of the total cover.
The NRI and NLCD datasets are fairly comparable (see Appendix A), making the county sprawl indices for
1990, 2000, and 2010 fairly comparable. However, NLCD is only available for the continental U.S.
Therefore counties and county equivalents from Alaska, Hawaii, and Puerto Rico, 72 in total, were lost to
the sample.
Once again, principal component analysis was used to reduce the six variables to a single index. This
index accounts for 59 percent of the variance in the original six variables. Factor loadings are shown in
Table 2.1.
Table 2.1. County Sprawl Index Variables and Factor Loadings in 2010
Observed variable
popden
lt1500
gt12500
urbden
avgblk
smlblk
Eigenvalue
Explained variance
* Correlation with county sprawl index

Factor loading*
0.858
-0.658
0.821
0.876
-0.664
0.711
3.56
59.3%

We transformed the overall compactness score into an index with a mean of 100 and a standard
deviation of 25. This was done for the sake of consistency and ease of understanding. With this
transformation, the more compact counties have index values above 100, while the more sprawling
have values below 100.
Appendix A contains county sprawl (compactness) indices for 994 county and county equivalents in
2010. The 10 most compact and 10 most sprawling counties are shown in Tables 2.2 and 2.3. The most
compact counties are as expected, central counties of large, older metropolitan areas. The most
sprawling counties are outlying counties of large metropolitan areas, or component counties of smaller
metropolitan areas. Values range from 54 for Jackson County outside Topeka, Kansas, the most
sprawling county in 2010, to 464 for New York County (Manhattan), the most compact county in 2010.
Appendix A also contains estimates of county sprawl in 2000, derived by applying the 2010 component
score coefficient values to data for counties in 2000. Finally, the appendix presents changes in county
sprawl, measured equivalently, between the two census years.
Table 2.2. 10 Most Compact Counties in 2010 According to the Six Variable Index
1
2
3

County
New York County, NY
Kings County, NY
Bronx County, NY

Metropolitan Area
New York-Northern New Jersey-Long Island, NY-NJ-PA
New York-Northern New Jersey-Long Island, NY-NJ-PA
New York-Northern New Jersey-Long Island, NY-NJ-PA

Index
463.9
341.4
331.5
15

4
5
6
7
8
9
10

County
Queens County, NY
San Francisco County, CA
Hudson County, NJ
Suffolk County, MA
Philadelphia County, PA
District of Columbia, DC
Richmond County, NY

Metropolitan Area
New York-Northern New Jersey-Long Island, NY-NJ-PA
San Francisco-Oakland-Fremont, CA
New York-Northern New Jersey-Long Island, NY-NJ-PA
Boston-Cambridge-Quincy, MA-NH
Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
Washington-Arlington-Alexandria, DC-VA-MD-WV
New York-Northern New Jersey-Long Island, NY-NJ-PA

Index
272.1
247.8
228.8
217.1
216.8
193.3
190.1

Table 2.3. 10 Most Sprawling Counties in 2010 According to the Six Variable Index
985
986
987
988
989
990
991
992
993
994

Ford County, IL
Osage County, KS
Jasper County, IN
Grant County, AR
Tipton County, IN
Chester County, TN
Morrow County, OH
Greene County, NC
Polk County, MN
Jackson County, KS

Metropolitan Area
Champaign-Urbana, IL
Topeka, KS
Chicago-Joliet-Naperville, IL-IN-WI
Little Rock-North Little Rock-Conway, AR
Kokomo, IN
Jackson, TN
Columbus, OH
Greenville, NC
Grand Forks, ND-MN
Topeka, KS

Index
67.3
66.9
66.8
66.8
66.4
65.4
63.4
63.3
61.1
54.6

Figure 2.1 is a plot of 2010 sprawl index values vs. 2000 sprawl index values computed with the same
component score coefficient values. As one would expect, the degree of county sprawl does not change
dramatically over a 10-year period. Figure 2.2 is a histogram of changes in county sprawl values
between 2000 and 2010, where 2000 sprawl values are computed using the 2010 component score
coefficient values. As one would expect, given their fixed boundaries, most counties become more
compact (denser and with smaller blocks) over the ten-year period. Sprawl occurs mainly as previously
rural counties (in 2000) outside metropolitan areas become low-density suburbs and exurbs of
metropolitan areas (in 2010).
Figure 2.1. Scatterplot of 2010 Sprawl Index vs. 2000 Sprawl Index (Estimated Equivalently)

16

Figure 2.3. Histogram of Changes in County Sprawl Index Between 2000 and 2010 (Estimated
Equivalently)

17

Chapter 3. Refined County Sprawl Measures


A literature review by Ewing (1997) found poor accessibility to be the common denominator of sprawl.
Sprawl is viewed as any development pattern in which related land uses have poor access to one
another, leaving residents with no alternative to long distance trips by automobile. Compact
development, the polar opposite, is any development pattern in which related land uses are highly
accessible to one another, thus minimizing automobile travel and attendant social, economic, and
environmental costs. The following patterns are most often identified in the literature: scattered or
leapfrog development, commercial strip development, uniform low-density development, or single-use
development (with different land uses segregated from one another, as in bedroom communities). In
scattered or leapfrog development, residents and service providers must pass by vacant land on their
way from one developed use to another. In classic strip development, the consumer must pass other
uses on the way from one store to the next; it is the antithesis of multipurpose travel to an activity
center. Of course, in low-density, single-use development, everything is far apart due to large private
land holdings and segregation of land uses.
While the technical literature on sprawl focuses on land use patterns that produce poor regional
accessibility, poor accessibility is also a product of fragmented street networks that separate urban
activities more than need be. When asked, planners now routinely associate sprawl with sparse street
networks as well as dispersed land use patterns.
The original county sprawl index operationalized only two dimensions of urban formresidential
density and street accessibility. Our grant from the National Institutes of Health (NIH) provides for the
development of refined measures of county compactness or, conversely, county sprawl. These
measures are modeled after the more complete metropolitan sprawl indices developed by Ewing et al.
(2002). The refined indices operationalize four dimensions, thereby characterizing county sprawl in all
its complexity. The four are density, mix, centering, and street accessibility. The dimensions of the new
county indices parallel the metropolitan indices, basically representing the relative accessibility provided
by the county.
The full set of variables was used to derive a refined set of compactness/sprawl factors using principal
component analysis. One principal component represents population density, another land use mix, a
third centering, and a fourth street accessibility. County principal component values, standardized such
that the mean value of each is 100 and the standard deviation is 25, are presented in Appendix B. The
simple structure of the original county sprawl index has become more complex, but also more nuanced
and comprehensive, in line with definitions of sprawl in the technical literature.

Density
Low residential density is on everyones list of sprawl indicators. Our first four density variables are the
same as in the original sprawl index, gross density of urban and suburban census tracts (popden),
percentage of the population living at low suburban densities (lt1500), percentage of the population

18

living at medium to high urban densities (gt12500), and urban density based on the National Land Cover
Database (urbden).
The fifth density variable is analogous to the first, except it is derived with employment data from the
Local Employment Dynamics (LED) database rather than population data from the 2010 Census. The LED
database is assembled by the Census Bureau through a voluntary partnership with state labor market
information agencies. The data provide unprecedented details about America's jobs, workers, and local
economies. The LED data, available from 2002 to 2010, are collected at census block geography level
and can be aggregated to any larger geography, in this case block groups. LED variables include total
number of jobs, average age of workers, monthly earnings, and as of 2009 sex, race, ethnicity, and
education levels. In this case, LED data were processed for the year 2010. The data were aggregated
from census block geography to census block group geography to generate total jobs by two-digit NAICS
code for every block group in the nation, except those in Massachusetts, which doesnt participate in
the program. The density variable derived from the LED database is:

gross employment density of urban and suburban census tracts (empden)

Principal components were extracted from the five density-related variables, and the principal
component that accounted for the greatest variance became the county density factor. Factor loadings
(that is, correlations of these variables with the density factor) are shown in Table 3.1. The eigenvalue
of the density factor is 3.56, which means that this one factor accounts for more of the variance in the
original dataset than three of the component variables combined. In other words, the density factor
accounts for more than 70 percent of the total variance in the data set. As expected, one of the
variables loads negatively on the density factor, that being the percentage of population living at less
than 1,500 persons per square mile. The rest load positively. Thus, for all component variables, higher
densities translate into higher values of the density factor.
Table 3.1. Variable Loadings on the County Density Factor for 2010
Observed variable
Factor loading*
popden
lt1500
gt12500
urbden
empden
Eigenvalue
Explained variance
* Correlation with the density factor

0.983
0.848
-0.440
0.850
0.977
3.56
71.1%

Mixed Use
Three types of mixed-use measures are found in the land use-travel literature: those representing
relative balance between jobs and population within subareas of a region; those representing the
diversity of land uses within subareas of a region; and those representing the accessibility of residential
19

uses to nonresidential uses at different locations within a region. In this study, all three types were
estimated for counties in our sample and became part of a mix factor.
The first two variables were calculated for each block group using block-level population data from the
2010 Census, and block-level employment data from the 2010 LED database. For the first variable, each
block group centroid was buffered with a one-mile ring, and jobs and population were summed for
blocks within the ring. One-mile rings were used to standardize geography for census block groups,
which vary widely in size, making balance easier to achieve in the larger block groups. The resulting job
and population totals were used to compute a job-population balance measure. 1 This variable equals 1
for block groups with the same ratio of jobs-to-residents within the one-mile ring as the metropolitan
area as a whole; 0 for block groups with only jobs or residents within the one-mile ring, not both; and
intermediate values for intermediate cases. All values were weighted by the sum of block group jobs and
residents as a percentage of the county total to obtain:

countywide average job-population balance (jobpop).

For the second mixed-use variable, each block group centroid was again buffered with a one-mile ring,
and jobs by sector were summed for blocks within the ring. An entropy formula was then used to
compute a measure of job mix. 2 The variable equals 1 for block groups with equal numbers of jobs in
each sector within the ring; 0 for block groups with all jobs in a single sector within the ring; and
intermediate values for intermediate cases. The sectors considered in this case were retail,
1

The equation used to calculate job-population balance was:

i =n

(1 ( ABS ( J
i =0

where:

JP * Pi )) /( J i + JP * Pi )) * (( BJ i + BPi ) /(TJ + TP ))

i = census tract number (excluding those with fewer than 100 persons per square mile)
n = number of census tracts in the county
J = jobs in the census tract
P = residents in the census tract
JP = jobs per person in the metropolitan area
TJ = total jobs in the county
TP = total residents in the county

The equation for this measure is:


n

(( P * LN ( P )) / LN ( j )) * (( BJ
i =1

where:

+ BPi ) /(TJ + TP ))

i = census tract number (excluding those with fewer than 100 persons per square mile)
n = number of census tracts in the county
j = number of sectors
Pj = proportion of jobs in sector j
JP = jobs per person in the metropolitan area
TJ = total jobs in the county
TP = total residents in the county

20

entertainment, health, education, and personal services. Values were weighted by the sum of block
group population and employment as a percentage of the county total to obtain:

countywide degree of job mixing (jobmix).

A third mixed-use variable uses data from Walk Score, Inc. to measure proximity to amenities, with
different amenities weighted differently and amenities discounted as the distance to them increases up
to one mile and a half, where they are assumed to be no longer accessible on foot. 3 Classic Walk Score
data were acquired for all urban census tracts in the United States. Year 2012 data were purchased to
reduce the cost of data acquisition. Values were weighted by the sum of block group population and
employment as a percentage of the county total to obtain:

countywide average Walk Score (walkscore)

Principal components were extracted from the three mix-related variables, and the principal component
that accounted for the greatest variance became the mix factor. Loadings of these variables on the mix
factor are shown in Table 3.2. The eigenvalue of the mix factor is 2.30, which means that this one factor
accounts for more than two-thirds of the total variance.
Table 3.2. Variable Loadings on the County Mix Factor for 2010
Observed variable
Factor loading*
jobpop
jobmix
walkscore
Eigenvalue
Explained variance
* Correlation with the mix use factor

0.891
0.942
0.784
2.30
76.6%

Centering
Urban centers are concentrations of activity that provide agglomeration economies, support alternative
modes and multipurpose trip making, create a sense of place in the urban landscape, and otherwise
differentiate compact urban areas from sprawling ones. Centeredness can exist with respect to
population or employment, and with respect to a single dominant center or multiple subcenters. The
technical literature associates compactness with centers of all types, and sprawl with the absence of
centers of any type.
Ewing et al. (2002) measured metropolitan centering, in part, in terms of concentrations of development
in or around historic central business districts (CBDs) of metropolitan areas. This concept of centering
does not make much sense when applied to the individual counties that make up a metropolitan area,
only one of which can contain the historic central business district. Other counties have their own
subcenters in the polycentric metropolitan areas of today, and the existence of and proximity to these
3

A grocery store, for example, gets three times the weight of a book score. The distance decay function starts with
a value of 100 and decays to 75 percent at a half mile, 12.5 percent at one mile, and zero at 1.5 miles.

21

are what distinguish counties with concentrations of activity from those without. Four measures of
centering were derived for metropolitan counties:
The first centering measure came straight out of the 2010 census:

coefficient of variation in census block group population densities, defined as the standard
deviation of block group densities divided by the average density of block groups. The more
variation in densities around the mean, the more centering and/or subcentering exists within
the county (varpop)

The second centering measure was derived from the LED database and is analogous to the first
measure, except for its use of employment density by block group rather than population density to
compute:

coefficient of variation in census block group employment densities, defined as the standard
deviation of block group densities divided by the average density of block groups. The more
variation in densities around the mean, the more centering and/or subcentering exists within
the county (varemp)

The last two centering variables measure the proportion of employment and population within CBDs
and employment sub-centers. We first identified the location of CBDs and employment sub-centers for
all metropolitan areas. For identifying CBDs, we ran a local spatial autocorrelation procedure using the
local Morans I statistic (Anselin, 1995). 4 With this procedure, it is possible to quantify the degree of
clustering of neighboring zones with high levels of density. This method has been used by Baumont & Le
Gallo (2003), Giullian et al. (2004) and Riguelle et al. (2007).

Local Morans I is defined as:

where Ii is the local Morans I coefficient, X is the value of the employment density, wij is the matrix of spatial
weights, and n is the number of observations. Through calculating z-values of the local Moran statistic (see Anselin,
1995; Getis and Ord, 1996) it is then possible to identify two types of spatial clusters, two types of outliers :

High-high
Low-low
High-low
Low-high

High values around neighbors with high values (cluster)


Low values around neighbors with low values (cluster)
High values around neighbors with low values (outlier)
Low values around neighbors with high values (outlier)

Using LED data of block groups, the Morans I analysis was done for all Metropolitan areas. The High-High clusters
with the highest employment density in each metropolitan statistical area (MSA) were considered as CBD
candidates. However not all of them are CBDs. We excluded the hot spots containing large firms such as hospitals,
malls and university campuses by applying the threshold of having employment share of no more than 75 percent
in each sector. We identified CBD for a total of 356 Metropolitan areas.

22

Having CBDs for 356 metropolitan areas, we identified employment sub-centers as the positive residuals
estimated from an exponential employment density function using Geographically Weighted Regression
method (GWR). 5 In the literature, urban sub-centers are areas with significantly higher employment
density than the surrounding areas (McDonald 1987). To identify sub-centers, researchers have used
several types of procedures: a minimum density procedure (Giuliano and Small 1991), identification of
local peaks (Craig & Ng, 2001 and Readfearn, 2007), and a nonparametric method (McMillen 2001). The
last of these methods works best, according to literature review by Lee (2007). Using this procedure, we
found 224 metropolitan areas to be monocentric (have only one center), 132 to be polycentric (have
more than one center), and 18 metropolitan areas to be dispersed (have no CBD and no sub-center).
This procedure resulted in two new centering variables. These findings were validated by inspecting
Google Earth satellite images to identify concentrations of activity, and see whether they corresponded
to our findings with GWR.

Percentage of county population in CBD or sub-centers (popcen)


Percentage of county employment in CBD or sub-centers (empcen)

Principal components were extracted from the set of centering variables, and the principal component
that accounted for the greatest variance became our centering factor. All component variables loaded
positively on the centering factor (see table 3.3). The eigenvalue of the centering factor is 1.96, which
means that this one factor accounts for just under half of the total variance in the data set.
Table 3.3. Variable Loadings on the County Centering Factor for 2010
Observed variable
Factor loading*
varpop
varemp
popcen
empcen
Eigenvalue
Explained variance
* Correlation with the centering factor

0.085
0.642
0.820
0.932
1.96
49.1%

GWR estimates a smoothed employment density surface using only nearby observations for any data point
(block groups), with more weights given to closer observations. The dependent variable of the GWR estimations is
employment density by block groups and the independent variable is the distance of the block group centroid from
the CBD. We used the Adaptive kernel type with 30 numbers of neighbors. The block groups with highest positive
residual (if residual is 4 times greater than predicted) are candidates for employment sub-centers. As with CBD
identification, we excluded the block groups containing large firms such as hospitals, regional malls, and university
campuses by applying the requirement that the employment share be no more than 75 percent in each sector.
Finally we excluded cases when their ratio of employment to population was less than 2.5 (Gordon et al 1986). We
identified a total of 451 sub-centers in 132 metropolitan areas.

23

Street Accessibility
In the refined sprawl indices, two street variables are the same as in the original county sprawl index:
average block size excluding rural blocks of more than one square mile (avgblk) and percentage of small
urban blocks of less than one hundredth of a square mile (smlblk). To these two street accessibility
variables were added. The two new street variables are:

intersection density for urban and suburban census tracts within the county, excluding rural
tracts with gross densities of less than 100 persons per square mile (intden)
percentage of 4-or-more-way intersections, again excluding rural tracts (4-way)

Intersection density captures both block length and street connectivity. Percentage of 4-or-more-way
intersections provides a pure measure of street connectivity, as 4-way intersections provide more
routing options than 3-way intersections.
Starting with a 2006 national dataset of street centerlines generated by TomTom that ships with ArcGIS,
we produced a national database of street intersection locations, including for each intersection feature
a count of streets that meet there. The TomTom dataset includes one centerline feature for each road
segment running between neighboring intersections; i.e. every intersection is the spatially coincident
endpoint of 3 or more road segments. 6
The resulting national intersection database contains 13.1 million features; 77% of these are three-way
intersections, and the remaining 23% are four- or more-way intersections. Total counts of 3- and 4-ormore-way intersections were tabulated for census tracts, and census tracts were aggregated to obtain
county-level data. For each county, the total number of intersections in urban and suburban tracts was
divided by the land area to obtain intersection density (intden), while the number of 4-or-more-way
intersections was multiplied by 100 and divided by the total number of intersections to obtain the
percentage of 4-or-more way intersections (4way).

Intersection features were created as follows: Using Census Feature Class Code (CFCC) values, we filtered out all
freeways, unpaved tracks, and other roadways that don't function as pedestrian routes. Divided roadways, which
from a pedestrian mobility perspective function similarly to undivided roadways of the same functional class, were
represented in the source data as pairs of (roughly) parallel centerline segments. These were identified by CFCC
value and merged into single segments using GIS tools. Streets intersecting the original divided roadways were
trimmed or extended to the new merged centerlines, and the new merged centerlines were split at each
intersection with side streets such that centerline features only intersect each other at feature endpoints.
Roundabouts were assumed to function similarly to single 4+-way intersections, rather than close-set clusters of
intersections joining the roundabout proper and the incoming streets. As such, centroids of roundabout circles
were located and assigned an assumed count of four incoming streets; endpoints of incoming street features were
ignored.

With the corrected street centerline data prepared, we generated point features at both endpoints of each street
segment. Points closer together than 12m were adjusted to be spatially coincident in order to control for any
possible remaining geometric errors related to divided roadways. We then used GIS tools to count the number of
points (representing ends of street segments) coinciding at any location. Locations with point counts of one (dead
ends) or two (locations where a roadway changes name, functional class, or other attribute) were discarded as
non-street intersections. Remaining locations were flagged with attributes indicating whether a point was a threeway or a four- or more-way intersection.

24

Principal components were extracted from the full set of street-related variables, and the principal
component that accounted for the greatest variance became our street accessibility factor. Loadings of
these variables on the street factor are shown in Table 3.4. The eigenvalue of the street factor is 2.39,
which means that this one factor accounts for more than half of the total variance in the data set. As
expected, one of the variables loads negatively on the street accessibility factor, that being the average
block size. The rest load positively. Thus, for all component variables, more accessibility translates into
higher values of the street factor.
Table 3.4. Variable Loadings on the County Street Factor for 2010
Observed variable
Factor loading*
avgblk
smlblk
inden
4-way
Eigenvalue
Explained variance
* Correlation with the street factor

-0.764
0.901
0.836
0.545
2.39
59.8%

Relationship Among Compactness Factors


It has been said that measures of the built environment are so highly correlated that they should not be
represented separately, but instead should be combined into a single index. Thus, for example, overall
measures of walkability have been advanced as an alternative to individual measures.
This position is not borne out by this study, at least not at the county level. While correlated, as one
might expect, the four compactness factors seem to represent distinct constructs. Their simple
correlation coefficients are shown in Table 3.5. The highest is 0.647, which means that each factor
explains less than 42 percent of the variation in the other.
Table 3.5. Simple Pearson Correlation between four factors
density factor

mix factor

centering factor

street factor

density factor
mix factor

1
0.399**

0.399**
1

0.523**
0.421**

0.583**
0.647**

centering factor
street factor

0.523**
0.583**

0.421**
0.647**

1
0.438**

0.438**
1

Composite Index
The next issue we had to wrestle with was how to combine the four factors into a single sprawl index. A
priori, there is no right way to do so, only ways that have more or less face validity.
Should the four factors be weighted equally, or should one or another be given more weight than the
others? Density has certainly received more attention as an aspect of sprawl than has, say, street
accessibility. However, beyond play in the literature, we could think of no rationale for differential
weights. The first three factors all contribute to the accessibility or inaccessibility of different
25

development patterns, none presumptively more than the others. Depending on their values, all move
a county along the continuum from sprawl to compact development to sprawl. Thus they were simply
summed, in effect giving each dimension of sprawl equal weight in the overall index.
As with the individual sprawl factors, we transformed the overall compactness score into an index with a
mean of 100 and a standard deviation of 25. This was done for the sake of consistency and ease of
understanding. With this transformation, the more compact counties have index values above 100,
while the more sprawling have values below 100.
Appendix B contains compactness factors and refined county sprawl (compactness) indices for 967
county and county equivalents in 2010. Note that Massachusetts counties are missing from the mix
factor and overall index for lack of LED data. The ten most compact and ten most sprawling counties are
shown in Tables 3.6 and 3.7. The rankings are similar to those with the original county sprawl index.
The most compact counties are central counties of large, older metropolitan areas. The most sprawling
counties are outlying counties of large metropolitan areas, or component counties of smaller
metropolitan areas. Values range from 42 for Oglethorpe County, GA outside Athens, the most
sprawling county in 2010.
Looking at Tables 2.2 and 3.6, it would seem that the original and new compactness indices are
measuring the same construct, but that is not quite true. Just compare Tables 2.3 and 3.7, where there
is no overlap in the most sprawling counties according to the two indices. The original compactness
index is dominated by density variables (four of six variables in the index) and only slightly diluted by
street variables (two of the six), which correlate strongly with density. The new compactness index
dilutes the role of density by adding two new factors (mix and centering). The simple correlation
coefficient between original and new indices is 0.865, which means that about 25 percent of the
variance in each index is unexplained by the other. We would expect that they have similar but not
identical relationships to outcome variables, and similar but not identical predictive power.
Table 3.6. 10 Most Compact Counties in 2010 According to the Four-Factor Index (excluding
Massachusetts counties)
1
2
3
4
5
6
7
8
9
10

County
New York County, NY
Kings County, NY
San Francisco County, CA
Bronx County, NY
Philadelphia County, PA
District of Columbia, DC
Queens County, NY
Baltimore city, MD
Norfolk city, VA
Hudson County, NJ

Metropolitan Area
New York-Northern New Jersey-Long Island, NY-NJ-PA
New York-Northern New Jersey-Long Island, NY-NJ-PA
San Francisco-Oakland-Fremont, CA
New York-Northern New Jersey-Long Island, NY-NJ-PA
Philadelphia-Camden-Wilmington, PA-NJ-DE-MD
Washington-Arlington-Alexandria, DC-VA-MD-WV
New York-Northern New Jersey-Long Island, NY-NJ-PA
Baltimore-Towson, MD
Virginia Beach-Norfolk-Newport News, VA-NC
New York-Northern New Jersey-Long Island, NY-NJ-PA

Index
425.2
265.2
251.3
224.0
207.2
206.4
204.2
190.9
179.6
178.7

Table 3.7. 10 Most Sprawling Counties in 2010 According to the Four-Factor Index (excluding
Massachusetts counties)

26

960
961
962
963
964
965
967
966
968
969

County
Spencer County, KY
Morrow County, OH
Brown County, IN
Blount County, AL
Greene County, NC
Harris County, GA
Macon County, TN
Elbert County, CO
Grant Parish, LA
Oglethorpe County, GA

Metropolitan Area
Louisville/Jefferson County, KY-IN
Columbus, OH
Indianapolis-Carmel, IN
Birmingham-Hoover, AL
Greenville, NC
Columbus, GA-AL
Nashville-Davidson--Murfreesboro--Franklin, TN
Denver-Aurora-Broomfield, CO
Alexandria, LA
Athens-Clarke County, GA

Index
60.4
58.8
58.5
56.6
56.6
55.1
54.3
54.3
53.8
45.5

Greater Validity of New Index


Compared to the original county compactness index, the new four-factor index has greater construct
and face validity. It has greater construct validity because it captures four different dimensions of the
construct compactness (density, mix, centering, and street accessibility), whereas the original index
captures only two dimensions (density and street accessibility).
The greater face validity of the new four-factor index requires some explanation. The very first county
compactness indices were derived for only 448 counties in the largest 101 metropolitan areas. The most
sprawling counties, such as Geauga County outside Cleveland, have classic sprawl patterns of lowdensity suburban development.
Expanding to 994 counties and adding smaller metropolitan areas, the picture becomes more
complicated. Tables 2.2 and 3.8 list the most compact counties as measured by both indices. The ten
most compact counties based on the original index largely overlap with the top ten based on the new
index (with the notable exception of Suffolk County (Boston), for which we dont have all required
variables). New York County (Manhattan) is the most compact according to both indices (see Figure
3.1). Kings County (Brooklyn) is the second most compact according to both indices (see Figure 3.2).
Figure 3.1. Most Compact County According to Both Indices (New York County, NY)

27

Figure 3.2. Second Most Compact County According to Both Indices (Kings County, NY)

However, the ten most sprawling counties are entirely different when measured by different indices
(see Tables 2.3 and 3.9). Which index has greater face validity? We reviewed satellite imagery for the
ten most sprawling counties, according to both indices, and found that the development patterns for
the new index are much more representative of classic suburban sprawl (see Tables 3.8 and 3.9). While
all 20 counties are part of metropolitan areas, many of the counties rated as most sprawling according
to the original index have different development patterns than expected. They would best be described
as exurban counties with small towns surrounded by farmlands (see Figures 3.3 and 3.4). The small
towns have moderate densities and gridded streets. The fact they are part of larger census tracts, our
28

units of analysis, depresses their densities and compactness scores. They are not examples of classic
suburban or exurban sprawl. On the other hand, the counties rated as most sprawling according to the
new four-factor index have census tracts with very low-density residential development.
Table 3.8. 10 Most Sprawling Counties in 2010 According to the Six-Variable Index
Ford County, IL
Osage County, KS
Jasper County, IN
Grant County, AR
Tipton County, IN
Chester County, TN
Morrow County, OH
Greene County, NC
Polk County, MN
Jackson County, KS

Development Pattern
Small town surrounded by rural development
Small town surrounded by rural development
Continuous low density suburban development
Continuous low density suburban development
Small town surrounded by rural development
Continuous low density suburban development
Continuous low density suburban development
Continuous low density suburban development
Small town surrounded by rural development
Small town surrounded by rural development

Index
67.3
66.9
66.8
66.8
66.4
65.4
63.4
63.3
61.1
54.6

Table 3.9. 10 Most Sprawling Counties in 2010 According to the Four-Factor Index (excluding
Massachusetts counties)
County
Spencer County, KY
Morrow County, OH
Brown County, IN
Blount County, AL
Greene County, NC
Harris County, GA
Macon County, TN
Elbert County, CO
Grant Parish, LA
Oglethorpe County, GA

Metropolitan Area
Louisville/Jefferson County, KY-IN
Columbus, OH
Indianapolis-Carmel, IN
Birmingham-Hoover, AL
Greenville, NC
Columbus, GA-AL
Nashville-Davidson--Murfreesboro--Franklin, TN
Denver-Aurora-Broomfield, CO
Alexandria, LA
Athens-Clarke County, GA

Index
60.4
58.8
58.5
56.6
56.6
55.1
54.3
54.3
53.8
45.5

Figure 3.3. Most Sprawling County According to Six-Variable Index (Jackson County, KS)

29

Figure 3.4. Second Most Sprawling County According to Six-Variable Index (Polk County, MN)

Figure 3.5. Most Sprawling County According to Four-Factor Index (Oglethorpe County, GA)

30

Figure 3.6. Second Most Sprawling County According to Four-Factor Index (Grant Parish, LA)

31

Chapter 4. Validation of County Sprawl Indices


The potential to moderate travel demand by changing the built environment is the most heavily
researched subject in urban planning. The effect of the built environment on transportation and travel
behavior is confirmed by more than 200 empirical studies (Ewing and Cervero, 2010). This literature is
summarized in recent reviews by Cao et al. (2009), Heath et al. (2006), Pont et al. (2009), Graham-Rowe
et al. (2011), Litman and Steele (2011), and Salon et al. (2012), and in meta-analyses by Leck (2006) and
Ewing and Cervero (2010).
In travel research, influences of the built environment on travel have often been named with words
beginning with D. The original three Ds, coined by Cervero and Kockelman (1997), are density,
diversity, and design, followed later by destination accessibility and distance to transit (Ewing and
Cervero, 2001; Ewing and Cervero, 2010; Ewing, Greenwald, Zhang, et al., 2011). While usually
measured at the neighborhood scale, the D variables roughly correspond to our conception of sprawl at
the larger scale of counties. Sprawl is any development pattern that is low on D variables (or high on
distance to transit). Compact development is any development pattern with low D values (or inversely,
high distance to transit). Development density, a component of county compactness, precisely
corresponds to density of neighborhoods, and is measured the same way. Land use mix, a component of
county compactness, precisely corresponds to land use diversity of neighborhoods, and is measured the
same way. Street accessibility, a component of county compactness, precisely corresponds to street
network design, the most common design variable measured in travel behavior studies. Even centering,
a component of county compactness, has a rough equivalency in the concept of destination accessibility.
Thus to the extent that the travel behavior literature has established D variables as predictors of travel
choices, we would expect to find a strong relationship between dimensions of sprawl and travel
outcomes at the more aggregate scale of counties.
The relationships between sprawl and travel outcomes can be used to validate our county sprawl
measures, and also to see if one measure has more predictive power than another other. If sprawl has
any consistently recognized outcome, it is automobile dependence. We would expect to find that, after
controlling for other relevant influences, compact counties have relatively low vehicle ownership, high
transit and walking commute mode shares, and short drive times to work.

Methods
Data and Variables
Using American Factfinder, we downloaded data from the 2010 American Community Survey (ACS), 5year estimates, and computed average vehicle ownership, mode shares, and average drive times by
county. We also downloaded socioeconomic data and computed percentages and mean values to
describe county populations. The ACS is an ongoing statistical survey which produces estimates of
demographic, social, economic, and housing characteristics at a variety of geographic scales. The 5-year
estimates are based on data collected over a 60-month period from 2006 through 2010. From the 2010
Census, we downloaded data on gender, age, race, household type and housing tenure, and computed
percentages and mean values to describe county populations.
We also downloaded data on crime, weather/climate, fuel price, and the size of metropolitan area in
which a county is located, all for 2010. Crime statistics came from the uniform crime report of the
Federal Bureau of Investigation (FBI). The FBI supplied crimes by type (violent and property) and
subtype (murder, rape, etc.) aggregated by county. Annual heating degree days (HDD), cooling degree
32

days (CDD), and annual precipitation data came from the National Oceanic and Atmospheric
Administration (NOAA)s National Climatic Data Center for 2009 for HDD and 2010 for the other two
variables. We recorded data for stations located in the same MSAs as the counties. In a few cases there
were not weather stations in MSAs, so we used data from the closest station to a particular MSA.
Gasoline price data at the MSA level were purchased from the Oil Price Information Service (OPIS). Retail
prices are average prices from samples of stations in each MSA and are reported with all relevant taxes
included. These prices are the true posted sign prices (as they would appear outside a gas station).
Table 4.1 provides a list of all dependent and independent variables. We estimated different models
using different measures of compactness. Our first set of regressions used the same six-variable
measure of compactness/sprawl as in 2000, but updated to 2010 (indexo). The second set of
regressions used the new measure of compactness/sprawl, with four distinct factors combined into an
overall compactness score (indexn). The third set of regressions used the four distinct factors
individually as regressors (denfac, mixfac, cenfac, strfac).
Table 4.1. Variables Used to Explain Travel Outcomes (all variables log transformed)
Level 1 Dependent variables (county level)
Variables
hhveh
pctwalk

average household vehicle ownership


percentage of commuters walking to work
percentage of commuters using public transportation
pcttrans
(excluding taxi)
drivetim
average journey-to-work drive time in minutes
Level 1 Independent variables (county level)
Variables
pctmale
pctwhite
pct1524
hhsize
hhinc
crime

percentage of male population


percentage of white population
percentage of population 15-24 years old
average household size
median household income

indexo

violent crime rate per 100,000 population


county compactness index for 2010 (using the same 2000
index variables)

denfac

density factor (a weighted combination of 5 density variables)

mixfac

mix factor (a weighted combination of 3 mixed-use variables)


centering factor (a weighted combination of 4 centering
variables)
street factor (a weighted combination of 4 street-related
variables)
county compactness index for 2010 (including additional
variables compared to 2000 index)

cenfac
strfac
indexn

Data Sources
ACS 2006-2010
ACS 2006-2010
ACS 2006-2010
ACS 2006-2010
Data Sources
Census 2010
Census 2010
Census 2010
Census 2010
ACS 2006-2010
FBI Uniform Crime
Statistics
computed
Census 2010, NLCD
database
Census 2010, LED 2010,
Walk Score
Census 2010, LED 2010
Census 2010, TomTom,
ESRI
computed

33

Level 2 Independent Variables (metropolitan level)


Variables
metpop
precip
hdd
cdd
fuel

metropolitan area population


annual precipitation (inches)
heating degree-days: the outside base temperature above
which a building needs heating
cooling degree-days: the outside base temperature above
which a building needs cooling
average fuel price for the metropolitan area

Data Sources
Census 2010
NOAA database 2010
NOAA database2010
NOAA database2010
OPIC database 2010

Our dependent variables were logged so as to be normally distributed and hence properly modeled with
regression analysis. Figures 4.1 and 4.2 show, respectively, histograms of the walk share of commute
trips for counties in our sample, and the natural logarithm of walk share. The latter clearly conforms
more closely to a normal distribution than does the former. The distribution is even more skewed for
the transit share of commute trips, and so it too was logged. The distribution of average drive times
comes closer to normality, but conforms more closely to a normal distribution when logged than when
linear.
Figure 4.1. Frequency Distribution of the Percentage of Commute Walk Trips by County

Figure 4.2. Frequency Distribution of the Natural Logarithm of the Percentage of Commute Walk Trips
by County

34

As for the independent variables, we focus on the compactness/sprawl measures, as the other variables
are simply used as controls. As shown in Figure 4.3, the natural logarithm of walk mode share, when
plotted against the linear version of the original compactness index, suffers from nonlinearity, outlying
values, and heteroskedasticity. All three statistical problems are moderated when, instead, the natural
logarithm of the walk mode share is plotted against the natural logarithm of the compactness index, and
the New York City cases are dropped from the sample, as in Figure 4.4. Therefore, in all regressions, we
transformed all variables into log form to achieve a better fit with the data, reduce the influence of
outliers, and adjust for nonlinearity of the data. The transformations had the added advantage of
allowing us to interpret regression coefficients as elasticities. An elasticity is a percentage change in one
variable that accompanies a one percent change in another variable. Elasticities are the most common
measures of effect size in both economics and planning.
Figure 4.3. Scatterplot of Natural Logarithm of the Commute Walk Mode Share vs. the Original
Compactness Index

35

Figure 4.4. Scatterplot of Natural Logarithm of the Commute Walk Mode Share vs. the Natural
Logarithm of Original Compactness Index

36

Analytical Method
As shown in Table 4.1, our independent variables come from two levels of geography, the county and
the metropolitan area or MSA. Since all counties located in a MSA share characteristics of that MSA
such as climate and fuel price, we can say that counties are nested within MSAs. This nesting tends to
produce dependence among cases, violating the independence assumption of ordinary least squares
regression or OLS. Standard errors of regression coefcients associated with MSA characteristics based
on OLS will consequently be underestimated, and regression coefficients themselves will be inefficient.
Whenever individuals or other entities are nested within higher-level units, hierarchical or multilevel
modeling (MLM) methods are preferred for explaining individual outcomes in terms of both individual
and group characteristics. Multilevel modeling overcomes the limitations of OLS, accounting for the
dependence among lower level units and producing more accurate coefcient and standard error
estimates. So MLM is used in this analysis for all transportation outcomes. Models were estimated with
HLM 6.08.

Results
Household Vehicle Ownership
Results for household vehicle ownership are shown in Table 4.2. Average vehicle ownership is, as
expected, directly related to household size, household income, and percentage of whites in the
population. Vehicle ownership is negatively related to average fuel price, doubtless because the
generalized cost of automobile use is higher in metropolitan areas with higher fuel prices. Vehicle
ownership is negatively related to cooling degree days. We would say this is because warm climates
support walking and bicycling, but this explanation does not square with later results.
The variables of greatest interestthe original compactness index, the new compactness index, and the
four individual compactness factorsare negatively related to household vehicle ownership (with the
exception of the street factor, which is insignificant). A household of comparable socioeconomic status
can make do with fewer cars in a compact county because vehicle trips are less numerous and shorter.
This phenomenon is known as car shedding. Of the four compactness factors, the density factor has the
strongest relationship to vehicle ownership. A dense urban environment usually offers alternatives to
the automobile; the automobile has less utility in such an environment due to traffic congestion and
parking problems. The centering factor is only slightly less significant than the density factor, and the
mix factor less significant still.
Table 4.2. Relationships to Average Household Vehicle Ownership (log-log form with robust standard
errors)

constant
hhsize
hhinc
pctwhite
cdd
fuel

Model 1
coeff.

Model 1
t-ratio

Model 1
p-value

Model 2
coeff.

Model 2
t-ratio

Model 2
p-value

Model 3
coeff.

Model 3
t-ratio

Model 3
p-value

0.88
0.49
0.17
0.099
-0.034
-0.39

2.39
10.19
8.53
4.74
-6.25
-4.41

0.017
<0.001
<0.001
<0.001
<0.001
<0.001

0.57
0.43
0.15
0.13
-0.049
-0.47

2.31
8.17
7.40
8.64
-6.51
-4.46

0.022
<0.001
<0.001
<0.001
<0.001
<0.001

2.42
0.42
0.16
0.069
-0.04
-0.18

6.39
10.4
14.1
4.46
-8.44
-2.46

<0.001
<0.001
<0.001
<0.001
<0.001
0.014
37

Model 1
coeff.

Model 1
t-ratio

indexo
-0.50
-5.29
indexn
denfac
mixfac
cenfac
strfac
pseudo-R2
NA*
* The null model doesnt converge.

Model 1
p-value

Model 2
coeff.

Model 2
t-ratio

Model 2
p-value

-0.38

-5.41

<0.001

Model 3
coeff.

Model 3
t-ratio

Model 3
p-value

-0.75
-0.036
-0.058
0.017

-7.69
-2.60
-5.14
0.70
NA*

<0.001
0.010
<0.001
0.49

<0.001

NA*

Walk Mode Shares


Results for walking to work are shown in Table 4.3. Among the level 1 variables, walking to work is
positively related to the percentage of young adults in the county population and the percentage of
males. These relationships are as expected, as young people are less likely to have private vehicles
available to them and more likely to be physically active. Males are less likely than females to be
intimidated by environmental conditions such as crime. Walking to work is negatively related to
household size and household income. These relationships are also as expected, since large households
can make more efficient use of automobiles through trip chaining and carpooling and higher income
households own more private vehicles. Finally, walking to work is negatively related to the violent crime
rate within the county. Crime has been found to discourage physical activity generally.
Among the level 2 variables, walking to work is positively related to metropolitan average fuel price and
negatively related to cooling degree days. The first relationship reflects the depressing effect of higher
fuel prices on auto use; there are persistent and significant differences in fuel prices across regions. The
second relationship indicates that hot climates and weather discourage walking to work. This seems
reasonable since the work trip tends to be longer than trips for other purposes, making environmental
conditions especially important.
As for the variables of greatest interest to us, those related to the built environment, the original
compactness index (indexo) in Model 1 has the expected strong positive relationship to walking.
Compact counties, with high values of the index, have relatively high walk shares. As noted, the
coefficients of these log-log models are elasticities. Every one percent increase in the original county
compactness index results in a 0.54 percent increase in the walk mode share. The new compactness
index (indexn) in Model 2 has an even stronger relationship to walking. The coefficient and t-statistic
are higher for the new index than the original index. Every one percent increase in this index results in a
0.66 percent increase in the walk mode share.
The new multidimensional compactness factors are significant in two cases and have expected positive
sign in three cases. The density factor, denfac, is the most important correlate of walking, followed by
the mix factor, mixfac. A one percent increase in the density factor is associated with a 0.72 percent
increase in walking, while a one percent increase in the mix factor is associated with a 0.25 percent
increase in walking. The centering factor has a negative but insignificant relationship to walking.
Population and employment centers may have less balanced jobs and housing, making walk trips less
practicable. The big surprise is that the street factor is not significant. It is widely assumed that small

38

blocks, frequent intersections, and high street connectivity translate into greater walkability. At least in
this aggregate analysis, this does not appear to be the case.
Table 4.3. Relationships to Walk Mode Share (log-log form with robust standard errors)

constant
hhsize
pctmale
pct1524
hhinc
crime
cdd
fuel
indexo
indexn
denfac
mixfac
cenfac
strfac
pseudo-R2

Model 1 Model 1 Model 1 Model 2 Model 2 Model 2 Model 3 Model 3 Model 3


coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
-11.2
-3.87
<0.001
-14.6
-5.17
<0.001
-15.8
-4.76
<0.001
-2.86
-10.0
<0.001
-2.56
-10.9
<0.001
-2.68
-10.0
<0.001
2.29
3.33
0.001
3.29
5.01
<0.001
3.31
4.20
0.001
1.29
18.4
<0.001
1.19
14.5
<0.001
1.21
15.7
<0.001
-0.12
-1.30
0.20
-0.22
-2.44
0.015
-0.25
-2.75
0.007
-0.051
-2.07
0.039
-0.082
-3.26
0.002
-0.079
-3.33
0.001
-0.18
-5.04
<0.001
-0.15
-4.42
<0.001
-0.16
-4.79
<0.001
2.70
5.27
<0.001
2.54
4.90
<0.001
2.40
4.83
<0.001
0.54
3.32
0.001
0.66
6.59
<0.001
0.72
2.58
0.010
0.25
2.14
0.033
-0.095
-1.08
0.28
0.16
1.28
0.20
0.37
0.37
0.38

Transit Mode Shares


Results for taking transit to work are shown in Table 4.4. Among the level 1 variables, transit use is
positively related to the percentage of young adults in the county population. This relationship is as
expected, as young people are less likely to have private vehicles available to them. Transit use is
negatively related to the percentage of whites in the population, which may suggest a particular bias
against transit for this racial group. Transit use is negatively related to average household size but
positively related to household income, both at significant levels. The first relationship is as expected,
since large households can make more efficient use of automobiles through trip chaining and
carpooling. The second relationship is unexpected. Our initial assumption was that income and transit
were positively related due to suburban commuters in large metropolitan areas such as New York and
San Francisco commuting in large numbers by rail into the city. However, we found that the positive
relationship extends to commuting by bus. So we must assume that the relationship is due to some
confounding influence. Alternatively, it may be that after controlling for race, higher income households
actually consume more transit service.
Among the level 2 variables, transit use is positively related to metropolitan average fuel price,
metropolitan population, and heating degree days. The first relationship reflects the depressing effect
of higher fuel prices on auto use; there are persistent and significant differences in fuel price across
regions. The second relationship reflects both transportation supply and demand. Large metropolitan
areas provide much more transit service than do small ones and are also more congested, which

39

discourages auto use. The relationship to climate and weather is plausible. Cold climates and weather
may encourage transit use as an alternative to walking greater distances.
As for the variables of greatest interest to us, the original compactness index (indexo) in Model 1 has the
expected strong positive relationship to commuting by transit. Compact counties, with high values of
the index, have much higher transit mode shares than do sprawling counties. The relationship is elastic.
Every one percent increase in the county compactness index results in an almost 3 percent increase in
the transit mode share. The new compactness index (indexn) also has a positive relationship to transit
use, but a weaker one than that of the original index. The original index is dominated by density
variables, and density is widely viewed as the sine qua non for high transit use. By contrast, the new
index dilutes the effect of density by giving equal weight to mix, centering, and street accessibility.
The new multidimensional compactness factors have the expected signs and are significant in three
cases. The density factor, denfac, has the highest elasticity and highest significance level. Every percent
increase in the density factor is associated with a 2.81 percent increase in transit mode share. The
centering factor, cenfac, and street factor, strfac, are also significant with elasticities of 0.67 and 0.63,
respectively. Activities are concentrated in centers, and hence easily served by transit. Grid-like street
networks are more easily served by transit routes, and tend to shorten walking access distances. The
mix factor is not significant, which flies in the face conventional wisdom but is consistent with some
empirical research.
Table 4.4. Relationships to Transit Mode Share (log-log form with robust standard errors)

constant
hhsize
hhinc
pctwhite
pct1524
crime
metpop
hdd
fuel
indexo
indexn
denfac
mixfac
cenfac
strfac
pseudo-R2

Model 1 Model 1 Model 1 Model 2 Model 2 Model 2 Model 3 Model 3 Model 3


coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
-34.1
-12.6
<0.001
-28.4
-10.5
<0.001
-38.9
-13.0
<0.001
-2.42
-4.33
<0.001
-2.07
-3.59
0.001
-1.93
-3.32
0.001
1.18
4.92
<0.001
1.02
4.14
<0.001
1.12
4.48
<0.001
-0.99
-5.66
<0.001
-1.09
-6.45
<0.001
-0.91
-5.13
<0.001
1.24
7.47
<0.001
1.09
6.20
<0.001
1.11
6.43
<0.001
0.11
2.35
0.019
0.053
1.09
0.28
0.074
1.57
0.12
0.12
3.97
<0.001
0.18
5.68
<0.001
0.13
3.83
<0.001
0.27
3.39
0.001
0.18
2.48
0.014
0.20
2.66
0.009
6.42
8.29
<0.001
5.71
7.19
<0.001
5.55
6.86
0.002
2.92
11.6
<0.001
2.38
9.58
<0.001
2.81
4.38
<0.001
0.21
0.93
0.35
0.67
4.41
<0.001
0.63
2.57
0.011
0.26
0.26
0.27

40

Drive Times
Results for average drive times to work are shown in Table 4.5. Among the level 1 variables, average
drive time to work is positively related to household size, negatively related to the percentage of young
adults, and negatively related to household income. These relationships may be explained by household
life styles and activity patterns. Larger households are more likely to trip chain and carpool on the way
to work, and more likely to live in the suburbs where single-family housing is less expensive and schools
are better. Young people living on their own can live close to work since they do not have to
accommodate a working spouse or children. The negative relationship to household income is difficult
to explain. We would have expected that higher income households consume more residential space
and hence live farther from work, consistent with economic theory. This result may be due to higher
income households being able to afford housing closer to work, whereas lower income households have
to drive until they qualify for a mortgage. Or it may due to higher income households actually living
farther from work in distance, but traveling on suburban streets and highways at higher speeds with
resulting shorter travel times to work.
Among the level 2 variables, larger metropolitan areas (measured in terms of population) generate
longer commutes to work. This is consistent with the literature and common sense. The amount of
annual precipitation is positively related to average drive time, which seems reasonable since vehicle
speeds are lower under wet or icy conditions. The positive relationship to average fuel price is likely due
to some confounding factor (perhaps a regional effect).
Also consistent with much of the literature is the negative relationship between average drive time and
the original compactness index. Every percent increase in the index is associated with a 0.40 percent
decrease in drive time. The new compactness index is actually more significant than the original, and
has a higher elasticity value. The new compactness index incorporates mixed use and centering
variables, which should translate into better jobs-housing balance.
In the more fully specified model, three of four compactness factors have negative signs, though only
two are statistically significant. The strongest influence is mixed use factor, followed by the centering
factor. Activity centers, mixed uses, and fine street networks bring homes and work places closer
together. The odd man out is density, which has a positive sign and is statistically significant. While
higher density also brings homes closer to work places, it apparently creates enough congestion to
partially offset that beneficial effect. This finding is consistent with some previous research.
Table 4.5. Relationships to Average Drive Time (log-log form with robust standard errors)

constant
hhsize
hhinc
pct1524
metpop
precip
fuel
indexo
indexn

Model 1 Model 1 Model 1 Model 2 Model 2 Model 2 Model 3 Model 3 Model 3


coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
coeff.
t-ratio
p-value
3.81
11.85
<0.001
3.98
13.0
<0.001
2.82
7.74
<0.001
0.51
6.58
<0.001
0.34
4.47
<0.001
0.38
4.93
<0.001
-0.077
-2.526
0.012
-0.072
-2.38
0.018
0.005
0.018
0.86
-0.21
-7.33
<0.001
-0.15
-5.77
<0.001
-0.12
-5.82
<0.001
0.083
14.6
<0.001
0.080
15.1
<0.001
0.068
12.9
<0.001
0.061
3.59
0.001
0.042
2.57
0.011
0.040
2.64
0.009
0.24
1.98
0.048
0.44
3.58
0.001
0.55
4.75
<0.001
-0.29
-3.59
0.001
-0.35
-7.09
<0.001
41

denfac
mixfac
cenfac
strfac
pseudo-R2

Model 1
coeff.

Model 1
t-ratio

NA*

Model 1
p-value

Model 2
coeff.

Model 2
t-ratio

NA*

Model 2
p-value

Model 3 Model 3 Model 3


coeff.
t-ratio
p-value
0.128
2.20
0.028
-0.36
-10.9
<0.001
-0.036
-1.67
0.096
-0.029
-1.08
0.28
NA*

* Pseudo-R2s are negative. A pseudo-R2 is not entirely analogous to R2 in linear regression, which can
only assume positive values. One standard text on multilevel modeling notes that the variance can
increase when variables are added to the null model. It goes on to say: This is counter-intuitive,
because we have learned to expect that adding a variable with decrease the error variance, or at least
keep it at its current level In general, we suggest not setting too much store by the calculation of
[pseudo-R2s] (Kreft and de Leeuw 1998, 119). For more discussion of negative pseudo-R2s, also see
Snijders and Bosker (1999).

Discussion
All measures of compactness have now been validated against journey-to-work data from the 2010
American Community Survey (2006-2010 average). As in 2003, we hypothesized that county
compactness in 2010 would be directly related to the walk share of commute trips and the transit share
of commute trips (after controlling for relevant covariates). County compactness was expected to be
inversely related to average household vehicle ownership and average drive time to work. These
relationships are confirmed. For all outcomes, sprawling counties perform less well than compact ones.
These results tend to validate our compactness/sprawl indices. They also suggest that the new
compactness index has stronger and more significant relationships to walking and drive time than does
the original compactness index. Only transit use favors the original index.
The four factors representing individual dimensions of sprawl have, in most cases, the expected
relationships to transportation outcomes. They also have not only face and construct validity, but a
measure of internal validity.
The main limitation of this analysis has to do with the data it employs. By necessity, the study uses
highly aggregate data from a variety of sources that are not always consistent as to the area under study
and time period. They are simply the best data available from national sources with sufficient breadth
to provide a panoramic view of sprawl in the United States. Results will have to be validated through
follow-up work of a more focused nature.

42

Chapter 5. Traffic Safety


The literature is replete with studies showing that areas with more residents, more employment, and
more arterial lane miles experience more crashes (Levine et al. 1995a, 1995b; Hadayeghi et al. 2003;
Kmet, Brasher and Macarthur 2003; Ladron de Guevara et al. 2004; Hadayeghi et al. 2006; and
Lovegrove et al. 2006). Such studies may be useful for crash prediction on individual facilities. However,
they do not explain the crash rates on a per capita basis, only overall crash frequency on specific
facilities or specific small areas. Where there are more people and jobs, there tends to be more of
everything, from traffic to crime to coffee shops.
Given the direct relationship between vehicle miles traveled (VMT) and crash exposure, development
patterns that generate lower VMT should also have lower traffic crash rates. If the relationship between
VMT and traffic fatalities is near-linear, then sprawling communities, which are known to generate
higher VMT per capita, should also report higher rates of traffic crashes and fatalities (Ewing et al. 2002).
Sprawl is partly characterized by strip commercial uses and out-sized big box retail outlets. Dumbaugh
and Li (2011) examined many characteristics of the built environment and correlated them to the
number of collisions involving pedestrians, cyclists, and motorists. They found major crash determinants
include the total miles of arterial roadways and the presence of strip commercial uses and big box
stores. On the other hand, pedestrian-scaled retail uses were associated with lower crash rates. Each
additional mile of arterial thoroughfare was associated with a 9.3 percent increase in motoristpedestrian crashes, each additional strip commercial use was associated with a 3 percent increase in
vehicle-pedestrian crashes, and each big box store was associated with an 8.7 percent increase in
vehicle-pedestrian crashes.
Sprawl is also characterized by wide arterials, long blocks, and poor street connectivity. Marshall and
Garrick (2011) analyzed 230,000 crashes occurring over 11 years in 24 cities in California to determine
associations between crashes and street network characteristics, including street network density and
street connectivity. Increasing street connectivity - normally associated with street grids - led to an
increase in automobile crashes. The authors hypothesized that increased street connectivity leads to
increased traffic conflicts and hence more crashes. On the other hand, the severity of crashes, and
incidence of fatal crashes, was lower in downtown areas despite their grids. The authors argued a
decrease in fatal crashes is a result of lower vehicle speeds on streets in downtown areas
In their 2003 article on sprawl vs. traffic fatalities, Ewing et al. (2003b) found that for every one percent
increase in the county compactness index, all-mode traffic fatality rates fell by 1.49 percent and
pedestrian fatality rates fell by 1.47 percent, after adjusting for pedestrian exposure. In this study we
seek to replicate these results using data for 2010. However, unlike the earlier study, we employ
structural equation modeling, use different database and dependent variables, include additional
control variables, and test different compactness measures.
We would expect compact development patterns to produce fewer traffic accidents, injuries, and
fatalities due to reduced vehicle miles driven and possibly also due to lower speeds of vehicle travel.
We would also expect the relationship between sprawl and traffic fatalities to be more significant than
in the original study due the larger sample of counties and greater variance in sprawl measures among
these counties.

43

Methods
Data
There is no national source of crash data comparable to the Fatality Analysis Reporting System (FARS)
database of fatalities. Instead, each state, through its department of transportation or department of
public safety, maintains a comprehensive database of crashes that result in a vehicle being towed away,
personal injury, or fatalities. Individual states establish their own reporting thresholds.
To test the theory that sprawl generates more crashes of all types, not just fatal crashes, we sought
crash data from all 50 states. Crash data were obtained from all states via online databases or per an
email/phone request. The survey years ranged from 2008 to 2011 with the majority between 2010 and
2011. The individual state crash data were compiled into a national database that includes 6 million
crashes, nearly 1.8 million injury crashes, and nearly 30,000 fatal crashes.
Table 5.1. Crash Database
fips
1
2
4
5
6
8
9
10
12
13
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

all crashes
Alabama
Alaska
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
Florida
Georgia
Hawaii
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska

AL
AK
AZ
AR
CA
CO
CT
DE
FL
GA
HI
ID
IL
IN
IA
KS
KY
LA
ME
MD
MA
MI
MN
MS
MO
MT
NE

124,258
12,462
103,423
59,076
416,490
101,574
101,625
20,872
305,887
306,174
7,940
21,410
281,878
181,452
54,804
59,740
127,524
34,467
32,770
89,985
115,641
284,049
72,117
22,519
141,615
19,747
29,735

injury crashes
26,943
3,659
33,028
17,759
161,094
9,616
25,391
5,204
195,096
77,150
4,816
8,036
60,057
31,413
16,957
13,325
24,196
34,007
8,215
30,414
30,312
52,487
21,662
7,542
35,279
5,352
6,519

fatal crashes
814
51
754
509
2,520
406
299
97
2,441
1,342
108
163
835
726
386
354
670
460
154
457
333
834
334
424
716
192
172
44

fips
32
33
34
35
36
37
38
39
40
41
42
44
45
46
47
48
49
50
51
53
54
55
56

all crashes
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
Total

NV
NH
NJ
NM
NY
NC
ND
OH
OK
OR
PA
RI
SC
SD
TN
TX
UT
VT
VA
WA
WV
WI
WY

50,461
31,512
293,595
46,156
439,660
208,509
18,823
296,170
68,701
49,053
108,929
41,786
107,673
17,362
195,799
430,226
46,272
10,279
211,054
98,878
29,946
112,516
14,112
6,056,706

injury crashes
18,744
6,165
64,345
13,120
131,131
67,983
3,548
73,427
23,683
23,887
48,902
7,927
31,152
3,973
48,293
143,142
14,153
1,862
43,072
32,725
9,050
28,965
3,643
1,788,421

fatal crashes
220
39
573
319
1,097
1,122
130
941
462
310
1,191
56
700
101
903
2,818
217
57
644
422
166
515
135
29,689

Variables
In this structural equation modeling (SEM) effort, we have four endogenous variables. County crash
rates per 100,000 population were computed by dividing frequency counts by population in 100,000s
obtained from the 2010 U.S. Census. The all-mode crash rates include all crashes involving private
motor vehicles, buses, trains, taxis, bicycles, and pedestrians.
County VMT estimates were obtained from the Environmental Protection Agency. EPA has a process
that uses surrogates such as population, roadway miles, and economic modeling to develop allocation
factors for distributing the statewide total VMT to individual counties. Total VMT was divided by the
number of households in each county in 2010 to obtain VMT per household. VMT per household is also
treated as endogenous.
Exogenous variables came from various sources. From the 2010 Census, we downloaded data on
population, households, gender, age, and race, and computed average household size, percentage of
the population that is male, percentage of the population that is white, and percentage of the
45

population of working age. Median household income for each county came from the 2010 American
Community Survey (ACS), 5-year estimates. The 5-year estimates are based on data collected over a 60month period from 2006 through 2010.
We also downloaded data on weather/climate and fuel price. Annual heating degree days (HDD), cooling
degree days (CDD), and annual precipitation data came from the National Oceanic and Atmospheric
Administration (NOAA)s National Climatic Data Center for 2009 for HDD and 2010 for the other two
variables. We recorded data for stations located in the same MSAs as the counties. In a few cases there
were not weather stations in MSAs, so we used data from the closest station to a particular MSA.
Gasoline price data at the MSA level were purchased from the Oil Price Information Service (OPIS). Retail
prices are average prices from samples of stations in each MSA and are reported with all relevant taxes
included. These prices are the true posted sign prices (as they would appear outside a gas station). See
Table 5.2 for the list of variables used in this chapter.
Table 5.2. Variables (variables log transformed)
Endogenous Variables
Variables
crash
traffic crash rate per 100,000 population
injury
injury crash rate per 100,000 population
fatal
fatal crash rate per 100,000 population
VMT
annual vehicle miles traveled per household
Exogenous Variables
Variables
hhsize
pct1564
hhinc
white
male
fuel
precp
hdd
cdd
indexo
indexn

average household size


percentage of working age population (15-64 years old)
median household income
percentage of white population
percentage of male population
average fuel price for the metropolitan area
annual precipitation (inches)
heating degree-days
cooling degree-days
county compactness index for 2010 (using the same 2000
index variables)
county compactness index for 2010 (including additional
variables compared to 2000 index)

Data Sources
States, Census 2010
States, Census 2010
States, Census 2010
EPA 2011
Data Sources
Census 2010
Census 2010
ACS 2006-2010
Census 2010
Census 2010
OPIC database 2010
NOAA database 2010
NOAA database2010
NOAA database2010

Method of Analysis
Models were estimated with structural equation modeling or SEM. SEM is a statistical methodology for
evaluating complex hypotheses involving multiple, interacting variables (Bollen 1989; Grace 2006). SEM
46

is a model-centered methodology that seeks to evaluate theoretically justified models against data.
The SEM approach is based on the modern statistical view that theoretically based models, when they
can be justified on scientific grounds; provide more useful interpretations than conventional methods
that simply seek to reject the null hypothesis of no effect.
There are several related and distinctive features of SEM (Grace, 2006).

Hypothesized path models are evaluated based on a priori knowledge about the processes
under investigation using all available information.
The investigator tests the degree to which the structure of one or more models is consistent
with the structure inherent in the data. Many models that might be envisioned commonly are
rejected because they are inconsistent with the data.
Probability statements about the model are reversed from those associated with null
hypotheses. Probability values (p-values) used in statistics are measures of the degree to which
the data are unexpected, given the hypothesis being tested. In null hypothesis testing, a finding
of a p-value <0.05 indicates that we can reject the null hypothesis because the data are very
unlikely to come from a random process. In SEM, we seek a model that has a large p-value
(>0.05) because that indicates that the data are not unlikely given that model (that is, the data
are consistent with the model).
Different processes operating in systems are distinguished by decomposing relationships into
direct and indirect pathways. Pathways can, thus, be either simple or compound, depending on
whether they pass through other variables or not. The total effect of one factor on another is
the cumulative impact summed over all the pathways connecting the two factors.

The estimation of structural equation (SE) models involves solving a set of equations. There is an
equation for each response or endogenous variable in the network. Variables that are solely
predictors of other variables are termed influences or exogenous variables. Typically, solution
procedures for SE models focus on the observed versus model-implied correlations in the data. The
unstandardized correlations or co-variances are the raw material for the analyses. Models are
automatically compared to a saturated model (one that allows all variables to inter-correlate), and this
comparison allows the analysis to discover missing pathways and, thereby, reject inconsistent models.
In this analysis, data first were examined for frequency distributions and simple bivariate relationships,
especially for linearity. This suggested the need for data transformation. To equalize and stabilize
variances, improve linearity, and still allow ready interpretations, all variables were log transformed.

Results
We have estimated three SE models with Amos 19.0, a popular SEM software package with a good
graphic display. Maximum likelihood methods are used in the estimations. Model evaluation are based
on four factors: 1) theoretical soundness; 2) chi-square tests of absolute model fit; 3) root-mean-square
errors of approximation (RMSEA), which unlike the chi-square, correct for sample size; and 4)
comparative fit indices (CFI).
The path diagram in Figure 5.1 is copied directly from Amos. Causal pathways are represented by
straight uni-directional arrows. Correlations are represented by curved bi-directional arrows (to simplify
the already complex causal diagram, some correlations are omitted). By convention, circles represent
error terms in the model, of which there is one for each endogenous (response) variable.
47

The main goodness-of-fit measure used to choose among models is the chi-square statistic. Probability
statements about an SEM model are reversed from those associated with null hypotheses. Probability
values (p-values) used in statistics are measures of the degree to which the data are unexpected, given
the hypothesis being tested. In null hypothesis testing, a finding of a p-value <0.05 indicates that we can
reject the null hypothesis because the data are very unlikely to come from a random process. In SEM,
we seek a model with a small chi-square and large p-value (>0.05) because that indicates that the data
are not unlikely given that model (that is, the data are consistent with the model).
The first analysis simply validates the findings of earlier analysis (Ewing et al., 2003b). The main
endogenous variable is the natural logarithm of the number of fatal crashes per 100,000 population.
Another endogenous variable, the natural logarithm of VMT per household, is a mediating variable on
the causal pathways between the exogenous variables and the fatal crash rate. The exogenous variables
are as above.
Judged by its significant coefficients, low model chi-square, and sample-size adjusted fit (the RMSEA),
the first model fits the data well. The comparative fit index (CFI) value shows that the model explains
most of the total discrepancy in the data (> 99%).
Figure 5.1. Causal Path Diagram for Fatal Crashes in Terms of County Compactness, VMT, and Other
Variables

Direct effects are presented in Table 5.3. Most of the causal paths shown in the path diagram are
statistically significant (have non-zero values). The exceptions are a couple paths that are theoretically
significant, though not statistically significant. Income and average household size are directly related to
VMT, while the percentage of whites in the population, annual precipitation, and average fuel price are
48

inversely related to VMT. Larger households have more complex activity patterns than do smaller
households. Higher income households own more cars, and consume more land at lower residential
densities. Due to patterns of housing segregation, whites may live closer to work and other common
destinations. Rain and snow may discourage long distance travel. High fuel prices increase the
generalized cost of travel, thereby depressing travel. County compactness is, naturally, inversely related
to VMT because origins and destinations are closer together in a compact county.
VMT is directly and significantly related to the fatal crash rate, as one would expect. This relationship
completes indirect pathways between our exogenous variables and the fatal crash rate. Some
exogenous variables also have direct, significant relationships to the fatal crash rate. Income is
negatively related to the fatal crash rate, perhaps because higher income households drive more
crashworthy vehicles. The amount of precipitation is directly related to the fatal crash rate, as rainy and
snow conditions are known to be implicated in crashes. Finally, the compactness index is inversely
related to the fatal crash rate, even after controlling for VMT. One possible explanation is that dense
areas have lower travel speeds, which lead to less severe crashes. Considering both direct and indirect
effects, the original compactness index has a greater effect on the fatal crash rate than does the new
compactness index (see Tables 5.4 and 5.5). The original compactness index features density variables,
which may do more to depress vehicular travel and speed than do the other elements of the new sprawl
index.
Table 5.3. Direct Effects of Variables on One Another in the Fatal Crash Model (log-log form)

y
VMT
VMT
VMT
VMT
VMT
fatal
fatal
fatal
VMT
fatal
VMT
fatal

x
<--<--<--<--<--<--<--<--<--<--<--<---

hhsize
hhinc
white
precip
fuel
hhinc
precip
VMT
indexo
indexo
indexn
indexn

chi-square
RMSEA
CFI
a
Original Indices
b
Refined Indices

Model 1a
coeff.

Model 1a
std. err.

Model 1a Model 1a Model 2b Model 2b Model 2b Model 2b


critical
p-value
coeff.
std. err.
critical
p-value
ratio
ratio
0.56
0.13
4.31
<0.001
0.34
0.14
2.44
0.02
0.12
0.04
2.71
0.01
0.08
0.04
1.78
0.08
-0.21
0.05
-4.56
<0.001
-0.15
0.05
-3.34
<0.001
-0.03
0.03
-1.22
0.22
-0.07
0.03
-2.71
0.01
-0.89
0.20
-4.38
<0.001
-0.74
0.21
-3.48
<0.001
-0.95
0.07
-13.38
<0.001
-0.98
0.07
-14.21
<0.001
0.17
0.05
3.72
<0.001
0.12
0.05
2.65
0.01
0.61
0.06
10.27
<0.001
0.55
0.06
9.32
<0.001
-0.95
0.06
-17.05
<0.001
-0.91
0.10
-8.90
<0.001
-0.78
0.05
-15.84
<0.001
-0.95
0.09
-10.94
<0.001
a
b
Model 1
Model 2
10.5
5.45
degrees of freedom = 3
degrees of freedom = 3
p-value = 0.015
p-value = 0.142
0.053
0.03
p-value = 0.38
p-value = 0.75
0.995
0.998

49

Table 5.4. Direct, Indirect, and Total Effects of the Original Compactness Index and Other Variables on
the Fatal Crash Rate
direct effect
indirect effect
total effect

hhsize

0
0.345
0.345

hhinc
white
precip
fuel
VMT
indexo
-0.953
0
0.169
0
0.614
-0.914
0.073
-0.126
-0.019
-0.548
0
-0.581
-0.881
-0.126
0.15
-0.548
0.614
-1.495

Table 5.5. Direct, Indirect, and Total Effects of the New Compactness Index and Other Variables on the
Fatal Crash Rate
hhsize
direct effect
indirect effect
total effect

hhinc

0
0.185
0.185

white

-0.983
0.043
-0.94

0
-0.082
-0.082

precip
0.119
-0.039
0.08

fuel

VMT
0
-0.404
-0.404

indexn

0.549
0
0.549

-0.946
-0.429
-1.375

The next analysis tests whether sprawling areas have higher or lower total crash rates than do compact
areas. Compact areas generate lower VMT and hence less crash exposure than sprawling areas. At the
same time, compact areas may have more fender benders due stop-and-go driving, even as they have
fewer serious crashes due to lower travel speeds. So a priori, our only expectation is that the
relationship between sprawl and crash rates will be weaker for total crashes than for fatal crashes.
Direct relationships between compactness indices and total crashes are different than they were for
fatal crashes (see Table 5.6). The original compactness index is no longer significantly related to the
crash rate, and the new compactness index actually has its sign reversed, now having a positive direct
relationship to the crash rate. We can envision concentrations of activity with lots of stop-and-go traffic
causing many nonfatal crashes. The new compactness index represents, in addition to density, other
components of sprawl. Strong population and employment centers, in particular, seem to generate
more crashes. However, considering both direct and indirect effects, the original compactness index has
a strong negative relationship to the total crash rate due to the negative indirect effect through VMT,
while the new compactness index has a slight positive relationship (see Tables 5.7 and 5.8).
Table 5.6. Direct Effects of Variables on One Another in the Total Crash Model (log-log form)

y
VMT
VMT
VMT
VMT
VMT
crash
crash

x
<--<--<--<--<--<--<---

hhsize
hhinc
white
precip
fuel
VMT
hhsize

Model 1a
coeff.
0.56
0.12
-0.21
-0.03
-0.89
0.26
-1.25

Model 1a
std. err.
0.13
0.04
0.05
0.03
0.20
0.06
0.23

Model 1a Model 1a Model 2b Model 2b Model 2b Model 2b


critical
p-value
coeff.
std. err.
critical
p-value
ratio
ratio
4.31
<0.001
0.34
0.14
2.44
0.02
2.71
0.01
0.08
0.04
1.78
0.08
-4.56
<0.001
-0.15
0.05
-3.34
<0.001
-1.22
0.22
-0.07
0.03
-2.71
0.01
-4.38
<0.001
-0.74
0.21
-3.48
<0.001
4.19
<0.001
0.31
0.06
5.13
<0.001
-5.36
<0.001
-1.14
0.24
-4.86
<0.001
50

y
crash
crash
VMT
crash
VMT
crash

x
<--<--<--<--<--<---

hhinc
precip
indexo
indexo
indexn
indexn

Model 1a
coeff.

Model 1a
std. err.

0.03
0.09
-0.95
0.14

0.08
0.05
0.06
0.10

Model 1a Model 1a Model 2b Model 2b Model 2b Model 2b


critical
p-value
coeff.
std. err.
critical
p-value
ratio
ratio
0.33
0.74
0.01
0.08
0.08
0.94
1.90
0.06
0.11
0.05
2.37
0.018
-17.05
<0.001
1.38
0.17
-0.78
0.05
-15.84
<0.001
0.28
0.09
3.13
0.002

Model 1a
chi-square
RMSEA

degrees of freedom = 2
p-value = 0.96
>0.001
p-value = 0.996

CFI
a
Original Index
b
New Index

0.74

Model 2b
degrees of freedom = 2
p-value = 0.58
>0.001
p-value = 0.924

1.09

Table 5.7. Direct, Indirect, and Total Effects of the Original Compactness Index and Other Variables on
the Total Crash Rate
direct effect
indirect effect
total effect

hhsize
-1.25
0.145
-1.105

hhinc
0.025
0.031
0.056

white
0
-0.053
-0.053

precip
0.089
-0.008
0.081

fuel
0
-0.23
-0.23

VMT
0.258
0
0.258

indexo
0.143
-0.244
-0.101

Table 5.8. Direct, Indirect, and Total Effects of the New Compactness Index and Other Variables on the
Total Crash Rate
direct effect
indirect effect
total effect

hhsize
-1.142
0.105
-1.037

hhinc
0.006
0.024
0.031

white
0
-0.047
-0.047

precip
0.112
-0.022
0.09

fuel
0
-0.23
-0.23

VMT
0.312
0
0.312

indexn
0.281
-0.244
0.037

The final analysis may be the most surprising. Nonfatal injury crash rates were modeled using the same
set of exogenous and endogenous variables as above (see Figure 5.3). Compactness indices have
positive signs and, in the case of the new compactness index, a significant relationship at the 0.05 level.
The direct effects of compactness indices on the injury crash rate actually overwhelm the indirect effects
through VMT, and the net effect is positive for both compactness indices. Due to the large number of
such crashes in our sample, and the intuitively plausible results for fatal crashes, we do not view this
result as spurious. Apparently the inherently large number of traffic conflicts in compact areas (mostly
at intersections) result in more crashes of a serious nature but not so serious as to be fatal.
51

Table 5.9. Direct Effects of Variables on One Another in the Non-fatal Injury Crash Model (log-log form)

y
VMT
VMT
VMT
VMT
VMT
injury
injury
injury
injury
injury
VMT
injury
VMT
injury

x
<--<--<--<--<--<--<--<--<--<--<--<--<--<---

hhsize
hhinc
white
precip
fuel
hhsize
hhinc
white
precip
VMT
indexo
indexo
indexn
indexn

chi-square
RMSEA
CFI
a
Original Indices
b
Refined Indices

Model 1a
coeff.
0.56
0.12
-0.21
-0.03
-0.89
-0.60
-0.22
-0.34
0.35
0.24
-0.95
0.36

Model 1a
std. err.
0.13
0.04
0.05
0.03
0.20
0.19
0.06
0.07
0.04
0.05
0.06
0.09

Model 1a Model 1a Model 2b Model 2b Model 2b Model 2b


critical
p-value
coeff.
std. err.
critical
p-value
ratio
ratio
4.31
<0.001
0.34
0.14
2.44
0.015
2.71
0.007
0.08
0.04
1.78
0.075
-4.56
<0.001
-0.15
0.05
-3.34
<0.001
-1.22
0.223
-0.07
0.03
-2.71
0.007
-4.38
<0.001
-0.74
0.21
-3.48
<0.001
-3.20
0.001
-0.56
0.19
-2.85
0.004
-3.63
<0.001
-0.20
0.06
-3.28
0.001
-5.15
<0.001
-0.38
0.06
-5.94
<0.001
9.90
<0.001
0.36
0.04
9.92
<0.001
5.12
<0.001
0.22
0.05
4.60
<0.001
-17.05
<0.001
3.94
<0.001
-0.78
0.05
-15.84
<0.001
0.23
0.08
3.02
0.003

Model 1a
0.061
degrees of freedom = 1
p-value = 0.80
<0.001
p-value = 0.93
1

Model 2b
0.091
degrees of freedom = 1
p-value = 0.76
<0.001
p-value = 0.92
1

Table 5.10. Direct, Indirect, and Total Effects of the Original Compactness Index and Other Variables on
the Non-fatal Injury Crash Rate
direct effect
indirect effect
total effect

hhsize
hhinc
white
precip
fuel
VMT
indexo
-0.603
-0.223
-0.335
0.354
0
0.243
0.355
0.137
0.029
-0.05
-0.007
-0.217
0
-0.231
-0.466
-0.194
-0.385
0.347
-0.217
0.243
0.124

Table 5.11. Direct, Indirect, and Total Effects of the New Compactness Index and Other Variables on the
Non-fatal Injury Crash Rate
direct effect
indirect effect
total effect

hhsize
hhinc
white
precip
fuel
VMT
indexn
-0.555
-0.2
-0.376
0.364
0
0.217
0.23
0.073
0.017
-0.032
-0.015
-0.159
0
-0.169
-0.482
-0.183
-0.408
0.349
-0.159
0.217
0.061
52

Limitations
The new sprawl measures have been tested against traffic crash data. As in 2003, we hypothesized that
county compactness would be inversely related to the total traffic fatality rate. This is due to the lower
vehicle miles traveled in a compact environment, and also possibly to the lower speeds of travel.
We confirmed this hypothesis. This study suggests that sprawl is a significant risk factor for traffic
fatalities. The recognition of this relationship is key, as it adds traffic safety to the other health risks
associated with urban sprawl, namely, obesity and air and water pollution. However, the strong
relationships observed for fatal crashes do not extend to injury crashes or to total crashes, including
those with property damage.
This study has weaknesses. The study design is ecological in nature. It treats each county as a
homogenous unit, and assigns to it a single fatality rate and compactness index, even though there are
likely to be large differences within its borders. In certain circumstances, such as the study of behaviors
antecedent to injury, the analysis may need to extend down to the individual level.
We recognize that the fatality data studied are based on place of crash, while the other data are based
on place of residence, which may be different. To the extent that fatalities occur during the morning or
evening commute, a (reassuring) bias towards the null may exist. In other words, because most
commuters who cross county borders live in lower-density bedroom communities and work in higherdensity central areas, the traffic fatality rate in urban counties would be inflated relative to the
population living there. Using these databases, we could not determine the extent to which such bias, if
any, exists. One solution would be to study the relationship at the (multi-county) metropolitan area
level, but this would be at the expense of desired precision in the measurement of differences within
metropolitan areas.
Additional studies are needed to confirm these findings and extend our knowledge in key areas. An
exploration of the relationship between vehicle speed, fatality rates, and specific street design features
common to urban sprawl (e.g., wide, long streets) would help guide countermeasures.

53

Chapter 6. Public Health


Obesity is a significant health problem and the prevalence of adult obesity and overweight in the United
States has risen significantly in the last 30 years (Khan et al., 2009). The rising incidence of obesity
presents serious long-term challenges including decreased life expectancy and the potential for negative
impacts on an individuals quality of life, the availability and cost of future health care and the viability
and productivity of future generations. There are many influences on weight and health including
genetics, socioeconomic status, race/ethnicity, gender, access to food, levels of physical activity, and
neighborhood environments. The fundamental cause of obesity and overweight is an imbalance
between calories consumed and calories expended and the two main modifiable risk factors are
unhealthy diets and physical inactivity (Black, J. B. & Macinko, J., 2008; Trost, S., Kerr, Ward, & Pate,
2001; Trost, S. G., Owen, Bauman, Sallis, & Brown, 2002).
It is commonly recognized that even a moderate amount of physical activity can result in significant
health benefits, yet current research indicates that physical activity levels have declined with many
adults (43 percent) failing to meet the recommended physical activity requirements(Trogdon,
Finkelstein, Hylands, Dellea, & Kamal-Bahl, 2008). In the last 50 years activity levels have dropped for a
variety of reasons including new technologies and automation that make our lives easier, television and
computer use that lead to sedentary life styles, automobile use that substitutes for active travel, and
urban sprawl that all but guarantees automobile dependence (Committee on Physical Activity, 2005).
Research indicates that the built environment is associated with active travel and leisure time physical
activity in adults (Goldstein, 2011; Wong, Faulkner, & Buliung, 2011). Socioeconomic factors as well as
characteristics of the built environment may also restrict access to healthy foods. All are components of
an ecological approach to obesity research. This ecological approach to research is a common thread
throughout the body of research investigating the influences of dietary intake, physical activity,
socioeconomic, and demographic variables as they relate to obesity (Galvez, Meghan, & Yen, 2010).
In 2003, Ewing et al. (2003b) published a landmark study that established a relationship between health
behaviors, health outcomes, and a county sprawl index and became the most widely cited academic
paper in the Social Sciences as of late 2005, according to Essential Science Indicators(Reuters). After
controlling for age, education, fruit and vegetable consumption, and other sociodemographic and
behavioral covariates, they found that adults living in sprawling counties had higher body mass indices
(BMIs) and were more likely to be obese (BMI >30) than were their counterparts living in compact
counties.
In the years since this original study was published, there has been a plethora of research studies in both
planning and public health investigating the unique relationship between the built environment and
health outcomes. A recent meta study identified 36 literature reviews; 26 with physical activity as the
outcome variable, five focused on obesity, and an additional five covering both. Youth were the focus of
11 studies, seven focused on either adults or seniors, and 17 did not specify age groups (Ding & Gebel,
2012). Walking is the most common form of physical activity (Lee & Moudon, 2004). Availability and
access to recreational facilities, high residential densities, mixed-land uses, perceived safety, and
aesthetics are most often associated with physical activity (Durand, Andalib, Dunton, Wolch, & Pentz,
2011; Humpel, Owen, & Leslie, 2002; Kaczynski & Henderson, 2007; Mccormack et al., 2004; Owen,
Humpel, Leslie, Bauman, & Sallis, 2004; Trost, S. G., et al., 2002; Vancauwenberg et al., 2011).

54

Surveys of the built environment-obesity research include four specific to youth (Casey et al., 2011;
Dunton, Kaplan, Wolch, Jerrett, & Reynolds, 2009; Finkelstein, Fiebelkorn, & Wang, 2005; Galvez, et al.,
2010) and four related to adults (Feng, Glass, Curriero, Stewart, & Schwartz, 2010; Lachowycz & Jones,
2011; Papas et al., 2007; Withrow & Alter, 2010). An additional search of published research studies
(May 2011) has resulted in a total inventory of 82 articles that relate some dimension of the built
environment to obesity. Most studies have focused on adults (53), been cross sectional in research
design (73), and addressed the built environment at the neighborhood scale (65). One study included
built environmental variables at both the meso (neighborhood) and macro (county) scales (Joshu,
Boehmer, Brownson, & Ewing, 2008). This growing body of research suggests that some built
environments may be more obesogenic than others through influences on dietary intake and physical
activity (Black & Macinko, 2008).
Seventeen of the 20 studies published between 2003 and 2007 established statistically significant links
between some aspect of the built environment and obesity (Papas et al. 2007), and correlations
between sprawl and obesity have been affirmed (Committee on Physical Activity, Health,
Transportation, and Land Use 2005; Frank et al. 2004; Kelly-Schwartz et al. 2004; Lopez 2004; Sturm and
Cohen 2004; Cho et al. 2006; Doyle et al. 2006; Ewing et al. 2006; Rundle et al. 2007; Joshu et al. 2008).
As with the land use-transportation literature, the possibility of self-selection has been raised in this
literature as well. Two studies have garnered media attention by contending that residential selfselection, not environmental determinism, accounts for the relationship between sprawl and obesity
(Plantigna and Bernell 2007; Eid et al. 2008). Both conclude that people with higher body mass indices
choose to live in sprawling neighborhoods. One can understand indifference to the built environment
on the part of overweight individuals, but actual preference for places that preclude physical activity,
when these people dont plan to engage in it anyway, defies logic.

Methods
This chapter updates the work of Ewing et al. (2003b). We model multiple health outcomes and
behaviors in terms of the updated and refined compactness/sprawl measures presented above, plus
appropriate control variables.
Geographic Scale
This study represents the built environment at the county scale rather than the smaller neighborhood
scale. The main reason is expediency, since the health database used in this study, for reasons of
confidentiality, only supplies geocodes for respondents by county. For the same reason, the earliest
studies of the built environment and obesity also represented the built environment with county or
metropolitan metrics.
Since then, most studies have chosen instead to characterize the built environments of individuals at the
neighborhood scale, whether in terms of census tracts, block groups, or small buffers around individuals
homes. There has been an implicit assumption that walking distance from home is the operative scale
at which the built environment affects physical activity, food availability, and ultimately weight. This is
just an assumption. Adults spend most of their waking hours away from home. Something like 30 to 40
percent of all trips are made from one place to another other than home.

55

The average trip length in the U.S., according to the National Household Travel Survey of 2009, is almost
ten miles, which takes the traveler well beyond the bounds of the neighborhood. Even trips to buy
goods and services, the shortest trips, average about five miles. Judged by the length of trips for grocery
shopping and eating out, the relevant food environment for an American at least five miles from home.
The relevant recreational environment is at least ten miles from home.
A sprawling built environment produces long trips by automobile, which cuts into time available for
leisure activities. There may be simply less time available any kind of physical activity for Americans
living in sprawl.
To our knowledge, only one study has compared the power of neighborhood and county environments
as predictors of obesity (Joshua). While this study found that perceived neighborhood characteristics
were more important than objectively measured county characteristics, the jury is still out on the
appropriate scale for such research.
Data and Variables
Our health-related data come from the Behavioral Risk Factor Surveillance System (BRFSS), a telephone
survey conducted by state health departments and managed by the Centers for Disease Control and
Prevention (CDC). Over 350,000 adults are interviewed nationally each year to collect detailed
information on health risk behaviors, preventive health practices, and health care access primarily
related to chronic disease and injury.
We use a subsample of individuals for which county geocodes of residence are available for public use.
Our data come from the Selected Metropolitan/Micropolitan Area Risk Trends (SMART) project, which is
populated with BRFSS data for metropolitan and micropolitan statistical areas with 500 or more
respondents. We have included data for survey years 2007 through 2010 (see Table 6.1 for a list of
variables). Different questions are asked in different survey years (all four years for most variables but
only two years for some). This accounts for the different sample sizes for different variables.
Our health outcome variables fall into three categories: weight status, physical activity, and chronic
diseases. Weight status variables are calculated from self-reported height and weight. Body mass index
(BMI) is a continuous variable defined as weight in kilograms divided by height in meters squared
(kg/m2). Obesity status is dichotomous, defined as having a BMI greater than or equal to 30.0. Two
physical activity outcomes are dichotomous: whether a respondent reported any physical activity and
whether a respondent met recommended aerobic physical activity requirements (calculated). A third is
continuous: minutes of moderate physical activity per week, which includes the kind of walking we
expect to see in compact areas. Chronic disease outcomes are the status conditions of hypertension,
diabetes, and coronary heart disease as diagnosed by a health care professional and reported by the
respondent.
Gender, age, race/ethnicity, income, and educational attainment are included as control variables
representing individual-level sociodemographics. The reference groups for these variables are females,
white non-Hispanics, college graduates, persons aged 18 to 30 years, and income $75,000 or greater.
Smoking status and fruit and vegetable consumption are also included as control variables representing
individual health behaviors. The control variables have very predictable results both within and
between models. They are summarized in this report. Complete results are presented in Appendix C.
56

Individual respondents are nested within counties and hence share the characteristics of the county
built, social, and natural environments. Having individuals nested within counties, this data set violates
the independence assumption of OLS, and instead is analyzed using multi-level modeling and HLM 6.08.
Hierarchical linear modeling is applied to the continuous outcomes (BMI and minutes of moderate
physical activity per week), while hierarchical nonlinear modeling (logistic modeling) is applied to the
dichotomous outcomes (all other outcome variables)
The variables of particular interest are the compactness measures: original county compactness index,
new county compactness index, and county density, mix, centering, and street factors. The control
variables are precipitation, heating degrees, cooling degree days, and the violent crime rate, all for year
2010.
We introduce a new variable into this analysis, to test for significant relationships to physical activity:
percentage of land area of the county occupied by parks, excluding rural census tracts. Low densities,
large blocks, segregated land uses, and weak centers all translate into poor accessibility. To get at
another dimension of sprawl, lack of functional open space, we needed a different approach and data
source. The open space metric was produced using data created by Tele Atlas North America, Inc. and
provided under license to ESRI for inclusion on ESRI Data & Maps for use with ESRI software. The data
were created for 2010 using ESRI ArcMap 10.0. U.S. Parks is a SDC Feature Database containing
thousands of polygons representing national, state, county, regional, and local level parks. For our
current work, the data set was filtered to remove national and state parks as well as golf courses.
Census geography (2010) for U.S. counties was connected with the park dataset using the union tool
within ArcMap. As before, rural census tracts with fewer than 100 persons per square mile were
excluded. Multiple park polygons were then aggregated resulting in total park land area in square miles.
Table 6.1. BRFSS Variables and Sample Sizes
Level 1 Dependent Variables
Variables
bmi
body mass index
obese
obese status
anypa
any physical exercise
minmod
minutes of moderate physical activity per
week
bphigh
diagnosed high blood pressure
chd
diagnosed heart disease
diabetes
diagnosed diabetes
Level 1 Independent Variables
Variables
male
age3044
age4564
age6574
age75+
black
otherrace

male
age 30-44
age 45-64
age 65-74
age 75+
black non-Hispanic
other non-Hispanic

scale
continuous
dichotomous
dichotomous

n
675,784
676,524
709,099

continuous
dichotomous
dichotomous
dichotomous

159,965
354,826
703,942
709,234

scale
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous

n
709,889
709,889
709,889
709,889
709,889
701,572
701,394
57

Variables
Hispanic
Hispanic
lesshs
less than high school
hsgrad
high school graduate
somecoll
some college
inclt25
income <$25,000
inc2550
income $25,000-$50,000
inc5075
income $50,000-$75,000
smoke
current smoker
recfv
recommended servings fruits/vegetables
Level 2 Independent Variables
Variables
crime
precip
hhd
cdd
park
indexo
indexn
denfac
mixfac
cenfac
strfac

violent crime rate per 100,000 population


annual precipitation
annual heating degree days
annual cooling degree days
percentage park land (relative to total land
area)
county compactness index for 2010 (using
the same 2000 index variables)
county compactness index for 2010
(including additional variables compared to
2000 index)
density factor (a weighted combination of 4
density variables)
mix factor (a weighted combination of 3
mixed-use variables)
centering factor (a weighted combination of
3 centering variables)
street factor (a weighted combination of 4
street-related variables)

scale
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
dichotomous
scale
continuous
continuous

n
701,572
709,889
709,889
709,889
709,889
615,801
615,801
705,549
343,695
n

continuous
continuous

316
316
316
316

continuous

316

continuous

316

continuous

316

continuous

316

continuous

316

continuous

316

continuous

316

BMI and Obesity


BMI data are available for all four survey years. As reported in Appendix C, BMI is higher for males than
females and for black non-Hispanics and Hispanics than white non-Hispanics. BMI is lower for other nonHispanics than whites, the former being mostly Asian. BMI is higher for those with less education,
though the pattern is complicated. BMI increases with age until age 65, when it begins to decline. BMI
declines with income following a smooth curve. BMI is lower for those who smoke and those who meet
the recommended servings of five fruit or vegetables a day. BMI is lower in counties with more land
devoted to parks, and higher in counties with more heating degree days. In both cases, there is a logical
connection to physical activity.

58

As for the variables of greatest interest, those related to the built environment, BMI is strongly and
negatively related to the original and new compactness indices (see Table 6.2). The new compactness
index has a slightly stronger relationship to BMI and a slightly higher significance level. By either
measure, controlling for sociodemographic and behavioral covariates, residents of more compact
counties have a lower BMIs. Of the four individual compactness factors, three have the expected
negative signs and the fourth, the streets factor, has no relationship. The most significant influence is
the mix factor (significant at the 0.004 level), followed by the density factor (significant at the 0.023
level), and then the centering factor (which falls just short of significance at the 0.10 level).
Comparatively, compactness measures have much higher significance levels in this study than in the
2003 study (Ewing et al. 2003). This is not due to the sample size. There are actually fewer counties
represented in the 2010 sample. Perhaps the reason for the added significance is the limitation of this
sample to larger and more urban counties with a minimum of 500 respondents. The earlier sample
included some counties that are much smaller and less urban.
Table 6.2. Relationships to BMI (with robust standard errors)

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac
psuedo-R2

Model 1 Model 1 Model 1


coeff.
t-ratio
p-value
25.14
95.14
<0.001
-0.00897
-4.25
<0.001

Model 2 Model 2 Model 2


coeff.
t-ratio
p-value
25.13
101.01
<0.001
-0.00910

0.079

-4.50

Model 3
coeff.

Model 3
t-ratio

Model 3
p-value

-0.0066
-0.0117
-0.00085
0.0033

-1.98
-3.33
-0.48
1.43
0.080

0.049
0.001
0.633
0.153

<0.001

0.079

As with BMI, males, black non-Hispanics, and Hispanics have higher probabilities of being obese than
their reference categories (see Appendix C). The likelihood of obesity declines with income and
generally with education. Obesity is less prevalent among current smokers and those who consume the
recommended servings of fruits and vegetables. In all models, the likelihood of obesity is higher for
residents of counties with more heating degree days. In two models, prevalence of obesity is higher in
counties with more violent crime. In another, the prevalence of obesity is lower in counties with more
park space. There are logical connections to physical activity.
Controlling for covariates, the original and new compactness indices are strongly and negatively related
obesity, with the new index having a slightly stronger relationship (see Table 6.3). In the more fully
specified model, the density and mix factors are negatively related to obesity status, in that order of
significance. The centering and street factors are not significant.
Table 6.3. Relationships to Obesity (with robust standard errors)

Constant
indexo

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
-1.68
-19.58
<0.001
-0.0035
-4.75
<0.001

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
-1.66
-19.65
<0.001

Model 3
Coeff.

Model 3
T-ratio

59

Model 3
P-value

indexn
denfac
mixfac
cenfac
strfac

Model 1
Coeff.

Model 1
T-ratio

Model 1
P-value

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
-0.00401
-6.11
<0.001

Model 3
Coeff.
-0.0028
-0.0036
-0.00075
0.0013

0.50

psuedo-R2

0.57

Model 3
T-ratio

Model 3
P-value

-2.29
-2.53
-0.93
1.17
0.58

0.023
0.012
0.351
0.243

Physical Activity
We modeled two physical activity (PA) variables in this study (Tables 6.4 and 6.5). The first PA variable
indicates whether a respondent engaged in any physical activity. The question reads: During the past
month, other than your regular job, did you participate in any physical activities or exercises such as
running, calisthenics, golf, gardening, or walking for exercise? This question is included in all four years
of the BRFSS survey.
Males are more likely to be physically active than females. The likelihood of any physical activity
generally increases with education and income, and declines with age. The two compactness indices are
not significantly related to engaging in any physical activity. This result parallels that of the 2003 study
(Ewing et al. 2003b).
It is not clear from the phrasing of the BRFSS question whether active travel (walking or bicycling to a
destination) is covered and will be captured by this variable. Thus it is possible that respondents walk or
bike more in compact areas but simply do not report it as exercise.
Results are mixed for the four individual factors. The degree of centering and the mix factor have a
positive relationship to physical activity, while density has a negative relationship. Assuming these
results arent spurious, their countervailing effects may explain the lack of significance of the overall
compactness indices.
Table 6.4. Relationships to Any Physical Activity (with robust standard errors)

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac
psuedo-R2

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
2.80
33.41
<0.001
0.000012
0.021
0.988

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
2.71
32.33
<0.001
0.00079

1.41

Model 3
Coeff.

0.67

Model 3
P-value

-1.94
1.69
2.60
0.35
0.70

0.053
0.091
0.010
0.726

0.160
-0.0019
0.0017
0.0016
0.00033

0.67

Model 3
T-ratio

60

The second physical activity variable is a calculated value for minutes of moderate physical activity per
week. The 2003 study found that minutes of leisure-time walking were positively related to county
compactness (Ewing et al. 2003). This was the only physical activity variable with a significant
relationship to compactness. More recent surveys have not asked about specific physical activities such
as walking and bicycling, but instead have asked about moderate and vigorous physical activity
generally. If any relationship is likely to show up between compactness and physical activity, it will be in
minutes of moderate activity.
In BRFSS 2007 and 2009, respondents were asked if they engaged in moderate physical activities outside
work for at least ten minutes at a time. The specific question included in the 2007 and 2009 surveys
was:
We are interested in two types of physical activity: vigorous and moderate. Vigorous activities
cause large increases in breathing or heart rate while moderate activities cause small increases
in breathing or heart rate. Now, thinking about the moderate activities you do when you are not
working (if employed) in a usual week, do you do moderate activities for at least 10 minutes at a
time, such as brisk walking, bicycling, vacuuming, gardening, or anything else that causes some
increase in breathing or heart rate?
If the answer was yes, follow-up questions were asked about amounts of physical activity. Some of
the values provided were unrealistically high. Therefore, values were truncated at 1,260 minutes a
week, which represents three hours per day seven days a week and included 99 percent of all
respondents.
Results are similar to those for other PA variables, except that respondents with less education get more
moderate physical activity than those with college degrees. As one might expect, moderate physical
activity is negatively related to precipitation, heating degree days, and cooling degree days, these
environmental conditions apparently discouraging outside activities. Surprisingly, after controlling for
other variables, moderate physical activity appears to be negatively related to county compactness. We
can think of no obvious explanation for these results, but would note that very little of the variance in
our data set is explained by these models. In a Bayesian statistical sense, this one result does not negate
all contrary evidence.
Table 6.6. Relationships to Minutes of Moderate Physical Activity per Week (with robust standard
errors)

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac
psuedo-R2

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
268.34
32.35
<0.001
-0.14
-2.432
0.016

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
269.40
28.90
<0.001
-0.154

-2.460

Model 3
Coeff.

0.16

Model 3
P-value

0.54
-0.68
-0.69
-0.36
0.17

0.592
0.500
0.493
0.723

0.015
0.046
-0.090
-0.050
-0.042

0.16

Model 3
T-ratio

61

Chronic Diseases
Three health outcomes were modeled in this study: the diagnosed conditions of high blood pressure,
heart disease, and diabetes (Tables 6.7 through 6.9). These are known to be related to obesity and
physical inactivity, and the former at least is related to sprawl. However, these three chronic conditions
are downstream outcomes of obesity and physical inactivity, and highly dependent on diet and
heredity. Thus, a priori, we cannot say whether they will have a relationship to the urban sprawl.
Largely tracking the findings for obesity, the incidence of diabetes, heart disease, and high blood
pressure increases with age, declines with income, and declines with education. All three conditions
are more prevalent in males than females. The effect of race is mixed.
More interestingly, the two overall compactness indices are negatively associated with all three chronic
diseases. The new compactness index has slightly stronger relationships to these conditions than does
the original compactness index. The mix factor is the most significantly related of the individual factors
with a negative relationship to chronic disease.
Table 6.7. Relationships to High Blood Pressure (with robust standard errors)

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac
psuedo-R2

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
-2.87
-37.11
<0.001
-0.0018
-3.15
0.002

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
-2.83
-37.14
<0.001
-0.0021

-3.62

Model 3
Coeff.

Model 3
P-value

-0.73
-1.25
-0.74
-0.53
0.57

0.463
0.214
0.462
0.595

Model 3
T-ratio

Model 3
P-value

0.001
-0.00068
-0.0011
-0.00044
-0.00047

0.54

Model 3
T-ratio

0.55

Table 6.8. Relationships to Coronary Heart Disease (with robust standard errors)

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac
psuedo-R2

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
-6.01
-20.86
<0.001
-0.0024
-1.79
0.073

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
-5.96
-20.73
<0.001
-0.0028

-2.12

Model 3
Coeff.

0.034
-0.0014
-0.0058
-0.0013
0.0025

0.51

0.58

-0.67
-3.07
-1.35
1.61
0.61

Table 6.9. Relationships to Diabetes (with robust standard errors)


62

0.504
0.003
0.179
0.107

Constant
indexo
indexn
denfac
mixfac
cenfac
strfac

Model 1 Model 1 Model 1


Coeff.
T-ratio
P-value
-5.18
-40.74
<0.001
-0.0015
-2.22
0.034

psuedo-R2

Model 2 Model 2 Model 2


Coeff.
T-ratio
P-value
-5.17
-40.10
<0.001
-0.0016

-2.27

Model 3
Coeff.

Model 3
P-value

-0.73
-2.46
-0.36
0.97
0.60

0.464
0.015
0.714
0.331

0.024
-0.00087
-0.0033
-0.00024
0.0011

0.57

Model 3
T-ratio

0.58

Limitations
The new sprawl measures have been tested against physical activity, obesity, and chronic disease data
from BRFSS. As in 2003, we hypothesized that county compactness would be inversely related to
moderate physical activity, BMI, obesity, and hypertension, and unrelated to overall physical activity,
diabetes, and coronary heart disease (after controlling for relevant covariates). With the exception of
physical activity, these relationships have been confirmed.
This analysis is subject to important limitations that call for additional research:

Because this study is ecological and cross sectional in nature, we cannot say that sprawl causes
obesity, high blood pressure, or any other health condition. Our study simply indicates that
sprawl is associated with these conditions. Future research using longitudinal data is needed to
tackle the more difficult job of testing for causality.
The presumptive relationships between sprawl and health are multiple and complex. In
particular, leisure time physical activity constitutes only one of four major sources of physical
activity. Greater precision in characterizing physical activity will help disentangle the effects of
sprawl on health.
We recognize that the relationships between sprawl and behavior or weight are probably not
completely linear. It may be that certain thresholds or critical levels of "compactness" are
needed before community design begins to have a palpable influence on physical activity
increasing density from one or two houses per acre to three or four may not meet the threshold
needed for change.
This study relates health to the built environment at the county scale, which is large compared
to the living and working environments of most residents. Geocodes are only available from
BRFSS down to the county level. If environmental effects are felt most strongly at the
community or neighborhood scale, these results may understate the effects of the built
environment on health.
Because they are not directly measured in any of the compactness/sprawl measures, this study
does not account for many other environmental variables that may act directly or interact to
influence physical activity and hence health such as sidewalks and topography.
People who have specific residential preferences may self-select to live in walkable or,
alternatively, auto-dominant environments rather than environments themselves affecting
63

physical activity and hence weight. The direction of causality cannot be determined from a cross
sectional database like that used in this chapter. However, the weight of evidence from the built
environment-travel literature indicates that the environment affects travel choices independent
of residential preferences, and that environmental effects are stronger than self-selection
effects (Ewing and Cervero, 2010).
By focusing on physical activity, this study largely ignores the other side of the energy equation,
calories consumed as opposed to calories expended. Only our fruit and vegetable consumption
variable begins to get at that dimension of the problem. Caloric intake may have a spatial
component. Future research could, for example, relate the density of fast food restaurants and
availability of food choices to diet and obesity.

64

Chapter 7. Derivation and Validation of Metropolitan Sprawl Indices


In 1958 William Whyte in his book The Exploding Metropolis referred to a new notion in planning,
Suburban Sprawl, and alerted Americans that their cities were becoming more sprawling. This began
the debate over sprawl and its impacts. There is still little agreement on the definition of sprawl or its
alternatives: compact development, pedestrian-friendly design, transit-oriented development, and the
catch-all term smart growth. There is also little consensus about how sprawl impacts everything from
housing affordability and traffic congestion to open space preservation and air quality (Burchell,1998
and 2002, Bruegmann, 2006; Duany et al., 2001; Ewing, 1997 and 2002; Gordon and Richardson, 1997;
Kahn, 2006; Hayden, 2004; Hirschhorn, 2005). Duany et al. (2001) use cultural, aesthetic and ecological
reasons to reject suburban sprawl as human habitat. At the other end of the spectrum, Bruegmann
(2006) describes suburban sprawl as a natural manifestation of the American Dream of a big house in
the suburbs.
In all of these classic writings, it is clear that sprawl is ordinarily conceptualized at the metropolitan
level, encompassing cities and their suburbs. When we say Atlanta sprawls badly, we are probably
referring to metropolitan Atlanta, not the city of Atlanta or Fulton County. The focus up to this point in
the report has been on counties, because counties are typically smaller than metropolitan areas and
more homogeneous than metropolitan areas. They more closely correspond to the environment in
which individuals live, work, and play on a daily basis, and hence are affected by the built environment.
But certain phenomena are manifested at the regional or metropolitan level, such as ozone levels and
racial segregation. So in this chapter we derive metropolitan sprawl indices.

Methods
Sample
The unit of analysis in this study is the metropolitan area. A metropolitan area is a region that consists of
a densely populated urban core and its less-populated surrounding territories that are economically and
socially linked to it. The criteria of defining metropolitan areas changed in 2003. Smaller MSAs remained
the same, but larger metropolitan areas, previously referred to as consolidated metropolitan statistical
areas (CMSAs) are now defined as MSAs. Different portions of CMSAs, previously referred to as primary
metropolitan statistical areas (PMSAs), have been redefined and reconfigured as metropolitan divisions.
For example, the old New York CMSA consisted of eleven counties in two states and four PMSAs: New
York PMSA, Nassau-Suffolk PMSA, Dutchess County PMSA and Newburgh, NY-PA PMSA. The current
New York MSA consists of twenty-three counties in three states and four metropolitan divisions. The
New York MSA now is strikingly heterogeneous, whereas the old New York PMSA contained only the five
boroughs that make up New York City. Metropolitan divisions do not perfectly substitute for PMSAs, as
they have different size thresholds (2.5 million vs. 1 million population), but they come as close to
representing homogenous units as we can come with current census geography. Metropolitan divisions
are designated for each of the eleven largest MSAs.7

The metropolitan divisions, as components of MSAs, somewhat resemble PMSAs under the old system. However,
PMSAs were much more common. The higher population threshold for establishing metropolitan divisions (at least
2.5 million), opposed to the threshold of at least 1 million to establish PMSAs, means that the new system contains
twenty-nine metropolitan divisions within eleven MSAs, compared to seventy-three PMSAs within eighteen CMSAs
under the old system.

65

The sample in this study is limited to medium and large metropolitan areas, and metropolitan divisions
where they are defined. It initially included a total of 228 areas with more than 200,000 population in
2010. The rationale for thus limiting our sample is simple: the concept of sprawl has particular
relevance to large areas where the economic, social, and environmental consequences of sprawl can be
significant. The concept of sprawl does not have much relevance to small MSAs such as Lewiston, ID and
Casper, WY.
Parenthetically, a total of seven metropolitan areas and divisions were ultimately dropped from our
sample due to the lack of local employment dynamics (LED) data, a key data source for measuring
sprawl. These metropolitan areas, or a portion of them, are located in Massachusetts, which does not
participate in the LED program. This reduces the final sample size to 221 MSAs and metropolitan
divisions.
Variables
Development Density
Low residential density is on everyones list of sprawl indicators. Our first five density variables are the
same as in the original sprawl index (Ewing et al., 2002): gross density of urban and suburban census
tracts (popden), percentage of the population living at low suburban densities (lt1500), percentage of
the population living at medium to high urban densities (gt12500), and urban density based on the
National Land Cover Database (urbden). These variables are measured the same way for metropolitan
areas as for counties (see Chapter 2).
A fifth variable is the estimated density at the center of the metropolitan area derived from a negative
exponential density function (dgcent). The function assumes the form:
where:

Di = Do exp (-b di).


Di = the density of census tract i
Do = the estimated density at the center of the metropolitan area
b = the estimated density gradient or rate of decline of density with distance
di = the distance of the census tract from the center of the principal city

The higher the central density, and the steeper the density function, the more compact the
metropolitan area (in a monocentric sense). 8
The sixth density variable, which is new, is analogous to the first, except it is derived with employment
data from the Local Employment Dynamics (LED) database (empden). The LED data were aggregated
from census block geography to generate total jobs by 2-digit NAICS code for every block group in the
nation. This was then divided by land area to produce a density measure.
8

The function was estimated as follows. The principal cities of the metro areas were identified as the first-named
cities in the 1990 definitions of those areas. Their centers were determined by locating central business district
tracts within the principal cities as specified in the 1980 STF3 file. 1980 designations were adopted because central
business districts have not been designated since then. The means of the latitudes and longitudes of the centroids
of those central business district tracts were taken as the metropolitan centers. The distances from the centers to
all tracts were calculated using an ArcGIS. Finally, a negative exponential density function was fit to the resulting
data points to estimate the intercept and density gradient.

66

The last two variables are related to employment centers identified by the authors as a part of this
study. For more information on how the centers were identified for MSAs see Activity Centering in
Chapter 3.The two variables are weighted average population density (popdcen) and weighted average
employment density (empdcen) of all centers within a metropolitan area. The average densities were
weighted by the sum of block group jobs and residents as a percentage of the MSA total.
Land Use Mix
Segregated land uses are also on most lists of sprawl development patterns. Conversely, mixed and
integrated land uses sit atop lists of pedestrian-friendly, transit-oriented, and smart growth patterns.
The first two variables of mix use factor represent the balance between jobs and population (jobpop);
and the diversity of land uses (entropy). Although using the same variables as Ewing et al. (2002) to
operationalize mixed use, we computed them differently using one mile buffers around the centers of
block groups rather than computing them within the boundaries of block groups. The reason is that the
latter are sensitive to the size of a block group. The larger the area, the higher the value of mixed use
variables because the block group will contain more activity in total. By using a uniform one-mile buffer,
we make the unit of analysis comparable for all block groups.
The two mixed-use variables were calculated for each block groups buffer using block-level population
data from the 2010 Census, and block-level employment data from the 2010 LED database. The first
variable is a job-population balance measure (jobpop). This variable equals 1 for block groups with the
same ratio of jobs-to-residents within the one-mile ring as the metropolitan area as a whole; 0 for block
groups with only jobs or residents within the one-mile ring, not both; and intermediate values for
intermediate cases. All values were weighted by the sum of block group jobs and residents as a
percentage of the MSA total. 9
We also derived a job mix variable (jobmix). The variable, an entropy measure, equals 1 for block groups
with equal numbers of jobs in each sector; 0 for block groups with all jobs in a single sector within the
ring; and intermediate values for intermediate cases. The sectors considered in this case were retail,
entertainment, health, education, and personal services. Values were weighted by the sum of block
group population and employment as a percentage of the MSA total.
A third mixed-use variable is metropolitan weighted average Walk Score (walkscore). It was computed
using data from Walk Score, Inc. to measure proximity to amenities, with different amenities weighted
differently and amenities discounted as the distance to them increases up to one mile and a half, where
they are assumed to be no longer accessible on foot.10 Classic Walk Score data were acquired for all
urban census tracts in the United States. Values were weighted by the sum of census tract population
and employment as a percentage of the MSA total.
Activity Centering
Commercial strip development is on most lists of sprawl development patterns. The antithesis of strips
are urban centers. Urban centers are concentrations of activity that provide economies of scale,
facilitate modal choices and multipurpose trip making, foster an identity in the urban landscape, and are
9

See land use mix section for the formula used for computing job-population balance and job mix measures.
A grocery store, for example, gets three times the weight of a book score. The distance decay function starts with
a value of 100 and decays to 75 percent at a half mile, 12.5 percent at one mile, and zero at 1.5 miles.
10

67

vastly different than commercial strips. This concentration, or centeredness, can relate to population or
employment, and may take the form of a single dominant center or multiple sub-centers. Compactness
is associated with centers of all types, and sprawl with the lack of centers of any type.
Centering is the compactness dimension with the most significant improvement compared to Ewing et
al.s indices. Ewing et al. (2002) measured metropolitan centering in terms of concentrations of
employment in or around (within three miles) historic central business districts (CBDs) of metropolitan
areas, and at a considerable distance from (more than 10 miles) from historic CBDs. This way of
measuring centering does not make much sense when applied to medium size MSAs because most of
the jobs and population fall within three miles of CBDs. It also doesnt make sense in large polycentric
metropolitan areas, where historic CBDs have sometimes been eclipsed by edge cities.
The first centering variable came straight out of Ewing et al. (2002) and the 2010 census. It is the
coefficient of variation in census block group population densities, defined as the standard deviation of
block group densities divided by the average density of block groups (varpop). The more variation in
population densities around the mean, the more centering and/or subcentering exists within the MSA.
The second centering variable is analogous to the first, except it is derived with employment data from
the LED database. It is the coefficient of variation in census block group employment densities, defined
as the standard deviation of block group densities divided by the average density of block groups
(varemp). The more variation in employment densities around the mean, the more centering and/or
subcentering exists within the MSAs.
The third variable contributing to the centering factor is the density gradient moving outward from the
CBD, estimated with a negative exponential density function. The faster density declines with distance
from the center, the more centered (in a monocentric sense) the metropolitan area will be (dgrad).
The next two centering variables measure the proportion of employment and population within CBDs
and employment sub-centers. For computing them, we first identified the location of CBDs and
employment sub-centers for all metropolitan areas (see Activity Centering section on Chapter 3).
Out of 374 metropolitan areas in the U.S., we found 224 MSAs to be monocentric (have only one
center), 132 to be polycentric (have more than one center), and 18 metropolitan areas to have no CBD
or sub-center. This procedure resulted in two new centering variables as the percentage of MSA
population (popcen) and employment (empcen) in CBDs and sub-centers.
Street Accessibility
Street accessibility is related to block size since smaller blocks translate into shorter and more direct
routes. Large block sizes indicate a lack of street connections and alternate routes. So, three street
accessibility variables were computed for each MSA based on blocks size: average block length
(avgblklngh), average block size (avgblksze) and the percentage of blocks that are less than 1/100 square
mile, which is the typical size of an urban block (smlblk).
These three variables were part of Ewing et al.s original sprawl metrics. To them, we have added two
new variables. They are intersection density and percentage of 4-or-more way intersections.
Intersections are where street connections are made and cars must stop to allow pedestrians to cross.
The higher the intersection density, the more walkable the city (Jacobs, 1993). Intersection density has

68

become the most common metric in studies of built environmental impacts on individual travel behavior
(Ewing and Cervero, 2010).
Another common metric in such studies is the percentage of 4-or-more-way intersections (Ewing and
Cervero, 2010). This metric provides the purest measure of street connectivity, as 4-way intersections
provide more routing options than 3-way intersections. A high percentage of 4-way intersections does
not guarantee walkability, as streets may connect at 4-way intersections in a super grid of arterials. But
it does guarantee routing options.
For each MSA, the total number of intersections in the urbanized portion of MSA was divided by the
land area to obtain intersection density (intden), while the number of 4-or-more-way intersections was
multiplied by 100 and divided by the total number of intersections to obtain the percentage of 4-ormore way intersections (4way).

Individual Compactness/Sprawl Factors


For each dimension of sprawl, we ran principal component analysis on the measured variables, and the
principal component that captured the largest share of common variance among the measured variables
was selected to represent that dimension. Factor loadings (the correlation between a variable and a
principal component), eigenvalues (the explanatory power of a single principal component), and
percentages of explained variance are shown in Table 7.1.
The eigenvalue of the density factor is 5.82, which indicates that this one factor accounts for about
three quarters of the total variance in the dataset. As anticipated, the percentage of the population
living at less than 1,500 persons per square mile loads negatively on the density factor. The rest load
positively.
The eigenvalue for the mix factor is 2.30, which indicates that this one factor accounts for more than
three quarter of the total variance in the dataset. All component variables load positively on the mix
factor.
The eigenvalue of the centering factor is 1.90, which indicates that this factor accounts for about 38% of
the total variance in the datasets. The density gradient loads negatively on centering factor as expected.
The rest load positively.
The eigenvalue of the street factor is 2.51, which indicates that this factor accounts for more than a half
of the total variance in the dataset. As expected, the average block size and average block length load
negatively on the street accessibility factor. The rest load positively.
Table 7.1: Variable Loadings of Four Factors for 2010
Component Matrix
Density Factor
popden
empden
lt1500

gross population density


gross employment density
percentage of the population living at low suburban
densities

Data Sources

Census 2010
LED 2010
Census 2010

Factor
Loadings
0.900
0.898
-0.597

69

Component Matrix
gt12500
urbden
dgcent

percentage of the population living at medium to high


urban densities
net population density of urban lands
estimated density at the center of the metro area
derived from a negative exponential density function
weighted average population density of centers
weighted average employment density of centers

popdcen
empdcen
Eigenvalue
Explained variance

Component Matrix
Mix use Factor
jobpop
job-population balance
jobmix
degree of job mixing (entropy)
walkscore
weighted average Walk Score
Eigenvalue
Explained variance
Component Matrix

Data Sources

Factor
Loadings

Census 2010

0.879

NLCD
Census 2010,
Tiger 2010
Census 2010
LED 2010

0.925
0.948

Data Sources

LED 2010
LED 2010
Walk Score Inc.

Data Sources

0.810
0.817
5.82
72.80%
Factor
Loadings
0.834
0.921
0.870
2.30
76.72%
Factor
Loadings

Centering Factor

varpop
varemp
dgrad
popcen
empcen

coefficient of variation in census block group


population densities
coefficient of variation in census block group
employment densities
density gradient moving outward from the CBD
percentage of MSA population in CBD or sub-centers
percentage of MSA employment in CBD or subcenters

Census 2010

0.495

LED 2010

0.313

Census 2010,
Tiger 2010
Census 2010
LED 2010

Eigenvalue
Explained variance

-0.375
0.833
0.847
1.90
37.89%

Component Matrix

Data Sources

Factor
Loadings

Census 2010
Census 2010
NAVTEQ 2012
TomTom 2007
TomTom 2007

0.871
-0.804
-0.649
0.729
0.380
2.51
50.03%

Street Factor

smlblk
percentage of small urban blocks
avgblksze
average block size
avgblklng
average block length
intden
intersection density
4way
percentage of 4-or-more-way intersections
Eigenvalue
Explained variance

70

Overall Compactness/Sprawl Index for 2010


Although density has received more attention as a dimension of sprawl than have other factors, similar
to Ewing et al. (2002) we could think of no rationale for giving different weights to the four factors. All
four factors affect the accessibility or inaccessibility of development patterns. Each factor can move a
MSA along the continuum from sprawl to compact development. Thus the four were simply summed, in
effect giving each dimension of sprawl equal weight in the overall index.
The second and more difficult issue was whether to, and how to, adjust the resulting sprawl index for
MSA size. As areas grow, so do their labor and real estate markets, and their land prices. Their density
gradients accordingly shift upward, and other measures of compactness (intersection density, for
example) follow suit. The simple correlation between the sum of the four sprawl factors and the
population of the MSA is 0.575, significant at .001 probability level. Thus, the largest urbanized areas,
perceived as the most sprawling by the public, actually appear less sprawling than smaller urbanized
areas when sprawl is measured strictly in terms of the four factors, with no consideration given to area
size.
We used the same methodology as Ewing et al (2002) to account for metropolitan area size. We
regressed the sum of the four sprawl factors on the natural logarithm of the population of the MSAs.
The standardized residuals became the overall measure of sprawl. As such, this index is uncorrelated
with population. However, the overall index still has a high correlation (r=0.866) with the sum of four
factors before adjustment.
We transformed the overall sprawl index into a metric with a mean of 100 and a standard deviation of
25 for ease of use and understanding. More compact metropolitans have index values above 100, while
the more sprawling have values below 100. Table 7.2 presents overall compactness scores and
individual component scores for the 10 most compact and the 10 most sprawling large metropolitan
areas. By these metrics, New York and San Francisco are the most compact large metropolitan divisions
(see Figures 7.1a&b), while Hickory, NC and Atlanta, GA are the most sprawling metropolitan areas (see
Figure 7.2a&b). These figures are at the same scale, and is clear that the urban footprints of the former
are more concentrated than those of the latter. Again all metropolitan areas and divisions in
Massachusetts, including the Boston metropolitan division, are not in the list due to the lack of available
employment data (LED) for this state.
Table 7.2. Compactness/Sprawl Scores for 10 Most Compact and 10 Most Sprawling metropolitan areas
and divisions in 2010
Ten Most Compact Metropolitan Areas
Rank
1
2
3
4
5
6
7
8

index
New York-White Plains-Wayne, NY-NJ Metro
Division
San Francisco-San Mateo-Redwood City, CA
Metro Division
Atlantic City-Hammonton, NJ Metro Area
Santa Barbara-Santa Maria-Goleta, CA Metro
Area
Champaign-Urbana, IL Metro Area
Santa Cruz-Watsonville, CA Metro Area
Trenton-Ewing, NJ Metro Area
Miami-Miami Beach-Kendall, FL Metro Division

denfac

mixfac

cenfac

strfac

203.4

384.3

159.3

213.5

193.8

194.3

185.9

167.2

230.9

162.8

150.4

112.3

148.9

109.5

122.1

146.6

100.8

93.7

137.3

94.1

145.2
145.0
144.7
144.1

160.2
96.3
98.9
100.0

136.4
100.1
146.2
123.3

117.9
154.5
107.9
153.6

166.9
130.7
112.2
82.8

71

Rank
9
10

Springfield, IL Metro Area


Santa Ana-Anaheim-Irvine, CA Metro Division

Ten Most Sprawling Metropolitan Areas


Rank
212
213
214
215
216
217
218
219
220
221

Kingsport-Bristol-Bristol, TN-VA Metro Area


Augusta-Richmond County, GA-SC Metro Area
Greenville-Mauldin-Easley, SC Metro Area
Riverside-San Bernardino-Ontario, CA Metro
Area
Baton Rouge, LA Metro Area
Nashville-Davidson--Murfreesboro--Franklin, TN
Metro Area
Prescott, AZ Metro Area
Clarksville, TN-KY Metro Area
Atlanta-Sandy Springs-Marietta, GA Metro Area
Hickory-Lenoir-Morganton, NC Metro Area

index
142.2
139.9

denfac
142.1
104.8

mixfac
105.0
117.8

cenfac
136.4
96.1

strfac
114.3
149.9

index
60.0
59.2
59.0

denfac
85.2
88.1
91.1

mixfac
60.7
60.6
71.7

cenfac
88.5
100.8
72.6

strfac
73.9
82.5
71.8

56.2

97.9

110.3

70.5

96.2

55.6

88.2

80.6

84.9

70.7

51.7

91.3

72.0

69.7

80.4

49.0
41.5
41.0
24.9

84.5
86.7
97.8
78.6

39.7
72.9
85.5
40.5

74.5
81.1
89.9
67.0

60.8
71.4
75.9
56.9

Figure 7.1. Most Compact Metropolitan Areas (New York and San Francisco)

Figure 7.2. Most Sprawling Metropolitan Areas (Atlanta and Hickory, NC)

72

Relationship to Transportation Outcomes


We used data from the 2011 American Community Survey (ACS), 5-year estimates to validate our sprawl
metrics. First, we computed average vehicle ownership per household, walk and transit mode shares,
and average drive times for census metropolitan areas and divisions. Then, from the 2010 Census, we
downloaded data on gender, age, race, household size, and computed percentages and mean values to
describe socio-economic status. These are control variables. We acquired gasoline price data at the MSA
level from the Oil Price Information Service (OPIS). Retail prices are average prices from samples of
stations in each MSA and are reported with all relevant taxes included. These prices are the true posted
sign prices (as they would appear outside a gas station). Table 7.3 shows a list of all dependent and
independent variables used in the validation. Average gasoline price is another control variable.
Table 7.3. Variables Used to Explain Transportation Outcomes (all variables log transformed)

Dependent variables
Variables
walkshr
transitshr
hhveh
drivetime

percentage of commuters walking to work


percentage of commuters using public transportation
Average vehicle ownership per household
average journey-to-work drive time in minutes

Data Sources
ACS 2007-2011
ACS 2007-2011
ACS 2007-2011
ACS 2007-2011

metropolitan population
average household size
percentage of population 15-24 years old
percentage of male population
percentage of white population
income per capita
average fuel price for the metropolitan area
MSA compactness index for 2010
density factor (a weighted combination of 8 density variables)
mix factor (a weighted combination of 4 mixed-use variables)
centering factor (a weighted combination of 5 centering variables)
street factor (a weighted combination of 6 street-related variables)

Data Sources
Census 2010
Census 2010
Census 2010
Census 2010
Census 2010
ACS 2007-2011
OPIC database 2010
computed
computed
computed
computed
computed

Independent variables
Variables
pop
hhsize
age1524
male
white
income
fuel
index
denfac
mixfac
cenfac
strfac

We estimated two sets of regression models. Our first set of regressions used the overall compactness
index for 2010 (index) as an independent variable. The second set of regressions used the four
compactness/sprawl factors individually (denfac, mixfac, cenfac, and strfac) as independent variables. In
both set of regressions, the dependent variables were logged so as to be normally distributed and all
independent variables were also transformed into log form to achieve a better fit with the data, reduce
the influence of outliers, and adjust for nonlinearity of the data. The transformations had the added
advantage of allowing us to interpret regression coefficients as elasticities.
Results for models with the overall compactness index are presented in Table 7.4 Control variables
mostly have the expected signs and often are significant. The compactness index (index) has the
expected strong positive relationships to walk and transit mode shares, and the expected negative
relationship to average drive time and average vehicle ownership per household. As noted, the
73

coefficients of these log-log models are elasticities. For every percentage increase in the compactness
index, the walk mode share increases by 0.39 percent, the transit mode share increases by 1.15 percent,
the vehicle ownership declines by 0.06 percent and the average drive time declines by 0.05 percent.
Based on these results, we consider the overall compactness index to be validated.
Table 7.4. Relationships of the Overall Indices to Transportation Outcomes (log-log transformed tstatistics in parentheses)
walkshr

transitshr

hhveh

drivetime

constant

-11.86

-0.62

-6.01

4.69

pop

0.01
(0.44)
-0.65
(-1.92)*
1.47
(9.19)***
-0.79
(-0.45)
0.55
(3.65)***
0.55
(3.33)***
2.45
(5.18)***
0.39
(4.22)***
0.54

0.42
(8.35)***
0.69
(1.24)
1.74
(6.61)***
-10.63
(-3.66)***
0.21
(0.83)
2.14
(7.80)***
3.35
(4.28)***
1.15
(7.57)***
0.73

-0.01
(-1.28)
0.80
(9.07)***
0.13
(3.02)**
0.82
(1.79)
0.12
(3.03)**
0.25
(5.80)***
-0.88
(-7.17)***
-0.06
(-2.45)**
0.42

0.04
(3.79) ***
0.40
(3.81) ***
-0.15
(-3.40) ***
-0.99
(-1.90)
-0.15
(-3.27) ***
0.25
(4.51) ***
0.21
(1.49)
-0.05
(-1.85)
0.48

hhsize
age1524
male
white
income
fuel
index
adjusted R2

* .05 probability level


** .01 probability level
*** .001 probability level
The new multidimensional compactness factors are mostly significant with the expected signs (see Table
7.5). The density factor, denfac, and centering factor, cenfac, are the most important correlates of
walking, followed by the mix factor, mixfac. The centering factor is also the most important correlate of
transit use, followed by the density factor, street factor and mix factor. All of the compactness factors
are inversely related to vehicle ownership per household, although only three, the density factor, mix
factor, and street factor, are statistically significant. All of compactness factors but street accessibility
have negative signs in the drive time equation. Mix and street factors are the most significant, followed
by the density factor. Street factor is positively related to average drive time. This may be due to the
increased wait time at intersections in a dense street network.
Table 7.5. Relationships of Individual Compactness Factors to Transportation Outcomes (log-log
transformed t-statistics in parentheses)

74

constant
pop
hhsize
age1524
male
white
income
fuel
denfac
mixfac
cenfac
strfac

walkshr

transitshr

hhveh

drivetime

-11.84
-0.09
(-2.51) **
-0.54
(-1.63)
1.25
(7.99) ***
-0.96
(-0.58)
0.63
(4.28) ***
0.44
(2.77) **
1.92
(4.15) ***
0.82
(3.69) ***
0.25
(2.13) **
0.36
(3.81) ***
-0.23
(-1.93)
0.605

-1.73
0.18
(3.09) ***
0.86
(1.50)
1.60
(5.91) ***
-10.85
(-3.82) ***
0.30
(1.20)
2.05
(7.49) ***
3.05
(3.80) ***
0.83
(2.14) *
0.60
(2.89) **
0.73
(4.41) ***
0.37
(1.81)
0.743

-4.04
0.03
(3.08)**
0.77
(9.57)***
0.11
(2.89)**
0.63
(1.59)
0.04
(1.10)
0.24
(6.23)***
-0.91
(-8.11)***
-0.37
(-6.94)***
-0.14
(-4.79)***
-0.01
(-0.52)
-0.06
(-2.24) *
0.564

2.62
0.06
(5.69)***
0.54
(5.41)***
-0.07
(-1.56)
0.62
(-1.30)
-0.13
(-3.05)**
0.32
(6.22)***
0.44
(3.31)**
-0.32
(-2.79)**
-0.14
(-3.84)***
-0.01
(-0.25)
0.16
(4.58)***
0.578

adjusted R2
* .05 probability level
** .01 probability level
*** .001 probability level

Theoretically, we do not expect four factors to be equally important for all travel outcomes. For
example, a concentration of activities is necessary to support investment and use of transit. So as one
would expect, the centering factor is the most significant correlate of transit mode share.

Discussion
This study used the same basic methodology as Ewing et al. (2002) to measure the sprawl for medium
and large metropolitan areas and divisions in 2010. We also expanded the sample size from 83
metropolitan areas in Ewing et al. (2002) to the 221 MSAs in this study.
For the 76 areas that are included in both studies, the compactness rankings are generally consistent
across years. The Spearman correlation between the compactness rankings in 2000 and 2010 is 0.635,
significant at .001 probability level which indicates, in general, the compact areas in 2000 are found to
be still compact in 2010; and the sprawling areas in 2000 are still sprawling. New York is the most
compact region followed by San Francisco in both years. Atlanta is the fourth most sprawling area in
2000 and the most sprawling area in 2010. Riverside-San Bernardino-Ontario, CA is the most sprawling
in 2000 and the third most sprawling area in 2010.
75

There are, however, metropolitan areas with significantly different ranking in 2010 than 2000. One of
the surprising cases is the Las Vegas-Paradise, NV metropolitan area. Its ranking rises from the 30th
most compact area in 2000 to the 16th in 2010 due to its moderate to high score in all four dimensions.
This is consistent with Fulton et al. (2001) study that found Las Vegas is getting more compact. Las
Vegas led the nation with an increase in its metropolitan density of 50 percent, thus rising in the overall
density rankings from 114th in 1982 to 14th in 1997 (Fulton et al. 2001, p: 7).
Refinements in operationalizing sprawl, is another reason for differences in rankings between years.
Land use mix and activity centering are the two dimensions with the most significant changes. As
contributors to centering, we now consider not only central business districts (CBDs) but employment
sub-centers. The existence of sub-centers is what distinguishes polycentric regions from monocentric
regions. The Washington DC metropolitan division is an example of polycentric region. As shown in
Figure 7.3, we identified 11 sub-centers (yellow color) in the metropolitan division. Out of 76
metropolitan areas with rankings in both years, the Washington DC metropolitan division has the 27th
highest score for activity centering in 2010 while it had the 41st highest score in 2000. Its overall
compactness ranking rises from 52nd most compact in 2000 to 27th most compact in 2010 due to its
change on the centering score.
Figure 7.3. Central Business District and Employment Sub-centers in Washington DC Metropolitan
Division

We also standardized the unit of analysis for mix use metrics by measuring them in half mile buffers
from the centroid of block groups. Out of 76 areas that are included in both years, Phoenix has the 19th
highest mix factor score in 2000 while it has the 24th lowest mix score in 2010. As a result, the Phoenix
76

metropolitan areas overall ranking drops from 18th most compact in 2000 to 14th most sprawling in
2010.
Finally, the changes in compactness score in some areas are due to changes in metropolitan boundaries.
Out of 76 metropolitan areas in both samples, Detroit moved up from 14th most sprawling in 2000 to
5th most compact in 2010. The 2010 Detroit, MI metropolitan division covers only about a fifth of the
area of the 2000 Detroit PMSA. The division is mostly limited to the Detroits downtown and
surroundings. The lowest density portions of Detroit PMSA are not included in 2010 metropolitan
division (see Figure 7.4). In particular, Warren-Troy-Farmington Hills, MI is now its own metropolitan
division, and a very sprawling one, the 20th most sprawling out of 221 metropolitan areas in 2010.
Figure 7.4. Detroit 2010 Metropolitan Division (dark) versus Detroit 2000 PMSA Boundary (light)

In 1997, the Journal of the American Planning Association published a pair of point-counterpoint articles
now listed by the American Planning Association as classics in the urban planning literature (APA,
2006). In the first article, Are Compact Cities Desirable? Peter Gordon and Harry Richardson argued in
favor of urban sprawl as a benign response to consumer preferences. In the counterpoint article, Is Los
Angeles-Style Sprawl Desirable?, Reid Ewing argued for compact cities as an alternative to sprawl. They
disagreed at the time about almost everything: the characteristics, causes, and costs of sprawl, and the
cures for any costs associated with sprawl.

77

In their original debate, Ewing challenged the notion that Los Angeles is compact, while Gordon and
Richardson labeled planning initiatives in Portland ineffective. The Los Angeles metropolitan division,
with a new compactness index of 136.7, ranks as the 21st most compact metropolitan area in the U.S.
among 221 MSAs in our sample. In a parallel analysis focusing on urbanized areas (see Chapter 9), Los
Angeles saw an impressive rise in its compactness ranking, from 18th in 2000 to 8th in 2010. The reason:
infill development. Contradicting metropolitan L.A.s reputation as the capital of unbridled sprawl,
roughly two-thirds of new housing built there between 2005 and 2009 was infill constructed in
previously developed areas rather than on raw land in the exurbs (Boxall , 2012). Los Angeles is
certainly more compact today than suggested by Ewing in 1997.
The change in Portlands compactness ranking is also interesting. Portland fell from the sixth most
compact in 2000 to 80th in 2010 because of boundary changes. The high ranking in 2000 applied to the
Portland PMSA, basically the area subject to an urban growth boundary. The metropolitan area now
includes Vancouver, WA, and its surroundings, in addition to the Portland PMSA (as shown in figure 7.5).
The new area is all outside the UGB.
Figure 7.5. Portland metropolitan area boundary in 2010 versus Portland PMSA boundary in 2000

78

Chapter 8. Relationship of Sprawl to Topical Outcomes


Urban planners are ultimately less interested in development patterns, per se, than in the costs and
benefits of one pattern versus another. That is to say, there are no inherently good or bad patterns,
only good or bad outcomes. The loaded term sprawl has come to be applied to certain development
patterns because of their documented negative outcomes.
The literature on impacts of sprawl is well-established. Researchers have studied impacts of sprawl on
quality-of-life outcomes such as housing affordability (Kahn, 2001; Kahn, 2007; and Wassmer and Baaas
2006); traffic congestion (Ewing et al, 2002; Zolnik, 2011; Holcombe and Williams, 2012); traffic safety
(Ewing et al., 2003; Trowbridge et al., 2009); open space preservation (Nelson 1992; Nelson & Sanchez
2005); physical activity and obesity (Ewing et al, 2003; Kelly-Schwartz et al, 2004; Sturm and Cohen,
2004; Doyle et al, 2006; Fan and Song, 2009; Plantinga and Bernell, 2007; Lee et al, 2009); social capital
(Kim et al, 2006; Nguyen, 2010); and air quality (Stone, 2008; Kahn, 2006; Stone et al, 2010; Schweitzer
and Zhou, 2010; Bereitschaft and Debbage, 2013).
There are, however, other potential outcomes of sprawl that have made the news recently. There is
little understanding on how urban sprawl may affect so called topical outcomes such as life
expectancy, upward mobility, and housing costs. In this chapter we seek to find out how urban sprawl
may affect these outcomes.

Life Expectancy
The literature points to several important sociodemographic variables that affect life expectancy,
including gender (women live longer than men), income (higher income people live longer), race (whites
live longer than nonwhites), and education (in part due to an income effect, those with more education
live longer).
Mirowsky and Ross (2000) hypothesized that American adults can expect longer lives the higher their
achieved socioeconomic status. Three aspects of socioeconomic status provide an outline for analysis:
education, employment and occupation, and economic well-being. Each has known or suspected effects
on health and survival that may shape subjective estimates of life expectancy.
Ezzati et al. (2008) analyzed the trends in county mortality and cross-county mortality disparities,
including the contributions of specific diseases to county level mortality trends. There are large
disparities in in health and mortality across population subgroups defined by race, income, geography,
social class, education, and community deprivation indices.
Education regulates access to occupations, and affects both personal earnings and household income.
Persons with less than a high school diploma have twice the age-adjusted mortality rate as those with
college degrees (Christenson and Johnson 1995; Elo and Preston 1996; Pappas et al. 1993; Rogers 1992).
Education also improves physical health and functioning net of employment status, occupational status,

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

job quality, earnings, household income, race, sex, age, and social origins (Johnson and Wolinsky 1993;
Reynolds and Ross 1998; Ross and Wu 1995).
Olshansky et al. (2005) analyzed the substantial rise in the prevalence of obesity and its life-shortening
implications. The study found that there are also other threats to life expectancy. Forces such as
infectious diseases, pollution, lack of regular exercise, ineffective blood-pressure screening, tobacco use,
and stress can be lead to a reduction in life expectancy.
Shaw et al. (2005) studied the determinants of life expectancy in developed countries. They argued that
life expectancy is predetermined by behavioral and policy variables in what can be loosely described as a
production function for health. While empirical results are mixed, the general consensus is that
population life expectancy is a function of environmental measures (e.g., wealth, education, safety
regulations, infrastructure), lifestyle measures (e.g., fruit and vegetable consumption; tobacco or alcohol
consumption), and health care consumption measures (e.g., medical or pharmaceutical expenditures).
Conceptual Framework
There are at least four plausible causal pathways between sprawl and life expectancyone through
obesity, which Chapter 6 determines to be higher in sprawling counties due, most likely, to reduced
levels of physical activity. Obesity is the second leading preventable cause of death in the U.S. (Mokdad
et al. 2004).
A second pathway is through traffic fatalities, which Chapter 5 finds to be higher in sprawling counties
due to increased driving and crash exposure. There are upwards of 32,000 traffic fatalities in the U.S.
each year, many involving young people who die prematurely. Traffic fatalities are the leading cause of
death for those 15 through 24 (Kochanek et al., 2011). Traffic fatalities are the sixth leading preventable
cause of death in the U.S. (Mokdad et al. 2004).
A third pathway is through air pollution. Four studies have related sprawl to poor air quality (Ewing
2002; Stone 2007; Schweitzer and Zhou, 2010; Bereitschaft and Debbage, 2013). The most recent and
most comprehensive found that metropolitan areas with lower levels of sprawl on average exhibit lower
concentrations of ozone (O3) and fine particulates (PM2.5) (Bereitschaft and Debbage, 2013). Countless
studies have related air quality to illness and premature death.
A fourth pathway is through violent crime, or specifically, homicide. Central cities are generally
considered more crime ridden than suburbs and exurbs. However, we know of no study that has
explicitly related compactness or sprawl to violent crime. One study did compare the Mortality Risk
Associated with Leaving Home across 15 metropolitan areas, and found that when traffic fatalities and
homicides-by-strangers were summed, central cities and inner suburbs were actually safer than outer
areas (Lucy, 2003). Another study reached a similar conclusion when it compared injury deaths (traffic,
homicide, suicide, etc.) across an urban-rural continuum, and found major cities to be safer than rural
areas (Myers et al., 2013).

80

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Air Pollution
Ozone and particulates are the most widespread air pollutants. Ground level ozone is not emitted
directly into the air, but is created by chemical reactions between nitrogen oxides (NOx) and volatile
organic compounds (VOC) in the presence of sunlight. Emissions from industrial facilities and electric
utilities, motor vehicle exhaust, gasoline vapors, and chemical solvents are some of the major sources of
NOx and VOC. The sources of PM2.5 include emissions from automobiles, power plants, wood burning,
industrial processes, and diesel powered vehicles. These fine particles are also formed in the
atmosphere when gases such as sulfur dioxide, NOx, and VOC (all of which are also products of fuel
combustion) are transformed in the air by chemical reactions. Since motor vehicles are a major source
of both ozone precursors and PM2.5 components and precursors, it is not unreasonable to expect some
relationship between sprawl and both ozone and PM2.5 concentrations, and hence between sprawl and
life expectancy.
Nearly 132 million people live in counties where monitors show unhealthy levels of one or both. Ozone
can harm lung function, irritate the respiratory system, and cause asthma, bronchitis, other
cardiopulmonary problems, and premature death (Levy et al. 2001; World Health Organization, 2003).
Particle pollution can increase the risk of heart disease, lung cancer, asthma attacks, and premature
death (Dockery 2009). A recent study estimates that ozone and PM2.5 cause about the same number of
deaths each year in North America (Silva et al., 2013). The estimated numbers are between 34,400 and
52,200 for ozone, and between 43,000 and 77,000 for PM2.5.
An air quality index has been estimated at the county level by EPA and includes an annual summary of
days with good, moderate, unhealthy for sensitive groups, unhealthy, and very unhealthy air at the
county level. The AQI takes into account all six criteria air pollutants: carbon monoxide, nitrogen dioxide,
ozone, sulfur dioxide, PM2.5, and PM10. Each row of the AQI Report lists summary values for one year
and one county. The summary values include both quantitative measures (days of the year having
"unhealthy" air quality, for example) and descriptive statistics (median AQI value, for example).
Crime
It may be thought that larger, denser cities are more plagued by crime than smaller cities and towns, but
research has shown that this is not always the case. According to Browning et al (2010), there may be
higher crime rates when density begins to increase initially, but once it reaches a certain level and there
is more pedestrian activity and traffic due to higher density, crime rates will likely decrease. However,
this decrease in crime at higher densities could also be due to increased policing that tends to come
along with mixed-use development, among other factors such as community cohesion, availability of
economic opportunities, and higher education and income levels of residents in larger cities (Browning
et al. 2010; Litman 2013). There seems to be more of a correlation between crime and income;
therefore, further research should be done on the effect of residential and commercial density in lowincome neighborhoods (Browning et al. 2010).

81

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Furthermore, a Brookings Institution Study analyzed data from the Uniform Crime Report (UCR) and the
American Community Survey (ACS) and found that the violent crime rate in cities dropped by 29 percent
compared to the violent crime rate in suburbs, which dropped only seven percent (Kneebone and
Raphael 2001). Although the overall rate of crime is higher in cities than in suburbs, the difference
between the two is narrowing.
While crime may be higher in cities, this doesnt mean that cities are less safe overall. Myers et al. (2013)
analyzed injury death rates in both urban and rural environments, and found that, overall, the
occurrence of serious injuries resulting in death was 20 percent higher in rural areas than urban areas.
The researchers do state that the risk of homicide is indeed higher in urban areas, but this is outweighed
by deaths resulting from motor vehicle crashes, among other injuries, in rural areas (Myers et al. 2013).
In fact, the risk of fatality due to motor vehicle crash is two times higher in rural than urban areas
(Myers et al. 2013). Similarly, another study looked at the dangers of leaving ones home in terms of rate
of homicide by strangers and motor vehicle or traffic fatalities and compared these risks in rural and
urban areas (Lucy 2003). The results again showed higher rates of traffic fatalities and homicides by
strangers in rural areas with lower population densities (Lucy 2003).
When looking at violent crime, its effect on life expectancy varies. Redelings et al (2010) found that
homicide reduced life expectancy by 0.4 years for residents in Los Angeles County, but it reduced life
expectancy for black males by 2.1 years, a stark difference. Similar results for life expectancy were seen
in low-income urban areas (Redelings et al. 2010), and the Centers for Disease Control finds that
discrepancies in life expectancy exist between difference races. The CDC reports that life expectancy for
the black population was 3.8 years lower than the white population (4.7 years lower for black males
compared to white males) due to factors such as health issues and homicide rates (Kochanek, 2013).
According to Haley et al. from the Center on Human Needs, high crime rates are consistently observed
in communities with lower educational attainment and less economic opportunity, as well as in highly
segregated communities (2012). Crime may play a role in life expectancy, but it is dependent on the
economic and social characteristics of a particular city or neighborhood.
Methods
In this section we use a recently released dataset of life expectancies by county (Wang et al., 2013) to
study the relationship between sprawl and life expectancy. The release of this dataset made the national
news, as the media realized that where you live could affect the length as well as the quality of your life
(NBC Nightly News, July 10, 2013).
Differences across counties were surprisingly large. The mean life expectancy for males and females
combined was just over 78 years. The standard deviation was more than two years. The range was 71
years at the bottom to 83 years at the top.

82

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

The dataset came with no explanation as to differences in life span from place to place. In this section,
we use structural equation modeling (SEM) to estimate a model of life expectancy. Readers are referred
to Chapter 5 for a description of SEM and its advantages, in cases like this with complex
interrelationships among variables.
The variables tested in the model are shown in Table 8.1. The variable of greatest interest is lifeexp,
county average life expectancy. There are four mediating (intermediate) variables: average county-level
vehicle miles traveled per capita (VMT), EPAs air quality index (AQI, a combination of six air quality
indicators), average county body mass index (BMI), and the violent crime rate (crime). They relate,
respectively, to four causes of premature death--traffic accidents, respiratory illnesses, obesity-related
chronic health conditions, and crime and its effects on physical and mental health.
Table 8.1. County Variables (variables log transformed)
Endogenous Variables
Variables
lifeexp
average life expectancy
VMT
annual vehicle miles traveled per household
AQI
air quality index
BMI
average body mass index
crime
violent crime rate per 100,000 population
Exogenous Variables
Variables
pop
metropolitan population
hhinc
median household income
white
percentage of white population
male
percentage of male population
county compactness index for 2010 (including additional
indexn
variables compared to 2000 index)

Data Sources
States, Census 2010
EPA 2011
EPA
BRFSS
FBI Uniform Crime Statistics
Data Sources
Census 2010
ACS 2006-2010
Census 2010
Census 2010

Results
We have estimated an SE model for life expectancy with Amos 19.0, a popular SEM software package
with a good graphic display. The fitted model is shown in Figure 8.1. The path diagram is copied directly
from Amos. Causal pathways are represented by straight uni-directional arrows. Correlations are
represented by curved bi-directional arrows (to simplify the already complex causal diagram, some
correlations are omitted). By convention, circles represent error terms in the model, of which there is
one for each endogenous (response) variable.
Direct relationships are presented in Table 8.2. Maximum likelihood methods were used in the
estimations. Model evaluation was based on four factors: 1) theoretical soundness; 2) chi-square tests

83

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

of absolute model fit; 3) root-mean-square errors of approximation (RMSEA), which unlike the chisquare, correct for sample size; and 4) comparative fit indices (CFI).
Relationships are highly significant and generally as expected. Goodness-of-fit measures at the bottom
of the table suggest that the model provides a good fit to the data (see Chapter 5 for a discussion of
these measures). The life expectancy model in Figure 8.1 has a chi-square of 5.9 with 8 model degrees of
freedom and a p-value of 0.653. The low chi-square relative to model degrees of freedom and a high
(>0.05) p-value are indicators of good model fit.
The county compactness index has a direct positive relationship to life expectancy, plus indirect effects
through mediating variables. Lets consider each indirect effect in turn. The compactness index is
negatively related to VMT per capita, which in turn is negatively related to life expectancy. This is as
expected from Chapter 5. Hence the indirect effect of compactness on life expectancy via this causal
pathway is positive. The compactness index is also negatively related to BMI, which is negatively related
to life expectancy. This is as expected from Chapter 6. Hence the indirect effect of compactness on life
expectancy via this causal pathway is also positive. The compactness index is positively related to the air
quality index, which is negatively related to life expectancy. This means that air quality is worse in
compact counties, and life expectancies may be shorter. However, the latter relationship is not
statistically significant. Hence the indirect effect of compactness on life expectancy via this causal
pathway, while negative, is weak. Finally, the compactness index is positively related to violent crime,
which is negatively related to life expectancy. Hence the indirect effect of compactness on life
expectancy via this causal pathway is also negative. This finding squares with popular notions that cities
are more dangerous than suburbs.
Direct, indirect, and total effects of variables on one another are shown in Table 8.3. The net indirect
effect of compactness on life expectancy is positive, and when added to the direct effect, is even more
positive. Using life expectancy data from a credible source, and our validated compactness/sprawl
index, we conclude that life expectancy is significantly higher in compact than sprawling counties. As the
compactness index doubles (increases by 100%), life expectancy increases by about 4 percent. For the
average American with a life expectancy of 78 years, this translates into a three year difference. That is
well within the range of life expectancy differences from county to county.
Figure 8.1. Causal Path Diagram for Life Expectancy in Terms of County Compactness and Other
Variables

84

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Table 8.2. Direct Effects of Variables on One Another in the Life Expectancy Model
coefficient
VMT
AQI
BMI
BMI
BMI
AQI
VMT
VMT
crime

<--<--<--<--<--<--<--<--<---

index
index
hhinc
index
white
pop
white
hhinc
white

-0.945
0.553
-0.064
-0.051
-0.029
0.13
-0.243
0.057
-1.367

std error

critical ratio

p-value

0.058
0.15
0.011
0.015
0.012
0.02
0.046
0.042
0.103

-16.243
3.682
-5.96
-3.5
-2.484
6.613
-5.313
1.337
-13.247

<0.001
<0.001
<0.001
<0.001
0.013
<0.001
<0.001
0.181
<0.001

85

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

coefficient
crime
crime
lifeexp
lifeexp
lifeexp
lifeexp
lifeexp
lifeexp
lifeexp
lifeexp

<--<--<--<--<--<--<--<--<--<---

index
hhinc
BMI
AQI
VMT
hhinc
white
index
pop
crime

std error

critical ratio

p-value

0.131
0.095
0.011
0.001
0.003
0.004
0.004
0.005
0.001
0.001

8.266
-11.582
-2.934
-0.83
-3.405
20.028
4.784
6.708
-3.945
-3.02

<0.001
<0.001
0.003
0.406
<0.001
<0.001
<0.001
<0.001
<0.001
0.003

1.083
-1.105
-0.033
-0.001
-0.01
0.082
0.018
0.034
-0.003
-0.004

5.9
degrees of freedom = 8
p-value = 0.653
0.0
p-value = 0.988
1.0

chi-square
RMSEA
CFI

Table 8.3. Direct, Indirect, and Total Effects of the County Compactness Index and Other Variables on
Life Expectancy

direct effect
indirect effect
total effect

index
0.034
0.006
0.04

pop
-0.003
0
-0.003

white
0.018
0.008
0.027

hhinc
0.082
0.006
0.087

crime
-0.004
0
-0.004

VMT
-0.01
0
-0.01

AQI
-0.001
0
-0.001

Housing Affordability (H+T)


Housing affordability has been one of the most persistent national concerns mainly because housing
costs are the biggest expenses in the budgets of most households. A typical American household spends
more than a third of its budget on housing while poor and near-poor households commonly devote
about half of their incomes to housing expenditures (Consumer Expenditure Survey, 2011).
The majority of studies of housing affordability focus on housing cost and its relationship to household
income as the sole indicator of affordability. The main providers of affordability indexes in the US are
real estate institutes and government agencies. The National Association of Realtors, for example,
publishes a Housing Affordability Index for existing single-family homes by metropolitan area. The NAR
affordability index measures whether or not a typical family could qualify for a mortgage loan on a
typical home.

86

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

A typical home is defined as the median-priced, existing single-family home as calculated by


NAR.

The typical family is defined as one earning the median family income as reported by the U.S.
Bureau of the Census.

The prevailing mortgage interest rate is the effective rate on loans closed on existing homes
from the Federal Housing Finance Board.

These components are used to determine if the median income family can qualify for a mortgage on a
typical home. Note that transportation costs associated with different development patterns are not
factored into this calculation.
Likewise, HUD has a standard of affordability that focuses exclusively on housing costs. If the ratio
housing costs to household income is less than 30 percent, then the dwelling is considered affordable
(Belsky et al. 2005; HUD, 2006). HUDs standard of affordability is the very definition of housing
affordability (Linneman & Megbolugbe, 1992; Belsky et al. 2005). The HUD measure is the legislative
standard used to qualify applicants for housing assistance. It is used in the administration of rental
housing subsidies, such as the Section 8 housing vouchers (Bogdon & Can, 1997). In addition to
qualifying ratios, it is often used to describe housing markets not only in the United States but also
internationally (Robinson et al. 2006).
Both indices and standards are structurally flawed in that they only consider costs directly related to
housing, ignoring those related to utilities and transportation. We know from the Consumer Expenditure
Survey that the typical American household spends about 26.3 percent of income on housing, excluding
utilities and public services costs. For the typical household, therefore, housing is affordable. But the
typical household also spends 16.7 percent for transportation. Housing plus transportation costs
consume 43 percent of household income in 2011. If a household's transportation costs were zero but
its housing costs were 35 percent of income, we would say that its housing was unaffordable, when in
fact the household would be no worse off than the typical American household.
At the same time, previous studies show that there is a clear tradeoff between the housing and
transportation expenses of working families. Families that spend more than half of their total household
expenditures on housing put 7.5 percent of their budget towards transportation. By contrast, families
that spend 30 percent or less of their total budget on housing spend nearly one-quarter of their budget
on transportation - three times as much as those in less affordable housing (Dietz, 1993 and Limpan,
2006)
The inherent weaknesses of traditional measures of housing affordability were revealed by the
Subprime Mortgage Crisis and ensuing wave of foreclosures that swept the United States in the late
2000s and directly helped precipitate the Global Financial Crisis or the Great Recession. Under
traditional metrics of affordability, lenders granted loans to families who were unable to maintain

87

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

mortgage payments, in many cases, because of the crushing costs of transportation in an environment
record high motor vehicle fuel prices. Foreclosures were centered in the Sunbelt states of Arizona and
Nevada, where rapid suburban and exurban development occurred in automobile dependent areas with
virtually no transit access.
Addressing this issue, the Center for Neighborhood Technology (CNT) and the Center for Transit
Oriented Development (CTOD) in 2006 developed an innovative tool that measured true housing
affordability called the Housing + Transportation Affordability Index. The H+T Affordability Index took
into account not only the cost of housing, but also the intrinsic value of location, as quantified through
transportation costs (Center for Transit-Oriented Development and Center for Neighborhood
Technology, 2006).
The H+T affordability index built on the analysis and theory of the location efficient mortgage (LEM), a
lending product that was developed by a group of researchers for Fannie Mae in 2000. The LEM was
rolled out in three regions. The LEM was very similar to the H+T affordability index in that it combined
the costs of housing and transportation, and presumed that homebuyers could afford a bigger mortgage
if they choose a neighborhood near public transit where they could realize significant savings on
transportation (Holtzclaw et al, 2002).
Later in 2010-13, the Departments of Transportation and Housing and Development funded the
development of refined H+T-like index, called the Location Affordability Index. The LAI is based on the
same methodology as H+T Affordability Index but uses the most recent and better quality data with
more coverage. 11
In the LAI Index, total housing costs are computed as current housing sales prices and rents from the
American Community Survey, and household transportation costs are estimated at the census block
group level as three separate cost components: costs of auto ownership, auto use, and transit use. Auto
use is modeled based on household VMT data from Chicago and St. Louis. Auto ownership is modeled
based on vehicle ownership data from the American Community Survey. Transit use and associated
costs are based on Google transit feeds. Housing is considered affordable if the sum of housing and
transportation costs is no more than 45 percent of household income.
Conceptual Framework
Our hypothesis is that urban sprawl can affect housing affordability in two ways. The first one is directly
through price of land and hence housing. The second is indirectly through the availability of
transportation choices and the amount of driving people do. In this chapter we account for both effects,
first individually and then together. We posit that due to higher land costs, compact areas will have
more expensive housing. However, due to the availability of alternative modes and the shortness of
auto trips, compact areas should entail lower transportation costs. A priori, it is impossible to say what
11

http://www.locationaffordability.info/

88

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

the net effect of sprawl on housing + transportation costs will be. Instead, we approach the subject
empirically in the sections that follow.
Method
Data and Variables
Our variables and data sources are shown in Table 8.4. Our data relate to approximately 150,000 census
block groups in 212 metropolitan areas and divisions. Most of the variables have been discussed and
used before. The new ones are the three housing variables in Table 8.4, medium year when housing
units were built, percentage of vacant housing units, and percentage of owner-occupied housing units.
Table 8.4. Variables Used to Explain Housing + Transportation Affordability (variables log transformed)
Level 1 Dependent variables (block group level)
Variables
hcost
housing costs as a percentage of income
tcost
transportation costs as a percentage of income
h+tcost
housing and transportation costs as a percentage of income
Level 1 Independent variables (block group level)

Data Sources
LAI2013
LAI2013
LAI2013

Variables
hhsize
average household size
hhinc
median household income
pctwhite percentage of white population
ybuilt
median year housing unit built
pctvac
percentage of vacant housing units
pctown
percentage of owner-occupied housing units
Level 2 Independent Variables (metropolitan level)

Data Sources
Census 2010
ACS 2007-2011
Census 2010
ACS 2007-2011
ACS 2007-2011
ACS 2007-2011

Variables
metpop
index

Data Sources
Census 2010
Computed

metropolitan area population


metropolitan compactness index for 2010

Analytical Method
Once again, the data used in this analysis are nested and must be analyzed accordingly. Census block
groups form Level 1 in our nesting structure. Metropolitan areas, within which census block groups fall,
form Level 2. Since all block groups located in an MSA share characteristics of that MSA such as fuel
price, we can say that block groups are nested within MSAs. This nesting tends to produce
dependence among cases, violating the independence assumption of ordinary least squares regression
or OLS. Standard errors of regression coefcients associated with MSA characteristics based on OLS will
consequently be underestimated, and regression coefficients themselves will be inefficient.

89

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Whenever individuals or other entities are nested within higher-level units, hierarchical or multilevel
modeling (MLM) methods are preferred for explaining individual outcomes in terms of both individual
and group characteristics. Multilevel modeling overcomes the limitations of OLS, accounting for the
dependence among lower level units and producing more accurate coefcient and standard error
estimates. So MLM is used in this analysis for all transportation outcomes. Models were estimated with
HLM 6.08.
Results
Results for simple housing affordability conform to our expectations (see Table 8.5). The block group
average percentage of income spent on housing increases with median household income, percentage
of whites in the population, the median year in which housing was built, and the percentage of housing
which is owner occupied. The percentage of income spent on housing declines with the percentage of
vacant units, as vacancies depress housing prices both directly by increasing supply relative to demand
and indirectly by having negative spillover effects. Among Level 2 variables, housing affordability is
greater in small than large metropolitan areas. That is a simple reflection of land prices and bid rent
curves.
Controlling for these variables, the average percentage of income spent on housing is greater in
compact than sprawling areas. This is to be expected. From the NAR Housing Affordability Index and
common knowledge, housing is most expensive in the most compact metropolitan areas such as the
New York and San Francisco areas. Each percent increase in the metropolitan compactness index is
associated with a 0.11 percent increase in housing costs relative to income.
Table 8.5. Relationships to Housing Affordability (log-log form with robust standard errors)
coeff.
constant
hhsize
hhinc
pctwhite
ybuilt
pctown
pctvac
metpop
index
pseudo-R2

t-ratio
-13.39
0.0052
0.17
0.012
1.79
0.14
-0.0094
0.019
0.113

p-value

NA 12

-5.93
0.183
23.19
2.744
5.937
14.16
-2.53
2.097
4.108

<0.001
0.855
<0.001
0.007
<0.001
<0.001
0.012
0.037
<0.001

12

Pseudo-R2s are negative. A pseudo-R2 is not entirely analogous to R2 in linear regression, which can
only assume positive values. One standard text on multilevel modeling notes that the variance can
increase when variables are added to the null model. It goes on to say: This is counter-intuitive,
because we have learned to expect that adding a variable with decrease the error variance, or at least
keep it at its current level In general, we suggest not setting too much store by the calculation of
[pseudo-R2s] (Kreft and de Leeuw 1998, 119). For more discussion of negative pseudo-R2s, also see
Snijders and Bosker (1999).

90

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Results for transportation affordability are less intuitive (see Table 8.6). Unlike housing expenditures, LAI
transportation expenditures are a modeled result, which may have introduced error. The block group
average percentage of income spent on transportation increases with household size, median
household income, and percentage of whites in the population. The relationship to household size is
surprising, since larger households can make more efficient use of vehicles through carpooling and trip
chaining. Among Level 2 variables, transportation affordability is greater in large than small
metropolitan areas. This is also surprising since large areas are known to generate longer trips.
Apparently, this effect is more than offset (in the LAI models) by greater availability of alternatives to
the automobile and, correspondingly, lower automobile ownership in large areas.
Controlling for these variables, the average percentage of income spent on transportation is smaller in
compact than sprawling areas. This is to be expected due to shorter trips and more transportation
options in compact areas. Each percent increase in the metropolitan compactness index is associated
with a 0.35 percent decrease in transportation costs relative to income.
Table 8.6. Relationships to Transportation Affordability (log-log form with robust standard errors)
constant
hhsize
hhinc
pctwhite
metpop
index
pseudo-R2

coeff.

5.23
0.27
0.079
0.042
-0.13
-0.34

t-ratio

NA12

24.70
7.583
8.368
7.713
-14.77
-10.31

p-value

<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

The final analysis relates at the sum of housing + transportation costs, again relative to income, to the
same set of variables (see Table 8.7). The sum increases with household size, income, and other
variables already discussed. This implies that other expenditures, such as food, utilities, and health care,
actually decline in relative to income, as these variables increase. Most interesting, the sum of housing +
transportation costs declines with metropolitan area compactness, though the relationship is significant
only at the 0.10 level. This means that the decline in transportation costs with greater compactness
more than offsets the rise in housing costs. This is a novel finding, conditioned only on quality of data
upon which the LAI is based.
Table 8.7. Relationships to Housing + Transportation Affordability (log-log form with robust standard
errors)

91

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

coeff.
constant
hhsize
hhinc
pctwhite
ybuilt
pctown
pctvac
metpop
index
pseudo-R2

t-ratio
-24.26
0.073
0.102
0.022
3.56
0.135
0.0083
-0.031
-0.034

p-value

NA12

-15.14
5.938
19.86
5.938
16.36
14.21
3.997
-4.357
-1.790

<0.001
0.003
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.074

Upward Mobility
This American ideal is rooted in the Declaration of Independence: hard work is enough to create upward
mobility, with greater opportunities, personal security, and affluence. But is the American ideal equally
achievable for all societal groups?
Recent studies show that the U. S., as the land of opportunity for all, has one of the lowest rates of
upward mobility in the developed world. Only a few citizens leave the class into which they are born for
a higher one (DeParle, 2012.). Downward mobility also is just as common as upward mobility,
particularly for African Americans, but also for the middle class in general. In the U. S., a person born in
the lowest economic class has only about a two percent chance of ending life anywhere near the top
(Mazumder, 2003, Jntii, et. al., 2006).
Social mobility is an often-researched topic with scholarly articles on the subject dating back to the
1950s and 1960s. Much of the research has focused on race (Hardaway and McLoyd, 2008), family
background (Jntii, et. al., 2006; Black and Devereux, 2010), income (Corak, 2006), family structure,
particularly divorce (DeLeire and Lopoo, 2010) as determinants of social mobility. Poorly staffed and
funded schools in poor and working-class neighborhoods, inadequate prenatal nutrition and health care,
environmental hazards and pollutions are some of the factors that affect the success of a poor youth
seeking a better life (Delgado, 2007).
Equality of Opportunity Project
A study of upward mobility by a team from Harvard and UC Berkeley made the news recently. Their
study is appropriately titled "The Equality of Opportunity Project." They found that one of the key
determinants of social mobility is geography; where a person grows up may dictate how likely that
person is to move out of the social class into which he or she was born (Chetty, et. al., 2013). They
found:
children from families at the 25th percentile in Seattle have outcomes comparable to children
from families at the median in Atlanta. Some cities such as Salt Lake City and San Jose have

92

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

rates of mobility comparable to countries with the highest rates of relative mobility, such as
Denmark. Other cities such as Atlanta and Milwaukee have lower rates of mobility than any
developed country for which data are currently available (Chetty et al., 2014a).
and
intergenerational mobility varies substantially across areas within the U.S. For example, the
probability that a child reaches the top quintile of the national income distribution starting from
a family in the bottom quintile is 4.4% in Charlotte but 12.9% in San Jose (Chetty et al., 2014b).
What struck us immediately about these findings is a possible connection to sprawl. In our metropolitan
rankings, Atlanta and Charlotte are at the sprawling end of the scale, while Salt Lake City and San Jose
are far more compact. Metropolitan sprawl may contribute to the low rates of upward mobility for
lower-income classes in sprawling metros. The most important indicator of sprawl is poor accessibility
(Ewing, 1997). Considering the fact that those with low-incomes have limited transportation mobility,
inaccessibility of job opportunities could affect their ability to get ahead. Other causal pathways may
also exist between sprawl and upward mobility (see below).
Chetty and colleagues have tested for correlations between upward mobility and possible causal factors
(computing bivariate correlations). They caution that all of their findings are correlational and cannot be
interpreted as causal effects. But they do find strong correlations between upward mobility and six
factors:

Income Growth Commuting zones (analogous to metropolitan areas) with low levels of
income growth have low rates of upward mobility
Racial Segregation Commuting zones with high levels of racial segregation have low rates
of upward mobility
Income Inequality Commuting zones with high levels of income inequality have low rates
of upward mobility
Quality of K-12 schools Commuting zones with poor schools have low rates of upward
mobility
Social Capital Commuting zones with low levels of social capital (poor social networks and
low community involvement) have low rates of upward mobility
Family Structure Commuting zones with high levels of single parenting have low rates of
upward mobility

Chetty and colleagues also speculate on a link between sprawl and upward mobility. They
operationalize sprawl in terms of commute times to work, but that is as far as they go. we also find
that upward mobility is higher in cities with less sprawl, as measured by commute times to work
(Chetty et al., 2014a). As we have shown in this report, there are much better ways to measure sprawl
than commute times to work. Indeed, some of the most compact metropolitan areas have some of

93

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

the longest commute times, by virtue of their size and heavy use of transit (which typically involves
longer travel times than automobiles). Think New York City and San Francisco.
Conceptual Framework
We can see at least four plausible causal pathways between sprawl and upward mobility. The first causal
pathway, as suggested in the previous section, is through job accessibility. It is reasonable to assume
that sprawl, as we have defined it (low-density, single-use, uncentered development) exacerbates the
mismatch of workers and jobs and results in unemployment of low-skill workers.
Another potential pathway between sprawl and social mobility may be through social capital. As a result
of increasing sprawl, space between residential and commercial development increases, forcing people
to drive more and to divert their time away from social-interaction activities (Nguyen, 2010). Also, the
strength of social networks and engagement in community organizations may be depressed by sprawl,
thereby affecting social and economic outcomes (Putnam, 1995).
A third potential pathway may be through racial segregation, which has proved in previous studies to be
influenced significantly by the degree of compactness in metropolitan areas (Khan, 2001; Galster, 1984;
Galster and Cutsinger, 2007; Nelson, Sanchez, and Dawkins, 2004). A fourth possible pathway may be
through income segregation. Income segregation has negative effects on economic mobility by reducing
exposure to successful peers and role models, decreasing funding for local public goods such as schools,
and hampering access to nearby jobs (Cutler and Glaeser 1997).
Job Inaccessibility
After World War II, many wealthy Americans decentralized out of the cities and into the suburbs.
Shopping and support services followed them, leaving poor and minority populations behind. In 1968,
John Kain formulated the spatial mismatch hypothesis, speculating that black workers left in central
cities were increasingly distant from and poorly connected to major centers of growth. The spatial
mismatch hypothesis has implications for inner city residents that are dependent on low level entry
jobs. Distance from work centers can lead to increasing unemployment rates among inner city residents
and increased poverty outcomes for the region as a whole.
Commuting cost is one obstacle to inner city workers landing jobs in the suburbs. Autos may be too
expensive for low-income workers, and public transportation is problematic in a sense that it is not
always reliable and may not stop at all job sites. Information access to jobs may also hinder matches
between inner city workers and suburban jobs. People who are living away from the job center are
generally less knowledgeable about potential openings than individuals who live closer to the job center.
Therefore, networking and information spillovers are of a major advantage in accessing information
about potential openings.
There were many empirical studies on the spatial mismatch hypothesis in the early 1970s, soon after
the hypothesis was advanced by Kain. There was resurgent interest in the hypothesis in the early 1990s,

94

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

when at least six review articles were published (Wheeler, 1990; Jencks and Mayer, 1990; Moss and
Tilly, 1991; Holzer, 1991; Kain, 1992; Ihlanfeldt, 1992). However, to our knowledge, the spatial
mismatch of low-paid jobs and low-skill workers has never been related to urban sprawl. Spatial
mismatch is represented in this study by job accessibility by transit, as measured in EPAs Smart Location
Database.
Social Capital
The majority of studies on the relationship between sprawl on social capital have focused on factors
such as trust and neighborhood ties (Brueckner and Largey, 2006; Freeman, 2001; Glaeser and Gottlieb,
2006; Leyden, 2003; Lund, 2003; Nguyen, 2010). In his 2001 study, Freeman successfully established a
negative relationship between the level of car usage and the level of social ties in neighborhoods.
Leyden (2003) found that high neighborhood walkability resulted in a higher likelihood of establishing
relationships with ones neighbors, general social interaction, and political participation. Lund (2003)
tested New Urbanist claims that placing amenities such as parks and retail shops within walking distance
of homes will increase pedestrian travel and thereby increase interaction among neighbors.
Meanwhile, researchers continue to debate the benefits of compact development, and some have
demonstrated that social capital is not diminished by suburban sprawl. Using DDB Needham Lifestyle
Survey data, Glaeser and Gottlieb (2006) found that the rates of four types of social-capital activities,
like attending church, volunteering for a community project, contacting a public official or other civic
engagement, and registering to vote, are lower among residents of center cities. In addition, Brueckner
and Largey (2006) regressed individuals' social-interaction variables on census-tract density in a study
using the Social Capital Benchmark Survey data (The Roper Center, 2005). Their analysis showed that
high density had a negative effect on all friendship and group-involvement variables.
Most recently, Nguyen (2010) related our earlier county compactness/sprawl index to social-capital
factors from the Social Capital Community Benchmark Survey. This study found that urban sprawl may
support some types of social capital while negatively affecting the others. So the evidence on the effects
of sprawl on social capital is clearly mixed, but it certainly could represent a pathway between sprawl
and upward mobility.
In this study we use the social capital index (SCI) from the Chetty et al. database (2014) as the proxy for
social capital. It was originally measured by Rupasingha and Goetz (2008) and employed by Putnam
(2007). This index is comprised of voter turnout rates, the fraction of people who return their census
forms, and various measures of participation in community organizations in the area.
Racial Segregation
The evidence regarding sprawls impact on racial segregation is mixed. Some studies point to the cost of
housing/land as the primary contributors to Black-White residential segregation. Kahn (2001) has shown
that, controlling for household income and metropolitan patterns of racial segregation, sprawl was
significant in closing the gap between black and white rates of suburban homeownership between 1980

95

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

and 1990. In other words, sprawling metropolitan areas provide greater opportunities for suburban
homeownership by black households. The study also found that blacks are more likely to own larger
homes in sprawling regions. This is said to be a direct result of more affordable housing in sprawling
metropolitan areas.
Galster and Cutsinger (2007) analyzed the relationship between land use patterns and levels of black
and white segregation in 50 U.S. metropolitan areas. They found a direct correlation between sprawl
and desegregation. They surmised that the dominant relationship observed is that, on several
measures, more sprawl-like land use patterns are associated with less segregation (Galster and
Cutsinger, 2007, p. 540). They identified the land/housing price effect as the dominant mechanism
through which land use patterns influence black and white segregation measures .
On the other hand, Nelson, Sanchez, and Dawkins (2004) found that urban containment measures were
effective at reducing black and white segregation levels in designated metropolitan areas between 1990
and 2000. In this study, racial segregation is represented by black isolation. The variable comes from
Chetty, et. al., (2013) study for commuting zones and is computed based on census 2000 data.
Income Segregation
A final possible causal pathway between sprawl and upward mobility is income segregation. Income
segregation is related to spatial mismatch and racial segregation, but is operationalized differently.
Margo (1992) argues that the movement of metropolitan populations in the U.S. toward suburban
locales over the latter half of the 20th century can be linked, to a significant degree, to the rise in
personal incomes. As individual incomes rose, so did the demand for land. As a result, higher-income
households moved to the outskirts, while lower-income households remained within central cities.
According to Jargowsky (2002) this movement of higher-income households isolated those in the central
cities, leading to concentrations of poverty and a lack of resources, such as employment and educational
opportunities. Furthermore, he argues that these spatial disparities increase poverty in the short run
and reduce equality of opportunity, therefore contributing to inequality in the long run (Jargowsky
2002).
Wheeler (2006) conducted a statistical analysis inquiring if urban decentralization and income inequality
were associated. The study found an inverse relationship between urban density and the degree of
income inequality within metropolitan areas, suggesting that, as cities spread out, they become
increasingly segregated by income (Wheeler, 2006).
A study by Reardon and Bischoff (2011) investigated how the growth in income inequality from 1970 to
2000 affected patterns of income segregation along three dimensions: the spatial segregation of poverty
and affluence, race-specific patterns of income segregation, and the geographic scale of income
segregation. The analysis yielded four findings.

96

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Increasing income inequality was responsible for 40-80% of the changes in income segregation
from 1970-2000.
Income inequality affects income segregation primarily by affecting the segregation of affluence
rather than the segregation of poverty.
Income inequality and income segregation differs for black and white families. The study relates
this to housing policies and housing discrimination seen throughout the last half century.
Income inequality appears to increase income segregation largely by inducing the highestearning families to move far away from lower-income households.

In this analysis, the segregation of poverty, as a proxy for income segregation, is borrowed from Chetty
et al. (2013) and is defined as the extent to which individuals in the bottom quartile are segregated from
those above the bottom quartile.
Data and Measures
We used data on upward mobility and covariates from the Equality of Opportunity databases of Chetty
et al. (2013, 2014a, 2014b), but added a sprawl metric from our own study and job accessibility from the
Smart Location Database (see Table 8.8). The measure of upward mobility we chose to use was the
probability that a child born to a family in the bottom quintile of the national income distribution
reached the top quintile of the national income distribution by age 30. That is the third of three
measures quantified in their study, and is highly correlated with the first measure, what they call relative
mobility, which is defined as the difference in the expected economic outcomes between children from
high-income and low-income families. We chose the third measure because it is simple to interpret.
Our exogenous variables (that drive the system) are the rate of income growth between 2000 and 2010,
the share of families with kids with a female head of household and no husband in 2000, the mean
school district expenditure per pupil in 1996, and the metropolitan compactness/sprawl index for 2010.
We would expect that upward mobility is positively related to the rate of income growth, school
spending, and the compactness index, and negatively related to the share of female headed households.
Regarding our metropolitan compactness index for 2010, we would have preferred using an index for a
year near the midpoint of the 30-year period over which upward mobility manifests itself, but did not
have such an index. Fortunately, from our longitudinal analysis of urbanized area sprawl in Chapter 9,
urban form does not change dramatically from decade to decade. Our compactness index is also for
metropolitan areas rather than commuting zones as defined by Chetty et al. Commuting zones include
rural counties and sometimes more than one metropolitan area. In the latter case, we computed
weighted averages of the compactness/sprawl index, weighting by population. Because boundaries of
commuting zones and metropolitan areas are not coincident, we dropped commuting zones from our
sample if their populations differed from the combined metropolitan areas' by more than 25 percent. It
is unfortunate that we lost observations in this manner, but we still had a sample of 126 commuting
zones from which to estimate our models.

97

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Table 8.8. Variables Used to Explain Upward Mobility (variables log transformed)
Endogenous Variables
Variables
Data Sources
The probability that a child born to a family in the bottom
quintile of the national income distribution in 1980-1982
reaches the top quintile of the national income
distribution in 2010-2011.
upward
EOP 2013
CZ population weighted Jobs within 45 minutes auto travel Smart Location Database
jobaccess time, time-decay in 1000s (network travel time)
2010
Index of social capital that aggregates various measures
Rupasingha and Goetz
socialcap identified by Putnam and collaborators
(2008); EOP 2013
Geographic Isolation of Blacks (Cutler, Glaeser, Vigdor
isolation index). Computed as the difference between the
average share of black residents in a black person's
neighborhood and the total share of black residents in the
CZ, normalized to fall between 0 and 1
blackiso
EOP 2013
Measure of how evenly those in the lower income quartile
are distributed across census tracts within a CZ. This is
Thiel's H measure for the two groups defined by splitting
the population at the national 25th percentile of income
povseg
EOP 2013
Exogenous Variables
Variables
Data Sources
Annualized growth rate (2000-2008) in real household
EOP 2013; Census 2000;
incgrowth income per working age capita (16-64)
ACS 2010
Share of families with kids with a female householder and
femkids
no husband
EOP 2013; Census 2000
schspend Mean district expenditures per student in 1996
EOP 2013
index
computed
Metropolitan compactness index for 2010
Model and Results
We have estimated a structural equation model of upward mobility. The model is shown in Figure 8.2.
To make this complex model a little less so, some correlational arrows have been omitted. Direct
relationships are presented in Table 8.9. Maximum likelihood methods were used in the estimations.
Model evaluation was based on four factors: 1) theoretical soundness; 2) chi-square tests of absolute
model fit; 3) root-mean-square errors of approximation (RMSEA), which unlike the chi-square, correct
for sample size; and 4) comparative fit indices (CFI).
Relationships are mostly significant and as expected. Goodness-of-fit measures at the bottom of the
table suggest that the model provides a good fit to the data (see Chapter 5 for a discussion of these

98

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

measures). The upward model in Figure 8.2 has a chi-square of 11.7 with 7 model degrees of freedom
and a p-value of 0.112. The low chi-square relative to model degrees of freedom and a high (>0.05) pvalue are indicators of good model fit.
The metropolitan compactness index has a strong direct relationship to upward mobility, plus indirect
relationships through two mediating variables. The direct relationship seems to be capturing some
phenomenon that is missed by the mediating variables. Lets consider each indirect effect in turn. The
compactness index is negatively related to black isolation, which in turn is negatively related to upward
mobility. Hence the indirect effect of compactness on upward mobility via this causal pathway is
positive. The compactness index is positively related to job accessibility, which is positively related to
upward mobility (though not a statistically significant level). The lack of significance may have to do with
the job accessibility variable itself. It applies to job accessibility generally, not to job accessibility for
lower income workers, and it applies to auto accessibility, not transit accessibility (for which there were
too many missing values in the Smart Location Database to be used in this analysis). Perhaps the poor
construct validity of the job accessibility variable accounts for the strong direct relationship between
compactness and upward mobility. The compactness index may be capturing some relationship missed
by general job accessibility.
The compactness index is unrelated to social capital and poverty segregation. Of these two variables,
poverty segregation has a significant, negative relationship to upward mobility, as expected, while social
capital has no relationship to upward mobility. Chetty et al. report a relationship between social capital
and upward mobility, but that is based on a simple correlation between the two rather than a fully
specified model.
Of the other exogenous variables in the model, income growth is positively related to upward mobility,
while the share of female headed families with kids is negatively related to upward mobility. Per pupil
spending is unrelated to upward mobility, which may indicate that spending is not a very good proxy for
school quality. Both income growth and female headed families have indirect relationships to upward
mobility through the mediating variable, black isolation. Income growth adds to upward mobility
indirectly, while female-headed families detract from upward mobility indirectly.
Direct, indirect, and total effects of variables on one another are shown in Table 8.10. The net indirect
effect of compactness on upward mobility is positive, though small compared to the direct effect. Using
upward mobility data from a credible source, and our validated compactness/sprawl index, we conclude
that upward mobility is significantly higher in compact than sprawling metropolitan areas/commuting
zones. The point elasticity of upward mobility with respect to compactness is 0.41. As the compactness
index doubles (increases by 100%), the probability that a child born to a family in the bottom quintile of
the national income distribution will reach the top quintile of the national income distribution by age 30
increases by about 41 percent. For the average poor kid in our sample, with an 8 percent chance of
moving up into the top quintile, this represents an increase of 3.2 percent in absolute terms, well within
the range of upward mobility differences from metropolitan area to metropolitan area.

99

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Figure 8.2. Causal Path Diagram for Life Expectancy in Terms of Metropolitan/Commuting Zone
Compactness and Other Variables

Table 8.9. Direct Effects of Variables on One Another in the Upward Mobility Model
coefficient
socialcap
povseg
blackiso
socialcap
blackiso

<--<--<--<--<---

index
index
index
femkids
femkids

0.001
0
-0.001
-3.438
2.574

std error
0.002
0
0.001
1.053
0.33

critical
p-value
ratio
0.556
0.578
1.053
0.292
-1.698
0.089
-3.266
0.001
7.801
<0.001

100

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

coefficient
povseg
socialcap
jobaccess
blackiso
blackiso
upward
upward
upward
upward
upward
upward
upward
upward

<--<--<--<--<--<--<--<--<--<--<--<--<---

femkids
schspend
index
incgrowth
schspend
blackiso
povseg
incgrowth
index
femkids
socialcap
schspend
jobaccess

chi-square
RMSEA
CFI

std error

0.319
0.044
0.189
0.038
0.368
0.222
-9.806
2.338
0.034
0.012
-1.589
0.902
-13.853
6.516
140.316
25.178
0.033
0.006
-27.799
4.478
-0.171
0.277
0.017
0.125
0.002
0.003
11.7
degrees of freedom = 7
p-value = 0.112
0.73
p-value = 0.260
0.989

critical
p-value
ratio
7.227
<0.001
5.025
<0.001
1.658
0.097
-4.193
<0.001
2.932
0.003
-1.762
0.078
-2.126
0.034
5.573
<0.001
5.912
<0.001
-6.208
<0.001
-0.617
0.537
0.138
0.89
0.64
0.522

Table 8.10. Direct, Indirect, and Total Effects of the Metropolitan Compactness Index and Other
Variables on Upward Mobility

direct
effect
indirect
effect
total
effect

index

schspend

femkids

incgrowth

jobaccess

blackiso

povseg

socialcap

0.033

0.017

-27.799

140.316

0.002

-1.589

-13.85

-0.171

0.001

-0.086

-7.924

15.579

0.033

-0.069

-35.723

155.895

0.002

-1.589

-13.85

-0.171

101

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Chapter 9. Urbanized Areas: A Longitudinal Analysis


In this chapter we seek to measure changes in sprawl by developing refined and enhanced
compactness/sprawl indices for 2000 and 2010 based on definitions and procedures in Ewing et al.
(2002, 2003), but refined and applied this time to urbanized areas (UZAs) rather than metropolitan areas
or counties. We chose census UZAs as our units of analysis because UZAs are the only census
geographies that expand systematically with urban development over time. Counties have fixed
boundaries and hence tend to appear more compact over time (except when counties are losing
population as in Detroit or New Orleans after Katrina). Metropolitan areas expand in large increments as
entire counties, both urban and rural portions, are added to core counties to reflect changing
commuting patterns and social and economic integration.
Census UZAs expand incrementally as rural areas are converted to urban uses and density thresholds are
exceeded. If expansion takes the form of low densities, segregated land uses, commercial strips, and
poorly connected streets, compactness scores will decrease. Conversely, if expansion occurs with
moderate to high densities, integrated land uses, activity centers, and interconnected streets,
compactness scores will increase. Likewise, if growth occurs though infill and redevelopment,
compactness scores will increase.
This chapter uses principal component analysis (PCA) and 2010 cross sectional data for large U.S.
urbanized areas (UZAs) to operationalize compactness/sprawl in each of four dimensionsdevelopment
density, land use mix, activity centering, and street accessibility. Higher values represent greater
compactness, lower values greater sprawl. The four factors are then combined into an overall
compactness/sprawl index.
The chapter then applies factor score coefficient values for 2010 to the same variables for 2000 to
create comparable metrics for 2000. Compactness scores for 2000 are compared to the same scores for
2010 to see which UZAs sprawled the most between censuses, and which sprawled the least or actually
became more compact.
Finally, the chapter validates the compactness index and its component factors against transportation
outcomes for 2010, specifically walk mode share, transit mode share, and average drive time on the
journey to work. We then seek further validation by relating changes in compactness between 2000 and
2010 to changes in commuting outcomes. This is a much more stringent test of association/causation,
since changes over a decade are small compared to sampling errors associated with commuting data
from the American Community Survey. Also, in a time-step analysis, we can only hope to capture short
term effects of compactness on commuting, whereas cross sectional analyses capture long-term effects.
Not surprisingly, results of the longitudinal analysis are disappointing.

102

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Methods
Sample
The term urbanized area as defined by the U.S. Census Bureau denotes an urban area of 50,000 or
more people. Urban areas are defined by core census block groups or blocks with population densities
of at least 1,000 people per square mile and surrounding census blocks with densities of at least 500
people per square mile. Urbanized areas often provide a more accurate gauge of city size than do the
incorporated political boundaries of cities.
This investigation is limited to large urbanized areas. Our sample consists of the 162 largest urbanized
areas in the United States, those with more than 200,000 population in 2010. The rationale for thus
limiting our sample is simple: the concept of sprawl has particular relevance to large areas where the
economic, social, and environmental consequences of sprawl can be significant. The concept of sprawl
does not have much relevance to small urbanized areas such as Pine Bluff, AR and Monroe, MI.
Variables
The urbanized area sprawl indices operationalize four dimensions of sprawl: development density, land
use mix, activity centering, and street accessibility. In all, we have measured 15 variables in four
dimensions for 162 urbanized areas in both 2000 and 2010.
Development Density
Our first four density variables are the same as in the original sprawl index, gross density of urban and
suburban census tracts (popden), percentage of the population living at low suburban densities (lt1500),
percentage of the population living at medium to high urban densities (gt12500), and urban density
based on the National Land Cover Database (urbden). The fifth density variable is analogous to the first,
except it is derived with employment data from the Local Employment Dynamics (LED) database rather
than population data (empden). In this case, LED data were processed for the years 2005 and 2010. Year
2005 is the earliest year that LED data is available for all states (except Massachusetts). The data were
aggregated from census block geography to census block group geography to generate total jobs by
two-digit NAICS code for every block group in the nation.
Land Use Mix
Although using the same variables as Ewing et al. (2002) to operationalize mixed use, we computed
them differently using one-mile buffers around the centers of block groups rather than computing them
within the boundaries of block groups. The reason is that measures of mixed use are sensitive to the size
of a block group. The smaller the area is, the lower the job-population balance is expected to be,
because small areas are usually homogeneous and the majority of them contain only employment or
population. Also the larger the area is, the higher the value of entropy because the block group will

103

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

contain more jobs. By using a uniform one-mile buffer, we make the unit of analysis comparable for all
block groups.
The two mixed use variables were calculated for each block groups buffer using block-level population
data from the 2010 Census, and block-level employment data from the 2010 LED database. The resulting
job and population totals were used to compute a job-population balance measure (jobpop). This
variable equals 1 for block groups with the same ratio of jobs-to-residents within the one-mile ring as
the urbanized area as a whole; 0 for block groups with only jobs or residents within the one-mile ring,
not both; and intermediate values for intermediate cases. All values were weighted by the sum of block
group jobs and residents as a percentage of the UZA total.
For the second mixed-use variable, each block group centroid was again buffered with a one-mile ring,
and jobs by sector were summed for blocks within the ring. An entropy formula was then used to
compute a measure of job mix (jobmix). The variable equals 1 for block groups with equal numbers of
jobs in each sector within the ring; 0 for block groups with all jobs in a single sector within the ring; and
intermediate values for intermediate cases. The sectors considered in this case were retail,
entertainment, health, education, and personal services. Values were weighted by the sum of block
group population and employment as a percentage of the urbanized areas total. 13
Unlike the mixed use factors at the county and metropolitan levels, the mixed use factor at the
urbanized area level does not include a third variables, Walk Score. The reason is simple. This is
longitudinal comparison of sprawl in 2000 and 2010, and Walk Score data were not available until 2007.
Activity Centering
Centering is the compactness dimension with the most significant improvement compared to Ewing et
al.s indices, as we identified not only central business districts but also employment sub-centers
nationwide as a part of this study. Ewing et al. (2002) measured metropolitan centering in terms of
concentrations of employment in or around (within three miles) historic central business districts (CBDs)
of metropolitan areas, and at a considerable distance from (more than ten miles) from historic CBDs.
This way of measuring centering does not make much sense when applied to small urbanized areas
because almost all of the jobs and population fall within three miles of CBDs. It also doesnt make sense
in large polycentric metropolitan areas, where historic CBDs have sometimes been eclipsed by edge
cities.
The first centering variable came straight out of Ewing et al. (2002, 2003) and the 2010 census. It is the
coefficient of variation in census block group population densities, defined as the standard deviation of
block group densities divided by the average density of block groups (varpop). The more variation in

13

See land use mix section for the formula used for computing job-population balance and job mix measures.

104

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

population densities around the mean, the more centering and/or subcentering exists within the
urbanized areas.
The second centering variable is analogous to the first, except it is derived with employment data from
the LED database. It is the coefficient of variation in census block group employment densities, defined
as the standard deviation of block group densities divided by the average density of block groups
(varemp). The more variation in employment densities around the mean, the more centering and/or
subcentering exists within the urbanized areas.
Ewing et al. (2002) used a 1980 shapefile as published by Census Bureau to identify CBDs. The shapefile
identified the location of CBDs only for 232 of the largest metropolitan areas. The shapefile is now more
than 30 years old and does not capture new CBDs that have emerged since 1980. Moreover, large subcenters are found in virtually all large metropolitan areas of today, and the existence of and proximity to
these are what distinguish polycentric urban areas.
The next two centering variables measure the proportion of employment and population within CBDs
and employment sub-centers. We first identified the location of CBDs and employment sub-centers for
all metropolitan areas (see Chapter 3). Using this procedure, we found 224 metropolitan areas to be
monocentric (have only one center), 132 to be polycentric (have more than one center), and 18
metropolitan areas to have no CBD or sub-center. This procedure resulted in two new centering
variables as the percentage of UZA population (popcen) and employment (empcen) in CBDs and subcenters.
Street Accessibility
Street accessibility is related to block size since smaller blocks translate into shorter and more direct
routes. Large block sizes indicate a relative paucity of street connections and alternate routes. So, two
street accessibility variables were computed for each urbanized area: average block size (avgblk) and
percentage of blocks with areas less than 1/100 square mile, the size of a typical traditional urban block
bounded by sides just over 500 feet in length (smlblk).
These two variables were part of Ewing et al.s original sprawl metrics. To them, we have added two
new variables. They are intersection density and percentage of 4-or-more way intersections.
For each UZA, the total number of intersections in the UZA was divided by the land area to obtain
intersection density (intden), while the number of 4-or-more-way intersections was multiplied by 100
and divided by the total number of intersections to obtain the percentage of 4-or-more way
intersections (4way).

105

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Statistical Methods
In this study we use two statistical methods. Principal component analysis (a type of factor analysis) is
used to derive individual compactness indices that represent the built environments of UZAs. Then
linear regression analysis is used to relate these indices to transportation outcomes, controlling for
influences other than the built environment.
For each dimension of sprawl, principal components were extracted from the component variables. The
principal component selected to represent the dimension was the one capturing the largest share of
common variance among the component variables, that is, the one upon which the observed variables
loaded most heavily. Because, in this study, the first component captured the majority of the combined
variance of these variables, no subsequent components were considered.
The other statistical method used in this study is linear regression (ordinary least squares or OLS). Our
dependent variables were logged so as to be normally distributed and hence properly modeled with
regression analysis. As for the independent variables (control variables), we transformed all variables
into log form to achieve a better fit with the data, reduce the influence of outliers, and adjust for
nonlinearity of the data. The transformations have the added advantage of allowing us to interpret
regression coefficients as elasticities. An elasticity is a percentage change in one variable that
accompanies a one percent change in another variable. Elasticities are the most common measures of
effect size in both economics and planning.

Results
Individual Compactness/Sprawl Factors
Factor loadings (that is, correlations of these variables with each factor), eigenvalues, and percentages
of explained variance are shown in Table 9.1. The eigenvalue of the density factor is 3.82, which means
that this one factor accounts for more of the total variance in the datasets than three component
variables combined, more than three quarters of the total variance. As expected, one of the variables
loads negatively on the density factor, that being the percentage of population living at less than 1,500
persons per square mile. The rest load positively. Thus, for all component variables, higher densities
translate into higher values of the density factor.
The eigenvalue of the mix factor is 1.54, which means that this one factor accounts for more than three
quarters of the total variance in the dataset. Both component variables load positively on the mix factor.
The eigenvalue of the centering factor is 2.20, which means that this one factor accounts for just over
half of the total variance in the datasets. All component variables load positively on the centering factor.
The eigenvalue of the street factor is 2.75, which means that this one factor accounts for two-thirds of
the total variance in the dataset. As expected, one of the variables loads negatively on the street

106

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

accessibility factor, that being the average block size. The rest load positively. Thus, for all component
variables, more street accessibility translates into higher values of the street factor.
Table 9.1. Variable Loadings on Four Factors for 2010
Component Matrix
Density Factor
popden
empden
lt1500
gt12500

gross population density


gross employment density
percentage of the population living at low suburban
densities
percentage of the population living at medium to high
urban densities
net population density of urban lands

urbden
Eigenvalue
Explained variance

Component Matrix
Mix use Factor
jobpop
job-population balance
jobmix
degree of job mixing (entropy)
Eigenvalue
Explained variance
Component Matrix

Data Sources

Factor
Loadings

Census 2010
LED 2010
Census 2010

0.970
0.891
-0.806

Census 2010

0.745

NLCD

0.941
3.82
76.5%

Data Sources

LED 2010
LED 2010

Data Sources

Factor
Loadings
0.879
0.879
1.54
77.2%
Factor
Loadings

Centering Factor

varpop
varemp

coefficient of variation in census block group


population densities
coefficient of variation in census block group
employment densities
percentage of UZA population in CBD or sub-centers
percentage of UZA employment in CBD or sub-centers

popcen
empcen
Eigenvalue
Explained variance

Component Matrix

Census 2010

0.661

LED 2010

0.749

Census 2010
LED 2010

0.757
0.790
2.20
54.8%

Data Sources

Factor
Loadings

Census 2010
Census 2010
TomTom 2007
TomTom 2007

0.844
-0.947
0.726
0.784

Street Factor

smlblk
avgblksze
intden
4way

percentage of small urban blocks


average block size
intersection density
percentage of 4-or-more-way intersections

107

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Eigenvalue
Explained variance

2.75
68.8%

Overall Compactness/Sprawl Index for 2010


Some of the technical literature on sprawl includes size in the definition. Certainly, sheer geographic size
is central to popular notions of sprawl. Despite their relatively high densities, urbanized areas such as
Los Angeles and Phoenix are perceived as sprawling because they go on forever. A sprawl index that
disregarded this aspect of urban form would never achieve face validity.
Accordingly, we sought a method of transforming the sum of the four sprawl factors into a sprawl index
that would be neutral with respect to population size. In this study, we use the exact same procedure
used with metropolitan area sprawl in the early 2000s (Ewing et al. 2002; Ewing et al. 2003). The
transformation was accomplished by regressing the sum of the four sprawl factors on the natural
logarithm of the population of the urbanized area. The standardized residuals (difference between
actual and estimated values divided by the standard deviation of the difference) became our overall
measure of sprawl. Given the way it was derived, this index is uncorrelated with population. Urbanized
areas that are more compact than expected, given their population size, have positive values.
Urbanized areas that are more sprawling than expected, again given their population size, have negative
values. This adjustment for population size still leaves the sprawl index highly correlated with the sum
of the four component factors (r = 0.87).
As with the individual sprawl factors, we transformed the overall sprawl index (index) into a metric with
a mean of 100 and a standard deviation of 25. This was done for the sake of consistency and ease of
understanding. With this transformation, the more compact urbanized areas have index values above
100, while the more sprawling have values below 100. Table 9.3 presents overall compactness scores
and individual component scores for the ten most compact and the ten most sprawling large urbanized
areas. By these metrics, San Francisco is the most compact large urbanized area, and Atlanta is the most
sprawling.
Table 9.3. Compactness/Sprawl Scores for 10 Most Compact and 10 Most Sprawling UZAs in 2010
Ten Most Compact UZAs
comfac
denfac
mixfac
cenfac
strfac
Rank
1
San Francisco-Oakland, CA
180.94
205.69
129.92
164.34
153.38
2
Reading, PA
169.32
127.71
150.87
124.45
147.46
3
Madison, WI
152.87
118.16
121.82
182.19
99.33
4
Eugene, OR
152.54
114.84
134.37
134.15
123.07
5
Laredo, TX
151.80
123.87
131.21
81.56
166.54
6
Oxnard, CA
146.19
147.55
137.14
82.42
135.08
7
Atlantic City, NJ
144.25
93.87
91.07
157.06
143.86
8
Los Angeles-Long Beach-Anaheim, CA
143.42
212.21
144.75
102.23
138.92

108

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Rank

comfac
denfac
mixfac
cenfac
strfac
143.38
118.63
127.46
97.02
141.77
142.71
197.50
106.80
179.10
125.06

Rank

comfac
denfac
mixfac
cenfac
strfac
64.38
81.92
75.30
77.21
77.61
61.05
79.40
73.65
67.16
64.43
60.96
68.92
54.18
97.03
70.33
60.57
67.92
75.26
89.88
57.88
60.27
87.51
47.43
111.18
70.03
57.41
82.95
64.56
115.94
53.01
55.56
66.31
68.97
88.15
54.29
54.15
82.38
67.79
57.01
61.88
48.64
46.92
78.41
72.20
44.94
37.45
84.64
75.63
107.29
36.84

9
Lincoln, NE
10
New York-Newark, NY-NJ-CT
Ten Most Sprawling UZAs
153
154
155
156
157
158
159
160
161
162

Baton Rouge, LA
Fayetteville, NC
Chattanooga, TN-GA
Greenville, SC
Nashville-Davidson, TN
Charlotte, NC-SC
Winston-Salem, NC
Victorville-Hesperia, CA
Hickory, NC
Atlanta, GA

Overall Compactness/Sprawl Index for 2000


To make apples to apples comparisons between two years (2000 and 2010), we applied the factor
coefficient matrices for four principal components in 2010 to built environmental variable values for
2000. This resulted in compactness factors for 2000 that are consistent with those for 2010.
Table 9.4 presents overall compactness scores and component scores for the ten most compact and the
ten most sprawling large urbanized areas in 2000. As one would expect, rankings did not change
dramatically in most cases over the ten years. San Francisco was the most compact in 2000, and has
remained so. Atlanta was the most sprawling in 2000, and has remained so.
Table 9.4. Compactness/Sprawl Scores for 10 Most Compact and 10 Most Sprawling UZAs in 2000
Ten Most Compact UZAs
comfac
denfac
mixfac
cenfac
strfac
Rank
1
San Francisco-Oakland, CA
184.06
219.66
128.39
162.41
149.84
2
Laredo, TX
174.12
134.65
148.02
86.2
189.55
3
Reading, PA
155.74
119.44
157.15
126.12
118.53
4
Eugene, OR
151.42
121.5
141.47
130.73
114.89
5
New Orleans, LA
149.64
161.24
106.84
95.97
181.06
6
Stockton, CA
147.55
134.42
145.18
104.41
124.09
7
Madison, WI
147.2
122.06
126.86
158.37
101.3
8
Visalia, CA
145.05
116.84
142.48
107.53
108.93
9
New York-Newark, NY-NJ-CT
141.75
197.18
115.6
170.57
120.19
10
Lincoln, NE
141.19
118.03
133.12
97.15
135.15
Ten Most Sprawling UZAs
comfac
denfac
mixfac
cenfac
strfac
Rank
153 Fayetteville, NC
64.13
78.97
98.97
62.63
56.65
154 Baton Rouge, LA
61.39
83.46
72.66
85.07
64.16

109

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Rank

155
156
157
158
159
160
161
162

Palm Bay-Melbourne, FL
Nashville-Davidson, TN
Victorville-Hesperia, CA
Winston-Salem, NC
Bonita Springs, FL
Chattanooga, TN-GA
Hickory, NC
Atlanta, GA

comfac
denfac
mixfac
cenfac
strfac
58.18
76.29
75.93
62.16
77.64
58.11
89.26
67.83
106.22
46.1
55.43
74.79
84.24
56.75
51.04
53.49
66.67
68.56
93.67
44.02
52.49
76.78
77.85
61.38
46.22
49.7
65.83
55.21
92.3
53.9
48.76
49.14
81.34
75.33
42.67
39.5
88.54
90.28
106.29
19.9

Validation Against Transportation Outcomes


The relationships between sprawl and travel outcomes can be used to validate the urbanized area
sprawl measures. If sprawl has any consistently recognized outcome, it is automobile dependence. We
would expect to find that, after controlling for other relevant influences, compact urbanized areas have
relatively high transit and walking commute mode shares and short drive times to work.
Table 9.5 provides a list of all dependent and independent variables used in the validation. With
American Factfinder, we downloaded data from the 2010 American Community Survey (ACS), 5-year
estimates, and computed walk and transit mode shares and average drive times for census urbanized
areas. From the 2010 Census, we downloaded data on gender, age, race, household size, and computed
percentages and mean values to describe urbanized area populations.
Table 9.5. Variables Used to Explain Travel Outcomes (all variables log transformed)
Dependent variables
Variables
Data Sources
walkshr
percentage of commuters walking to work
ACS 2007-2011
transitshr percentage of commuters using public transportation (excluding taxi) ACS 2007-2011
drivetime average journey-to-work drive time in minutes
Independent variables
Variables
pop
urbanized area population
hhsize
average household size
age1524 percentage of population 15-24 years old
male
percentage of male population
white
percentage of white population
income
income per capita
fuel
average fuel price for the metropolitan area
index
urbanized areas compactness index for 2010
denfac
density factor (a weighted combination of 5 density variables)
mixfac
mix factor (a weighted combination of 2 mixed-use variables)
cenfac
centering factor (a weighted combination of 4 centering variables)

ACS 2007-2011
Data Sources
Census 2010
Census 2010
Census 2010
Census 2010
Census 2010
ACS 2007-2011
OPIC 2010
computed
computed
computed
computed

110

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Variables
strfac

street factor (a weighted combination of 4 street-related variables)

Data Sources
computed

We estimated two sets of regression models. Our first set of regressions used the overall compactness
index for 2010 (index) as an independent variable. The second set of regressions used the four
compactness/sprawl factors individually (denfac, mixfac, cenfac, strfac) as independent variables. Both
sets of regressions included control variables.
Also, in both sets of regressions, the dependent variables were logged so as to be normally distributed
and hence properly modeled with ordinary least squares (OLS) regression. All independent variables
were also transformed into log form to achieve a better fit with the data, reduce the influence of
outliers, and adjust for nonlinearity of the data. The transformations had the added advantage of
allowing us to interpret regression coefficients as elasticities.
Results for models with the overall compactness index are presented in Table 9.6. Control variables
mostly have the expected signs and often are significant. The compactness index (index) has the
expected strong positive relationships to walk and transit mode shares, and the expected negative
relationship to average drive time. As noted, the coefficients of these log-log models are elasticities.
For every percentage increase in the compactness index, the walk mode share increases by 0.74
percent, the transit mode share increases by 1.04 percent, and the average drive time declines by 0.12
percent. Based on these results, we declare the overall compactness index to be validated.
Table 9.6. Relationships of UZA Compactness Index in 2010 to Transportation Outcomes (log-log
transformed t-statistics in parentheses)
walkshr
transitshr
drivetime
constant

2.44

-6.47

2.12

pop

.13
(3.80)***
-1.21
(-3.16)**
1.16
(8.18)***
-3.12
(-1.74)
0.45
(2.66) **
0.065
(0.31)
1.17
(1.93)*

.42
(7.706)***

.033
(3.11)**

1.18
(1.87)
0.86
(3.68)***
-7.33
(-2.48) *
0.17
(0.62)
1.83
(5.42)***
2.66
(2.67)**

0.89
(7.19)***
-0.13
(-2.79)**
-0.92
(-1.59)
-0.084
(-1.53)
0.39
(5.87)***
0.52
(2.69)**

hhsize
age1524
male
white
income
fuel

111

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

index
adjusted R2

walkshr

transitshr

drivetime

0.74
(6.38)***
0.53

1.04
(5.44)***
0.62

-0.12
(-3.17)***
0.61

* .05 probability level


** .01 probability level
*** .001 probability level
The new multidimensional compactness factors are mostly significant with the expected signs (see Table
9.7). The centering factor, cenfac, is the most important correlate of walking, followed by the mix
factor, mixfac, and the density factor, denfac. The street factor has a negative sign but is not significant.
The centering factor is also the most important correlate of transit use, followed by the density factor.
The other two compactness factors are not significant. All of the compactness factors have negative
signs in the drive time equation, but only the density factor is significant. We declare the density and
centering factors to be validated. The mix factor has promise based on its correct sign and significance in
the walk share equation. The street factor is never significant and has the opposite signs to what we
expected. Given the fact that all four street variables load as expected on the street factor, we are
confident that we properly operationalized street accessibility. But we have no explanation for the
disappointing regression results.
Table 9.7. Relationships of 2010 UZA Individual Compactness Factors to Transportation Outcomes (loglog transformed t-statistics in parentheses)
walkshr
transitshr
drivetime
constant
pop
hhsize
age1524
male
white
income
fuel
denfac
mixfac

3.76
-0.015
(-.37)
-0.93
(-2.21)*
0.89
(5.78)***
-3.18
(-1.85)
0.46
(2.83)**
-0.18
(-.85)
0.91
(1.52)
0.53
(2.03)*
0.36
(2.98)**

.326
0.16
(2.23)*
1.33
(1.88)
0.41
(1.59)
-8.74
(-3.05)**
0.29
(1.08)
1.50
(4.28)***
2.39
(2.39)*
1.33
(3.05)***
0.087
(0.43)

1.62
0.056
(3.94)***
0.99
(6.86)***
-0.094
(-1.78)
-0.904
(-1.54)
-0.087
(-1.56)
0.43
(5.96) ***
0.64
(3.09)**
-0.17
(-1.92)*
-0.057
(-1.38)

112

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

cenfac
strfac
adjusted R2

walkshr

transitshr

0.79
(6.05)***
-0.08
(-0.63)
0.60

1.06
(4.93)***
-0.082
(-0.37)
0.66

drivetime
-0.026
(-0.59)
0.016
(0.35)
0.61

* .05 probability level


** .01 probability level
*** .001 probability level
Attempted Validation Using Change Variables
As noted in the introduction, a longitudinal analysis is a more stringent test of association/causation
than is a cross sectional analysis. In our longitudinal analyses, we related changes in compactness and
socioeconomic variables between 2000 and 2010 to changes in commuting outcomes during the same
time period. Such analyses control for unmeasured characteristics of urbanized areas that may confound
results in cross sectional analyses, but largely cancel out in longitudinal analyses.
If our compactness measures are valid, we would expect to see some change in transportation
outcomes associated with a change in compactness measures for individual urbanized areas. However,
we would expect much weaker relationships in longitudinal than cross sectional analyses for two
reasons: first, changes over a decade are small compared to sampling errors associated with commuting
data from the American Community Survey, and relatedly, we can only hope to capture short term
effects of compactness on commuting in a 10-year longitudinal analysis.
We looked at changes in three commuting variableswalk mode share, transit mode share, and average
drive time on the journey to work. The most significant relationship is between the change in the
density factor and the change in the transit mode share (p=<0.001). The other significant relationship is
between the change in the centering factor and the change in the transit mode share (p=0.014). These
relationships mirror those of the cross sectional analysis. No other relationships proved significant in any
of our models.

Discussion
This chapter developed and sought to validate an overall measure of compactness/sprawl for U.S.
urbanized areas in 2010. It also sought to validate individual factors representing four dimensions of
compactness/sprawl: development density, land use mix, activity centering, and street accessibility. By
these measures, San Francisco is the most compact urbanized area in the nation, and Atlanta is the most
sprawling.

113

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

We were able to validate the overall index against transportation outcome measures, specifically walk
mode share, transit mode share, and average drive time on the journey to work. We were also able to
validate the factors representing density and centering. The factor representing land mix was less
predictive of transportation outcomes, but did have a significant relationship to walking, as one would
expect. Only the factor representing street accessibility could not be validated.
Once we had measures of compactness for 2010, we were able to apply the same factor coefficients to
data for 2000, thus generating consistent measures of compactness for 2000 and allowing longitudinal
comparisons. Generalizing across the entire universe of large urbanized areas, compactness decreased
and sprawl increased between the two census years, but only slightly. Summing the four indices of
compactness (each with an average score of 100 in 2010), the average combined score was 405.8 in
2000, dropping to 400 in 2010, a relatively small change. This means that that on average, urbanized
areas became less compact between 2000 and 2010. However, the average case obscures some
interesting stories.
Los Angeles saw an impressive rise in its compactness ranking, from 18th in 2000 to 8th in 2010. The
reason: infill development. Contradicting metropolitan L.A.s reputation as the capital of unbridled
sprawl, roughly two-thirds of new housing built there between 2005 and 2009 was infill constructed in
previously developed areas rather than on raw land in the exurbs (Boxall , 2012).
New Orleans dropped seven places between 2000 and 2010, from 5th to 12th most compact, which is a
result of population loss following Katrina. The drop would have been greater in the immediate
aftermath of the hurricane. The magnitude of population loss after Katrina means the city has
needed to add relatively few people to rank among the fastest-growing places in the nation since 2007.
It also has been aided by billions in federal reconstruction dollars and a healthy tourism industry (Bass,
2012).
Stockton was 6th most compact in 2000 and dropped to 16th in 2010. This is presumably due to the
housing bubble, which burst in Stockton perhaps more than any other city in the nation. Stockton has
been referred to as Foreclosureville due to its record high home foreclosure rate (Nieves, 2010).
Charlotte rose from 11th most sprawling to 5th most sprawling during the decade, now competing to be
the next Atlanta. A recent study by Northwest Environment Watch, a nonprofit research and
communication center based in Seattle, shows Charlotte as having the worst urban sprawl out of 15
cities across the nation. Charlotte ranked last in every category, including land converted from rural area
to suburban and urban areas, average metropolitan density and growth in compact neighborhoods
(Tompkins, 2009).
So the compactness/sprawl measures have the additional quality of face validity. They paint a plausible
picture of sprawl in the U.S.

114

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Chapter 10. Conclusion


This study has updated a county and metropolitan compactness/sprawl indices, widely used by planning
and public researchers since their release in 2002 and 2003. The updated indices reflect conditions on
the ground circa 2010.
This study has also developed new measures of compactness/sprawl that incorporate additional
dimensions of the construct sprawl, and used additional variables to operationalize these dimensions.
The four dimensions, measured individually and with a composite index, are development density, land
use mix, activity centering, and street accessibility. Measures, presented in the Appendices, are
immediately available to study the costs and benefits of different urban forms. They will be posted on
an NIH website and available for download.
Using updated and enhanced measures of compactness/sprawl, this study has validated both the
original and new indices, and largely validated the individual measures representing the four dimensions
of sprawl. These new results mirror and confirm the earlier findings of Ewing et al. (2002, 2003a, 2003b,
2003c). If anything, relationships of sprawl to important quality-of-life outcomes are stronger than in
the original studies.
An obvious question is whether the new measures have more face, construct, and internal validity than
the original compactness/sprawl indices, and thus should substitute for the original indices in future
research. They have more face validity because places that fit the definition of sprawl in satellite
imagery rank lowest on compactness. They have more construct validity because they capture more
aspect of sprawl. As for internal validity, they generally outperform the original county sprawl indices as
predictors of negative outcomes.
The new multi-dimensional factors representing density, mix, centering, and streets are somewhat
correlated, of course, but still quite distinct in their relationships to outcomes. We can see these being
used to determine which specific aspects of sprawl result in costs and benefits.

115

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

References
Anselin, L. (1995). Local indicators of spatial associationLISA. Geographical analysis, 27(2), 93-115.
Anthony, J. (2004). Do state growth management regulations reduce sprawl? Urban Affairs Review,
39(3), 376-397.
Badoe, D. A. & Miller, E. J. (2000). Transportationland-use interaction: Empirical findings in North
America, and the implications for modeling. Transportation Research D, 5(4), 235-263.
Bass, F. (2012). Katrina Comeback Makes New Orleans Fastest-Growing City. Bloomberg, 27 June,
http://www.bloomberg.com/news/2012-06-28/katrina-comeback-makes-new-orleans-fastest-growingcity.html.
Baumont, C., Ertur, C., Le Gallo, J. (2003). 4 Spatial Convergence Clubs and the European Regional
Growth Process, 1980. European regional growth, 131.
Bereitschaft, B., and K. Debbage. 2013. Urban Form, Air Pollution, and CO2 Emissions in Large U.S.
Metropolitan Areas. The Professional Geographer, 65(4), 612635.

Black, J. L., & Macinko, J. (2008). Neighborhoods and obesity. Nutrition Reviews, 66(1), 2-20.
Boxall, B. (2012). Infill Housing Development Rises in Los Angeles Region. Los Angeles Times, 20
December, http://articles.latimes.com/2012/dec/20/science/la-sci-sn-infill-housing-increases20121220.
Bruegmann, R. (2006). Sprawl: A compact history. University of Chicago press, Chicago, IL.
Burchell, R.W., Lowenstein, G., Dolphin, W.R., Galley, C.C., Downs, A., Seskin, S., Gray Still, K., Moore, T.
(2002). Costs of Sprawl 2000. TCRP Report 74, Transit Cooperative Research Program, Transportation
Research Board, Washington, D.C.
Burchell, R.W., Shad, N.A., Listokin, D., Phillips, H., Seskin, S., Davis, J.S., Moore, T., Helton, D., Gall, M.
(1998). The Costs of Sprawl-Revisited. Transportation Research Board Report 39, National Academy
Press, Washington, D.C.
Cao, X., Mokhtarian, P. L., & Handy, S. L. (2009a). Examining the impacts of residential self-selection on
travel behaviour: A focus on empirical findings. Transport Reviews, 29(3), 359-395.
Casey, R., Oppert, J.-M., Weber, C., Charreire, H., Salze, P., Badariotti, D., . . . Simon, C. (2011).
Determinants of childhood obesity: What can we learn from built environment studies? Food Quality
and Preference.
CDC (2009). Fact Sheet for Health Professionals on Physical Activity Guidelines for Adults. In CDC (Ed.):
Centers for Disease Control and Prevention,Division of Nutrition, Physical Activity, and Obesity.

116

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Cervero, R. (2001). Walk-and-ride: Factors influencing pedestrian access to transit. Journal of Public
Transportation, 3(4), 1-23.
Cervero, R. (2003). The built environment and travel: Evidence from the United States. European Journal
of Transport and Infrastructure Research, 3(2), 119-137.
Cho, S., Z. Chen, S.T. Yen, and D.B. Eastwood. 2006. The Effects of Urban Sprawl on Body Mass Index:
Where People Live Does Matter. The 52nd Annual ACCI Conference, Baltimore, Maryland, March 1518.
Committee on Physical Activity, H., Transportation,and Land Use. (2005). Does the Built Environment
Influence Physical Activity? Examining the Evidence -- Special Report 282. Transportation Research
Board Institute of Medicine.
Crane, R. (2000). The influence of urban form on travel: An interpretive review. Journal of Planning
Literature, 15(1), 3-23.
Cutsinger, J., Galster, G., Wolman, H., Hanson, R., Towns, D. (2005). Verifying the MultiDimensional
Nature of Metropolitan Land Use: Advancing the Understanding and Measurement of Sprawl. Journal of
Urban Affairs, 27(3), 235-259.
Ding, D., & Gebel, K. M. (2012). Built environment,physical activity,and obesity:What have we learned
from reviewing the literature? Health &Place, 18, 100-105.
Doyle, S., A. Kelly-Schwartz, M. Schlossberg, and J. Stockard. 2006. Active Community Environments and
Health: The Relationship of Walkable and Safe Communities to Individual Health. Journal of the
American Planning Association. 72(1), 1931.
Duany, A., Plater-Zyberk, E., Speck, J. (2001). Suburban nation: The rise of sprawl and the decline of the
American dream. North Point Press.
Dumbaugh, E., & Li, W. (2011). Designing for the safety of pedestrians, cyclists, and motorists in the built
environment. Journal of the American Planning Association, 77(1), 69-88.
Dunton, G. F., Kaplan, J., Wolch, J., Jerrett, M., & Reynolds, K. D. (2009). Physical environmental
correlates of childhood obesity: a systematic review. Obesity reviews, 10, 393-402.
Durand, C. P., Andalib, M., Dunton, G. F., Wolch, J., & Pentz, M. A. (2011). A systematic review of built
environment factors related to physical activity and obesity risk: implications for smart growth urban
planning. Obesity Revews, 12(173-182).
Ewing, R. (1997). Is Los Angeles-style sprawl desirable? Journal of the American Planning Association,
63(1), 107-126.

117

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Ewing, R. and Cervero, R. (2001). Travel and the built environment. Transportation Research Record,
1780, 87-114.
Ewing R, Pendall R, Chen D. 2002. Measuring Sprawl and Its Impacts. Washington, DC: Smart Growth
America.
Ewing, R., R. Pendall, and D. Chen. 2003a. Measuring Sprawl and Its Transportation Impacts.
Transportation Research Record. 1832: 175-183.
Ewing R.,T. Schmid, R. Killingsworth, A. Zlot, and S. Raudenbush. 2003b. Relationship Between Urban
Sprawl and Physical Activity, Obesity, and Morbidity. American Journal of Health Promotion.18:47-57.
Ewing, R., R. Schieber and C. Zegeer. 2003c. Urban Sprawl as a Risk Factor in Motor Vehicle Occupant
and Pedestrian Fatalities, American Journal of Public Health 93: 1541-1545.
Ewing R., R. Brownson, and D. Berrigan. 2006. Relationship Between Urban Sprawl and Weight of U.S.
Youth. American Journal of Preventive Medicine. 31: 464-474.
Ewing, R. and F. Rong. 2008. Impact of Urban Form on U.S. Residential Energy Use. Housing Policy
Debate. 19: 1-30.
Ewing, R., Cervero, R.: Travel and the built environment: a meta-analysis. J. Am. Plan. Assoc. 76(3), 265
294 (2010)
Fan, Y. and Y. Song. 2009. Is Sprawl Associated with a Widening UrbanSuburban Mortality Gap? Journal
of Urban Health: Bulletin of the New York Academy of Medicine. 86(5): 708-728.
Feng, J., Glass, T., Curriero, F., Stewart, W., & Schwartz, B. (2010). The built environment and obesity: A
systematic review of the epidemiologic evidence. Health and Place, 16, 175-190.
Finkelstein, E., Fiebelkorn, I., & Wang, G. (2005). The Costs of Obesity Among Full-time Employees.
American Journal of Health Promotion, 20(1), 45-53.
Frenkel, A., and Ashkenazi, M. 2008. Measuring urban sprawl: how can we deal with it?. Environment
and Planning B: Planning and Design, 35(1), 56-79.
Fulton, W. Pendall, R, Nguyen. M. and Harrison, A. 2001. Who Sprawls Most? How Growth Patterns
Differ Across the U.S., Center for Urban & Metropolitan Policy, The Brookings Institution, Washington,
D.C.
Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., and Wickham, J., 2011.
Completion of the 2006 National Land Cover Database for the Conterminous United States, PE&RS, Vol.
77(9):858-864.

118

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Galster, G., Hanson, R., Ratcliffe, M.R., Wolman, H., Coleman, S., Freihage, J. (2001). Wrestling sprawl to
the ground: defining and measuring an elusive concept. Housing policy debate, 12(4), 681-717.
Galvez, M. P., Meghan, P., & Yen, I. (2010). Childhood Obesity and the Built Environment: A Review of
the Literature from 2008-2009. Current Opinion in Pediatrics, 22(2), 202-207.
Getis, A., and Ord, J.K. (1996). Local spatial statistics: an overview. Spatial analysis: modelling in a GIS
environment, Wiley, Inc., 261-279.
Glaeser, E. Kahn, M. and Chu, C. (2001). Job Sprawl: Employment Location in U.S. Metropolitan Areas,
Center for Urban & Metropolitan Policy, The Brookings Institution, Washington, D.C.
Goldstein, H. (2011). Multilevel Statistical Models Wiley Series in Probablity and Statistics
Gordon, P., and Richardson, H.W. (1997). Are compact cities a desirable planning goal? Journal of the
American Planning Association, 63(1), 95-106.
Griffin, B.A., Eibner, C., Bird, C.E., Jewell, A., Margolis, K., Shih, R., Escarce, J.J. (2013). The Relationship
between Urban Sprawl and Coronary Heart Disease in Women. Health & place, 20, 5161.
Hadayeghi, A., A.S. Shalaby, and B.N. Persaud. 2003. Macro Level Accident Prediction Models for
Evaluating Safety of Urban Transportation Systems. Transportation Research Record: Journal of the
Transportation Research Board, 1840, pp. 87-95
Hadayeghi, A., A.S. Shalaby, B.N. Persaud, and C. Cheung.2006. Temporal Transferability and Updating of
Zonal Level Accident Prediction Models. Accident Analysis and Prevention, 38:3, pp. 579-589.
Handy, S. L. (2005). Critical assessment of the literature on the relationships among transportation, land
use, and physical activity. (Special Report 282: Does the built environment influence physical activity?
Examining the evidence.) Washington, DC: Transportation Research Board and Institute of Medicine
Committee on Physical Activity, Health, Transportation, and Land Use.
Hayden D, 2004 A field guide to sprawl (WW Norton & Company)
Heath, G. W., Brownson, R. C., Kruger, J., Miles, R., Powell, K. E., Ramsey, L. T., & the Task Force on
Community Preventive Services. (2006). The effectiveness of urban design and land use and transport
policies and practices to increase physical activity: A systematic review. Journal of Physical Activity and
Health, 3(1), 55-76.
Hirschhorn, J.S. (2005). Sprawl Kills: how bland burbs steal your time, health and money (Sterling & Ross)
Holcombe, R.G., and Williams, D.W. (2012). Urban Sprawl and Transportation Externalities. The Review
of Regional Studies, 40(3), 257-272.

119

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Holzer, H.J. (1991). The Spatial Mismatch Hypothesis: What Has the Evidence Shown? Urban Studies,
28:105122.

Huang, J., Lu, X.X., Sellers, J.M. (2007). A global comparative analysis of urban form: Applying spatial
metrics and remote sensing. Landscape and Urban Planning, 82(4), 184-197.
Humpel, N., Owen, N., & Leslie, E. (2002). Environmental factors associated with adults participation in
physical activity. American Journal of Preventive Medicine, 26, 119125.
Ihlanfeldt, K. (1994). The spatial mismatch between jobs and residential locations within urban areas.
Cityscape, 1(1), 219-244.
Jaeger, J.A., Bertiller, R., Schwick, C., Kienast, F. (2010). Suitability criteria for measures of urban
sprawl. Ecological Indicators, 10(2), 397-406.
Jencks, C. and Mayer, S.E. (1990). Residential Segregation, Job Proximity, and Black Job Opportunities. In
Inner-City Poverty in the United States, edited by L.E. Lynn, Jr., and M.M. McGreary, 187222.
Washington, D.C.: National Academic Press.
Joshu, C. E., Boehmer, T. K., Brownson, R. C., & Ewing, R. (2008). Personal, neighbourhood and urban
factors associated with obesity in the United States. Journal of Epidemiology and community Health, 62,
202-208.
Kaczynski, A. T., & Henderson, K. A. (2007). Environmental Correlates of Physical Activity: A Review of
Evidence about Parks and Recreation. Leisure Sciences, 29, 315-354.
Kahn, M.E. 2006. The Quality of Life in Sprawled versus Compact Cities, prepared for the OECD ECMT
Regional Round, Berkeley, California, March 2006, Table 137, 2728.
Kahn, M.E. (2006). Green cities: urban growth and the environment. Brookings Institution Press
Washington, D.C., p 160.
Kain, J.F (1992). The Spatial Mismatch Hypothesis: Three Decades Later. Housing Policy Debate, 3(2),
371462.
Khan, L. K., Sobush, K., Keener, D., Goodman, K., Lowry, A., Kakietek, J., & Zaro, S. (2009). Centers for
Disease Control and Prevention. Recommended Community Strategies and Measurements to Prevent
Obesity in the United States. MMWR, 58(RR-7), 1-29.
Kelly-Schwartz, A., J. Stockard, S. Doyle, and M. Schlossberg (2004). Is Sprawl Unhealthy? A Multilevel
Analysis of the Relationship of Metropolitan Sprawl to the Health of Individuals. Journal of Planning
Education and Research. 24: 184196.

120

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Kim, D., Subramanian, S.V., Gortmaker, S.L., Kawachi, I. (2006). U.S. state-and county-level social capital
in relation to obesity and physical inactivity: a multilevel, multivariable analysis. Social science &
medicine, 63(4), 1045-1059.
Kochanek, K. D., Xu, J., Murphy, S. L., Minio, A. M., & Kung, H. C. (2011). National vital statistics reports.
National Vital Statistics Reports, 59(4), 1.
Kmet, L, P Brasher, & C. Macarthur (2003). A Small Area Study of Motor Vehicle Crash Fatalities in
Alberta, Canada. Accident Analysis and Prevention, 35:2, pp. 177-182.
Kneebone, E. (2009) Job Sprawl Revisited: The Changing Geography of Metropolitan Employment. Metro
Economy Series for the Metropolitan Policy Program at Brookings.
Lachowycz, K., & Jones, A. P. (2011). Greenspace and obesity: a systematic review of the evidence. Obes
Reviews, 12, 183-189.
Ladron de Guevara, F. S.P. Washington, and J. Oh. 2004. Forecasting Travel Crashes at the Planning
Levels: Simultaneous Negative Binomial Crash Model Applied in Tucson, Arizona. Transportation
Research Record: Journal of the Transportation Research Board, 1897, pp. 191-199.
Lang, R. (2003). Edgeless cities: Exploring the elusive metropolis. Brookings Institution Press.
Lee, C., & Moudon, A. V. (2004). Physical Activity and Environment Research in the Health Field:
Implications for Urban and Transportation Planning Practice and Research. Journal of Planning
Literature, 19(2), 147-181.
Lee, I.M., R. Ewing, and H.D. Sesso. 2009. The Built Environment and Physical Activity Levels: The
Harvard Alumni Health Study. American Journal of Preventive Medicine, 37(4): 293-298.
Levine, N., K.E. Kim, and L.H. Nitz. 1995a. Spatial Analysis of Honolulu Motor Vehicle Crashes: I. Spatial
Analysis. Accident Analysis and Prevention, 27:5, pp. 663-674.
Levine, N., K.E. Kim, and L.H. Nitz. 1995b. Spatial Analysis of Honolulu Motor Vehicle Crashes: II. Zonal
Generators. Accident Analysis and Prevention, 27:5, pp. 675-685.
Levy, J.I., Carrothers, T.J., Tuomisto, J.T., Hammitt, J.K., Evans, J.S., 2001. Assessing the public health
benefits of reduced ozone concentrations. Environmental Health Perspectives 109 (12), 920.
Lopez, R., Hynes, H.P. (2003). Sprawl in the 1990s Measurement, Distribution, and Trends. Urban Affairs
Review, 38(3), 325-355.
Lovegrove, G.R. and T. Sayed. 2006. Macro-level Collision Prediction Models for Evaluating
Neighborhood Traffic Safety. Canadian Journal of Civil Engineering, 33:5, pp. 609621.
Lund, H. 2003. Testing the Claims of New Urbanism: Local Access, Pedestrian Travel, and Neighboring
Behaviors. Journal of the American Planning Association, 69(4) 414-429.

121

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Malpezzi, S.,Wen, Kai G. (2001). Measuring ``Sprawl'': Alternative Measures of Urban Form in U.S.
Metropolitan Areas Center for Urban Land Economics Research, University of Wisconsin, Madison, WI.
Marshall, W. E., & Garrick, N. W. (2011). Does street network design affect traffic safety? Accident
Analysis and Prevention, 43(3), 769-781.
Martellozzo, F., Clarke, K.C. (2011). Measuring urban sprawl, coalescence, and dispersal: a case study of
Pordenone, Italy. Environment and Planning B: Planning and Design. 38(6), 1085-1104.
McCormack, G., Giles-Corti, B., Lange, A., Smith, T., Martin, K., & Pikora, T. (2004). An update of recent
evidence of the relationship between objective and self-report measures of the physical environment
and physical activity behaviours. Journal of Science and Medicine in Sport, 7(1), 81-92.
McDonald, N., Trowbridge, M. (2009). Does the built environment affect when American teens become
drivers? Evidence from the 2001 National Household Travel Survey. Journal of Safety Research, 40(3),
177-183.
McMillan, T. E. (2005). Urban form and a child's trip to school: The current literature and a framework
for future research. Journal of Planning Literature, 19(4), 440-456.
McMillan T. E. (2007). The relative influence of urban form on a child's travel mode to school.
Transportation Research A, 41(1), 69-79.
Moss, P. and Tilly, C. (1991). Why Black Men Are Doing Worse in the Labor Market. New York: Social
Science Research Council.
Mubareka, S., Koomen, E., Estreguil, C., Lavalle, C. (2011). Development of a composite index of urban
compactness for land use modelling applications. Landscape and Urban Planning, 103(3), 303-317.
Nieves, E. (2010). Stockton, California Is Foreclosureville, USA, Has One of the Worst Foreclosure Rates
In The United States, Huffington Post Business, 10 January,
http://www.huffingtonpost.com/2010/01/10/stockton-california-is-fo_n_417704.html
Nguyen D, 2010. Evidence of the impacts of urban sprawl on social capital. Environment and Planning B:
Planning and Design 37(4) 610 627.
Owen, N. P., Humpel, N. P., Leslie, E. P., Bauman, A. P., & Sallis, J. F. P. (2004). Understanding
Environmental Influences on Walking Review and Research Agenda. American Journal of Preventative
Medicine, 27(1), 67-76.
Papas, M. A., Alberg, A. J., Ewing, R., Helzlsouer, K. J., Gary, T. L., & C.Klassen, A. (2007). The Built
Environment and Obesity. Epidemiologic Reviews. 29(1), 129-143.
Pendall, R. (1999) Do Land-Use Controls Cause Sprawl? Environment and Planning B. 26(4), 555-571.

122

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Pendall, R., Carruthers, J.I. (2003). Does density exacerbate income segregation? Evidence from US
metropolitan areas, 1980 to 2000. Housing Policy Debate, 14(4), 541-589.
Plantinga, A. and S. Bernell. 2007. The Association between Urban Sprawl and Obesity: Is It a Two-Way
Street? Journal of Regional Science, 47(5): 857879.
Poelmans, L., Van Rompaey, A. (2009). Detecting and modelling spatial patterns of urban sprawl in
highly fragmented areas: A case study in the FlandersBrussels region. Landscape and Urban Planning,
93(1), 10-19.
Pont, K., Ziviani, J., Wadley, D., Bennett, S., & Abbott, R. (2009). Environmental correlates of childrens
active transportation: A systematic literature review. Health & Place, 15(3), 827-840.
Riguelle, F., Thomas, I., Verhetsel, A. (2007). Measuring urban polycentrism: a European case study and
its implications. Journal of Economic Geography, 7(2), 193-215.
Ryan, S. & Frank, L. D. (2009). Pedestrian Environments and Transit Ridership. Journal of Public
Transportation 12(1), 39-57.
Schweitzer, L. and Zhou, J. 2010. Neighborhood Air Quality Outcomes in Compact and Sprawled Regions,
Journal of the American Planning Association, 76(3): 363-371
Saelens, B. E., Sallis, J. F., & Frank, L. D. (2003). Environmental correlates of walking and cycling: Findings
from the transportation, urban design, and planning literatures. Annals of Behavioral Medicine, 25(2),
80-91.
Stead, D., & Marshall, S. (2001). The relationships between urban form and travel patterns. An
international review and evaluation. European Journal of Transport and Infrastructure Research, 1(2),
113-141.
Stone, B. 2008. Urban Sprawl and Air Quality in Large U.S. Cities. Journal of Environmental Management,
86:688-698.
Stone, B., J. Hess, H. Frumkin. 2010. Urban Form and Extreme Heat Events: Are Sprawling Cities More
Vulnerable to Climate Change than Compact Cities? Environmental Health Perspectives, 118(10): 14251428.
Sturm, R. and D. Cohen. 2004. Suburban Sprawl and Physical and Mental Health, Public Health
118(7) 488-496.
Tompkin, K. (2009). Charlotte Sprawl among Worst in Country. The Daily Tar Heel, 22 August,
http://www.dailytarheel.com/article/2004/11/charlotte_sprawl_among_worst_in_country
Torrens, P. (2008). A Toolkit for Measuring Sprawl. Applied Spatial Analysis and Policy, 1(1), 5-36.

123

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Trogdon, J. G., Finkelstein, E. A., Hylands, T., Dellea, P. S. , & Kamal-Bahl, A. S. J. (2008). Indirect costs of
obesity: a review of the current literature. Obes Rev, 9, 489-500.
Trost, S., Kerr, L., Ward, D., & Pate, R. (2001). Physical activity and determinants of physical activity in
obese and non-obese children. International Journal of Obesity, 25, 822-829.
Trost, S. G., Owen, N., Bauman, A. E., Sallis, J. F., & Brown, W. (2002). Correlates of adults? participation
in physical activity: review and update. Medicine and Science in Sports and Exercise 34(12), 1996-2001
Trowbridge, M. J. and N. C. McDonald. 2008. Urban Sprawl and Miles Driven Daily by Teenagers in the
United States. American Journal of Preventive Medicine 34(3): 202-206.
Trowbridge, MJ, Gurka, MJ, OConnor, R. 2009. Urban Sprawl and Delayed Ambulance Arrival in the
United States. American Journal of Preventive Medicine. 37(5), 428-432.
Vancauwenberg, J., De Bourdeaudhuij, I., Demeester, F., Van Dyke, D., Salmon, J., Clarys, P., & Deforche,
B. (2011). Relationship between the physical environment and physical activity in older adults: A
systematic review. Health &Place, 17, 458-469.
Wheeler, L.A. (1990). A Review of the Spatial Mismatch Hypothesis: Its Impact on the Current Plight of
Central City in the United States. Occasional paper no. 137. Metropolitan Studies Program, The Maxwell
School of Citizenship and Public Affairs, Syracuse University.
Withrow, D., & Alter, D. A. (2010). The economic burden of obesity worldwide: a systematic review of
the direct costs of obesity. Obes Rev, 12, 131-141.
Wong, B., Faulkner, G., & Buliung, R. (2011). GIS measured environmental correlates of active school
transport: A systematic review of 14 studies. International Journal of Behavioral Nutrition and Physical
Activity, 8(39), 22.
World Health Organization, 2003. Health aspects of air pollution with particulate matter, ozone and
nitrogen dioxide. Report on a WHO working group, Bonn, Germany, 1315 January 2003.
Zolnik E J, 2011, "The effect of sprawl on private-vehicle commuting outcomes" Environment and
Planning A 43(8) 1875 1893.

124

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix A. County Compactness Indices for 2010, 2000, and Changes


New compactness index 2010 is the county compactness/sprawl index for 2010, using the six variables
that make up the original county sprawl index. New compactness index 2000 is the analogous county
sprawl index for 2000 obtained by applying component score coefficient values for 2010 to data for
2000. Change in new compactness index is the change in the index between 2000 and 2010 measured
as above. Original compactness index is the county sprawl index for 2000 based on component score
coefficient values for 2000. And change in original compactness index is the change in the index
between 2000 and 2010 using the indices for each year respectively, based on component score
coefficient values for each year.
fips

county

1009
1015
1021
1033
1051
1055
1069
1073
1077
1079
1081
1083
1089
1097
1101
1103
1113
1115
1117
1125
1127
4005
4013
4015
4019
4021
4025
4027

Blount County, AL
Calhoun County, AL
Chilton County, AL
Colbert County, AL
Elmore County, AL
Etowah County, AL
Houston County, AL
Jefferson County, AL
Lauderdale County, AL
Lawrence County, AL
Lee County, AL
Limestone County, AL
Madison County, AL
Mobile County, AL
Montgomery County, AL
Morgan County, AL
Russell County, AL
St. Clair County, AL
Shelby County, AL
Tuscaloosa County, AL
Walker County, AL
Coconino County, AZ
Maricopa County, AZ
Mohave County, AZ
Pima County, AZ
Pinal County, AZ
Yavapai County, AZ
Yuma County, AZ

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
76.9
93.9
88.8
5.0
95.0
74.2
64.4
9.8
74.1
103.6
97.6
5.9
103.2
89.9
79.4
10.5
84.6
92.9
89.0
3.9
94.7
89.0
83.0
6.0
90.1
113.9
108.9
5.0
113.1
90.5
80.4
10.1
87.3
75.7
64.9
10.8
72.2
93.6
85.4
8.3
91.1
85.2
75.2
9.9
81.6
108.9
89.2
19.7
95.4
109.9
98.8
11.0
103.4
105.3
101.6
3.7
107.3
100.0
89.6
10.5
95.1
92.9
86.2
6.7
92.8
89.7
82.2
7.5
87.5
96.1
85.4
10.7
91.5
105.0
94.4
10.5
99.9
89.2
83.1
6.1
89.1
87.3
74.0
13.4
88.5
116.5
111.8
4.7
119.4
91.3
104.0
103.2
0.9
106.6
100.7
79.8
21.0
87.5
93.5
83.1
10.4
90.5
104.7
92.6
12.1
101.1

change in
original
compactness
index
-1.1
0.1
0.3
5.3
-1.8
-1.1
0.8
3.2
3.4
2.5
3.6
13.5
6.5
-1.9
4.9
0.1
2.1
4.6
5.0
0.1
-1.1
-2.8
-2.6
13.2
3.1
3.6

125

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

5007
5031
5033
5035
5045
5051
5053
5069
5079
5085
5087
5091
5111
5119
5125
5131
5143
6001
6007
6013
6017
6019
6025
6029
6031
6037
6039
6041
6047
6053
6055
6059
6061
6065
6067
6069

county

Benton County, AR
Craighead County, AR
Crawford County, AR
Crittenden County, AR
Faulkner County, AR
Garland County, AR
Grant County, AR
Jefferson County, AR
Lincoln County, AR
Lonoke County, AR
Madison County, AR
Miller County, AR
Poinsett County, AR
Pulaski County, AR
Saline County, AR
Sebastian County, AR
Washington County, AR
Alameda County, CA
Butte County, CA
Contra Costa County, CA
El Dorado County, CA
Fresno County, CA
Imperial County, CA
Kern County, CA
Kings County, CA
Los Angeles County, CA
Madera County, CA
Marin County, CA
Merced County, CA
Monterey County, CA
Napa County, CA
Orange County, CA
Placer County, CA
Riverside County, CA
Sacramento County, CA
San Benito County, CA
San Bernardino County,
6071 CA
6073 San Diego County, CA
6075 San Francisco County, CA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
94.0
86.9
7.0
92.8
87.4
79.6
7.8
87.1
84.9
81.1
3.8
88.9
91.6
89.7
1.9
100.2
87.4
81.1
6.3
88.0
94.5
91.6
2.9
95.7
66.8
64.3
2.5
73.6
100.2
98.1
2.1
104.4
72.3
70.9
1.4
81.4
82.4
77.6
4.7
85.9
79.2
98.9
98.8
0.1
106.4
73.6
67.9
5.7
80.8
114.1
108.7
5.4
112.6
85.3
80.1
5.2
86.0
102.8
100.4
2.4
105.1
101.1
90.7
10.5
98.3
153.3
145.3
7.9
152.4
97.5
95.5
2.0
103.7
121.7
118.1
3.6
123.6
89.7
85.0
4.7
90.0
100.0
96.0
4.0
104.2
89.7
88.1
1.6
95.8
96.6
92.4
4.3
100.8
96.8
85.2
11.6
95.6
160.6
155.9
4.7
161.5
82.6
80.3
2.3
88.5
115.9
111.8
4.1
119.2
94.1
89.7
4.4
99.2
110.6
108.7
1.8
120.3
110.7
107.0
3.7
112.3
145.8
140.6
5.1
146.2
102.3
94.6
7.6
100.5
105.8
100.2
5.7
107.3
124.4
118.0
6.5
124.5
101.5
97.0
4.5
107.5
102.4
126.0
247.8

99.2
122.6
246.1

3.2
3.4
1.7

change in
original
compactness
index
1.2
0.2
-4.0
-8.6
-0.6
-1.2
-6.8
-4.2
-9.1
-3.5
-7.5
-7.3
1.5
-0.7
-2.4
2.8
0.9
-6.2
-1.9
-0.3
-4.3
-6.1
-4.2
1.3
-0.9
-5.9
-3.3
-5.1
-9.7
-1.6
-0.4
1.7
-1.5
-0.1
-5.9

106.7
130.7
257.6

-4.2
-4.6
-9.8

126

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

6077 San Joaquin County, CA


San Luis Obispo County,
6079 CA
6081 San Mateo County, CA
6083 Santa Barbara County, CA
6085 Santa Clara County, CA
6087 Santa Cruz County, CA
6089 Shasta County, CA
6095 Solano County, CA
6097 Sonoma County, CA
6099 Stanislaus County, CA
6101 Sutter County, CA
6107 Tulare County, CA
6111 Ventura County, CA
6113 Yolo County, CA
6115 Yuba County, CA
8001 Adams County, CO
8005 Arapahoe County, CO
8013 Boulder County, CO
8014 Broomfield County, CO
8019 Clear Creek County, CO
8031 Denver County, CO
8035 Douglas County, CO
8039 Elbert County, CO
8041 El Paso County, CO
8059 Jefferson County, CO
8069 Larimer County, CO
8077 Mesa County, CO
8101 Pueblo County, CO
8119 Teller County, CO
8123 Weld County, CO
9001 Fairfield County, CT
9003 Hartford County, CT
9007 Middlesex County, CT
9009 New Haven County, CT
9011 New London County, CT
9013 Tolland County, CT
10001 Kent County, DE
10003 New Castle County, DE
11001 District of Columbia, DC

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
117.3
110.1
7.2
118.3
100.6
141.8
124.8
138.1
113.1
94.1
114.9
104.2
111.0
92.6
99.0
119.6
108.8
93.6
117.7
122.3
114.7
117.9
96.0
144.4
104.1
68.4
117.7
115.0
105.0
104.8
107.6
97.1
97.3
115.0
107.5
93.1
112.2
95.7
84.3
93.0
119.4
193.3

change in
original
compactness
index
-1.0

94.2
138.6
117.3
133.5
111.9
82.3
110.9
102.1
108.6
86.0
91.7
112.5
105.8
89.5
127.9
117.3
111.3

6.4
3.2
7.5
4.6
1.3
11.7
4.0
2.1
2.4
6.6
7.3
7.1
3.0
4.2
-10.2
5.0
3.4

101.8
146.0
125.7
140.5
118.1
88.6
117.5
106.6
117.4
95.1
103.1
121.6
113.1
96.5
130.3
122.0
113.9

-1.2
-4.2
-0.9
-2.4
-5.0
5.4
-2.6
-2.4
-6.5
-2.5
-4.1
-1.9
-4.3
-2.9
-12.6
0.3
0.8

93.4

2.6

97.2

-1.2

95.7

8.4

99.0

5.1

107.4
115.9
100.2
97.5
103.1
91.9
88.9
111.3
104.5
90.0
108.9
93.0
80.9
85.9
117.4

10.3
-0.9
4.9
7.3
4.5
5.2
8.4
3.8
3.0
3.0
3.3
2.7
3.4
7.1
1.9

110.2
117.8
105.2
105.7
110.7
93.6
99.6
115.0
110.1
94.6
113.7
98.9
86.4
92.1
122.3

7.5
-2.9
-0.1
-0.9
-3.1
3.5
-2.3
0.0
-2.5
-1.6
-1.5
-3.2
-2.1
0.9
-2.9

127

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

12001
12003
12005
12009
12011
12015
12019
12021
12031
12033
12035
12039
12053
12057
12061
12069
12071
12073
12081
12083
12085
12086
12089
12091
12095
12097
12099
12101
12103
12105
12109
12111
12113
12115
12117
12127
12129
13013
13015
13021

county

Alachua County, FL
Baker County, FL
Bay County, FL
Brevard County, FL
Broward County, FL
Charlotte County, FL
Clay County, FL
Collier County, FL
Duval County, FL
Escambia County, FL
Flagler County, FL
Gadsden County, FL
Hernando County, FL
Hillsborough County, FL
Indian River County, FL
Lake County, FL
Lee County, FL
Leon County, FL
Manatee County, FL
Marion County, FL
Martin County, FL
Miami-Dade County, FL
Nassau County, FL
Okaloosa County, FL
Orange County, FL
Osceola County, FL
Palm Beach County, FL
Pasco County, FL
Pinellas County, FL
Polk County, FL
St. Johns County, FL
St. Lucie County, FL
Santa Rosa County, FL
Sarasota County, FL
Seminole County, FL
Volusia County, FL
Wakulla County, FL
Barrow County, GA
Bartow County, GA
Bibb County, GA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
106.6
105.7
0.9
109.6
72.7
70.0
2.7
77.4
102.1
103.0
-0.9
107.4
108.4
105.1
3.3
109.6
133.0
131.6
1.3
136.5
100.3
96.1
4.2
100.1
98.7
85.2
13.5
92.0
104.5
93.9
10.6
99.2
117.4
116.0
1.4
120.3
106.4
106.3
0.2
111.0
99.2
91.0
86.6
4.4
91.8
98.7
94.1
4.6
98.6
119.1
115.9
3.2
119.6
111.1
103.6
7.5
107.5
108.0
101.6
6.5
105.1
107.0
100.5
6.5
104.8
102.8
97.0
5.8
102.0
116.3
114.2
2.2
118.0
93.5
92.1
1.4
96.0
108.1
100.6
7.4
105.3
152.1
146.4
5.7
92.9
79.3
13.7
84.1
101.3
102.9
-1.6
108.5
119.4
115.8
3.6
121.2
106.9
105.1
1.7
109.6
114.2
112.4
1.8
115.8
110.8
110.1
0.7
114.9
133.1
131.6
1.5
134.0
110.0
104.9
5.1
109.2
104.1
99.9
4.2
103.0
110.1
100.0
10.1
104.5
84.6
77.5
7.1
84.0
111.0
110.2
0.8
114.0
116.8
113.8
3.0
117.5
107.4
105.0
2.4
109.2
83.9
73.8
10.1
80.7
81.6
75.8
5.9
81.8
86.6
79.0
7.7
85.2
105.2
103.6
1.6
107.1

change in
original
compactness
index
-2.9
-4.8
-5.3
-1.2
-3.6
0.2
6.7
5.2
-2.9
-4.5
-0.9
0.1
-0.5
3.5
2.9
2.2
0.7
-1.7
-2.5
2.7
8.9
-7.2
-1.8
-2.7
-1.6
-4.0
-0.9
0.8
1.1
5.6
0.6
-3.1
-0.7
-1.8
3.2
-0.2
1.5
-1.9

128

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

13029
13035
13045
13047
13051
13053
13057
13059
13063
13067
13073
13077
13083
13085
13089
13095
13097
13103
13113
13115
13117
13121
13127
13135
13139
13143
13145
13151
13153
13169
13171
13177
13179
13185
13189
13195
13199
13207
13213

county

Bryan County, GA
Butts County, GA
Carroll County, GA
Catoosa County, GA
Chatham County, GA
Chattahoochee County,
GA
Cherokee County, GA
Clarke County, GA
Clayton County, GA
Cobb County, GA
Columbia County, GA
Coweta County, GA
Dade County, GA
Dawson County, GA
DeKalb County, GA
Dougherty County, GA
Douglas County, GA
Effingham County, GA
Fayette County, GA
Floyd County, GA
Forsyth County, GA
Fulton County, GA
Glynn County, GA
Gwinnett County, GA
Hall County, GA
Haralson County, GA
Harris County, GA
Henry County, GA
Houston County, GA
Jones County, GA
Lamar County, GA
Lee County, GA
Liberty County, GA
Lowndes County, GA
McDuffie County, GA
Madison County, GA
Meriwether County, GA
Monroe County, GA
Murray County, GA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
80.9
70.0
10.9
77.2
79.9
76.9
3.0
82.9
75.3
71.9
3.4
77.9
88.5
84.4
4.1
89.5
113.3
113.0
0.3
115.8
92.0
94.9
100.6
107.2
111.8
88.1
84.0
81.5
80.4
112.0
102.7
86.3
83.2
85.2
91.7
84.9
112.3
99.2
104.0
93.3
85.1
78.6
87.2
100.2
73.6
77.0
77.8
97.8
95.4
80.3
74.7
74.1
76.4
77.9

change in
original
compactness
index
3.7
-2.9
-2.6
-1.0
-2.5

84.0
96.6
101.2
102.8
87.2
79.7
75.1
72.5
107.4
95.3
79.5
79.6
74.8
89.5
70.6
107.6
96.7
94.5
87.6
76.4

11.0
4.1
6.0
9.0
0.9
4.2
6.4
7.9
4.6
7.4
6.7
3.6
10.4
2.2
14.4
4.7
2.5
9.5
5.6
8.7

90.0
100.4
105.1
106.4
91.5
84.7
81.3
79.9
110.3
99.9
84.4
85.8
79.8
94.3
76.2
111.3
100.0
98.9
91.7
82.5

4.9
0.2
2.1
5.4
-3.4
-0.7
0.2
0.4
1.7
2.7
1.9
-2.7
5.4
-2.6
8.7
1.0
-0.8
5.1
1.6
2.6

72.5
95.9
70.8
69.7
74.0
88.9
92.7
75.7
62.8
69.2
73.3
74.8

14.7
4.3
2.8
7.3
3.8
8.9
2.6
4.7
11.9
4.9
3.0
3.1

78.5
99.6
77.6
77.0
80.9
93.4
97.1
81.8
73.5
76.3
79.9
80.6

8.7
0.5
-4.0
0.0
-3.1
4.5
-1.7
-1.5
1.2
-2.2
-3.5
-2.7

129

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

13215
13217
13219
13221
13223
13227
13231
13245
13247
13255
13273
13295
13297
13313
13321
16001
16005
16019
16027
16045
16051
16055
16069
17003
17005
17007
17019
17027
17031
17037
17043
17053
17063
17073
17083
17089
17091
17093
17097
17111

county

Muscogee County, GA
Newton County, GA
Oconee County, GA
Oglethorpe County, GA
Paulding County, GA
Pickens County, GA
Pike County, GA
Richmond County, GA
Rockdale County, GA
Spalding County, GA
Terrell County, GA
Walker County, GA
Walton County, GA
Whitfield County, GA
Worth County, GA
Ada County, ID
Bannock County, ID
Bonneville County, ID
Canyon County, ID
Gem County, ID
Jefferson County, ID
Kootenai County, ID
Nez Perce County, ID
Alexander County, IL
Bond County, IL
Boone County, IL
Champaign County, IL
Clinton County, IL
Cook County, IL
DeKalb County, IL
DuPage County, IL
Ford County, IL
Grundy County, IL
Henry County, IL
Jersey County, IL
Kane County, IL
Kankakee County, IL
Kendall County, IL
Lake County, IL
McHenry County, IL

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
108.0
108.3
-0.3
113.0
88.9
78.7
10.2
83.4
79.2
72.6
6.6
78.9
67.7
85.6
80.9
4.7
86.4
76.5
74.7
1.8
80.8
71.5
105.7
102.2
3.5
106.9
94.5
82.4
12.1
86.9
87.7
85.4
2.3
88.9
76.8
82.2
-5.4
90.7
83.6
79.7
3.8
85.1
73.0
67.2
5.7
74.0
94.7
92.5
2.2
97.4
73.3
64.8
8.5
73.8
109.3
103.1
6.2
108.0
112.3
100.2
12.1
111.5
105.1
95.0
10.0
102.6
104.6
92.2
12.4
99.4
94.7
72.4
22.3
81.3
83.0
63.6
19.4
79.0
101.1
94.9
6.2
98.8
98.3
101.7
-3.3
107.9
89.3
83.4
77.9
5.6
86.3
92.4
89.7
2.7
97.5
109.0
101.1
7.8
111.4
83.2
82.6
0.6
91.2
167.6
165.2
2.4
171.3
97.8
90.8
7.0
100.2
118.8
117.0
1.8
121.3
67.3
70.1
-2.9
82.3
100.4
84.7
15.8
92.2
79.9
79.2
0.7
88.3
82.2
113.1
109.9
3.2
115.1
92.0
90.4
1.6
98.7
95.5
87.8
7.7
95.2
112.9
110.6
2.3
114.5
100.9
99.9
1.0
104.7

change in
original
compactness
index
-5.1
5.5
0.3
-0.8
-4.3
-1.2
7.6
-1.2
-13.9
-1.6
-1.0
-2.7
-0.4
1.3
0.8
2.4
5.2
13.4
4.0
2.2
-9.6
-2.9
-5.1
-2.4
-8.0
-3.6
-2.3
-2.5
-15.0
8.3
-8.4
-1.9
-6.7
0.3
-1.6
-3.8

130

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

17113
17115
17117
17119
17123
17129
17131
17133
17143
17147
17161
17163
17167
17179
17183
17197
17201
17203
18003
18005
18011
18013
18015
18019
18021
18029
18035
18039
18043
18047
18051
18055
18057
18059
18061
18063
18067
18073
18081
18089

McLean County, IL
Macon County, IL
Macoupin County, IL
Madison County, IL
Marshall County, IL
Menard County, IL
Mercer County, IL
Monroe County, IL
Peoria County, IL
Piatt County, IL
Rock Island County, IL
St. Clair County, IL
Sangamon County, IL
Tazewell County, IL
Vermilion County, IL
Will County, IL
Winnebago County, IL
Woodford County, IL
Allen County, IN
Bartholomew County, IN
Boone County, IN
Brown County, IN
Carroll County, IN
Clark County, IN
Clay County, IN
Dearborn County, IN
Delaware County, IN
Elkhart County, IN
Floyd County, IN
Franklin County, IN
Gibson County, IN
Greene County, IN
Hamilton County, IN
Hancock County, IN
Harrison County, IN
Hendricks County, IN
Howard County, IN
Jasper County, IN
Johnson County, IN
Lake County, IN

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
102.8
100.8
2.0
110.1
93.3
97.4
-4.1
104.3
93.2
88.6
4.7
99.1
105.5
101.2
4.3
107.4
88.9
77.7
11.1
87.7
78.8
78.6
0.2
97.4
74.9
72.2
2.7
82.5
87.9
81.6
6.2
90.2
104.3
104.6
-0.2
110.0
76.4
75.8
0.6
87.8
107.8
107.0
0.8
111.9
104.7
103.4
1.3
109.0
103.1
101.3
1.8
108.8
103.0
94.7
8.3
100.8
98.1
85.1
13.1
93.8
103.9
98.7
5.2
103.5
110.1
106.1
4.0
111.3
88.7
77.5
11.1
87.3
101.3
96.4
4.8
102.8
106.2
92.7
13.5
99.0
91.8
73.4
18.4
84.0
80.8
81.0
-0.2
86.1
81.3
67.2
14.1
76.9
105.1
102.0
3.2
107.5
86.2
76.4
9.8
86.8
94.1
81.9
12.2
89.1
107.3
98.9
8.4
105.5
104.8
92.2
12.5
98.7
105.2
100.5
4.6
102.1
82.2
78.0
4.2
87.0
90.0
78.2
11.9
88.2
78.3
79.2
-1.0
87.3
101.1
93.2
7.9
98.5
87.5
79.1
8.3
87.3
76.1
71.7
4.4
78.8
94.2
83.9
10.4
90.0
105.1
93.2
11.9
101.1
66.8
61.7
5.1
72.8
104.9
96.0
8.9
101.1
115.4
112.5
2.9
116.9

change in
original
compactness
index
-7.3
-11.0
-5.9
-1.9
1.1
-18.6
-7.7
-2.3
-5.6
-11.4
-4.1
-4.3
-5.8
2.2
4.3
0.4
-1.1
1.4
-1.6
7.2
7.8
-5.3
4.3
-2.3
-0.7
5.0
1.9
6.1
3.0
-4.8
1.8
-9.1
2.6
0.2
-2.8
4.2
3.9
-6.0
3.8
-1.6

131

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

18091
18095
18097
18105
18109
18115
18119
18127
18129
18133
18141
18145
18153
18157
18159
18163
18165
18167
18173
18175
18179
18183
19011
19013
19017
19049
19061
19085
19103
19105
19113
19121
19129
19153
19155
19163
19169
19181
19183

county

LaPorte County, IN
Madison County, IN
Marion County, IN
Monroe County, IN
Morgan County, IN
Ohio County, IN
Owen County, IN
Porter County, IN
Posey County, IN
Putnam County, IN
St. Joseph County, IN
Shelby County, IN
Sullivan County, IN
Tippecanoe County, IN
Tipton County, IN
Vanderburgh County, IN
Vermillion County, IN
Vigo County, IN
Warrick County, IN
Washington County, IN
Wells County, IN
Whitley County, IN
Benton County, IA
Black Hawk County, IA
Bremer County, IA
Dallas County, IA
Dubuque County, IA
Harrison County, IA
Johnson County, IA
Jones County, IA
Linn County, IA
Madison County, IA
Mills County, IA
Polk County, IA
Pottawattamie County,
IA
Scott County, IA
Story County, IA
Warren County, IA
Washington County, IA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
92.9
90.4
2.5
96.8
101.5
99.8
1.7
105.7
116.5
114.6
1.9
119.2
105.5
104.3
1.1
106.3
99.4
86.4
13.1
93.0
88.0
80.1
7.9
89.0
90.0
71.0
19.0
81.8
94.4
93.1
1.4
98.6
84.0
82.3
1.7
91.6
75.5
72.3
3.1
80.8
115.2
106.0
9.2
112.6
98.8
84.7
14.1
93.7
74.1
71.9
2.2
80.8
106.1
104.1
2.1
110.0
66.4
64.5
1.9
77.1
109.1
107.2
1.9
111.5
110.8
102.0
8.8
109.5
109.6
101.3
8.3
107.3
92.2
83.6
8.6
91.1
89.6
77.4
12.2
87.9
70.5
65.3
5.2
76.1
68.6
67.4
1.2
77.7
80.1
66.9
13.3
78.2
104.1
101.6
2.5
108.1
75.5
72.4
3.1
83.4
87.0
76.2
10.8
87.1
102.0
101.7
0.3
108.2
69.0
65.3
3.6
76.7
97.8
94.0
3.8
102.6
83.9
70.1
13.9
81.6
100.1
99.0
1.2
106.0
77.7
64.5
13.3
76.6
81.6
68.2
13.4
81.0
108.5
105.1
3.3
110.9
95.3
115.0
97.0
82.7
72.4

89.3
103.4
90.8
73.2
70.0

6.0
11.7
6.2
9.4
2.4

change in
original
compactness
index
-3.9
-4.1
-2.7
-0.8
6.4
-0.9
8.2
-4.2
-7.6
-5.3
2.6
5.1
-6.8
-3.9
-10.6
-2.4
1.3
2.2
1.1
1.7
-5.6
-9.1
1.9
-4.0
-7.9
-0.1
-6.2
-7.8
-4.8
2.4
-5.8
1.2
0.7
-2.4

99.6
109.5
102.2
86.5
82.9

-4.3
5.5
-5.2
-3.8
-10.5

132

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

19193
20015
20045
20059
20061
20079
20085
20091
20103
20121
20139
20149
20161
20173
20177
20191
20209
21015
21017
21019
21029
21037
21047
21049
21059
21067
21081
21089
21093
21101
21103
21111
21113
21117
21123
21163
21179
21185
21209
21211

Woodbury County, IA
Butler County, KS
Douglas County, KS
Franklin County, KS
Geary County, KS
Harvey County, KS
Jackson County, KS
Johnson County, KS
Leavenworth County, KS
Miami County, KS
Osage County, KS
Pottawatomie County, KS
Riley County, KS
Sedgwick County, KS
Shawnee County, KS
Sumner County, KS
Wyandotte County, KS
Boone County, KY
Bourbon County, KY
Boyd County, KY
Bullitt County, KY
Campbell County, KY
Christian County, KY
Clark County, KY
Daviess County, KY
Fayette County, KY
Grant County, KY
Greenup County, KY
Hardin County, KY
Henderson County, KY
Henry County, KY
Jefferson County, KY
Jessamine County, KY
Kenton County, KY
Larue County, KY
Meade County, KY
Nelson County, KY
Oldham County, KY
Scott County, KY
Shelby County, KY

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
104.8
97.0
7.8
105.4
81.5
76.1
5.4
84.8
99.1
93.7
5.4
100.8
84.9
62.6
22.3
72.3
107.6
70.9
69.1
1.8
79.3
54.6
47.4
7.2
61.5
104.9
103.1
1.8
108.9
92.8
91.4
1.3
100.2
87.8
65.7
22.1
76.1
66.9
64.1
2.8
74.2
85.9
99.5
108.0
106.4
1.6
111.3
102.3
99.5
2.8
105.4
77.9
61.6
16.3
73.3
114.9
111.7
3.2
116.4
95.6
92.0
3.6
96.3
95.7
80.3
15.4
88.7
101.5
101.0
0.5
106.0
95.4
86.1
9.3
90.3
112.6
110.7
1.9
115.5
100.1
86.2
13.9
93.5
97.3
91.0
6.4
97.3
103.9
99.3
4.7
107.7
115.7
105.6
10.1
109.2
87.1
78.3
8.8
84.0
102.0
97.3
4.7
103.3
95.9
85.7
10.3
92.4
93.9
87.1
6.8
94.9
89.9
75.0
14.9
87.5
118.4
112.4
5.9
115.9
94.9
87.0
7.9
91.3
118.4
111.0
7.4
114.2
76.9
88.9
78.5
10.4
83.7
90.0
78.4
11.6
84.5
90.9
85.3
5.6
87.9
99.4
88.9
10.5
96.3
94.7
85.4
9.3
92.1

change in
original
compactness
index
-0.7
-3.2
-1.7
12.7
-8.4
-7.0
-4.0
-7.4
11.7
-7.3

-3.4
-3.1
4.6
-1.5
-0.7
7.0
-4.4
5.1
-2.9
6.6
0.0
-3.8
6.5
3.1
-1.3
3.6
-1.1
2.4
2.4
3.6
4.3
5.2
5.5
3.0
3.1
2.6

133

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

21215
21227
21239
22005
22015
22017
22019
22031

Spencer County, KY
Warren County, KY
Woodford County, KY
Ascension Parish, LA
Bossier Parish, LA
Caddo Parish, LA
Calcasieu Parish, LA
De Soto Parish, LA
East Baton Rouge Parish,
LA
Grant Parish, LA
Iberville Parish, LA
Jefferson Parish, LA
Lafayette Parish, LA
Lafourche Parish, LA
Livingston Parish, LA
Orleans Parish, LA
Ouachita Parish, LA
Plaquemines Parish, LA
Pointe Coupee Parish, LA
Rapides Parish, LA
St. Bernard Parish, LA
St. Charles Parish, LA
St. John the Baptist
Parish, LA
St. Martin Parish, LA
St. Tammany Parish, LA
Terrebonne Parish, LA
Union Parish, LA
West Baton Rouge
Parish, LA
Androscoggin County,
ME
Cumberland County, ME
Penobscot County, ME
Sagadahoc County, ME
York County, ME
Allegany County, MD
Anne Arundel County,
MD

22033
22043
22047
22051
22055
22057
22063
22071
22073
22075
22077
22079
22087
22089
22095
22099
22103
22109
22111
22121
23001
23005
23019
23023
23031
24001
24003

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
86.5
106.1
94.7
11.4
101.3
91.2
82.5
8.7
89.0
91.4
85.7
5.6
91.2
94.9
93.7
1.2
100.0
105.5
102.6
2.9
107.9
95.4
94.0
1.3
101.5
81.2
78.1
3.1
85.2

change in
original
compactness
index
4.8
2.2
0.2
-5.1
-2.4
-6.2
-4.0

110.9
76.4
94.1
130.5
105.3
95.7
87.5
144.7
102.9
92.2
83.2
95.0
112.5
100.6

106.1

4.8

111.0

-0.1

95.8
128.0
101.3
92.0
80.6
160.3
99.8
86.4
79.0
96.4
116.5
88.8

-1.7
2.5
4.1
3.7
6.9
-15.6
3.1
5.9
4.2
-1.4
-4.0
11.9

101.8
132.7
105.9
98.7
87.1
165.7
104.8
92.6
86.4
103.4
119.9
95.8

-7.7
-2.2
-0.6
-3.0
0.4
-21.0
-1.8
-0.3
-3.2
-8.4
-7.4
4.8

105.3
88.9
103.8
102.1
75.6

96.7
81.2
96.3
101.8
67.1

8.5
7.7
7.6
0.3
8.5

103.0
87.8
100.5
106.8
76.8

2.3
1.2
3.4
-4.7
-1.2

95.6

90.5

5.1

95.3

0.3

96.0
100.3
84.9
90.5
88.4
106.7

90.7
95.4
80.1
82.8
84.9
98.9

5.3
4.9
4.8
7.7
3.5
7.7

95.9
100.5
89.1
88.8
91.3
103.1

0.1
-0.1
-4.2
1.7
-2.9
3.6

115.6

109.3

6.3

113.3

2.3

134

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

24005
24009
24013
24015
24017
24021
24025
24027
24031

Baltimore County, MD
Calvert County, MD
Carroll County, MD
Cecil County, MD
Charles County, MD
Frederick County, MD
Harford County, MD
Howard County, MD
Montgomery County, MD
Prince George's County,
MD
Queen Anne's County,
MD
Somerset County, MD
Washington County, MD
Wicomico County, MD
Baltimore city, MD
Barnstable County, MA
Berkshire County, MA
Bristol County, MA
Essex County, MA
Franklin County, MA
Hampden County, MA
Hampshire County, MA
Middlesex County, MA
Norfolk County, MA
Plymouth County, MA
Suffolk County, MA
Worcester County, MA
Barry County, MI
Bay County, MI
Berrien County, MI
Calhoun County, MI
Cass County, MI
Clinton County, MI
Eaton County, MI
Genesee County, MI
Ingham County, MI
Ionia County, MI
Jackson County, MI

24033
24035
24039
24043
24045
24510
25001
25003
25005
25009
25011
25013
25015
25017
25021
25023
25025
25027
26015
26017
26021
26025
26027
26037
26045
26049
26065
26067
26075

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
121.6
109.3
12.3
113.4
104.4
90.1
14.2
95.3
97.1
81.5
15.6
86.4
99.2
85.6
13.6
91.1
108.6
88.4
20.2
94.6
103.1
85.5
17.6
92.2
105.8
91.6
14.2
97.6
113.7
94.7
19.0
98.9
122.7
116.8
5.9
120.8

change in
original
compactness
index
8.2
9.1
10.6
8.0
14.0
10.9
8.3
14.8
2.0

125.4

114.7

10.7

120.5

4.9

83.7
99.4
99.1
106.7
179.6
102.2
97.6
121.5
128.3
91.5
116.5
101.9
132.5
121.6
107.3
217.1
107.4
79.8
96.5
95.7
92.6
79.5
75.3
82.5
101.4
110.4
82.2
91.7

74.9
83.7
92.7
92.2
182.0
96.6
86.9
117.5
123.4
83.5
110.4
87.1
127.7
116.8
101.6
206.5
100.4
68.9
92.5
90.4
87.7
72.6
61.7
74.6
98.7
103.5
73.0
84.2

8.8
15.7
6.4
14.5
-2.4
5.6
10.7
4.0
4.9
8.0
6.1
14.7
4.8
4.8
5.6
10.6
7.0
10.8
4.0
5.3
4.9
6.8
13.6
7.9
2.7
6.9
9.2
7.4

81.4
89.8
98.7
98.0
187.6
100.0
93.6
121.6
127.7
90.0
114.5
92.6
132.3
121.9
105.8
212.7
105.8
76.5
100.2
96.9
94.9
81.3
72.0
82.2
103.7
110.3
82.3
90.6

2.3
9.6
0.4
8.8
-8.1
2.2
4.0
-0.1
0.6
1.5
2.0
9.3
0.2
-0.3
1.4
4.4
1.6
3.3
-3.7
-1.2
-2.3
-1.8
3.3
0.3
-2.3
0.2
-0.2
1.1

135

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

26077
26081
26087
26093
26099
26115
26121
26123
26125
26139
26145
26147
26159
26161
26163
27003
27009
27013
27017
27019
27025
27027
27037
27039
27053
27055
27059
27103
27109
27119
27123
27137
27139
27141
27145
27157
27163
27171
28029
28033

Kalamazoo County, MI
Kent County, MI
Lapeer County, MI
Livingston County, MI
Macomb County, MI
Monroe County, MI
Muskegon County, MI
Newaygo County, MI
Oakland County, MI
Ottawa County, MI
Saginaw County, MI
St. Clair County, MI
Van Buren County, MI
Washtenaw County, MI
Wayne County, MI
Anoka County, MN
Benton County, MN
Blue Earth County, MN
Carlton County, MN
Carver County, MN
Chisago County, MN
Clay County, MN
Dakota County, MN
Dodge County, MN
Hennepin County, MN
Houston County, MN
Isanti County, MN
Nicollet County, MN
Olmsted County, MN
Polk County, MN
Ramsey County, MN
St. Louis County, MN
Scott County, MN
Sherburne County, MN
Stearns County, MN
Wabasha County, MN
Washington County, MN
Wright County, MN
Copiah County, MS
DeSoto County, MS

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
96.4
94.0
2.5
100.3
101.6
97.7
3.9
103.4
76.4
68.8
7.6
76.1
88.5
80.9
7.6
86.8
112.6
108.1
4.5
113.6
83.9
80.6
3.2
87.9
100.7
98.3
2.4
104.1
80.2
65.3
14.9
74.6
108.3
106.2
2.1
111.2
92.2
85.0
7.2
92.3
96.9
93.6
3.2
101.5
89.6
87.1
2.6
92.9
77.1
72.3
4.7
80.1
102.8
100.2
2.6
106.1
126.3
126.6
-0.3
132.0
106.7
96.9
9.8
100.5
92.3
93.3
-1.0
102.3
90.0
83.3
75.8
7.5
83.4
98.4
84.3
14.1
90.0
84.0
77.0
7.0
83.6
84.2
79.8
4.5
93.2
106.2
98.5
7.7
103.6
80.3
61.6
18.7
73.2
123.7
124.0
-0.3
128.0
88.3
82.9
5.4
91.5
84.9
65.4
19.5
75.2
93.8
100.1
90.1
10.0
97.2
61.1
128.9
126.2
2.7
131.6
93.1
92.6
0.5
99.3
91.2
89.2
2.0
94.8
84.3
76.9
7.4
83.1
94.2
85.9
8.3
94.2
93.8
84.4
9.4
93.5
109.1
97.6
11.5
101.5
81.6
77.8
3.8
84.1
80.6
64.5
16.1
76.6
89.1
80.2
8.9
86.3

change in
original
compactness
index
-3.8
-1.7
0.3
1.7
-1.0
-4.1
-3.4
5.6
-2.8
-0.1
-4.7
-3.3
-3.1
-3.3
-5.7
6.2
-10.0
-0.1
8.4
0.5
-9.0
2.7
7.1
-4.3
-3.3
9.6
2.9
-2.7
-6.2
-3.6
1.2
0.0
0.3
7.5
-2.5
4.1
2.7

136

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

28035
28039
28045
28047
28049
28059
28073
28089
28093
28121
28127
28131
28137
28143
29003
29013
29019
29021
29027
29031
29037
29043
29047
29049
29051
29055
29071
29077
29095
29097
29099
29107
29113
29135
29145
29165
29177
29183

county

Forrest County, MS
George County, MS
Hancock County, MS
Harrison County, MS
Hinds County, MS
Jackson County, MS
Lamar County, MS
Madison County, MS
Marshall County, MS
Rankin County, MS
Simpson County, MS
Stone County, MS
Tate County, MS
Tunica County, MS
Andrew County, MO
Bates County, MO
Boone County, MO
Buchanan County, MO
Callaway County, MO
Cape Girardeau County,
MO
Cass County, MO
Christian County, MO
Clay County, MO
Clinton County, MO
Cole County, MO
Crawford County, MO
(pt.)*
Franklin County, MO
Greene County, MO
Jackson County, MO
Jasper County, MO
Jefferson County, MO
Lafayette County, MO
Lincoln County, MO
Moniteau County, MO
Newton County, MO
Platte County, MO
Ray County, MO
St. Charles County, MO

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
99.0
91.1
7.9
97.5
82.9
63.7
19.2
73.5
93.6
84.9
8.7
90.2
106.1
98.4
7.7
102.8
103.4
96.7
6.7
102.7
100.5
92.2
8.3
97.2
80.2
66.5
13.7
74.1
92.8
76.7
16.1
85.4
84.7
71.3
13.3
78.7
87.6
79.2
8.4
86.3
83.3
69.5
13.8
78.5
84.6
64.9
19.8
73.4
87.1
64.1
23.0
73.6
74.7
65.5
9.2
74.9
77.4
61.4
16.0
72.8
84.6
72.1
12.5
83.9
103.6
93.5
10.1
99.4
115.7
107.2
8.5
112.8
84.9
71.6
13.4
79.8

change in
original
compactness
index
1.4
9.4
3.4
3.3
0.7
3.3
6.1
7.4
6.0
1.3
4.8
11.2
13.5
-0.1
4.6
0.8
4.2
2.8
5.1

98.0
87.3
92.0
100.2
88.7
89.4

80.5
79.9
97.3
79.0
88.3

6.9
12.1
2.9
9.7
1.1

88.5
87.2
103.0
86.5
95.4

-1.2
4.8
-2.9
2.2
-5.9

82.6
91.8
108.1
115.1
96.2
101.6
87.9
92.4
75.4
90.5
97.4
69.4
116.8

82.2
96.1
113.4
95.8
94.5
80.8
78.7
70.8
76.9
88.6
70.5
108.6

9.6
12.0
1.7
0.4
7.1
7.1
13.7
4.6
13.6
8.8
-1.1
8.2

87.9
101.6
119.7
101.3
98.2
89.5
86.3
80.4
87.7
94.7
79.2
113.7

3.9
6.5
-4.6
-5.1
3.4
-1.6
6.0
-5.0
2.7
2.6
-9.8
3.1

137

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

29189
29219
29221
29225
29510
30009
30013
30063
30111
31025
31043
31055
31109
31153
31155
31159
31177
32003
32031
32510
33011
33015
33017
34001
34003
34005
34007
34009
34011
34013
34015
34017
34019
34021
34023
34025
34027
34029
34031
34033

St. Louis County, MO


Warren County, MO
Washington County, MO
Webster County, MO
St. Louis city, MO
Carbon County, MT
Cascade County, MT
Missoula County, MT
Yellowstone County, MT
Cass County, NE
Dakota County, NE
Douglas County, NE
Lancaster County, NE
Sarpy County, NE
Saunders County, NE
Seward County, NE
Washington County, NE
Clark County, NV
Washoe County, NV
Carson City, NV
Hillsborough County, NH
Rockingham County, NH
Strafford County, NH
Atlantic County, NJ
Bergen County, NJ
Burlington County, NJ
Camden County, NJ
Cape May County, NJ
Cumberland County, NJ
Essex County, NJ
Gloucester County, NJ
Hudson County, NJ
Hunterdon County, NJ
Mercer County, NJ
Middlesex County, NJ
Monmouth County, NJ
Morris County, NJ
Ocean County, NJ
Passaic County, NJ
Salem County, NJ

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
116.9
121.0
-4.1
124.2
89.8
77.8
12.0
83.8
83.3
76.7
6.6
83.8
91.3
69.4
21.9
79.9
145.7
77.9
72.9
5.0
86.9
101.9
99.6
2.4
106.0
101.7
97.5
4.3
102.6
103.5
99.8
3.6
105.4
84.7
70.4
14.3
81.2
104.9
97.8
7.1
105.1
122.1
120.9
1.2
124.5
112.1
106.9
5.2
112.8
113.6
101.4
12.2
107.0
75.2
61.5
13.7
72.4
73.2
62.6
10.6
112.0
86.8
71.7
15.2
81.6
123.5
117.8
5.7
123.4
106.6
103.0
3.6
113.4
112.6
111.2
1.3
120.0
103.8
102.3
1.5
106.8
90.4
87.0
3.3
92.3
93.6
87.7
5.9
93.5
112.9
112.2
0.7
116.7
140.1
138.8
1.3
142.1
104.1
101.4
2.7
106.6
130.0
129.2
0.9
132.2
115.6
114.1
1.6
116.4
101.2
93.1
8.1
99.0
168.9
170.0
-1.1
175.4
105.6
102.5
3.1
106.2
228.8
225.1
3.7
230.1
87.0
80.6
6.4
85.3
123.8
121.4
2.4
124.9
130.4
127.1
3.3
130.3
117.5
113.5
4.0
116.6
107.2
103.3
3.9
106.0
119.0
114.1
4.9
117.6
158.1
154.0
4.1
157.1
94.0
88.1
5.9
94.1

change in
original
compactness
index
-7.3
6.0
-0.5
11.4
-9.0
-4.0
-0.9
-1.9
3.5
-0.2
-2.4
-0.7
6.6
2.9
-38.8
5.2
0.1
-6.8
-7.5
-3.0
-2.0
0.1
-3.8
-2.0
-2.5
-2.2
-0.8
2.2
-6.5
-0.5
-1.3
1.7
-1.1
0.1
0.9
1.2
1.4
1.0
-0.1

138

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

34035
34037
34039
34041
35001
35013
35043
35045
35049
35061
36001
36005
36007
36015
36027
36029
36043
36047
36051
36053
36055
36059
36061
36063
36065
36067
36069
36071
36073
36075
36079
36081
36083
36085
36087
36091
36093
36095
36103
36107

Somerset County, NJ
Sussex County, NJ
Union County, NJ
Warren County, NJ
Bernalillo County, NM
Dona Ana County, NM
Sandoval County, NM
San Juan County, NM
Santa Fe County, NM
Valencia County, NM
Albany County, NY
Bronx County, NY
Broome County, NY
Chemung County, NY
Dutchess County, NY
Erie County, NY
Herkimer County, NY
Kings County, NY
Livingston County, NY
Madison County, NY
Monroe County, NY
Nassau County, NY
New York County, NY
Niagara County, NY
Oneida County, NY
Onondaga County, NY
Ontario County, NY
Orange County, NY
Orleans County, NY
Oswego County, NY
Putnam County, NY
Queens County, NY
Rensselaer County, NY
Richmond County, NY
Rockland County, NY
Saratoga County, NY
Schenectady County, NY
Schoharie County, NY
Suffolk County, NY
Tioga County, NY

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
106.1
98.1
8.0
102.0
95.0
92.2
2.8
95.9
151.0
146.8
4.2
150.7
98.6
97.0
1.7
100.7
120.2
116.3
3.9
117.7
102.1
95.3
6.8
100.3
91.5
86.7
4.8
95.6
84.0
76.8
7.2
84.5
97.6
93.7
3.9
97.9
86.8
82.8
3.9
86.9
111.2
106.5
4.8
111.8
331.5
328.8
2.7
322.7
102.0
98.4
3.6
103.6
99.9
97.0
2.8
103.2
94.0
87.2
6.8
92.2
107.6
108.8
-1.1
114.9
88.5
85.4
3.2
92.5
341.4
341.3
0.1
341.5
73.9
71.8
2.1
82.2
77.8
73.6
4.2
80.4
107.5
105.5
2.0
110.2
144.2
143.4
0.9
149.4
463.9
459.5
4.4
478.8
100.3
99.1
1.2
104.5
96.8
94.2
2.7
98.5
106.9
103.3
3.6
107.8
80.7
77.0
3.7
84.8
101.7
98.1
3.6
104.2
73.7
76.1
-2.5
85.1
84.8
82.6
2.2
89.0
95.1
93.2
1.8
96.6
272.1
269.1
2.9
272.6
99.6
99.5
0.1
104.9
190.1
188.2
1.9
188.4
123.6
113.3
10.3
117.3
91.0
87.2
3.8
93.8
111.8
109.9
1.9
115.4
75.9
77.2
-1.3
85.5
113.7
111.2
2.6
115.5
83.3
80.2
3.1
86.6

change in
original
compactness
index
4.1
-0.8
0.3
-2.1
2.5
1.8
-4.1
-0.5
-0.4
-0.1
-0.6
8.8
-1.6
-3.3
1.8
-7.2
-4.0
0.0
-8.3
-2.7
-2.7
-5.2
-14.9
-4.1
-1.6
-0.9
-4.0
-2.5
-11.4
-4.2
-1.6
-0.5
-5.4
1.7
6.3
-2.8
-3.6
-9.6
-1.8
-3.3

139

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

36109
36111
36113
36115
36117
36119
37001
37003
37007
37019
37021
37023
37025
37027
37035
37037
37051
37053
37059
37063
37065
37067
37069
37071
37079
37081
37087
37089
37093
37101
37115
37119
37127
37129
37133
37135
37141
37145
37147
37151

Tompkins County, NY
Ulster County, NY
Warren County, NY
Washington County, NY
Wayne County, NY
Westchester County, NY
Alamance County, NC
Alexander County, NC
Anson County, NC
Brunswick County, NC
Buncombe County, NC
Burke County, NC
Cabarrus County, NC
Caldwell County, NC
Catawba County, NC
Chatham County, NC
Cumberland County, NC
Currituck County, NC
Davie County, NC
Durham County, NC
Edgecombe County, NC
Forsyth County, NC
Franklin County, NC
Gaston County, NC
Greene County, NC
Guilford County, NC
Haywood County, NC
Henderson County, NC
Hoke County, NC
Johnston County, NC
Madison County, NC
Mecklenburg County, NC
Nash County, NC
New Hanover County, NC
Onslow County, NC
Orange County, NC
Pender County, NC
Person County, NC
Pitt County, NC
Randolph County, NC

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
92.6
90.5
2.1
96.7
92.5
88.8
3.8
94.0
94.2
92.1
2.1
97.5
78.0
75.1
2.9
83.7
74.8
72.0
2.8
79.6
140.2
138.0
2.2
141.6
97.5
88.9
8.6
94.0
73.6
72.3
1.3
78.8
67.9
64.0
4.0
72.7
84.8
79.7
5.1
84.9
97.8
93.6
4.2
98.4
82.8
82.8
0.1
87.4
94.2
88.5
5.7
93.6
86.8
86.1
0.7
90.8
92.1
90.0
2.1
93.9
75.2
73.1
2.1
79.6
99.7
96.8
2.9
102.3
87.0
76.1
68.0
8.1
75.4
105.5
99.5
6.0
104.0
85.2
85.8
-0.6
91.7
98.7
95.9
2.8
101.2
78.3
73.5
4.8
80.9
96.3
92.6
3.7
97.2
63.3
60.3
3.0
70.2
101.6
96.9
4.6
101.8
95.4
95.5
-0.1
99.5
93.7
92.6
1.1
96.5
78.3
76.2
2.1
82.8
77.7
78.1
-0.4
84.0
93.1
92.4
0.8
99.6
107.0
97.4
9.6
101.7
83.2
81.1
2.1
87.8
113.3
106.0
7.3
108.7
90.2
91.2
-1.0
96.4
91.5
85.8
5.6
90.7
75.5
71.3
4.3
79.5
74.8
73.3
1.5
80.6
95.1
93.9
1.2
98.7
75.4
75.5
0.0
81.5

change in
original
compactness
index
-4.1
-1.4
-3.2
-5.7
-4.8
-1.4
3.6
-5.1
-4.8
-0.1
-0.6
-4.6
0.6
-4.0
-1.9
-4.4
-2.7
0.7
1.5
-6.5
-2.5
-2.6
-0.9
-6.9
-0.2
-4.1
-2.9
-4.5
-6.3
-6.4
5.3
-4.5
4.6
-6.2
0.8
-4.0
-5.8
-3.6
-6.1

140

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

37157
37169
37179
37183
37191
37197
38015
38017
38035
38059
39003
39013
39015
39017
39019
39023
39025
39035
39041
39043
39045
39049
39051
39055
39057
39061
39081
39085
39087
39089
39093
39095
39097
39099
39103
39109
39113
39117
39123
39129

Rockingham County, NC
Stokes County, NC
Union County, NC
Wake County, NC
Wayne County, NC
Yadkin County, NC
Burleigh County, ND
Cass County, ND
Grand Forks County, ND
Morton County, ND
Allen County, OH
Belmont County, OH
Brown County, OH
Butler County, OH
Carroll County, OH
Clark County, OH
Clermont County, OH
Cuyahoga County, OH
Delaware County, OH
Erie County, OH
Fairfield County, OH
Franklin County, OH
Fulton County, OH
Geauga County, OH
Greene County, OH
Hamilton County, OH
Jefferson County, OH
Lake County, OH
Lawrence County, OH
Licking County, OH
Lorain County, OH
Lucas County, OH
Madison County, OH
Mahoning County, OH
Medina County, OH
Miami County, OH
Montgomery County, OH
Morrow County, OH
Ottawa County, OH
Pickaway County, OH

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
82.7
81.6
1.1
87.4
78.4
67.7
10.7
75.9
91.8
73.8
17.9
80.4
103.6
96.6
7.0
100.3
88.4
86.7
1.7
92.9
69.7
65.1
4.6
73.9
92.6
85.1
7.5
93.5
95.6
76.8
18.8
89.4
92.3
78.0
14.3
90.8
85.0
80.3
4.7
89.4
104.8
89.5
15.3
96.4
103.6
93.8
9.8
100.7
82.0
75.7
6.3
83.5
103.9
101.2
2.8
108.1
76.0
75.0
1.0
83.8
100.8
95.0
5.8
100.9
95.6
86.9
8.7
91.6
114.5
119.6
-5.1
125.5
95.7
79.9
15.8
86.3
101.1
94.0
7.0
99.4
93.0
84.7
8.4
90.2
123.2
119.5
3.8
123.5
88.4
62.7
25.7
71.8
70.0
60.1
10.0
67.6
97.7
89.2
8.5
95.8
116.5
115.3
1.1
119.1
104.0
99.7
4.3
105.9
100.4
96.2
4.2
101.8
99.3
96.1
3.2
101.6
102.4
83.3
19.2
89.0
98.2
93.2
4.9
99.7
113.3
112.8
0.6
117.2
86.3
79.9
6.4
87.7
101.7
97.2
4.6
103.2
78.3
74.7
3.6
81.1
91.1
84.9
6.2
91.5
111.3
109.1
2.1
113.9
63.4
61.3
2.1
71.2
92.3
88.4
3.9
93.7
86.9
82.0
4.9
89.6

change in
original
compactness
index
-4.7
2.5
11.4
3.3
-4.5
-4.1
-0.9
6.2
1.6
-4.4
8.5
2.9
-1.5
-4.1
-7.9
-0.2
4.0
-11.0
9.4
1.7
2.8
-0.3
16.6
2.4
1.9
-2.6
-1.9
-1.4
-2.3
13.5
-1.6
-3.8
-1.4
-1.5
-2.8
-0.5
-2.6
-7.8
-1.5
-2.7

141

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

39133
39135
39139
39151
39153
39155
39159
39165
39167
39173
40017
40027
40031
40037
40051
40079
40083
40087
40109
40111
40113
40117
40131
40135
40143
40145
41003
41005
41009
41017
41029
41039
41047
41051
41053
41067
41071
42003
42005
42007

county

Portage County, OH
Preble County, OH
Richland County, OH
Stark County, OH
Summit County, OH
Trumbull County, OH
Union County, OH
Warren County, OH
Washington County, OH
Wood County, OH
Canadian County, OK
Cleveland County, OK
Comanche County, OK
Creek County, OK
Grady County, OK
Le Flore County, OK
Logan County, OK
McClain County, OK
Oklahoma County, OK
Okmulgee County, OK
Osage County, OK
Pawnee County, OK
Rogers County, OK
Sequoyah County, OK
Tulsa County, OK
Wagoner County, OK
Benton County, OR
Clackamas County, OR
Columbia County, OR
Deschutes County, OR
Jackson County, OR
Lane County, OR
Marion County, OR
Multnomah County, OR
Polk County, OR
Washington County, OR
Yamhill County, OR
Allegheny County, PA
Armstrong County, PA
Beaver County, PA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
99.7
82.1
17.6
88.7
91.9
72.4
19.5
81.1
99.5
84.5
15.0
90.7
108.5
106.5
2.0
111.3
108.4
107.2
1.3
111.8
95.8
92.4
3.4
98.1
92.6
76.1
16.6
84.5
96.9
89.2
7.8
94.2
90.2
83.8
6.5
91.0
88.5
81.4
7.1
89.2
94.8
77.2
17.7
86.9
106.3
94.3
12.0
100.1
104.0
94.2
9.8
102.5
94.3
88.6
5.7
95.4
87.9
78.4
9.6
89.1
88.9
81.8
7.1
90.0
84.6
76.9
7.7
85.5
83.1
76.2
6.9
84.5
111.9
107.2
4.6
111.4
93.3
86.1
7.2
92.8
97.1
95.9
1.2
104.3
87.8
79.4
8.4
88.8
93.7
84.9
8.9
91.0
92.0
79.7
12.4
88.1
110.5
109.6
0.9
113.6
95.3
86.4
8.9
93.2
99.2
97.6
1.6
105.7
103.6
98.1
5.4
103.8
90.1
86.8
3.3
95.8
86.7
77.1
9.5
85.5
96.8
90.1
6.7
96.2
104.1
99.2
4.9
106.0
106.1
99.9
6.2
105.8
138.5
134.4
4.1
140.8
91.5
87.0
4.5
95.4
115.9
109.1
6.7
114.7
97.1
94.5
2.7
104.0
125.3
124.4
0.8
128.1
88.9
85.8
3.1
93.0
103.0
105.4
-2.4
109.7

change in
original
compactness
index
11.0
10.8
8.8
-2.8
-3.4
-2.3
8.1
2.7
-0.8
-0.7
7.9
6.2
1.5
-1.1
-1.2
-1.1
-0.9
-1.4
0.5
0.4
-7.1
-1.0
2.7
3.9
-3.1
2.1
-6.5
-0.3
-5.7
1.2
0.5
-1.9
0.2
-2.3
-3.9
1.1
-6.9
-2.9
-4.1
-6.8

142

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

42011
42013
42017
42019
42021
42025
42027
42029
42041
42043
42045
42049
42051
42069
42071
42075
42077
42079
42081
42085
42091
42095
42099
42101
42103
42125

Berks County, PA
Blair County, PA
Bucks County, PA
Butler County, PA
Cambria County, PA
Carbon County, PA
Centre County, PA
Chester County, PA
Cumberland County, PA
Dauphin County, PA
Delaware County, PA
Erie County, PA
Fayette County, PA
Lackawanna County, PA
Lancaster County, PA
Lebanon County, PA
Lehigh County, PA
Luzerne County, PA
Lycoming County, PA
Mercer County, PA
Montgomery County, PA
Northampton County, PA
Perry County, PA
Philadelphia County, PA
Pike County, PA
Washington County, PA
Westmoreland County,
PA
Wyoming County, PA
York County, PA
Bristol County, RI
Kent County, RI
Newport County, RI
Providence County, RI
Washington County, RI
Aiken County, SC
Anderson County, SC
Berkeley County, SC
Charleston County, SC
Darlington County, SC

42129
42131
42133
44001
44003
44005
44007
44009
45003
45007
45015
45019
45031

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
116.5
108.3
8.2
113.4
111.2
110.0
1.1
115.4
105.9
100.8
5.1
105.3
87.0
85.2
1.8
91.0
105.7
105.3
0.4
109.4
95.2
93.3
1.9
97.6
109.1
101.4
7.7
108.2
96.7
90.1
6.5
94.4
109.3
97.0
12.3
102.4
116.1
115.2
0.9
120.3
132.5
131.9
0.6
135.0
103.0
99.8
3.2
104.7
101.2
96.4
4.8
103.4
114.8
113.9
0.9
116.0
103.5
94.8
8.7
100.4
105.8
101.6
4.1
106.8
127.5
124.0
3.5
126.9
108.5
107.6
0.9
112.1
105.0
96.0
9.0
102.7
91.0
87.1
3.9
93.7
112.1
108.4
3.6
113.4
115.9
112.0
3.8
116.4
83.2
80.3
2.9
86.9
216.8
217.1
-0.4
225.0
91.2
84.2
7.0
88.0
98.1
99.3
-1.2
105.8
103.9
84.2
101.2
124.8
117.9
111.1
137.5
98.2
92.3
86.4
93.9
113.6
81.6

100.4
74.7
95.0
121.7
116.9
108.1
136.5
91.4
85.3
80.4
88.9
111.5
82.9

3.5
9.5
6.2
3.1
0.9
3.0
1.0
6.8
7.0
6.0
5.1
2.1
-1.4

change in
original
compactness
index
3.1
-4.2
0.6
-4.0
-3.7
-2.4
1.0
2.3
6.9
-4.2
-2.6
-1.6
-2.2
-1.2
3.1
-1.0
0.5
-3.5
2.3
-2.7
-1.3
-0.5
-3.7
-8.2
3.1
-7.8

104.9
83.5
100.3
125.3
121.7
111.0
141.5
96.8
90.3
85.8
95.0
115.3
88.3

-1.0
0.7
0.9
-0.5
-3.9
0.0
-4.0
1.4
2.0
0.6
-1.0
-1.6
-6.7

143

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

45035
45037
45039
45041
45045
45051
45055
45059
45063
45077
45079
45083
45085
45091
46083
46093
46099
46103
47001
47009
47011
47019
47021
47023
47037
47043
47047
47057
47063
47065
47073
47089
47093
47105
47111
47113
47115
47125
47147
47149

Dorchester County, SC
Edgefield County, SC
Fairfield County, SC
Florence County, SC
Greenville County, SC
Horry County, SC
Kershaw County, SC
Laurens County, SC
Lexington County, SC
Pickens County, SC
Richland County, SC
Spartanburg County, SC
Sumter County, SC
York County, SC
Lincoln County, SD
Meade County, SD
Minnehaha County, SD
Pennington County, SD
Anderson County, TN
Blount County, TN
Bradley County, TN
Carter County, TN
Cheatham County, TN
Chester County, TN
Davidson County, TN
Dickson County, TN
Fayette County, TN
Grainger County, TN
Hamblen County, TN
Hamilton County, TN
Hawkins County, TN
Jefferson County, TN
Knox County, TN
Loudon County, TN
Macon County, TN
Madison County, TN
Marion County, TN
Montgomery County, TN
Robertson County, TN
Rutherford County, TN

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
98.1
87.0
11.2
92.8
75.5
78.2
77.6
0.6
84.8
88.6
84.5
4.1
91.0
97.2
93.9
3.2
98.9
96.5
92.1
4.5
97.8
72.8
73.1
-0.3
79.0
82.7
77.2
5.4
84.3
89.9
85.2
4.7
90.6
86.9
82.9
4.0
87.8
107.2
102.0
5.2
106.6
91.8
85.8
6.0
90.7
90.3
84.9
5.4
90.0
88.9
83.0
5.9
88.1
81.5
89.1
85.9
3.2
91.3
102.9
100.0
2.9
106.6
95.5
94.1
1.4
99.5
92.0
88.8
3.3
94.1
94.4
89.3
5.1
93.4
94.4
90.3
4.1
95.9
97.3
97.3
0.0
102.2
80.6
74.5
6.1
79.0
65.4
58.6
6.7
68.4
112.1
102.6
9.5
106.1
83.4
78.2
5.2
85.4
70.7
80.5
76.3
4.2
84.5
98.4
95.0
3.4
99.5
103.8
99.8
4.0
104.5
86.0
85.0
1.0
90.8
86.6
81.0
5.6
86.4
100.8
99.0
1.8
103.9
93.3
84.7
8.7
90.2
68.3
91.0
79.6
11.5
87.6
84.8
89.3
83.9
5.4
90.2
77.7
72.8
4.9
82.5
93.6
84.4
9.1
89.6

change in
original
compactness
index
5.3
-6.6
-2.4
-1.8
-1.3
-6.2
-1.6
-0.7
-0.9
0.6
1.1
0.3
0.8
-2.2
-3.7
-3.9
-2.1
1.1
-1.5
-4.8
1.5
-3.1
5.9
-2.0
-4.0
-1.1
-0.7
-4.8
0.2
-3.1
3.1
3.5
-0.9
-4.8
3.9

144

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

47153
47157
47159
47163
47165
47167
47169
47171
47173
47179
47187
47189
48007
48013
48015
48019
48021
48027
48029
48037
48039
48041
48051
48055
48057
48061
48071
48077
48085
48091
48099
48113
48119
48121
48135
48139
48141
48157
48167
48181

county

Sequatchie County, TN
Shelby County, TN
Smith County, TN
Sullivan County, TN
Sumner County, TN
Tipton County, TN
Trousdale County, TN
Unicoi County, TN
Union County, TN
Washington County, TN
Williamson County, TN
Wilson County, TN
Aransas County, TX
Atascosa County, TX
Austin County, TX
Bandera County, TX
Bastrop County, TX
Bell County, TX
Bexar County, TX
Bowie County, TX
Brazoria County, TX
Brazos County, TX
Burleson County, TX
Caldwell County, TX
Calhoun County, TX
Cameron County, TX
Chambers County, TX
Clay County, TX
Collin County, TX
Comal County, TX
Coryell County, TX
Dallas County, TX
Delta County, TX
Denton County, TX
Ector County, TX
Ellis County, TX
El Paso County, TX
Fort Bend County, TX
Galveston County, TX
Grayson County, TX

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
72.3
111.6
105.2
6.4
109.7
82.9
81.7
1.2
89.5
97.8
93.1
4.7
97.3
91.3
86.2
5.1
91.5
79.8
75.7
4.1
81.8
76.4
73.9
2.5
81.2
106.3
103.5
2.8
108.1
82.7
81.2
1.5
88.5
96.5
92.2
4.3
96.4
94.7
81.8
12.8
87.4
83.8
77.3
6.5
82.7
101.2
100.9
0.2
104.4
87.2
85.3
1.8
93.4
82.8
85.6
80.7
4.9
85.9
88.3
86.8
1.6
92.4
106.8
99.8
7.1
105.0
117.6
113.9
3.7
118.8
92.9
89.5
3.4
95.3
99.2
95.7
3.5
99.8
110.1
106.7
3.5
110.9
90.5
83.6
6.8
89.8
88.1
82.2
5.9
90.6
108.0
98.5
9.5
103.9
107.4
101.1
6.3
107.9
84.1
85.7
-1.6
91.1
82.5
75.5
7.0
82.9
116.4
101.4
15.0
106.3
94.1
91.3
2.8
96.7
95.4
92.6
2.8
100.4
126.9
119.4
7.5
122.9
92.3
88.4
3.8
97.3
115.1
99.2
15.8
103.6
101.9
100.3
1.6
103.2
94.8
87.3
7.5
92.6
118.1
112.8
5.3
116.2
113.4
101.0
12.4
105.7
115.7
111.1
4.6
114.2
93.9
92.3
1.6
97.8

change in
original
compactness
index
2.0
-6.6
0.5
-0.2
-1.9
-4.8
-1.8
-5.8
0.1
7.3
1.1
-3.2
-6.3
-0.2
-4.0
1.8
-1.3
-2.4
-0.6
-0.8
0.7
-2.5
4.1
-0.5
-6.9
-0.4
10.1
-2.6
-5.0
4.0
-5.1
11.4
-1.3
2.2
1.9
7.7
1.5
-3.9

145

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

48183
48187
48199
48201
48209
48215
48231
48245
48251
48257
48259
48281
48291
48303
48309
48325
48329
48339
48355
48361
48367
48375
48381
48397
48401
48409
48423
48439
48451
48453
48459
48469
48473
48479
48485
48491
48493
48497
49005
49011

Gregg County, TX
Guadalupe County, TX
Hardin County, TX
Harris County, TX
Hays County, TX
Hidalgo County, TX
Hunt County, TX
Jefferson County, TX
Johnson County, TX
Kaufman County, TX
Kendall County, TX
Lampasas County, TX
Liberty County, TX
Lubbock County, TX
McLennan County, TX
Medina County, TX
Midland County, TX
Montgomery County, TX
Nueces County, TX
Orange County, TX
Parker County, TX
Potter County, TX
Randall County, TX
Rockwall County, TX
Rusk County, TX
San Patricio County, TX
Smith County, TX
Tarrant County, TX
Tom Green County, TX
Travis County, TX
Upshur County, TX
Victoria County, TX
Waller County, TX
Webb County, TX
Wichita County, TX
Williamson County, TX
Wilson County, TX
Wise County, TX
Cache County, UT
Davis County, UT

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
99.4
95.1
4.3
100.0
98.0
88.2
9.8
95.6
83.4
77.9
5.4
84.0
126.4
116.2
10.3
120.4
93.6
88.6
4.9
92.7
106.1
98.8
7.3
105.7
91.7
86.4
5.3
92.4
114.8
111.6
3.2
115.6
93.7
88.8
4.9
94.0
95.4
85.6
9.8
92.8
87.5
97.5
-10.0
101.9
85.4
85.9
-0.5
91.5
82.0
82.8
-0.8
88.6
106.0
104.7
1.3
111.3
101.8
99.6
2.2
106.4
81.2
109.3
109.9
-0.6
113.7
91.9
87.0
4.9
91.8
112.8
109.8
3.1
114.9
97.3
93.2
4.0
97.4
83.6
79.1
4.4
84.6
110.2
108.3
1.9
111.7
104.8
105.9
-1.1
111.6
98.8
90.2
8.7
95.1
75.3
72.1
3.1
79.2
94.2
92.0
2.2
97.1
99.5
93.4
6.1
98.7
120.0
113.0
7.0
116.3
97.3
91.3
6.0
98.5
114.6
109.6
5.0
112.6
81.0
73.1
7.8
81.2
102.3
103.0
-0.7
108.3
99.8
94.2
5.6
98.6
105.9
103.8
2.0
114.0
100.4
98.6
1.8
103.5
105.2
97.1
8.1
103.5
79.6
73.3
6.3
82.2
81.8
79.9
1.9
87.4
91.5
86.5
5.0
93.1
110.9
106.5
4.3
113.5

change in
original
compactness
index
-0.6
2.5
-0.6
6.1
0.9
0.5
-0.7
-0.8
-0.4
2.6
-14.4
-6.1
-6.6
-5.2
-4.6
-4.4
0.1
-2.0
-0.1
-1.0
-1.5
-6.8
3.8
-3.9
-2.8
0.8
3.7
-1.2
2.0
-0.3
-6.0
1.2
-8.1
-3.1
1.6
-2.6
-5.6
-1.6
-2.7

146

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

49023
49035
49043
49045
49049
49053
49057
50007
50011
50013
51003
51009
51011
51013
51019
51023
51031
51033
51041
51043
51053
51059
51061
51065
51067
51069
51073
51075
51079
51085
51087
51093
51095
51101
51107
51115
51121
51127
51143
51145

Juab County, UT
Salt Lake County, UT
Summit County, UT
Tooele County, UT
Utah County, UT
Washington County, UT
Weber County, UT
Chittenden County, VT
Franklin County, VT
Grand Isle County, VT
Albemarle County, VA
Amherst County, VA
Appomattox County, VA
Arlington County, VA
Bedford County, VA
Botetourt County, VA
Campbell County, VA
Caroline County, VA
Chesterfield County, VA
Clarke County, VA
Dinwiddie County, VA
Fairfax County, VA
Fauquier County, VA
Fluvanna County, VA
Franklin County, VA
Frederick County, VA
Gloucester County, VA
Goochland County, VA
Greene County, VA
Hanover County, VA
Henrico County, VA
Isle of Wight County, VA
James City County, VA
King William County, VA
Loudoun County, VA
Mathews County, VA
Montgomery County, VA
New Kent County, VA
Pittsylvania County, VA
Powhatan County, VA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
72.2
67.4
4.8
81.7
118.3
118.0
0.4
121.5
83.4
77.9
5.5
84.2
86.9
82.4
4.5
93.1
114.4
107.3
7.1
116.5
94.3
88.9
5.4
95.5
109.7
107.1
2.6
111.7
103.4
96.8
6.6
103.1
88.3
81.0
7.4
89.5
91.4
72.4
19.1
79.5
90.0
75.2
14.8
82.0
82.7
75.3
7.3
83.6
73.8
176.8
84.1
59.2
24.9
66.5
86.4
80.1
6.3
88.1
101.9
73.0
28.9
79.3
83.8
70.6
13.2
78.4
105.0
94.3
10.8
98.2
87.5
79.3
66.7
12.6
77.9
120.5
84.7
68.5
16.2
75.4
81.0
71.8
9.3
77.6
87.2
73.9
13.3
80.3
92.7
69.5
23.2
75.6
97.2
82.6
14.6
86.8
84.6
64.2
20.4
71.9
85.6
72.2
13.4
78.2
91.3
73.8
17.5
79.0
114.6
102.1
12.5
105.9
82.2
70.4
11.8
79.7
101.7
90.2
11.5
93.9
88.9
81.1
7.8
87.4
112.8
93.4
19.4
99.6
90.4
78.3
12.0
82.0
101.0
74.3
26.7
80.8
82.8
73.6
9.2
80.3
79.5
66.1
13.4
73.1
82.4
72.8
9.6
76.3

change in
original
compactness
index
-9.4
-3.2
-0.8
-6.2
-2.1
-1.1
-2.0
0.3
-1.2
12.0
8.0
-0.9

17.6
-1.7
22.6
5.5
6.8
1.4
9.3
3.4
6.9
17.1
10.4
12.7
7.4
12.3
8.7
2.5
7.9
1.5
13.2
8.3
20.2
2.5
6.4
6.1

147

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

Prince George County,


51149 VA
Prince William County,
51153 VA
51155 Pulaski County, VA
51161 Roanoke County, VA
51165 Rockingham County, VA
51169 Scott County, VA
51177 Spotsylvania County, VA
51179 Stafford County, VA
51183 Sussex County, VA
51187 Warren County, VA
51191 Washington County, VA
51199 York County, VA
51510 Alexandria city, VA
51515 Bedford city, VA
51520 Bristol city, VA
51540 Charlottesville city, VA
51550 Chesapeake city, VA
51570 Colonial Heights city, VA
51590 Danville city, VA
51600 Fairfax city, VA
51610 Falls Church city, VA
51630 Fredericksburg city, VA
51650 Hampton city, VA
51660 Harrisonburg city, VA
51670 Hopewell city, VA
51680 Lynchburg city, VA
51683 Manassas city, VA
51685 Manassas Park city, VA
51700 Newport News city, VA
51710 Norfolk city, VA
51730 Petersburg city, VA
51735 Poquoson city, VA
51740 Portsmouth city, VA
51750 Radford city, VA
51760 Richmond city, VA
51770 Roanoke city, VA
51775 Salem city, VA
51800 Suffolk city, VA

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index

change in
original
compactness
index

84.4

105.5

-21.1

108.6

-24.2

116.4
98.9
101.0
89.5
88.9
99.2
98.7
87.2
96.5
92.0
105.1
181.3
100.5
120.4
138.2
108.6
121.8
109.0
115.8
134.6
137.2
127.5
132.6
135.0
118.6
125.9
134.2
125.3
148.1
118.0
106.4
130.6
126.4
135.9
127.7
121.8
97.2

101.8
89.1
76.8
71.3
86.6
72.2
87.3

14.6
9.8
24.2
18.1
2.3
27.0
11.4

94.5
82.0
78.4
93.5
78.0
92.4

4.4
19.0
11.0
-4.6
21.2
6.3

90.1
92.7
111.0

6.5
-0.7
-5.8

96.1
97.7
112.2

0.4
-5.7
-7.1

104.7

3.9

108.0

0.6

138.4

9.7

141.6

6.5

128.0

2.5

132.4

-1.8

132.1

3.8

135.8

0.1

87.0

10.2

93.8

3.4

148

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

51810
51830
51840
53003
53005
53007
53011
53015
53017
53021
53033
53035
53053
53057
53061
53063
53067
53073
53077
54003
54005
54009
54011
54029
54037
54039
54051
54057
54061
54065
54069
54077
54079
54099
54107
55009
55015
55017
55021
55025

Virginia Beach city, VA


Williamsburg city, VA
Winchester city, VA
Asotin County, WA
Benton County, WA
Chelan County, WA
Clark County, WA
Cowlitz County, WA
Douglas County, WA
Franklin County, WA
King County, WA
Kitsap County, WA
Pierce County, WA
Skagit County, WA
Snohomish County, WA
Spokane County, WA
Thurston County, WA
Whatcom County, WA
Yakima County, WA
Berkeley County, WV
Boone County, WV
Brooke County, WV
Cabell County, WV
Hancock County, WV
Jefferson County, WV
Kanawha County, WV
Marshall County, WV
Mineral County, WV
Monongalia County, WV
Morgan County, WV
Ohio County, WV
Preston County, WV
Putnam County, WV
Wayne County, WV
Wood County, WV
Brown County, WI
Calumet County, WI
Chippewa County, WI
Columbia County, WI
Dane County, WI

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
124.5
117.4
7.1
120.6
123.3
128.4
113.9
107.6
6.2
113.0
99.5
93.7
5.8
99.2
98.4
91.6
6.7
104.0
107.2
103.1
4.1
108.5
96.6
88.3
8.2
96.9
97.5
91.0
6.5
102.1
103.2
94.5
8.6
102.1
127.1
120.5
6.6
126.0
100.9
96.7
4.2
102.3
112.3
107.1
5.2
113.0
98.1
91.9
6.1
98.3
107.2
99.9
7.3
106.3
111.9
109.9
2.0
113.8
100.6
93.6
7.0
99.3
98.6
91.3
7.3
97.9
94.2
90.4
3.9
100.4
97.1
89.2
7.9
92.4
102.4
99.5
2.9
105.1
104.2
99.1
5.1
104.0
113.0
110.2
2.8
115.5
107.3
108.2
-0.9
112.2
95.5
89.8
5.6
94.4
113.5
108.6
4.8
113.7
105.9
98.1
7.9
104.1
96.6
90.3
6.3
96.9
110.8
105.6
5.2
110.6
83.3
77.4
6.0
83.2
114.8
111.3
3.6
116.2
84.7
75.3
9.4
82.5
100.7
95.8
5.0
100.7
102.2
98.4
3.8
108.4
111.9
102.7
9.2
108.8
95.2
94.2
1.0
99.8
87.6
74.5
13.2
83.1
84.6
79.1
5.5
86.5
85.7
74.2
11.5
82.7
109.9
102.4
7.5
109.7

change in
original
compactness
index
3.9

0.8
0.3
-5.6
-1.3
-0.3
-4.7
1.1
1.0
-1.4
-0.7
-0.3
0.9
-1.9
1.3
0.8
-6.2
4.7
-2.7
0.2
-2.5
-4.9
1.1
-0.2
1.8
-0.3
0.2
0.1
-1.4
2.2
0.1
-6.2
3.1
-4.7
4.5
-1.9
3.0
0.2

149

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

55031
55035
55039
55049
55059
55061
55063
55073
55079
55083
55087
55089
55093
55101
55105
55109
55117
55131
55133
55139
56021
56025

county

Douglas County, WI
Eau Claire County, WI
Fond du Lac County, WI
Iowa County, WI
Kenosha County, WI
Kewaunee County, WI
La Crosse County, WI
Marathon County, WI
Milwaukee County, WI
Oconto County, WI
Outagamie County, WI
Ozaukee County, WI
Pierce County, WI
Racine County, WI
Rock County, WI
St. Croix County, WI
Sheboygan County, WI
Washington County, WI
Waukesha County, WI
Winnebago County, WI
Laramie County, WY
Natrona County, WY

new
new
change in
original
compactness compactness
new
compactness
index 2010
index 2000 compactness
index
index
90.2
87.2
2.9
95.1
97.1
99.7
-2.6
106.8
94.3
85.6
8.6
92.6
74.8
73.7
1.1
82.6
110.2
106.5
3.7
111.2
67.8
65.3
2.5
76.4
108.2
105.2
3.0
110.5
85.3
81.9
3.3
90.7
139.8
141.2
-1.4
145.3
72.1
68.6
3.5
77.1
99.5
96.7
2.8
102.9
92.1
86.7
5.4
93.0
87.5
80.5
7.1
90.5
105.5
103.7
1.8
108.1
97.7
95.2
2.5
100.5
77.5
72.1
5.4
81.4
97.5
92.1
5.4
98.2
86.2
79.0
7.1
85.3
99.5
90.3
9.2
94.5
107.0
104.6
2.4
109.7
99.4
95.7
3.7
102.0
105.3
94.6
10.7
102.4

change in
original
compactness
index
-4.9
-9.7
1.6
-7.8
-1.0
-8.6
-2.3
-5.4
-5.5
-4.9
-3.4
-0.9
-3.0
-2.6
-2.9
-3.9
-0.7
0.8
5.0
-2.8
-2.7
2.8

150

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix B. County Compactness Factors and Composite Indices for


2010
fips

county

1009

Blount County

1015
1021
1033
1051
1055
1069
1073
1077
1079
1081
1083
1089
1097
1101
1103
1113
1115
1117
1125
1127
4005
4013
4015
4019
4021
4025
4027
5007
5031
5033
5035
5045
5051

Calhoun County
Chilton County
Colbert County
Elmore County
Etowah County
Houston County
Jefferson County
Lauderdale County
Lawrence County
Lee County
Limestone County
Madison County
Mobile County
Montgomery County
Morgan County
Russell County
St. Clair County
Shelby County, AL
Tuscaloosa County, AL
Walker County, AL
Coconino County, AZ
Maricopa County, AZ
Mohave County, AZ
Pima County, AZ
Pinal County, AZ
Yavapai County, AZ
Yuma County, AZ
Benton County, AR
Craighead County, AR
Crawford County, AR
Crittenden County, AR
Faulkner County, AR
Garland County, AR

density
mix
centering street
composite
factor
factor
factor
factor
index
90.36
37.85
74.28
60.14
56.6
91.58
89.98
95.11
91.59
93.78
94.83
99.01
94.46
89.38
96.48
91.62
97.61
99.06
102.14
96.47
94.83
91.04
94.43
96.71
90.6
95.58
110.5
96.2
102.91
96.42
96
99.68
95.22
95.83
92.25
96.93
95.11
92.69

86.7
52.55
104.27
60.63
91.28
102.37
110.72
84.43
51.74
87.9
58.45
98.59
108.17
120.67
95.35
90.91
55.96
91.33
101.44
65.74
105.89
118.07
90.76
109.55
74.63
89.71
105.56
95.05
97.46
90.19
115.43
92.1
89.51

117.7
81.61
76.99
86.59
116.86
98.64
122.44
105.63
86.98
104.17
89.78
103.31
93.94
118.34
116.51
78.65
81.95
88.2
136.82
86.66
159.7
118.48
97.35
129.25
93.08
88.28
142.91
104.81
113.68
82.88
79.24
83.67
116.53

104.38
62.37
124.68
85.71
93.1
88.97
126.81
88.5
66.67
84.55
82.64
114.82
113.78
105.98
101.04
93.54
84.47
92.91
110.56
92.5
80.11
118.04
95.37
101.54
100.74
86.4
107.38
89.33
76.68
80.03
89.18
74.78
103.18

100.11
64.14
100.33
76.15
98.43
95.2
118.64
91.48
66.75
91.5
75.51
104.53
104.72
114.89
102.96
86.71
72.65
89.53
114.39
79.62
113.04
120.56
93.58
113.66
88.9
87.49
117.54
95.07
94.83
82.74
93.93
82.83
100.6

151

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

5053
5069
5079
5085
5087
5091
5111
5119
5125
5131
5143
6001
6007
6013
6017
6019
6025
6029
6031
6037
6039
6041
6047
6053
6055
6059
6061
6065
6067
6069
6071
6073
6075
6077
6079
6081
6083
6085

Grant County, AR
Jefferson County, AR
Lincoln County, AR
Lonoke County, AR
Madison County, AR
Miller County, AR
Poinsett County, AR
Pulaski County, AR
Saline County, AR
Sebastian County, AR
Washington County, AR
Alameda County, CA
Butte County, CA
Contra Costa County, CA
El Dorado County, CA
Fresno County, CA
Imperial County, CA
Kern County, CA
Kings County, CA
Los Angeles County, CA
Madera County, CA
Marin County, CA
Merced County, CA
Monterey County, CA
Napa County, CA
Orange County, CA
Placer County, CA
Riverside County, CA
Sacramento County, CA
San Benito County, CA
San Bernardino County, CA
San Diego County, CA
San Francisco County, CA
San Joaquin County, CA
San Luis Obispo County, CA
San Mateo County, CA
Santa Barbara County, CA
Santa Clara County, CA

density
mix
centering street
composite
factor
factor
factor
factor
index
89.11
79.34
77.98
60.72
70.67
94.66
97.82
96.55
113.66
100.85
88.97
51.59
72.47
62.71
60.74
91.76
79.64
91.84
75.65
80.69
88.44
61.16
73.67
72.44
67.05
97.29
106.83
82.03
115.58
100.54
89.31
105.78
77.99
71.03
82.34
100.95
111.48
116.72
127.01
117.74
92.78
80.99
106.43
75.8
86.1
97.44
103.71
93.42
108.24
100.89
98.58
104.46
109.89
91.83
101.5
137.65
143.4
115.28
151.09
146.57
99.2
121.87
106.28
91.9
106.08
112.02
128.7
100.81
121.28
119.84
96.18
88.17
84.58
77.8
83.17
103.35
127.85
104.03
94.25
109.31
99.38
132.78
99.61
82.71
104.58
102.91
121.33
99.62
92.21
105.08
100.77
115.21
108.98
90.98
105.04
152.55
145.2
121.62
141.02
150.67
96.68
110.34
104.67
69.69
94.12
109.25
141.52
96.85
111.15
118.57
100.54
122.04
112.8
85.94
106.74
109.05
122.36
110.26
101.72
113.71
102.69
135.45
131.01
110.28
125.09
134.15
142.55
95.13
144.21
136.66
101.97
116.93
90.93
98.05
102.49
105.36
117.55
108.49
98.38
109.41
115.28
128.54
135.7
129.68
134.5
103.1
115.79
78.56
105.1
100.81
106.82
122.13
95.87
92.42
105.45
118.35
129.64
121.82
116.14
127.15
250.84
153.79
258.47
215.72
251.27
106.5
132.92
104.79
118.62
119.85
97.52
124.79
111.43
102.74
111.53
130.72
144.53
93.82
131.35
131.72
116.62
139.7
112.02
116.13
126.69
131.02
139.68
107.58
132.85
135.11

152

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

6087
6089
6095
6097
6099
6101
6107
6111
6113
6115
8001
8005
8013
8014
8019
8031
8035
8039
8041
8059
8069
8077
8101
8119
8123
9001
9003
9007
9009
9011
9013
10001
10003
11001
12001
12003
12005
12009

Santa Cruz County, CA


Shasta County, CA
Solano County, CA
Sonoma County, CA
Stanislaus County, CA
Sutter County, CA
Tulare County, CA
Ventura County, CA
Yolo County, CA
Yuba County, CA
Adams County, CO
Arapahoe County, CO
Boulder County, CO
Broomfield County, CO
Clear Creek County, CO
Denver County, CO
Douglas County, CO
Elbert County, CO
El Paso County, CO
Jefferson County, CO
Larimer County, CO
Mesa County, CO
Pueblo County, CO
Teller County, CO
Weld County, CO
Fairfield County, CT
Hartford County, CT
Middlesex County, CT
New Haven County, CT
New London County, CT
Tolland County, CT
Kent County, DE
New Castle County, DE
District of Columbia, DC
Alachua County, FL
Baker County, FL
Bay County, FL
Brevard County, FL

density
mix
centering street
composite
factor
factor
factor
factor
index
104.2
138.71
114.16
107.34
120.35
96
110.79
114.25
88.66
103.07
106.86
130.6
103.94
114.95
117.8
100.37
131.12
101.87
97.67
109.81
107.86
135.71
94.54
107.84
114.52
98.92
119.22
126.45
82.89
108.68
100.44
117.82
102.53
93.41
104.49
110.13
131.48
99.8
114.98
117.82
107.3
126.92
98.5
110.1
113.53
97.57
95.43
82.17
89.37
88.8
106.63
122.25
82.26
122.37
110.59
114.44
124.3
102.43
134.2
123.81
107.71
122
111.33
115.52
117.87
105.87
113.8
83.11
129.14
110.09
90.58
67.38
117.81
129.34
137.67
174.54
181.54
170.48
102.77
97.61
92.17
97.77
96.94
88.27
44.14
72.69
50.26
54.3
104.62
119.18
95.89
123.96
113.79
106.94
125.25
90.89
112.99
111.4
100.68
117.76
111.95
103.05
110.57
101.69
113.73
124.35
107.33
114.88
100.43
112.15
112.96
121.67
114.91
94.68
82.25
81.88
108.04
89.53
97.29
114.35
111.18
95.06
105.65
110.88
131.47
125.41
101.99
122.04
107.85
126.56
138.02
92.46
120.5
95.74
116.02
98.9
81.98
97.68
107.16
128.91
137.15
102.88
124.04
96.76
106.51
131.52
85.24
106.33
96.05
89.61
97.77
63.29
83.17
94.72
97.37
102.26
89.82
95
108.44
126.15
111.75
121.39
121.4
193.52
138.05
219.97
185.15
206.37
100.66
110.17
115.43
107.74
110.74
89.21
63.21
89.68
61.02
69.39
99.21
105.55
93.7
115.16
104.31
102.39
103.2
86.39
110.4
100.75

153

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

12011
12015
12019
12021
12031
12033
12035
12039
12053
12057
12061
12069
12071
12073
12081
12083
12085
12086
12089
12091
12095
12097
12099
12101
12103
12105
12109
12111
12113
12115
12117
12127
12129
13013
13015
13021
13029
13035

Broward County, FL
Charlotte County, FL
Clay County, FL
Collier County, FL
Duval County, FL
Escambia County, FL
Flagler County, FL
Gadsden County, FL
Hernando County, FL
Hillsborough County, FL
Indian River County, FL
Lake County, FL
Lee County, FL
Leon County, FL
Manatee County, FL
Marion County, FL
Martin County, FL
Miami-Dade County, FL
Nassau County, FL
Okaloosa County, FL
Orange County, FL
Osceola County, FL
Palm Beach County, FL
Pasco County, FL
Pinellas County, FL
Polk County, FL
St. Johns County, FL
St. Lucie County, FL
Santa Rosa County, FL
Sarasota County, FL
Seminole County, FL
Volusia County, FL
Wakulla County, FL
Barrow County, GA
Bartow County, GA
Bibb County, GA
Bryan County, GA
Butts County, GA

density
mix
centering street
composite
factor
factor
factor
factor
index
120.61
133.24
95.43
148.86
131.01
94.98
97.96
103.74
114.83
103.64
97.16
92.55
98.14
95.4
94.71
99.42
104.7
83.67
105.06
97.74
106.31
113.1
118.71
125.06
119.96
99.94
109.08
100.14
116.67
108.16
96.82
82.32
79.96
99.05
86.78
90.27
57.12
83.72
95.13
76.69
96.2
80.29
108.25
102.08
95.84
106.16
115.63
127.6
128.18
124.51
97.1
101.81
112.72
132.01
113.79
95.53
87.32
121.33
116.84
106.64
98.87
104.6
119.36
121.83
114.11
102.05
106.83
149.96
99.11
118.31
102.17
114.33
112.33
129.01
118.27
93.51
83.3
140.38
98.85
105.07
98.62
110.16
106.69
113.84
109.26
137.38
132.85
131.33
156.48
149.93
93.25
78.04
98.01
97.21
89.42
100.2
113.18
109.67
105.87
109.14
108.01
110.76
118.48
124.47
119.5
98.45
86.64
87.23
114.77
95.92
107.77
125.08
107.06
118.32
118.4
99.18
100.48
84.02
117.84
100.48
114.66
132.11
93.74
163.76
132.94
96.76
90.29
115.86
120.94
107.53
97.43
86.85
85.06
106.86
92.48
100.74
97.46
102.45
120.07
106.54
92.28
93.99
81.78
80.59
83.78
101.61
116.04
113.62
124.42
117.59
105.12
116.39
81.81
121.13
107.72
99.33
107.91
100.7
115.72
107.47
89.66
45.54
78.68
79.41
66.29
92.36
70.78
85.3
72.18
74.92
90.76
77.69
86.6
80.47
79.63
98.07
113.15
103.59
112.7
108.69
89.84
61.04
81.95
71.54
69.79
91.1
82.26
87.09
67.51
77.24

154

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

13045
13047
13051
13053
13057
13059
13063
13067
13073
13077
13083
13085
13089
13095
13097
13103
13113
13115
13117
13121
13127
13135
13139
13143
13145
13151
13153
13169
13171
13177
13179
13185
13189
13195
13199
13207
13213
13215

Carroll County, GA
Catoosa County, GA
Chatham County, GA
Chattahoochee County, GA
Cherokee County, GA
Clarke County, GA
Clayton County, GA
Cobb County, GA
Columbia County, GA
Coweta County, GA
Dade County, GA
Dawson County, GA
DeKalb County, GA
Dougherty County, GA
Douglas County, GA
Effingham County, GA
Fayette County, GA
Floyd County, GA
Forsyth County, GA
Fulton County, GA
Glynn County, GA
Gwinnett County, GA
Hall County, GA
Haralson County, GA
Harris County, GA
Henry County, GA
Houston County, GA
Jones County, GA
Lamar County, GA
Lee County, GA
Liberty County, GA
Lowndes County, GA
McDuffie County, GA
Madison County, GA
Meriwether County, GA
Monroe County, GA
Murray County, GA
Muscogee County, GA

density
mix
centering street
composite
factor
factor
factor
factor
index
92.24
80.47
108.64
59.41
81.28
93.34
79.45
88.25
78.55
80.91
99.64
117.03
126.17
126.88
122.03
97.14
100.48
70.87
98.62
89.61
97.06
94.58
80.91
83.44
86.1
100.91
115.76
98.31
92.89
102.49
106.35
106.15
84.62
98.1
98.49
106.99
116.91
91.39
107.76
107.28
96.83
95.43
80.24
72.04
82.48
92.69
85.33
81.74
72.61
78.64
89.57
56.36
80.64
69.91
67.3
89.94
63.53
86.08
69.43
71.24
111.99
120.73
96.18
100.65
109.34
97.65
109.27
95.6
107.9
103.3
95.83
89.53
103.33
70.96
87.25
91.03
60.74
84.13
75.9
72.13
93.23
94.36
100.88
78.34
89.51
92.92
90.67
103.37
89.35
92.52
96.31
91.93
97.11
68.48
85.41
107.63
122.6
146.48
108.57
126.94
92.87
102
95.73
111.38
100.62
106.36
111.94
88.7
89.68
98.95
94.45
89.1
139.3
87.59
103.3
90.08
73.41
78.3
82.15
75.97
89.51
34.28
71.89
62.25
55.12
95.26
81.75
86.07
74.28
80.21
99.67
97.7
89.66
91.56
93.23
90.26
80.32
81.59
59.82
72.19
90.01
68.75
79.24
69.42
70.75
90.74
63.81
80.13
67.38
69.06
96.95
85.66
100.72
88.85
91.21
95.78
102.08
106.87
91.72
98.88
89.94
68.85
78.49
72.18
71.4
89.81
53.09
73.41
61.79
61.49
89.17
52.92
79.4
65.55
64.31
89.72
49.47
77.43
66.44
63.06
90.63
57.18
84.75
68.86
68.85
103.92
119.01
133.98
108.41
120.64

155

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

13217
13219
13221
13223
13227
13231
13245
13247
13255
13273
13295
13297
13313
13321
16001
16005
16019
16027
16045
16051
16055
16069
17003
17005
17007
17019
17027
17031
17037
17043
17053
17063
17073
17083
17089
17091
17093
17097

Newton County, GA
Oconee County, GA
Oglethorpe County, GA
Paulding County, GA
Pickens County, GA
Richmond County, GA
Rockdale County, GA
Spalding County, GA
Terrell County, GA
Walker County, GA
Walton County, GA
Whitfield County, GA
Worth County, GA
Ada County, ID
Bannock County, ID
Bonneville County, ID
Canyon County, ID
Gem County, ID
Jefferson County, ID
Kootenai County, ID
Nez Perce County, ID
Alexander County, IL
Bond County, IL
Boone County, IL
Champaign County, IL
Clinton County, IL
Cook County, IL
DeKalb County, IL
DuPage County, IL
Ford County, IL
Grundy County, IL
Henry County, IL
Jersey County, IL
Kane County, IL
Kankakee County, IL
Kendall County, IL
Lake County, IL

density
mix
centering street
composite
factor
factor
factor
factor
index
94.48
61.24
123.65
77.77
86.46
90.84
85.05
74.86
69.72
74.87
88.61
22.76
70.81
45.28
45.49
93.49
68.19
83.49
74.96
74.76
90.19
68.61
81.67
61.08
68.89
99.09
95.92
93.04
88.84
91.84
91.96
94.64
88.76
103.58
101.28
98.84
98.64
92.23
89.1
97.55
99.34
89.05
91.76
96.36
109.28
89.17
151.4
99.94
111.41
90
92.99
90.62
89.46
108.34
95.65
94.3
103.98

111.4
93.91
83.74
78.95
77.95
71.8
87.29
52.25
124.6
123.06
118.52
112.28
83.41
69.82
113.96
116.89
87.79
95.37
127.58
87.01
141.34
111.36
135.96
136.48
101.16
116.08
78.12
120.57
119.77
90.54
121.02

124.13
82.64
102.12
78.22
88.88
87.33
115.72
84.69
102.02
128.18
99.62
90.6
76.44
83.29
122.32
92.82
70.12
129.58
81.63
141.54
82.04
155.66
84.27
88.41
78.31
86.63
84.59
85.72
90.86
105.98
82.01
97.08

104.91
86.78
85.73
74.53
75.62
54.96
88.51
68.22
108.68
124.04
109.57
106.1
113.29
88.98
101.44
113.12
121.33
109.49
85.74
107.66
94.5
170.12
93.39
126.48
83.16
110.27
81.22
85.66
109.06
97.47
95.42
118.15

112.49
87.13
88.83
74.9
79.24
70.32
95.63
66.48
112.28
124.18
108.39
102.41
89.06
78.26
111.14
107
105.89
87.08
127.19
85.06
169.04
96.51
119.67
96.19
97.17
91.31
80.72
109.11
105.96
88.08
112.71

156

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

17111
17113
17115
17117
17119
17123
17129
17131
17133
17143
17147
17161
17163
17167
17179
17183
17197
17201
17203
18003
18005
18011
18013
18015
18019
18021
18029
18035
18039
18043
18047
18051
18055
18057
18059
18061
18063
18067

McHenry County, IL
McLean County, IL
Macon County, IL
Macoupin County, IL
Madison County, IL
Marshall County, IL
Menard County, IL
Mercer County, IL
Monroe County, IL
Peoria County, IL
Piatt County, IL
Rock Island County, IL
St. Clair County, IL
Sangamon County, IL
Tazewell County, IL
Vermilion County, IL
Will County, IL
Winnebago County, IL
Woodford County, IL
Allen County, IN
Bartholomew County, IN
Boone County, IN
Brown County, IN
Carroll County, IN
Clark County, IN
Clay County, IN
Dearborn County, IN
Delaware County, IN
Elkhart County, IN
Floyd County, IN
Franklin County, IN
Gibson County, IN
Greene County, IN
Hamilton County, IN
Hancock County, IN
Harrison County, IN
Hendricks County, IN
Howard County, IN

density
mix
centering street
composite
factor
factor
factor
factor
index
98.53
105.24
83.23
95.57
94.49
104.94
120.63
110.85
102.41
112.27
95.56
114.15
112.75
97.28
106.24
92.2
111.71
78.1
115.16
99.1
96.83
119.34
103.17
114.28
110.62
89.56
95.57
68.03
113.51
89.47
88.81
90.2
83.8
84.09
83.22
88.81
97.3
71.15
95.19
84.98
89.84
90.63
77.62
91.7
84.14
100.95
120.84
143.87
112.87
124.81
88.83
107.89
81.61
83.39
87.9
101.09
128.28
104.97
116.1
115.93
96.6
114.62
90.19
113.08
104.58
97.54
115.25
157.52
108.44
124.88
96.01
107.55
85.37
110.59
99.85
91.84
99.84
112.75
117.88
107.05
101.35
114.01
92.55
100.58
102.68
100.8
123.79
117.91
120.01
119.75
89.23
111.21
85.84
94.01
93.77
100.69
113.3
110.06
100.51
107.76
96.38
101.42
108.25
114.65
106.54
94.39
103.9
79.83
90.61
90.12
92.73
36.11
76.3
63.42
58.47
89.42
86.26
86.24
85.98
83.54
97.57
113.96
86.06
107.2
101.51
91.51
101.15
76.58
109.38
93.25
91.96
82.67
89.51
96.29
87.5
103.15
118.8
91.63
109.13
107.18
94.95
104.81
89.66
114.82
101.34
101.1
121.02
86.15
99.15
102.35
90.85
54.82
78.33
95.48
74.56
92.92
109.39
77.46
124.54
101.36
90.44
93.15
82.02
88.86
85.62
99.85
104.3
81.69
94.95
93.93
93.31
95.1
82.93
84.8
86.14
91.11
56.7
85.5
61.31
66.71
95.72
91.32
79.42
89.16
85.98
98.37
114.28
95.94
109.61
105.75

157

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

18073
18081
18089
18091
18095
18097
18105
18109
18115
18119
18127
18129
18133
18141
18145
18153
18157
18159
18163
18165
18167
18173
18175
18179
18183
19011
19013
19017
19049
19061
19085
19103
19105
19113
19121
19129
19153
19155

Jasper County, IN
Johnson County, IN
Lake County, IN
LaPorte County, IN
Madison County, IN
Marion County, IN
Monroe County, IN
Morgan County, IN
Ohio County, IN
Owen County, IN
Porter County, IN
Posey County, IN
Putnam County, IN
St. Joseph County, IN
Shelby County, IN
Sullivan County, IN
Tippecanoe County, IN
Tipton County, IN
Vanderburgh County, IN
Vermillion County, IN
Vigo County, IN
Warrick County, IN
Washington County, IN
Wells County, IN
Whitley County, IN
Benton County, IA
Black Hawk County, IA
Bremer County, IA
Dallas County, IA
Dubuque County, IA
Harrison County, IA
Johnson County, IA
Jones County, IA
Linn County, IA
Madison County, IA
Mills County, IA
Polk County, IA
Pottawattamie County, IA

density
mix
centering street
composite
factor
factor
factor
factor
index
89.52
90.18
73.22
51.82
69.9
98.31
116.23
81.08
102.48
99.4
102.28
124.13
124.4
126.26
124.35
95.04
104.81
108.11
96.11
101.29
96.4
113.83
107.92
112.32
109.63
108.62
123.19
125.02
127.04
126.5
104.36
112.59
163.85
98.52
125.06
94.61
85.99
85.6
99.46
89.15
91.06
97.13
78.9
99.39
89.41
91.06
35.65
78.62
99.32
69.87
96.95
108.4
88.88
87.95
94.37
92.19
75.2
81.37
81.92
78.1
91.01
96.03
82.78
73.04
81.95
100.67
117.65
124.8
131.2
123.48
98.24
116
82.26
97.84
98.21
89.97
94.33
85.42
79.03
83.81
104.58
112.14
101.52
96
104.5
89.55
85.73
80.1
62.84
74.17
101.79
119.7
120.43
116.35
118.41
103.23
90.48
79.32
155.06
108.87
96.9
111.19
114.75
128.65
116.27
99.66
102.11
81.65
82.32
89.18
94.15
67.81
80.3
87.16
77.7
89.98
90.1
83.04
70.18
78.93
90.31
89.14
84.12
56.3
74.69
88.87
108.97
90.6
97.81
95.65
99.1
129.91
94.2
118.5
113.18
89
112.79
82.24
77.7
87.91
95.45
106.94
79.89
91.67
91.77
100.57
130.56
115.08
106.99
116.81
89.16
113.13
76.21
76.79
85.87
103.02
124.12
157.95
85.78
122.39
89.77
115.53
71.55
95.83
91.37
100.19
118.29
121.29
103.21
113.58
90.62
124.56
70.25
103.16
96.4
89.93
84.78
77.08
92.04
82.25
102.96
129.31
116.94
112.82
119.6
97.53
120.78
95.92
99.22
104.25

158

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

19163
19169
19181
19183
19193
20015
20045
20059
20061
20079
20085
20091
20103
20121
20139
20149
20161
20173
20177
20191
20209
21015
21017
21019
21029
21037
21047
21049
21059
21067
21081
21089
21093
21101
21103
21111
21113
21117

Scott County, IA
Story County, IA
Warren County, IA
Washington County, IA
Woodbury County, IA
Butler County, KS
Douglas County, KS
Franklin County, KS
Geary County, KS
Harvey County, KS
Jackson County, KS
Johnson County, KS
Leavenworth County, KS
Miami County, KS
Osage County, KS
Pottawatomie County, KS
Riley County, KS
Sedgwick County, KS
Shawnee County, KS
Sumner County, KS
Wyandotte County, KS
Boone County, KY
Bourbon County, KY
Boyd County, KY
Bullitt County, KY
Campbell County, KY
Christian County, KY
Clark County, KY
Daviess County, KY
Fayette County, KY
Grant County, KY
Greenup County, KY
Hardin County, KY
Henderson County, KY
Henry County, KY
Jefferson County, KY
Jessamine County, KY
Kenton County, KY

density
mix
centering street
composite
factor
factor
factor
factor
index
100.21
128.03
85.19
130.22
113.79
96.6
115.01
125.73
97.63
111.05
93.98
105.61
82.31
83.56
89.09
90
104.89
78.56
86.53
87.36
97.33
125.17
117.13
122.41
119.6
95.93
116.69
81.59
76.86
90.86
100.21
127.37
99.68
98.22
108.05
89.92
101.1
85.19
101.84
93.07
96.96
84.76
128.69
90.56
115.17
75.64
73.36
85.7
88.64
77.77
79.63
44.65
65.47
104.45
125.43
86.47
101.88
105.76
95.13
99.39
87.24
93.72
92.25
89.17
87.98
79.03
102.93
87.08
89.37
97.03
68.66
75.02
77.91
89
81.55
95.3
98.61
93.38
105.56
102.93
118.91
117.57
112.3
116.34
98.59
111.59
125.79
108.8
114.14
88.32
98.41
84.72
92.96
88.76
101.91
113.88
103.1
127.92
114.79
99.7
101.93
95.37
84.83
94.26
97.22
93.99
80.83
92.96
88.94
94.45
98.55
126.68
104.55
107.65
95.94
83.26
81.17
86.62
83.25
102.73
124.27
85.29
109.72
106.95
97.34
94.37
87.11
104.06
94.59
93.45
102
79.27
98.84
91.64
99.18
109.86
121.56
106.12
111.6
110.05
128.66
134.26
116.37
128.22
90.59
52.57
80.01
76.95
68.44
94.52
87.52
78.55
112.22
91.41
95.48
90.76
131.65
93.87
103.72
99.09
105.95
76.39
103.24
95.15
89.37
76.6
77.64
85.73
77.68
109.11
119.34
118.64
123.85
122.42
94.35
102.5
84.93
91.02
91.41
104.06
117.51
88.49
119.32
109.28

159

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

21123
21163
21179
21185
21209
21211
21215
21227
21239
22005
22015
22017
22019
22031
22033
22043
22047
22051
22055
22057
22063
22071
22073
22075
22077
22079
22087
22089
22095
22099
22103
22109
22111
22121
23001
23005
23019
23023

Larue County, KY
Meade County, KY
Nelson County, KY
Oldham County, KY
Scott County, KY
Shelby County, KY
Spencer County, KY
Warren County, KY
Woodford County, KY
Ascension Parish, LA
Bossier Parish, LA
Caddo Parish, LA
Calcasieu Parish, LA
De Soto Parish, LA
East Baton Rouge Parish, LA
Grant Parish, LA
Iberville Parish, LA
Jefferson Parish, LA
Lafayette Parish, LA
Lafourche Parish, LA
Livingston Parish, LA
Orleans Parish, LA
Ouachita Parish, LA
Plaquemines Parish, LA
Pointe Coupee Parish, LA
Rapides Parish, LA
St. Bernard Parish, LA
St. Charles Parish, LA
St. John the Baptist Parish, LA
St. Martin Parish, LA
St. Tammany Parish, LA
Terrebonne Parish, LA
Union Parish, LA
West Baton Rouge Parish, LA
Androscoggin County, ME
Cumberland County, ME
Penobscot County, ME
Sagadahoc County, ME

density
mix
centering street
composite
factor
factor
factor
factor
index
89.43
63.3
84.72
65.93
69.47
93.39
46.63
84.9
78.41
69.46
91.95
66.86
78.24
89.54
76.81
94.48
74.42
80.9
81.7
78.36
95.24
97.32
80.79
97.28
90.72
95.85
91.76
112.29
86.78
95.79
91.13
31.97
75.02
76.42
60.36
101.86
102.72
124.59
100.77
109.46
93.43
105.61
79.51
90.95
90.36
92.32
90.2
93.22
86.92
88.2
95.13
94.84
83.39
90.35
88.54
98.39
108.22
98.44
110.2
104.82
95.68
105.58
123.81
94.14
106.07
89.07
61.88
140.34
77.66
90.19
103.91
113.92
97.85
114.04
109.39
88.67
34.23
66.17
64.67
53.79
93.41
93.69
84.62
92.02
88.54
113.17
132.12
84.47
148.19
124.62
99.95
114.45
110.96
106.53
110.08
95.04
99.35
143.72
98.05
111.43
93.18
62.05
84.88
75.38
73.3
121.91
137.94
153.63
214.43
172.01
95.23
94.61
111.6
108.52
103.15
90.01
91.73
81.72
104.87
90
91.55
71.09
98.29
93.23
98.11
100.74
101.17
97.87
100.03
121.48
80.94
130.72
110.48
93.42
97.97
81.23
108.41
94.01
97.39
101.63
88.78
109.44
99.13
90.6
70.42
94.32
86.13
81.51
95.66
94.37
97.06
109.33
98.87
96.62
103.72
99.01
107.65
102.21
89.87
71.18
70.25
78.43
71.48
92.8
93.51
81.41
106.35
91.81
94.76
103.78
136.26
91.39
108.27
98.75
114.38
138.89
90.26
113.36
92.4
98.83
131.29
77.32
99.95
91.37
75.85
95.72
87.89
84.47

160

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

23031
24001
24003
24005
24009
24013
24015
24017
24021
24025
24027
24031
24033
24035
24039
24043
24045
24510
25001
25003
25005
25009
25011
25013
25015
25017
25021
25023
25025
25027
26015
26017
26021
26025
26027
26037
26045
26049

York County, ME
Allegany County, MD
Anne Arundel County, MD
Baltimore County, MD
Calvert County, MD
Carroll County, MD
Cecil County, MD
Charles County, MD
Frederick County, MD
Harford County, MD
Howard County, MD
Montgomery County, MD
Prince George's County, MD
Queen Anne's County, MD
Somerset County, MD
Washington County, MD
Wicomico County, MD
Baltimore city, MD
Barnstable County, MA
Berkshire County, MA
Bristol County, MA
Essex County, MA
Franklin County, MA
Hampden County, MA
Hampshire County, MA
Middlesex County, MA
Norfolk County, MA
Plymouth County, MA
Suffolk County, MA
Worcester County, MA
Barry County, MI
Bay County, MI
Berrien County, MI
Calhoun County, MI
Cass County, MI
Clinton County, MI
Eaton County, MI
Genesee County, MI

density
mix
centering street
composite
factor
factor
factor
factor
index
92.68
89.8
93.72
78.52
85.7
94.56
117.81
106.32
116.79
111.21
105.04
115.29
100.72
118.53
112.5
109.47
130.43
100.71
118.19
118.58
95.09
73.94
82.27
107.81
87.08
95.33
95.07
100.64
94.25
95.35
93.63
88.61
89.42
100.5
91.2
97.94
88.84
83.65
107.96
93.17
97.32
108.73
104.01
100.82
103.44
100.16
109.82
96.6
99.78
102.01
104.93
128.35
97.95
107.27
112.17
117.8
129.94
123.29
116.7
127.72
112.7
124.13
90.27
125.16
116.51
91.01
67.98
77.17
76.61
72.44
91.18
73.8
82.53
110.34
86.69
97.32
110.91
127.59
95.52
109.9
96
106.22
124.92
114.15
113.05
163.61
143.97
183.84
196.44
190.94
119.45
95.18
33.82
120.97
36.98
122.2
83.51
32.99
112.97
85.5
38.77
122.51
34.74
117.59
104.2
53.29
201.99
30.9
98.17
90.18
57.23
87.88
75.47
71.8
96.11
112.33
108.4
104.1
106.61
94.04
108.26
90.63
99.01
97.45
95.5
103.98
103.91
94.09
99.21
89.45
65.94
94.7
73.69
75.91
91.92
77.85
131.4
63.62
88.88
94.44
101.46
85.64
72.87
85.6
97.37
109.34
123.51
103.52
110.66

161

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

26065
26067
26075
26077
26081
26087
26093
26099
26115
26121
26123
26125
26139
26145
26147
26159
26161
26163
27003
27009
27013
27017
27019
27025
27027
27037
27039
27053
27055
27059
27103
27109
27119
27123
27137
27139
27141
27145

Ingham County, MI
Ionia County, MI
Jackson County, MI
Kalamazoo County, MI
Kent County, MI
Lapeer County, MI
Livingston County, MI
Macomb County, MI
Monroe County, MI
Muskegon County, MI
Newaygo County, MI
Oakland County, MI
Ottawa County, MI
Saginaw County, MI
St. Clair County, MI
Van Buren County, MI
Washtenaw County, MI
Wayne County, MI
Anoka County, MN
Benton County, MN
Blue Earth County, MN
Carlton County, MN
Carver County, MN
Chisago County, MN
Clay County, MN
Dakota County, MN
Dodge County, MN
Hennepin County, MN
Houston County, MN
Isanti County, MN
Nicollet County, MN
Olmsted County, MN
Polk County, MN
Ramsey County, MN
St. Louis County, MN
Scott County, MN
Sherburne County, MN
Stearns County, MN

density
mix
centering street
composite
factor
factor
factor
factor
index
109.11
118.48
141.89
104.33
123.32
92.27
71.44
96.34
76.97
80.1
94.83
98.29
137.01
86.66
105.3
97.5
106.35
113.21
90.33
102.33
99.67
119.56
128.07
96.76
113.92
92.22
70.09
131.99
63.03
86.52
92.3
81.87
104.2
80.88
87.13
107.83
131.48
92.09
106.26
111.9
92.58
95.56
109.24
75.47
91.42
96.94
110.29
96.74
107.62
103.66
89.64
63.71
82.85
79.68
73.43
103.79
122.43
99.39
107.48
110.46
96.62
104.73
106.96
84.83
97.83
96.26
111.36
121.05
101.28
109.46
95.48
93.49
115.33
87.56
97.42
90.64
78.99
85.3
71.88
76.88
105.17
117.06
155.39
87.03
120.43
112.5
126.5
136.09
148.34
139
101.07
111.72
98.03
105.23
105.07
99.34
111.8
83.26
89.21
94.82
97.06
81.38
83.73
89.72
89.44
86.19
89.97
85.88
94.8
100.1
82.7
100.41
93.05
91.23
72.57
80.16
79.33
75.77
101.35
118.95
84.41
81.24
95.56
104.83
115.9
86.85
107.32
104.71
90.15
114.35
78.13
95.81
93.19
114.74
127.82
151.96
129.69
139.24
89.84
94.39
70.75
100.51
85.94
91.07
89.01
80.16
86.9
83.3
97.81
77.6
107.27
98.99
108.08
166.15
100.7
123.35
89.65
106.65
85.6
58.59
81.2
117.31
135.35
105.13
148.75
133.66
95.96
113.02
140.27
103.63
116.7
96.04
104.74
81.51
85.26
89.75
92.57
80.55
85.4
79.35
80.37
95.49
112.29
109.13
96.54
104.25

162

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

27157
27163
27171
28029
28033
28035
28039
28045
28047
28049
28059
28073
28089
28093
28121
28127
28131
28137
28143
29003
29013
29019
29021
29027
29031
29037
29043
29047
29049
29051
29055
29071
29077
29095
29097
29099
29107
29113

Wabasha County, MN
Washington County, MN
Wright County, MN
Copiah County, MS
DeSoto County, MS
Forrest County, MS
George County, MS
Hancock County, MS
Harrison County, MS
Hinds County, MS
Jackson County, MS
Lamar County, MS
Madison County, MS
Marshall County, MS
Rankin County, MS
Simpson County, MS
Stone County, MS
Tate County, MS
Tunica County, MS
Andrew County, MO
Bates County, MO
Boone County, MO
Buchanan County, MO
Callaway County, MO
Cape Girardeau County, MO
Cass County, MO
Christian County, MO
Clay County, MO
Clinton County, MO
Cole County, MO
Crawford County, MO (pt.)*
Franklin County, MO
Greene County, MO
Jackson County, MO
Jasper County, MO
Jefferson County, MO
Lafayette County, MO
Lincoln County, MO

density
mix
centering street
composite
factor
factor
factor
factor
index
89.66
101.77
80.16
119.28
97.11
100.91
108.44
82.51
109.35
100.38
92.03
88.12
85.17
74.14
80.87
90.59
89.53
72.41
81.93
79.29
95.25
88.58
99.48
78.18
87.83
95.34
105.53
96.31
100.75
99.35
90.76
69.74
77.91
92.68
78.23
92.04
77.68
80.99
112.7
88.44
97.88
105.23
107.35
113.32
107.51
100.02
107.02
141.59
102.57
116.18
95.32
88.99
120.77
104.57
103.05
90.94
85.24
82.62
69.99
77.5
96.21
91.29
91.18
87.79
89.4
89.58
45.7
77.07
80.95
66.29
94.27
82.77
81.61
77.7
79.89
89.83
72.44
81.01
94.49
80.34
90.38
88.05
70.63
94.96
82.31
92.63
63.13
71.62
95.88
75.76
88.41
60.42
81.24
70.41
68.56
88.73
86.17
72.6
76.11
75.86
89.22
111.73
80.53
106.69
96.26
98.98
107.9
126.76
103.07
111.6
101.7
120.56
95.28
141.17
118.55
90.4
82.96
97.28
84.65
85.87
95.78
114.42
102.52
94.15
94.94
79.62
83.45
84.89
91.93
89.25
81.1
90.63
85.12
97.62
113.96
88.28
98.64
99.52
90.37
103.72
78.89
114.83
96.15
94.77
101.06
122.96
85.07
101.22
89.11
71.96
88.13
91.1
94.49
82.43
93.59
87.87
100.74
119.9
88.95
115.29
107.86
105.14
126.53
136.74
127.96
130.44
94.9
113.72
88.44
114.86
103.76
96.02
87.54
85.42
99.04
89.9
89.16
87.92
74.98
94.53
83.13
90.59
52.94
85.39
93.02
75.34

163

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

29135
29145
29165
29177
29183
29189
29219
29221
29225
29510
30009
30013
30063
30111
31025
31043
31055
31109
31153
31155
31159
31177
32003
32031
32510
33011
33015
33017
34001
34003
34005
34007
34009
34011
34013
34015
34017
34019

Moniteau County, MO
Newton County, MO
Platte County, MO
Ray County, MO
St. Charles County, MO
St. Louis County, MO
Warren County, MO
Washington County, MO
Webster County, MO
St. Louis city, MO
Carbon County, MT
Cascade County, MT
Missoula County, MT
Yellowstone County, MT
Cass County, NE
Dakota County, NE
Douglas County, NE
Lancaster County, NE
Sarpy County, NE
Saunders County, NE
Seward County, NE
Washington County, NE
Clark County, NV
Washoe County, NV
Carson City, NV
Hillsborough County, NH
Rockingham County, NH
Strafford County, NH
Atlantic County, NJ
Bergen County, NJ
Burlington County, NJ
Camden County, NJ
Cape May County, NJ
Cumberland County, NJ
Essex County, NJ
Gloucester County, NJ
Hudson County, NJ
Hunterdon County, NJ

density
mix
centering street
composite
factor
factor
factor
factor
index
90.4
117.93
68.41
89.59
89.37
92.11
83.25
102.74
93.49
91.02
98.15
104.96
79.77
94.12
92.73
89.65
108.59
73.35
65.04
79.98
104.37
118.4
86.54
121.39
109.7
107.75
126.19
95.35
120.59
115.76
90.25
65.09
88.5
88.94
78.76
89.88
65.15
71.89
94.61
75.21
89.7
58.65
78.35
95.58
75.45
126.98
137.55
194.29
185.95
177.33
88.78
68.92
85.23
93.01
79.76
97.85
123.74
127.17
118.61
121.28
98.92
119.3
111.04
110.74
112.64
103.87
120.17
119.97
115.07
118.66
89.1
86.96
86.25
95.22
86.59
98.92
114.43
75.16
122.4
103.44
110.08
132.45
125.37
138.38
133.58
109.75
133.02
115.33
121.45
125.13
101.37
112.49
87.29
118.08
106.08
88.71
95.5
88.74
85.06
86.74
89.14
99.79
77.47
81.06
83.4
89.99
86.51
117.82
94.88
96.59
119.01
116.44
140.45
122.06
130.94
103.05
110.72
131.45
103.68
115.45
104.88
133.53
80.1
118.62
111.73
101.22
116.91
121.07
97.04
111.45
94
101.41
97.51
82.02
92.08
95.77
105.8
88.23
82.45
91.23
103
114.8
142.81
120.73
125.7
128.56
150.29
86.86
143.25
134.43
100.52
120.12
99.61
99.94
106.38
115.67
137.68
105.55
141.06
131.58
97.81
117.44
101.22
145.73
119.65
99.51
113.21
119.51
98.78
109.8
161.02
146.99
128.46
148.71
158.5
100.59
121.22
87.46
104.71
104.41
223.23
156.67
92.82
176.49
178.73
93.84
90.14
95.2
74
85.21

164

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

34021
34023
34025
34027
34029
34031
34033
34035
34037
34039
34041
35001
35013
35043
35045
35049
35061
36001
36005
36007
36015
36027
36029
36043
36047
36051
36053
36055
36059
36061
36063
36065
36067
36069
36071
36073
36075
36079

Mercer County, NJ
Middlesex County, NJ
Monmouth County, NJ
Morris County, NJ
Ocean County, NJ
Passaic County, NJ
Salem County, NJ
Somerset County, NJ
Sussex County, NJ
Union County, NJ
Warren County, NJ
Bernalillo County, NM
Dona Ana County, NM
Sandoval County, NM
San Juan County, NM
Santa Fe County, NM
Valencia County, NM
Albany County, NY
Bronx County, NY
Broome County, NY
Chemung County, NY
Dutchess County, NY
Erie County, NY
Herkimer County, NY
Kings County, NY
Livingston County, NY
Madison County, NY
Monroe County, NY
Nassau County, NY
New York County, NY
Niagara County, NY
Oneida County, NY
Onondaga County, NY
Ontario County, NY
Orange County, NY
Orleans County, NY
Oswego County, NY
Putnam County, NY

density
mix
centering street
composite
factor
factor
factor
factor
index
114.81
128.87
109.53
119.34
122.92
118.29
135.37
114.47
132.03
131.64
105.74
133.26
84.28
121.16
114.04
103
125.29
87.76
100.05
105.09
105.44
110.28
91.35
129.32
111.5
143.82
148.45
101.63
135.66
140.93
94.41
98
80.11
92.91
89.08
101.83
120.78
86.24
103.35
103.86
95.74
89.17
86.54
87.85
87.14
140.17
153.96
89.87
148.9
141.99
95.86
119.17
85.21
97.52
99.29
110.26
122.46
113.45
131.01
124.38
99.2
106.04
114.72
103.66
107.46
97.97
91.24
110.1
85.16
95.09
93.52
88.26
135.96
78.81
98.91
99.91
106.29
116.83
88.05
103.5
94.94
85.92
108.47
76.38
89.17
107.1
128.39
135.96
104.63
124.04
336.7
143.95
100.25
211.61
224.01
99.92
115.8
121.53
93.89
109.84
98.96
117.49
130.79
99.06
114.63
97.07
110.29
128.55
81.19
105.4
109.71
131.45
111.78
93.59
114.7
96.91
100.82
82.72
80.37
87.62
355.5
142.16
199.99
225.25
265.2
93.13
102.59
78.75
53.09
77.11
94.67
96.7
85.84
57.89
79.49
106.45
123.67
121.06
93.28
114.04
128.98
149.38
111.6
160.85
147.65
654.01
144.57
400.25
230.33
425.15
100.04
115.62
92.59
94.32
100.81
101.65
107.32
112.12
84.48
101.76
104.46
122.19
142.75
96.45
120.8
94.36
101.34
91.19
62.58
84.03
101.31
113.59
90.33
87.33
97.65
94.19
97.46
78.22
53.47
75.78
96.64
90.83
108.43
70.57
89.4
94.19
95.77
83.82
88.92
88.21

165

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

36081
36083
36085
36087
36091
36093
36095
36103
36107
36109
36111
36113
36115
36117
36119
37001
37003
37007
37019
37021
37023
37025
37027
37035
37037
37051
37053
37059
37063
37065
37067
37069
37071
37079
37081
37087
37089
37093

Queens County, NY
Rensselaer County, NY
Richmond County, NY
Rockland County, NY
Saratoga County, NY
Schenectady County, NY
Schoharie County, NY
Suffolk County, NY
Tioga County, NY
Tompkins County, NY
Ulster County, NY
Warren County, NY
Washington County, NY
Wayne County, NY
Westchester County, NY
Alamance County, NC
Alexander County, NC
Anson County, NC
Brunswick County, NC
Buncombe County, NC
Burke County, NC
Cabarrus County, NC
Caldwell County, NC
Catawba County, NC
Chatham County, NC
Cumberland County, NC
Currituck County, NC
Davie County, NC
Durham County, NC
Edgecombe County, NC
Forsyth County, NC
Franklin County, NC
Gaston County, NC
Greene County, NC
Guilford County, NC
Haywood County, NC
Henderson County, NC
Hoke County, NC

density
mix
centering street
composite
factor
factor
factor
factor
index
266.34
147.42
91.93
224.01
204.16
99.2
109.08
97.62
92.25
99.41
175.08
131.67
78.94
179.98
152.34
117.77
134.18
81.37
105.52
112.27
95.36
98.37
102.26
80.9
92.7
107.32
130.66
104.18
110.94
116.78
90.59
78.79
84.01
56.05
71.39
105.86
126.74
94.53
115.53
113.48
94.68
75.76
82.48
64.79
74
102.44
95.84
144.53
72.43
104.82
95.12
96.8
124.18
81.42
99.22
94.99
105.93
183.56
89.94
123.51
92.47
80.23
80.51
59.21
72.33
92.68
85.72
85.91
55.37
74.62
129.24
146.99
93.74
123.66
129.58
95.78
102.85
94.52
96.28
96.66
91.03
78.52
79.96
55.54
70
89.44
65.32
80.36
52.48
64.49
90.81
69.18
88.65
85.96
79.34
95.14
101.18
126.22
94.85
105.5
90.8
78.73
87.53
75.57
78.72
96.2
97.46
88.76
88
90.65
92.41
74.22
123.75
80.6
90.83
93.56
91.54
85.36
88.36
86.99
91.14
56.42
79.76
62.63
65.23
100.01
104.64
91.45
90.81
95.86
90.42
69.81
77.63
76.98
73.1
91.08
61.13
81.22
60.37
66.45
102.68
108.43
103.83
103.7
105.89
91.45
83.77
99.4
93.79
90.02
98.47
107.56
110.15
95.01
103.53
91.13
52.43
78.63
63.74
63.96
95.33
103.37
110.64
94.2
101.12
90.47
47.46
83.61
40.96
56.56
100.36
113.56
102.77
95.45
103.84
91.09
79.15
80.84
102.68
85.39
92.12
98.21
84.83
93.59
90.13
91.51
57.98
83.07
70.19
69.27

166

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

37101
37115
37119
37127
37129
37133
37135
37141
37145
37147
37151
37157
37169
37179
37183
37191
37197
38015
38017
38035
38059
39003
39013
39015
39017
39019
39023
39025
39035
39041
39043
39045
39049
39051
39055
39057
39061
39081

Johnston County, NC
Madison County, NC
Mecklenburg County, NC
Nash County, NC
New Hanover County, NC
Onslow County, NC
Orange County, NC
Pender County, NC
Person County, NC
Pitt County, NC
Randolph County, NC
Rockingham County, NC
Stokes County, NC
Union County, NC
Wake County, NC
Wayne County, NC
Yadkin County, NC
Burleigh County, ND
Cass County, ND
Grand Forks County, ND
Morton County, ND
Allen County, OH
Belmont County, OH
Brown County, OH
Butler County, OH
Carroll County, OH
Clark County, OH
Clermont County, OH
Cuyahoga County, OH
Delaware County, OH
Erie County, OH
Fairfield County, OH
Franklin County, OH
Fulton County, OH
Geauga County, OH
Greene County, OH
Hamilton County, OH
Jefferson County, OH

density
mix
centering street
composite
factor
factor
factor
factor
index
93.03
70.6
103.97
64.44
78.53
89.4
44.18
77.93
90.45
69.03
105.91
115.35
135.51
101.84
118.52
91.58
88.78
88.52
79.45
83.68
102.34
118.86
107.7
121.5
115.92
94.97
82.72
104.59
82.75
88.95
99.4
106.99
120.04
75.56
100.63
91.15
64.41
81.67
60.61
67.72
91.24
74.11
81.98
61.12
71.08
98.36
104.23
117.55
87.14
102.3
92.22
84.74
100.63
57.18
79.39
90.85
72.36
83.7
76.47
75.79
90.59
52.98
81.84
64.72
65.29
94.98
81.73
100.88
84.45
88.01
103.07
115.17
134.61
96.6
115.62
93.55
78.79
130.76
84.88
96.2
90.06
70.68
79.45
49.29
65.08
96.52
118.46
128.76
90.68
110.87
99.52
125.9
113.31
97.15
111.34
104.24
124.99
97.01
96.71
107.25
91.13
108.21
82.17
85.86
89.69
95.85
114.27
117.83
118.07
114.54
92.89
98.58
83.73
112.11
95.99
90.42
54.19
85.62
78.68
71.22
101.42
116.84
94.22
101.13
104.3
89.77
69.05
94.41
68.25
75.19
96.98
111.55
97.15
102.52
102.6
98.23
97.66
83.05
84.14
88.34
112.92
133.64
119.54
109.64
123.93
97.21
109.37
84.07
87.68
93.15
96.77
121.77
104.84
102.29
108.11
95.2
100.29
89.76
89.15
91.91
111.37
131.41
124.87
127.88
130.18
90.59
113.35
82.43
93.65
93.69
90.84
82.83
86.85
50.2
71.79
97.09
114.93
85.08
94.01
97.19
110.12
134.12
141.56
113.68
131.43
95.1
103.84
109.52
107.8
105.14

167

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

39085
39087
39089
39093
39095
39097
39099
39103
39109
39113
39117
39123
39129
39133
39135
39139
39151
39153
39155
39159
39165
39167
39173
40017
40027
40031
40037
40051
40079
40083
40087
40109
40111
40113
40117
40131
40135
40143

Lake County, OH
Lawrence County, OH
Licking County, OH
Lorain County, OH
Lucas County, OH
Madison County, OH
Mahoning County, OH
Medina County, OH
Miami County, OH
Montgomery County, OH
Morrow County, OH
Ottawa County, OH
Pickaway County, OH
Portage County, OH
Preble County, OH
Richland County, OH
Stark County, OH
Summit County, OH
Trumbull County, OH
Union County, OH
Warren County, OH
Washington County, OH
Wood County, OH
Canadian County, OK
Cleveland County, OK
Comanche County, OK
Creek County, OK
Grady County, OK
Le Flore County, OK
Logan County, OK
McClain County, OK
Oklahoma County, OK
Okmulgee County, OK
Osage County, OK
Pawnee County, OK
Rogers County, OK
Sequoyah County, OK
Tulsa County, OK

density
mix
centering street
composite
factor
factor
factor
factor
index
100.55
123.58
82.99
88.29
98.55
93.75
81.53
83.82
104.35
88.45
95.01
99.59
98.19
106.48
99.77
98.61
117.13
93.18
95.05
101.26
105.01
131.81
114.29
116.4
121.33
92.38
85.12
84.52
84.97
83.25
98.98
121.53
107.96
102.09
109.66
96.03
105.54
93.2
57.23
84.83
92.97
103.49
85.25
95.62
92.84
102.99
130.21
114.82
117.4
120.67
89.85
49.6
83.41
46.82
58.82
93.01
98.23
86.34
94.39
91.15
95.16
82.72
83.74
78.2
80.99
94.89
103.8
90.32
100.22
96.6
90.05
70.46
86.69
100.99
83.63
94.98
105.89
118.65
103.59
107.3
98.73
120.66
98.8
120.61
112.26
101.67
125.68
109.41
114.42
116.17
95.85
111.81
91.49
95.52
98.31
94.04
77.41
81.94
86.51
81.01
97.43
106.62
84.37
88.63
92.75
93.06
88.2
86.67
83.86
84.77
94.89
111.78
91.96
82.11
93.91
97.03
97.68
82.74
92.01
90.35
101.04
107.98
106.44
102.24
105.59
99.03
118.45
98.2
116.33
110.11
90.09
85.48
84.46
104.69
88.85
91.37
75.37
86.82
102.85
86.23
89.15
67.37
83.45
99.19
80.78
89.7
68.27
90.56
98.34
83.21
89.63
80.94
81.73
88.92
81.43
103.44
120.48
122.5
117.89
120.32
89.76
90.51
83.84
122.81
95.86
93.63
66.07
86.07
96.84
81.87
88.73
75.14
77.53
99.62
81.37
92.33
79.74
87.59
95.45
85.82
89.78
72.88
91.9
101.22
86.03
102.6
121.46
117.13
113.15
117.17

168

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

40145
41003
41005
41009
41017
41029
41039
41047
41051
41053
41067
41071
42003
42005
42007
42011
42013
42017
42019
42021
42025
42027
42029
42041
42043
42045
42049
42051
42069
42071
42075
42077
42079
42081
42085
42091
42095
42099

Wagoner County, OK
Benton County, OR
Clackamas County, OR
Columbia County, OR
Deschutes County, OR
Jackson County, OR
Lane County, OR
Marion County, OR
Multnomah County, OR
Polk County, OR
Washington County, OR
Yamhill County, OR
Allegheny County, PA
Armstrong County, PA
Beaver County, PA
Berks County, PA
Blair County, PA
Bucks County, PA
Butler County, PA
Cambria County, PA
Carbon County, PA
Centre County, PA
Chester County, PA
Cumberland County, PA
Dauphin County, PA
Delaware County, PA
Erie County, PA
Fayette County, PA
Lackawanna County, PA
Lancaster County, PA
Lebanon County, PA
Lehigh County, PA
Luzerne County, PA
Lycoming County, PA
Mercer County, PA
Montgomery County, PA
Northampton County, PA
Perry County, PA

density
mix
centering street
composite
factor
factor
factor
factor
index
93.2
77.7
83.08
102.13
86.14
100.72
123.18
126.52
95.34
114.46
101.8
126.17
90.03
96.25
104.5
93.28
102.74
80.42
84.73
87.73
95.73
115.65
115.3
80.19
102.17
97.76
122.2
122.65
91.71
110.84
101.73
127.48
138.05
98.88
120.9
101.62
130.36
123.77
101.1
117.96
120.53
142.82
150.58
166.68
157.06
94.97
105.79
80.13
83.85
88.86
110.39
132.91
85.02
113.1
113.09
99.08
122.85
81.32
93.49
98.97
109.54
133.89
145.4
135.7
139.34
92.89
85.75
101.54
84.86
88.95
95.17
110.16
84.42
111.13
100.28
108.58
126.11
116
110.71
119.4
97.22
121.95
124.31
123.01
121.01
102.39
126.03
79.87
99.58
102.49
93.68
105.26
120.02
79.27
99.44
95.43
107.43
120.16
119.48
113.43
93.36
98.43
90.96
97.65
93.81
110.1
115.7
149.49
91.83
121.21
98.81
117.12
91.2
89.11
98.81
98.59
111.24
85.52
112.72
102.55
104.58
124.71
129.24
125.68
126.61
119.69
141.69
83.25
137.9
126.07
102.74
130.88
122.48
102.4
118.48
93.03
102.25
96.86
108.42
100.17
101.86
133.13
134.53
123.5
129.39
102.63
119.9
128.6
94.47
114.41
96.31
122.77
84.72
116.98
106.56
111.48
134.36
115.73
137.75
131.38
99.44
121.47
93.27
114.55
109.08
97.09
120.85
113.98
117.91
115.74
95.34
106.25
83.44
87.04
91.17
107.67
136.32
85.84
109.26
112.35
103.88
133.01
101.8
124.28
119.89
89.79
63.67
91.33
79.02
75.93

169

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

42101
42103
42125
42129
42131
42133
44001
44003
44005
44007
44009
45003
45007
45015
45019
45031
45035
45037
45039
45041
45045
45051
45055
45059
45063
45077
45079
45083
45085
45091
46083
46093
46099
46103
47001
47009
47011
47019

Philadelphia County, PA
Pike County, PA
Washington County, PA
Westmoreland County, PA
Wyoming County, PA
York County, PA
Bristol County, RI
Kent County, RI
Newport County, RI
Providence County, RI
Washington County, RI
Aiken County, SC
Anderson County, SC
Berkeley County, SC
Charleston County, SC
Darlington County, SC
Dorchester County, SC
Edgefield County, SC
Fairfield County, SC
Florence County, SC
Greenville County, SC
Horry County, SC
Kershaw County, SC
Laurens County, SC
Lexington County, SC
Pickens County, SC
Richland County, SC
Spartanburg County, SC
Sumter County, SC
York County, SC
Lincoln County, SD
Meade County, SD
Minnehaha County, SD
Pennington County, SD
Anderson County, TN
Blount County, TN
Bradley County, TN
Carter County, TN

density
mix
centering street
composite
factor
factor
factor
factor
index
206.38
144.48
178.43
209.98
207.19
91.08
56.19
144.75
90.61
94.51
95.07
106.69
93.55
102.25
99.23
95.84
111.77
104.88
108.5
106.63
90.4
51.38
86.24
74.76
69.28
99.69
112.24
115.21
96.33
107.42
109.79
144.16
83.56
135.16
122.96
103.82
122.09
81.7
122.57
109.54
99.45
121.07
99.03
118.74
112.1
121.1
142.01
141.75
134.74
144.11
94.03
102.13
88.56
97.1
94.26
93.29
79.37
103.25
96.65
91.33
92.29
82.54
110.42
81.7
89.56
98.3
88.34
80.72
78.85
83
103.2
119.32
138.48
116.56
124.5
91.78
86.08
84.55
73.08
79.62
103.61
98.38
81.02
84.79
89.83
89.95
55.96
76.27
60.96
63.08
89.55
49.53
76.12
74.02
65
96.07
90.47
109.63
83.71
93.64
98.68
106.59
100.39
91.07
98.97
94.78
90.85
112.78
101.88
100.09
90.43
61.7
129.24
61.49
81.95
89.91
59.53
87.21
79.89
73.63
94.92
94.04
88
80.44
86.54
92.45
92.02
97.27
82.26
88.63
101.53
109.51
144.33
110.91
120.94
93.37
97.98
112.28
90.54
98.16
93.59
86.69
119.72
90.32
96.94
95.01
95.83
94.28
80.22
89.05
92.75
107.03
82.73
77.53
87.38
89.23
75.07
81.4
103.16
83.84
102.86
120.06
105.9
107.25
111.4
96.18
101.49
117.26
95.04
103.15
92.32
81.1
121.37
89.51
95.04
94.52
79.63
87.08
89.16
84.33
94.75
85.38
114.48
87.22
94.26
93.3
77.41
129.08
96.48
98.82

170

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

47021
47023
47037
47043
47047
47057
47063
47065
47073
47089
47093
47105
47111
47113
47115
47125
47147
47149
47153
47157
47159
47163
47165
47167
47169
47171
47173
47179
47187
47189
48007
48013
48015
48019
48021
48027
48029
48037

Cheatham County, TN
Chester County, TN
Davidson County, TN
Dickson County, TN
Fayette County, TN
Grainger County, TN
Hamblen County, TN
Hamilton County, TN
Hawkins County, TN
Jefferson County, TN
Knox County, TN
Loudon County, TN
Macon County, TN
Madison County, TN
Marion County, TN
Montgomery County, TN
Robertson County, TN
Rutherford County, TN
Sequatchie County, TN
Shelby County, TN
Smith County, TN
Sullivan County, TN
Sumner County, TN
Tipton County, TN
Trousdale County, TN
Unicoi County, TN
Union County, TN
Washington County, TN
Williamson County, TN
Wilson County, TN
Aransas County, TX
Atascosa County, TX
Austin County, TX
Bandera County, TX
Bastrop County, TX
Bell County, TX
Bexar County, TX
Bowie County, TX

density
mix
centering street
composite
factor
factor
factor
factor
index
93.65
56.61
86.41
61.81
67.92
91.73
79.08
69.11
55.42
66.93
104.68
111.86
121.78
111.57
115.76
91.19
65.43
90.57
73.7
75.01
89.34
50.43
89.51
51.46
62.32
89.49
45.66
74.08
70.51
62.01
95.73
85
142.29
95.5
105.85
98.48
101.33
119.36
103.4
107.13
90.78
69.01
90.1
81.51
78.33
91.49
63.63
91.38
79.72
76.69
99.46
102.38
136.24
96.83
111.03
90.6
74.46
83.62
96.59
82.71
90.08
45.11
73.25
47.03
54.34
95.08
104.99
108.51
91.26
99.95
89.77
69.94
73.16
87.72
74.91
97.02
80.87
113.11
75.99
89.57
91.68
72.06
85.62
63.1
72.35
97.98
90.6
108.29
83.25
93.72
90.25
76.45
78.98
57.33
69.36
105.33
109.94
122.61
114.9
116.68
90.53
70.87
66.08
83.13
71.76
93.76
86.37
119.66
101.34
100.36
97.36
86.46
115.6
76.15
92.28
92.75
59.76
87.84
64.39
69.9
90.52
71.81
67.37
64.82
66.68
94.94
90.3
80.78
113.03
93.38
89.52
50.58
82.78
73.69
67.32
94.93
91.12
94.03
93.77
91.74
97
85.43
133.03
87.19
100.84
93.71
71.92
85.24
70.33
75.1
91.9
104.27
84.03
122.27
100.78
89.05
79.5
85.77
94.63
83.87
88.89
64.78
86.07
82.34
75.38
89.19
38.15
69.25
101.83
67.91
89.76
76.25
87.26
96.1
84.01
99.9
110.3
106.9
110.75
108.8
107.69
116.02
115.57
118.94
118.4
93.73
106.36
80.75
99.24
93.71

171

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

48039
48041
48051
48055
48057
48061
48071
48077
48085
48091
48099
48113
48119
48121
48135
48139
48141
48157
48167
48181
48183
48187
48199
48201
48209
48215
48231
48245
48251
48257
48259
48281
48291
48303
48309
48325
48329
48339

Brazoria County, TX
Brazos County, TX
Burleson County, TX
Caldwell County, TX
Calhoun County, TX
Cameron County, TX
Chambers County, TX
Clay County, TX
Collin County, TX
Comal County, TX
Coryell County, TX
Dallas County, TX
Delta County, TX
Denton County, TX
Ector County, TX
Ellis County, TX
El Paso County, TX
Fort Bend County, TX
Galveston County, TX
Grayson County, TX
Gregg County, TX
Guadalupe County, TX
Hardin County, TX
Harris County, TX
Hays County, TX
Hidalgo County, TX
Hunt County, TX
Jefferson County, TX
Johnson County, TX
Kaufman County, TX
Kendall County, TX
Lampasas County, TX
Liberty County, TX
Lubbock County, TX
McLennan County, TX
Medina County, TX
Midland County, TX
Montgomery County, TX

density
mix
centering street
composite
factor
factor
factor
factor
index
96.54
96.26
92.15
97.38
94.42
105.72
112.86
101.13
110.13
109.43
89.32
100.91
77.93
109.68
93
89.63
89.32
84.6
100.93
88.78
97.89
104.62
74.17
145.39
106.98
100.34
102.76
87.93
110.32
100.42
88.91
43.66
75.63
77.45
63.87
88.03
67.28
76.56
111.02
81.95
106.24
114.06
85.45
118.59
107.69
93.66
86.53
108.62
88.26
92.76
97.23
77.14
87.93
86.13
83.7
116.03
123.21
125.52
139.21
132.85
88.85
80.3
68.73
127.14
88.95
104.96
107.37
91.25
114.16
105.61
101.41
123.37
112.23
111.89
115.45
92.65
86.97
84.65
100.12
88.75
109.16
113.33
102.45
125.22
115.85
104.19
96.2
101.96
111.59
104.41
100.94
113.67
106.27
130.51
116.24
93.05
103.96
92.14
102.59
97.39
96.1
114.14
103.02
99.15
103.92
96.53
93.38
84.13
94.73
90.14
89.38
75.62
84.66
83.17
78.78
112.9
122.96
115.12
138.63
128.31
95.58
87.83
131.77
84.13
99.78
100.21
101.69
104.76
109.1
104.98
91.85
76.8
100.17
94.77
88.5
99.99
118.66
127.39
137.42
126.37
94.62
85
88.74
91.72
87.39
91.56
77.63
83.06
108.05
87.46
94.46
97.53
79.63
72.72
82.42
89.18
74.92
86.25
95.76
82.98
89.41
54.79
90.7
83.18
74.12
101.82
123.12
97.75
110.77
110.57
96.64
112.13
100.28
109.99
106.02
88.53
55.51
85.3
81.66
71.88
103.45
123.85
110.9
119.62
118.27
95.68
87.52
111.61
84.05
93.32

172

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

48355
48361
48367
48375
48381
48397
48401
48409
48423
48439
48451
48453
48459
48469
48473
48479
48485
48491
48493
48497
49005
49011
49023
49035
49043
49045
49049
49053
49057
50007
50011
50013
51003
51009
51011
51013
51019
51023

Nueces County, TX
Orange County, TX
Parker County, TX
Potter County, TX
Randall County, TX
Rockwall County, TX
Rusk County, TX
San Patricio County, TX
Smith County, TX
Tarrant County, TX
Tom Green County, TX
Travis County, TX
Upshur County, TX
Victoria County, TX
Waller County, TX
Webb County, TX
Wichita County, TX
Williamson County, TX
Wilson County, TX
Wise County, TX
Cache County, UT
Davis County, UT
Juab County, UT
Salt Lake County, UT
Summit County, UT
Tooele County, UT
Utah County, UT
Washington County, UT
Weber County, UT
Chittenden County, VT
Franklin County, VT
Grand Isle County, VT
Albemarle County, VA
Amherst County, VA
Appomattox County, VA
Arlington County, VA
Bedford County, VA
Botetourt County, VA

density
mix
centering street
composite
factor
factor
factor
factor
index
104.85
127.12
106.59
121.3
118.91
90.28
87.97
84.52
104.13
89.54
90.72
77.89
87.88
79
79.62
101.4
118.2
99.33
132.71
116.32
101.51
122.09
78.97
110.72
104.2
97.13
97.42
79.27
94.18
89.89
89.28
80.54
82.05
67.69
74.59
93.48
114.78
84.07
111.29
101.14
95.5
100.31
119.02
100.6
104.88
108.94
119.35
100.17
128.9
118.12
97.73
119.81
103.96
106.9
108.97
108.45
120.81
148.98
110.66
128.09
90.15
67.18
79.57
86.71
75.86
103.1
120.55
119.38
119.7
119.82
95.59
60.29
82.16
92.14
77.94
101.78
122.77
102.69
121.89
115.53
98.04
121.94
121.17
110.29
116.25
101.28
106.24
98.74
101.69
102.51
89.22
46.7
88.44
72.24
67.33
89.07
68.46
80.23
80.04
74.03
100.03
120.88
128.98
82.21
110.14
103.45
125.21
80.47
105.19
104.52
88.62
93.3
78.14
83.59
82.2
112.04
129.1
106.26
116.3
120.12
90.7
90.55
91.28
75.6
83.61
97.75
102.75
79.12
75.88
85.94
108.21
127.19
89.82
106.36
109.98
95.06
98.96
84.85
91.6
90.67
105.74
124.44
97.16
108.01
111.17
101.56
121.65
152.59
89.97
120.78
92.87
95.99
82.45
75.67
83.25
89.13
86.07
69.37
90.87
79.6
95.3
102.67
87.34
78.58
88.59
89.69
70.62
84.6
75.08
74.72
89.68
39.87
90.05
58.37
61.45
174.41
153.2
95.54
177.13
163.28
89.97
55.41
91.02
73.51
71.54
89.85
72
83.63
88.06
79

173

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

51031
51033
51041
51043
51053
51059
51061
51065
51067
51069
51073
51075
51079
51085
51087
51093
51095
51101
51107
51115
51121
51127
51143
51145
51149
51153
51155
51161
51165
51169
51177
51179
51183
51187
51191
51199
51510
51515

Campbell County, VA
Caroline County, VA
Chesterfield County, VA
Clarke County, VA
Dinwiddie County, VA
Fairfax County, VA
Fauquier County, VA
Fluvanna County, VA
Franklin County, VA
Frederick County, VA
Gloucester County, VA
Goochland County, VA
Greene County, VA
Hanover County, VA
Henrico County, VA
Isle of Wight County, VA
James City County, VA
King William County, VA
Loudoun County, VA
Mathews County, VA
Montgomery County, VA
New Kent County, VA
Pittsylvania County, VA
Powhatan County, VA
Prince George County, VA
Prince William County, VA
Pulaski County, VA
Roanoke County, VA
Rockingham County, VA
Scott County, VA
Spotsylvania County, VA
Stafford County, VA
Sussex County, VA
Warren County, VA
Washington County, VA
York County, VA
Alexandria city, VA
Bedford city, VA

density
mix
centering street
composite
factor
factor
factor
factor
index
91.88
77.31
83.38
109.02
87.87
89.04
40.8
74.87
77.09
62.65
100.63
98.15
114.36
102.77
105.03
89.87
79.72
79.01
86.65
79.55
90.02
49.1
78.23
71.08
64.75
117.83
123.7
113.17
114.82
121.96
90.61
73.98
90.24
80.5
79.57
92.01
71.24
75.82
69.22
71.02
91.3
47.21
88.85
77.48
69.94
93.79
81.33
87.14
85.85
83.61
92.66
69.24
89.69
99.14
84.43
90.23
55.11
75.26
78.66
68.17
90.55
59.72
70.1
78.44
68.03
94.37
84.41
82.56
88.35
84.1
105.97
114.27
86.41
123.03
109.38
90.76
75.64
77.65
79.82
75.95
93.7
97.02
79.6
106.28
92.61
90.95
56.69
79.27
102.1
77.57
102.68
116.85
81.49
113.55
104.6
92.2
52.08
72.32
78.22
66.77
95.29
95.57
85.4
102.18
93.19
89.75
43.95
80.36
72.4
64.13
89.61
42.72
80.8
66.85
62.08
94.07
44.51
74.52
65.38
61.61
90.96
66.68
75.53
81.97
73.19
106.28
106.57
94.52
115.14
107.11
91.55
84.58
83.02
103.9
88.33
96.03
110.04
80.69
98.89
95.46
90.09
73.51
86.01
88.97
80.6
89.25
50.38
78.01
92.28
71.54
97.94
84.86
88.47
92.55
88.57
98.78
84.11
81.07
88.85
85.09
102.08
63.8
93.5
92.21
88.78
94.07
90.07
90.49
77.47
81.92
90.19
81.06
97.29
99
86.14
108.5
97.13
176.94
154.32
115.16
173.76
169.56
94.78
123.63
72.04
113.62
101.29

174

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

51520
51540
51550
51570
51590
51600
51610
51630
51650
51660
51670
51680
51683
51685
51700
51710
51730
51735
51740
51750
51760
51770
51775
51800
51810
51830
51840
53003
53005
53007
53011
53015
53017
53021
53033
53035
53053
53057

Bristol city, VA
Charlottesville city, VA
Chesapeake city, VA
Colonial Heights city, VA
Danville city, VA
Fairfax city, VA
Falls Church city, VA
Fredericksburg city, VA
Hampton city, VA
Harrisonburg city, VA
Hopewell city, VA
Lynchburg city, VA
Manassas city, VA
Manassas Park city, VA
Newport News city, VA
Norfolk city, VA
Petersburg city, VA
Poquoson city, VA
Portsmouth city, VA
Radford city, VA
Richmond city, VA
Roanoke city, VA
Salem city, VA
Suffolk city, VA
Virginia Beach city, VA
Williamsburg city, VA
Winchester city, VA
Asotin County, WA
Benton County, WA
Chelan County, WA
Clark County, WA
Cowlitz County, WA
Douglas County, WA
Franklin County, WA
King County, WA
Kitsap County, WA
Pierce County, WA
Skagit County, WA

density
mix
centering street
composite
factor
factor
factor
factor
index
105
130.6
82.35
145.26
119.97
128.8
148.33
210.83
152.37
175.93
103.4
108.24
88.28
109.52
102.98
108.95
135.66
77.65
153.6
123.97
99.84
126.2
121.82
120.33
121.54
116.97
152.84
73
131.05
123.34
127.12
177.53
72.72
164.07
144.69
120.16
145.13
97.72
154.28
137.06
110.55
123.19
114.92
150.96
131.48
122.83
143.99
144.42
131.8
145.19
112.29
124.58
79.39
185.81
132.25
104.8
130.42
104.85
132.31
122.87
115.54
140.36
76.57
150.36
126.17
129.66
128.88
82.19
133.5
123.45
112.21
121.94
86.53
137.18
118.28
129.98
131.46
210.96
179.44
179.57
101.48
127
104.35
144.23
124.34
97.09
105.92
77.55
104.32
95.22
111.16
129.35
88.86
163.76
129.42
105.79
135.4
81.24
156.21
124.84
120.46
133.06
160.69
172.23
158.9
109.84
129.71
120.97
155.62
136.69
107.3
128.88
76.93
140.41
116.91
95.77
99.14
103.14
98.02
98.76
111.75
123.1
86.61
137.93
118.77
108.92
118.37
158.9
136.03
138.61
114.03
135.13
133.91
150.19
142.1
106.62
134.33
77
134.97
116.72
98.56
118.73
109.61
97.28
107.64
97.97
126.31
120.3
99.04
113.78
102.63
123.4
89.55
105.28
106.59
96.07
103.4
128.01
99
108.37
103.94
116.98
82.17
91.3
98.23
101.59
119.22
82.23
111.14
104.48
114.85
128.93
159.34
131.7
142.6
98.92
107.82
115.62
96.04
105.81
103.02
117.02
126.32
119.43
120.78
96.68
112.71
101.76
99.87
103.48

175

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

53061
53063
53067
53073
53077
54003
54005
54009
54011
54029
54037
54039
54051
54057
54061
54065
54069
54077
54079
54099
54107
55009
55015
55017
55021
55025
55031
55035
55039
55049
55059
55061
55063
55073
55079
55083
55087
55089

Snohomish County, WA
Spokane County, WA
Thurston County, WA
Whatcom County, WA
Yakima County, WA
Berkeley County, WV
Boone County, WV
Brooke County, WV
Cabell County, WV
Hancock County, WV
Jefferson County, WV
Kanawha County, WV
Marshall County, WV
Mineral County, WV
Monongalia County, WV
Morgan County, WV
Ohio County, WV
Preston County, WV
Putnam County, WV
Wayne County, WV
Wood County, WV
Brown County, WI
Calumet County, WI
Chippewa County, WI
Columbia County, WI
Dane County, WI
Douglas County, WI
Eau Claire County, WI
Fond du Lac County, WI
Iowa County, WI
Kenosha County, WI
Kewaunee County, WI
La Crosse County, WI
Marathon County, WI
Milwaukee County, WI
Oconto County, WI
Outagamie County, WI
Ozaukee County, WI

density
mix
centering street
composite
factor
factor
factor
factor
index
103.47
116.86
122.73
100.03
113.62
101.37
122.39
122.32
127.12
123.13
97.83
103.71
132.9
95.16
109.35
95.83
110.62
115.26
99
106.54
98.64
124.46
128.18
89.38
112.84
94.85
90.23
97.7
94.03
92.67
90.83
61.03
123.52
91.02
93.32
87.28
116.81
96.34
98.52
112.81
183.48
119.12
135.99
94.13
110.72
86.79
118.07
103.07
91.79
75.67
87.64
98.81
85.44
96.1
108.14
147.64
125.6
124.48
92.36
89.16
137.78
120.37
112.53
90.81
75.55
159.67
111.67
111.91
98.42
117.16
120.1
115.01
116.02
89.5
67.7
90
74.66
75.31
95.76
115.77
150.91
129.79
129.14
88.93
44.98
90.63
80.67
70.06
93.37
87.87
78.21
99.34
86.98
93.73
81.82
84.99
106.16
89.48
96.66
116.84
107.75
121.08
113.37
99.46
115.4
101.3
91.01
102.26
94.95
80.84
87.75
80.59
82.35
92.19
85.15
89.4
88.5
85.86
90.01
92.46
87.63
90.9
87.68
106.96
126.2
153.67
106.96
129.63
95.01
99.68
81.91
108.53
95.3
98.55
115.5
116.85
96.62
108.7
95.54
109.78
153.06
94.09
116.57
89.19
78
83.48
83.09
79.07
100.8
119.03
123.52
118.9
119.67
92.15
103.67
77.23
79.49
85.01
98.49
119.38
88.95
117.4
107.65
94.14
102.58
121.29
83.21
100.38
128.75
139.35
178.96
155.69
164.06
88.82
49.35
77.77
66.91
62.99
99.06
120.79
164.21
97.96
125.91
95.11
116.53
106.77
87.76
101.95

176

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

fips

county

55093
55101
55105
55109
55117
55131
55133
55139
56021

Pierce County, WI
Racine County, WI
Rock County, WI
St. Croix County, WI
Sheboygan County, WI
Washington County, WI
Waukesha County, WI
Winnebago County, WI
Laramie County, WY

density
mix
centering street
composite
factor
factor
factor
factor
index
94.38
92.07
143.31
81.67
103.61
100.48
122.63
111.62
107.68
113.4
97.51
113.9
108.04
98.59
105.7
92.02
87.72
93.45
67.27
81.19
97.6
115.59
94.01
98.77
101.88
94.74
96.05
128.67
75.35
98.36
96.89
112.13
147.79
101.06
118.28
100.65
118.29
97.48
113.49
109.45
100.71
112.98
132.64
114.68
119.28

177

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix C. Detailed Model Results for Public Health Outcomes


Table C.1. Relationships to BMI (with robust standard errors
Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black non-Hispanic
Other non-Hispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Percent Parks
Heating Degree
Days
Original
Compactness Index
New Compactness
Index

Original
Model
Coeff.
25.14
0.96
1.99
2.55
1.94
0.36
1.81
-1.05
0.78
0.99

Original
Model
T-ratio
95.14
15.22
23.42
36.59
19.17
3.01
19.68
-5.83
5.78
11.89

Original
Model
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.003
<0.001
<0.001
<0.001
<0.001

Refined
Model
Coeff.
25.13
0.96
1.99
2.55
1.94
0.36
1.81
-1.05
0.77
0.99

Refined
Model
T-ratio
101.01
15.22
23.40
36.58
19.15
3.01
19.65
-5.84
5.77
11.91

Refined
Model
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.003
<0.001
<0.001
<0.001
<0.001

Model
Three
Coeff.
25.83
0.96
1.99
2.55
1.94
0.35
1.81
-1.05
0.78
0.99

Model
Three
T-ratio
85.86
15.21
23.42
36.61
19.19
3.02
19.71
-5.82
5.79
11.90

Model
Three
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.003
<0.001
<0.001
<0.001
<0.001

0.92

14.29

<0.001

0.92

14.28

<0.001

0.92

14.28

<0.001

1.04
0.98
0.68

18.78
8.38
10.17

<0.001
<0.001
<0.001

1.04
0.98
0.68

18.77
8.39
10.19

<0.001
<0.001
<0.001

1.04
0.98
0.67

18.75
8.38
10.14

<0.001
<0.001
<0.001

0.57

7.61

<0.001

0.57

7.62

<0.001

0.57

7.60

<0.001

-0.82
-0.52

-8.49
-12.28

<0.001
<0.001

-0.82
-0.52

-8.48
-12.28

<0.001
<0.001

-0.82
-0.52

-8.50
-12.27

<0.001
<0.001

-0.028
0.000042

-2.18
2.18

0.030
0.030

-0.029
0.00005

-2.02
2.58

0.044
0.011

-0.021
0.000065

-1.63
3.33

0.105
0.001

-0.00897

-4.25

<0.001
-0.00910

-4.50

<0.001
-0.0066

-1.98

0.049

-0.0117

-3.33

0.001

-0.00085

-0.48

0.633

0.0033

1.43

0.153

Density factor
Mix factor
Centering factor
Street factor
Pseudo R2

0.079

0.079

0.080

178

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Table C.2. Relationships to Obesity (with robust standard errors)


Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black non-Hispanic
Other nonHispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Percent Parks
Crime
Heating Degree
Days
Original
Compactness
Index
Refined
Compactness
Index

Original
Model
Coeff.
-1.68
0.10
0.59
0.73
0.52
-0.10
0.55
-0.44

Original
Model
T-ratio
-19.58
5.71
17.86
24.80
13.16
-1.78
12.86
-6.19

Original
Model
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.075
<0.001
<0.001

Refined
Model
Coeff.
-1.66
0.10
0.59
0.73
0.52
-0.10
0.54
-0.44

Refined
Model
T-ratio
-19.65
5.70
17.76
24.68
13.07
-1.76
12.78
-6.20

Refined
Model
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.077
<0.001
<0.001

Model
Three
Coeff.
-1.44
0.10
0.59
0.73
0.52
-0.10
0.55
-0.44

Model
Three
T-ratio
-11.56
5.71
17.86
24.79
13.15
-1.78
12.89
-6.17

Model
Three
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
0.074
<0.001
<0.001

0.24
0.39

4.31
12.72

<0.001
<0.001

0.24
0.39

4.30
12.66

<0.001
<0.001

0.24
0.39

4.34
12.75

<0.001
<0.001

0.40

15.19

<0.001

0.40

15.10

<0.001

0.40

15.14

<0.001

0.43
0.35
0.23

15.52
8.73
7.43

<0.001
<0.001
<0.001

0.43
0.35
0.22

15.44
8.73
7.43

<0.001
<0.001
<0.001

0.43
0.35
0.23

15.49
8.75
7.43

<0.001
<0.001
<0.001

0.20

8.40

<0.001

0.20

8.41

<0.001

0.20

8.41

<0.001

-0.26
-0.21

-6.74
-14.11

<0.001
<0.001

-0.26
-0.21

-6.70
-14.10

<0.001
<0.001

-0.26
-0.21

-6.74
-14.09

<0.001
<0.001

-0.010
0.00015
0.000020

-2.11
2.40
2.36

0.035
0.021
0.019

-0.0091
0.00019
0.000024

-1.72
3.25
2.91

0.086
0.002
0.004

-0.0083
0.00013
0.000027

-1.60
2.09
3.17

0.111
0.037
0.002

-0.0035

-4.75

<0.001
-0.00401

-6.11

<0.001
-0.0028

-2.29

0.023

-0.0036

-2.53

0.012

-0.00075

-0.93

0.351

0.0013

1.17

0.243

Density factor
Mix factor
Centering factor
Street factor
pseudo R2

0.50

0.57

0.58

179

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Table C.3. Relationships to Any Physical Activity (with robust standard errors)
Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black nonHispanic
Other nonHispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Crime
Precipitation
Original
Compactness
Index
Refined
Compactness
Index
Density factor
Mix factor
Centering factor
Street factor
Pseudo R2

Model 1
Coeff.
2.80
0.27
-0.42
-0.63
-0.62
-0.97
-0.30

Model 1
T-ratio
33.41
10.76
-6.98
-12.34
-13.77
-20.92
-7.87

Model 1
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model
2 Coeff.
2.71
0.27
-0.42
-0.63
-0.62
-0.97
-0.30

Model 2
T-ratio
32.33
10.79
-6.97
-12.34
-13.79
-20.92
-7.88

Model 2
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model 3
Coeff.
2.59
0.26
-0.42
-0.63
-0.61
-0.97
-0.29

Model 3
T-ratio
22.55
10.76
-6.97
-12.34
-13.76
-20.90
-7.87

Model 3
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

-0.28

-6.10

<0.001

-0.28

-6.14

<0.001

-0.27

-6.10

<0.001

-0.31
-0.78

-8.53
-11.78

<0.001
<0.001

-0.31
-0.78

-8.60
-11.83

<0.001
<0.001

-0.31
-0.77

-8.50
-11.77

<0.001
<0.001

-0.54

-23.27

<0.001

-0.54

-23.24

<0.001

-0.54

-23.26

<0.001

-0.26
-0.91
-0.63

-8.88
-18.67
-22.99

<0.001
<0.001
<0.001

-0.26
-0.91
-0.63

-8.93
-18.67
-22.99

<0.001
<0.001
<0.001

-0.26
-0.91
-0.62

-8.87
-18.59
-22.98

<0.001
<0.001
<0.001

-0.40

-15.20

<0.001

-0.40

-15.19

<0.001

-0.40

-15.18

<0.001

-0.34
0.57

-11.87
23.46

<0.001
<0.001

-0.34
0.58

-11.86
23.60

<0.001
<0.001

-0.34
0.57

-11.89
23.47

<0.001
<0.001

0.00012
-0.0074
0.000012

-2.05
-6.41
0.021

0.041
<0.001
0.988

0.00015
-0.0074

-2.36
-6.35

0.019
<0.001

-0.00018
-0.0071

-2.90
-6.16

0.005
0.019

0.00079

1.41

0.160
-0.0019
0.0017
0.0016
0.00033

-1.94
1.69
2.60
0.35
0.70

0.053
0.091
0.010
0.726

0.67

0.67

180

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Table C.4. Relationships to Minutes of Moderate Physical Activity (with robust standard errors)
Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black nonHispanic
Other nonHispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Precipitation
Cooling Degree
Days
Heating Degree
Days
Original
Compactness
Index
Refined
Compactness
Index
Density factor
Mix factor
Centering factor
Street factor
Pseudo R2

Model 1
Coeff.
268.34
22.11
-25.35
-31.25
-23.07
-58.41
-26.60

Model 1
T-ratio
32.35
13.87
-11.34
-13.50
-7.77
-19.11
-12.97

Model 1
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model
2 Coeff.
269.40
22.11
-25.35
-31.25
-23.07
-58.41
-26.60

Model 2
T-ratio
28.90
13.87
-11.34
-13.50
-7.77
-19.11
-12.97

Model 2
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model 3
Coeff.
260.86
15.59
-36.92
-39.69
-30.68
-30.68
-64.64

Model 3
T-ratio
18.75
4.84
-8.58
-11.45
-6.83
-14.73
-7.04

Model 3
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

-9.06

-3.00

0.003

-9.06

-3.00

0.003

-27.14

-4.71

<0.001

-15.49
12.16

-4.82
4.55

<0.001
<0.001

-15.49
12.16

-4.82
4.55

<0.001
<0.001

-28.75
-25.42

-5.08
4.45

<0.001
<0.001

24.56

14.70

<0.001

24.56

14.69

<0.001

21.58

13.50

<0.001

20.36
-24.12
-3.27

15.25
-12.74
-2.15

<0.001
<0.001
0.031

20.36
-24.12
-3.27

15.25
-12.74
-2.15

<0.001
<0.001
0.031

34.60
-28.28
-14.04

10.03
-4.76
-2.87

<0.001
<0.001
0.005

-1.10

-0.70

0.484

-1.10

-0.70

0.484

-5.27

-1.68

0.092

38.61
52.58

19.15
43.12

<0.001
<0.001

38.61
52.58

19.15
43.12

<0.001
<0.001

45.18
57.76

10.82
23.23

<0.001
<0.001

-0.22
-0.0065

-1.96
-4.78

0.050
<0.001

-0.23
-0.0064

-2.05
-4.47

0.041
<0.001

-0.37

-3.01

0.003

-0.0031

-4.01

<0.001

-0.0029

-3.54

0.001

-0.14

-2.432

0.016
-0.154

-2.460

0.015
0.046
-0.090
-0.050
-0.042

0.54
-0.68
-0.69
-0.36
0.17

0.592
0.500
0.493
0.723

0.16

0.16

181

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Table C.5. Relationships to High Blood Pressure (with robust standard errors)
Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black nonHispanic
Other nonHispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Crime
Original
Compactness
Index
Refined
Compactness
Index
Density factor
Mix factor
Centering factor
Street factor
Pseudo R2

Model 1
Coeff.
-2.87
0.26
0.76
1.93
2.77
2.90
0.61

Model 1
T-ratio
-37.11
16.19
15.72
49.48
68.74
62.62
19.24

Model 1
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model 2
Coeff.
-2.83
0.26
0.75
1.93
2.77
2.89
0.61

Model 2
T-ratio
-37.14
16.20
15.73
49.57
68.76
62.77
19.16

Model 2
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

Model 3
Coeff.
-2.76
0.26
0.75
1.93
2.77
2.90
0.61

Model 3
T-ratio
-30.79
16.18
15.72
49.48
68.73
62.66
19.17

Model 3
P-value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001

-0.038

-0.56

0.577

-0.039

-0.56

0.573

-0.038

-0.56

0.575

-0.19
0.21

-2.96
4.77

0.004
<0.001

-0.19
0.21

-3.00
4.76

0.003
<0.001

-0.18
0.20

-2.97
4.77

0.003
<0.001

0.25

8.13

<0.001

0.25

8.11

<0.001

0.25

8.12

<0.001

0.20
0.46
0.27

8.77
16.12
12.89

<0.001
<0.001
<0.001

0.20
0.46
0.27

8.77
16.14
12.94

<0.001
<0.001
<0.001

0.20
0.46
0.27

8.75
16.06
12.86

<0.001
<0.001
<0.001

0.17

6.34

<0.001

0.17

6.35

<0.001

0.17

6.32

<0.001

-0.016
-0.049

-0.67
-2.17

0.504
0.030

-0.016
-0.049

-0.66
-2.18

0.507
0.029

-0.015
-0.049

-0.67
-2.17

0.504
0.030

0.00015
-0.0018

2.66
-3.15

0.008
0.002

0.00017

3.00

0.003

0.00015

2.47

0.014

-0.0021

-3.62

0.001
-0.00068
-0.0011
-0.00044
-0.00047

-0.73
-1.25
-0.74
-0.53
0.57

0.463
0.214
0.462
0.595

0.54

0.55

182

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Table C.6. Relationships to Coronary Heart Disease (with robust standard errors)
Model 1 Model 1 Model 1 Model 2 Model 2 Model 2 Model 3
Model 3 Model 3
Fixed Effect
Coeff.
T-ratio
P-value Coeff.
T-ratio P-value Coeff.
T-ratio P-value
Constant
-6.01
-20.86 <0.001
-5.96 -20.73 <0.001
-5.59 -18.25
<0.001
Male
0.68
14.65 <0.001
0.68 14.77 <0.001
0.68
14.62
<0.001
Age 30-44
0.79
3.11
0.002
0.80
3.17
0.002
0.79
3.09
0.002
Age 45-64
2.37
10.53 <0.001
2.37 10.75 <0.001
2.37
10.50
<0.001
Age 65-74
3.30
14.70 <0.001
3.30 15.00 <0.001
3.29
14.65
<0.001
Age 75+
3.66
16.27 <0.001
3.66 16.61 <0.001
3.66
16.20
<0.001
Black non-0.18
-2.74
0.007
-0.18
-2.69
0.007
-0.18
-2.81
0.005
Hispanic
Other non0.032
0.27
0.787
0.031
0.26
0.797
0.034
0.28
0.777
Hispanic
Hispanic
-0.11
-1.09
0.267
-0.11
-1.13
0.259
-0.11
-1.15
0.252
Less than high
0.23
4.50 <0.001
0.23
4.54 <0.001
0.23
4.50
<0.001
school
High school
0.067
1.45
0.143
0.067
1.46
0.145
0.066
1.42
0.157
graduate
Some College
0.19
5.03 <0.001
0.19
5.08 <0.001
0.18
4.98
<0.001
Income < $25000
0.78
14.39 <0.001
0.78 14.57 <0.001
0.78
14.23
<0.001
Income
0.45
7.85 <0.001
0.45
7.94 <0.001
0.44
7.76
<0.001
$25000<$50000
Income
0.26
4.65 <0.001
0.26
4.71 <0.001
0.26
4.60
<0.001
$50000<$75000
Current Smoker
0.14
2.05
0.040
0.14
2.06
0.039
0.14
2.05
0.040
Recommended
-0.026
-0.73
0.463
-0.026
-0.74
0.461
-0.025
-0.70
0.480
Servings
Fruits/Vegetables
Original
-0.0024
-1.79
0.073
Compactness
Index
Refined
-0.0028
-2.12
0.034
Compactness
Index
Density factor
-0.0014
-0.67
0.504
Mix factor
-0.0058
-3.07
0.003
Centering factor
-0.0013
-1.35
0.179
Street factor
0.0025
1.61
0.107
Pseudo R2
0.51
0.58
0.61

183

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES


Table C.7. Relationships to Diabetes (with robust standard errors)
Fixed Effect
Constant
Male
Age 30-44
Age 45-64
Age 65-74
Age 75+
Black non-Hispanic
Other nonHispanic
Hispanic
Less than high
school
High school
graduate
Some College
Income < $25000
Income
$25000<$50000
Income
$50000<$75000
Current Smoker
Recommended
Servings
Fruits/Vegetables
Original
Compactness
Index
Refined
Compactness
Index
Density factor
Mix factor
Centering factor
Street factor
Pseudo R2

Model 1
Coeff.
-5.18
0.28
1.26

Model 1
T-ratio
-40.74
10.92
12.59

Model 1
P-value
<0.001
<0.001
<0.001

Model 2
Coeff.
-5.17
0.28
1.26

Model 2
T-ratio
-40.10
10.92
12.57

Model 2
P-value
<0.001
<0.001
<0.001

Model 3
Coeff.
-4.97
0.28
1.26

Model 3
T-ratio
-33.83
10.87
12.51

Model 3
P-value
<0.001
<0.001
<0.001

2.56
3.16
3.01
0.64
0.33

28.13
32.76
30.69
17.27
4.13

<0.001
<0.001
<0.001
<0.001
<0.001

2.56
3.16
3.01
0.63
0.33

28.13
32.77
30.68
17.21
4.09

<0.001
<0.001
<0.001
<0.001
<0.001

2.56
3.16
3.01
0.63
0.33

27.96
32.59
30.50
17.40
4.13

<0.001
<0.001
<0.001
<0.001
<0.001

0.42
0.27

8.34
4.23

<0.001
<0.001

0.42
0.27

8.42
4.23

<0.001
<0.001

0.42
0.27

8.11
4.23

<0.001
<0.001

0.20

4.78

<0.001

0.20

4.77

<0.001

0.20

4.75

<0.001

0.24
0.84
0.46

5.92
15.59
9.21

<0.001
<0.001
<0.001

0.24
0.84
0.46

5.90
15.57
9.23

<0.001
<0.001
<0.001

0.24
0.84
0.46

5.88
15.34
9.03

<0.001
<0.001
<0.001

0.30

5.08

<0.001

0.30

5.09

<0.001

0.30

5.00

<0.001

-0.25
-0.020

-6.10
-0.60

<0.001
0.550

-0.25
-0.020

-6.10
-0.60

<0.001
0.548

-0.25
-0.020

-6.15
-0.58

<0.001
0.560

-0.0015

-2.22

0.034
-0.0016

-2.27

0.024
-0.00087
-0.0033
-0.00024
0.0011

-0.73
-2.46
-0.36
0.97
0.60

0.464
0.015
0.714
0.331

0.57

0.58

184

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix D. 2010 Metropolitan Indices


MSAc Code Geoid
Name
10420 10420 Akron, OH MSA
Albany-Schenectady-Troy, NY
10580 10580 MSA
10740 10740 Albuquerque, NM MSA
Allentown-Bethlehem-Easton,
10900 10900 PA-NJ MSA
11100 11100 Amarillo, TX MSA
11460 11460 Ann Arbor, MI MSA
11540 11540 Appleton, WI MSA
11700 11700 Asheville, NC MSA
Atlanta-Sandy Springs12060 12060 Marietta, GA MSA
Atlantic City-Hammonton, NJ
12100 12100 MSA
Augusta-Richmond County, GA12260 12260 SC MSA
Austin-Round Rock-San
12420 12420 Marcos, TX MSA
12540 12540 Bakersfield-Delano, CA MSA
12580 12580 Baltimore-Towson, MD MSA
12940 12940 Baton Rouge, LA MSA
13140 13140 Beaumont-Port Arthur, TX MSA
13380 13380 Bellingham, WA MSA
13780 13780 Binghamton, NY MSA
13820 13820 Birmingham-Hoover, AL MSA
14260 14260 Boise City-Nampa, ID MSA
14500 14500 Boulder, CO MSA
14740 14740 Bremerton-Silverdale, WA MSA
Bridgeport-Stamford-Norwalk,
14860 14860 CT MSA
15180 15180 Brownsville-Harlingen, TX MSA
15380 15380 Buffalo-Niagara Falls, NY MSA
Burlington-South Burlington,
15540 15540 VT MSA
15940 15940 Canton-Massillon, OH MSA
15980 15980 Cape Coral-Fort Myers, FL MSA
16300 16300 Cedar Rapids, IA MSA
16580 16580 Champaign-Urbana, IL MSA

Lsad density
10
factor
M1
94.55

mix
centering
factor
factor
113.13
90.69

street composite
factor
index
106.81
103.15

M1
M1

95.40
103.60

105.96
102.57

108.19
99.36

86.04
97.51

95.12
98.07

M1
M1
M1
M1
M1

98.76
96.16
103.27
90.65
80.71

128.59
109.27
105.04
99.81
64.12

101.10
76.98
123.11
156.72
97.61

135.97
91.56
89.95
79.92
88.53

124.40
107.49
122.76
132.69
76.52

M1

97.80

85.47

89.89

75.92

40.99

M1

96.33

100.10

154.52

130.71

150.36

M1

85.25

60.69

88.47

73.85

59.18

M1
M1
M1
M1
M1
M1
M1
M1
M1
M1
M1

100.42
101.29
115.97
91.27
85.37
85.29
89.70
86.67
95.80
106.89
90.48

99.66
114.13
123.21
72.03
88.45
92.75
88.92
67.88
110.45
115.32
87.55

138.78
76.82
123.12
69.74
112.62
113.43
102.07
99.52
75.15
100.09
112.87

102.88
73.14
136.35
80.40
113.76
96.89
69.84
105.21
91.88
118.95
86.20

102.44
81.78
115.62
55.60
111.54
118.01
95.97
73.55
91.06
133.68
108.86

M1
M1
M1

110.63
90.92
107.94

132.86
77.74
127.67

118.02
51.43
102.46

100.81
105.96
95.10

121.64
74.69
106.36

M1
M1
M1
M1
M1

88.32
90.54
91.87
92.94
100.00

102.21
106.64
81.41
105.64
123.27

168.79
76.45
91.52
104.67
153.64

70.68
117.92
126.34
81.25
82.81

135.06
106.99
99.22
111.81
145.16

185

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code Geoid


Name
16620 16620 Charleston, WV MSA
Charleston-North Charleston16700 16700 Summerville, SC MSA
Charlotte-Gastonia-Rock Hill,
16740 16740 NC-SC MSA
16820 16820 Charlottesville, VA MSA
16860 16860 Chattanooga, TN-GA MSA
17020 17020 Chico, CA MSA
Cincinnati-Middletown, OH-KY17140 17140 IN MSA
17300 17300 Clarksville, TN-KY MSA
Cleveland-Elyria-Mentor, OH
17460 17460 MSA
17780 17780 College Station-Bryan, TX MSA
17820 17820 Colorado Springs, CO MSA
17900 17900 Columbia, SC MSA
17980 17980 Columbus, GA-AL MSA
18140 18140 Columbus, OH MSA
18580 18580 Corpus Christi, TX MSA
Davenport-Moline-Rock Island,
19340 19340 IA-IL MSA
19380 19380 Dayton, OH MSA
Deltona-Daytona Beach19660 19660 Ormond Beach, FL MSA
Denver-Aurora-Broomfield, CO
19740 19740 MSA
Des Moines-West Des Moines,
19780 19780 IA MSA
20260 20260 Duluth, MN-WI MSA
20500 20500 Durham-Chapel Hill, NC MSA
21340 21340 El Paso, TX MSA
21500 21500 Erie, PA MSA
21660 21660 Eugene-Springfield, OR MSA
21780 21780 Evansville, IN-KY MSA
22020 22020 Fargo, ND-MN MSA
22180 22180 Fayetteville, NC MSA
Fayetteville-Springdale-Rogers,
22220 22220 AR-MO MSA
22420 22420 Flint, MI MSA

Lsad density
10
factor
M1
83.81

mix
centering
factor
factor
67.01
136.80

street composite
factor
index
112.05
115.68

M1

95.29

89.19

108.94

99.03

98.53

M1
M1
M1
M1

94.55
91.16
86.14
91.18

84.71
86.08
61.15
114.46

103.05
141.81
94.27
88.79

86.93
71.77
72.90
79.93

70.45
119.08
63.63
109.94

M1
M1

98.75
84.48

107.80
39.67

98.95
74.47

93.67
60.83

80.75
41.49

M1
M1
M1
M1
M1
M1
M1

105.11
102.49
102.94
89.63
94.45
101.58
98.68

123.72
94.65
108.37
69.14
84.78
112.24
118.31

95.54
91.03
75.94
108.38
125.19
95.56
90.15

84.96
91.47
121.76
66.63
77.79
112.19
110.41

85.62
111.72
106.33
67.45
108.38
93.00
117.29

M1
M1

91.78
93.65

121.21
114.40

70.03
95.13

102.95
105.55

105.59
101.48

M1

91.35

88.02

66.48

116.35

89.68

M1

118.31

119.44

109.11

125.16

107.10

M1
M1
M1
M1
M1
M1
M1
M1
M1

97.68
85.24
91.59
114.90
97.73
95.35
91.57
99.18
91.13

120.63
89.56
74.84
99.42
130.61
125.70
92.59
118.65
71.69

99.46
117.03
80.27
73.41
113.69
116.84
86.07
106.96
72.57

82.83
77.22
84.98
128.66
88.92
91.29
84.34
73.56
71.77

104.90
103.14
73.84
105.64
130.39
125.63
91.67
121.82
66.02

M1
M1

84.55
89.57

67.95
90.58

80.67
114.82

81.81
97.49

66.26
106.48

186

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code Geoid


Name
22500 22500 Florence, SC MSA
22660 22660 Fort Collins-Loveland, CO MSA
22900 22900 Fort Smith, AR-OK MSA
23060 23060 Fort Wayne, IN MSA
23420 23420 Fresno, CA MSA
23540 23540 Gainesville, FL MSA
Grand Rapids-Wyoming, MI
24340 24340 MSA
24540 24540 Greeley, CO MSA
24580 24580 Green Bay, WI MSA
Greensboro-High Point, NC
24660 24660 MSA
Greenville-Mauldin-Easley, SC
24860 24860 MSA
25060 25060 Gulfport-Biloxi, MS MSA
Hagerstown-Martinsburg, MD25180 25180 WV MSA
25420 25420 Harrisburg-Carlisle, PA MSA
Hartford-West Hartford-East
25540 25540 Hartford, CT MSA
Hickory-Lenoir-Morganton, NC
25860 25860 MSA
26100 26100 Holland-Grand Haven, MI MSA
Houma-Bayou Cane-Thibodaux,
26380 26380 LA MSA
Houston-Sugar Land-Baytown,
26420 26420 TX MSA
Huntington-Ashland, WV-KY26580 26580 OH MSA
26620 26620 Huntsville, AL MSA
26900 26900 Indianapolis-Carmel, IN MSA
27140 27140 Jackson, MS MSA
27260 27260 Jacksonville, FL MSA
28020 28020 Kalamazoo-Portage, MI MSA
28140 28140 Kansas City, MO-KS MSA
Kennewick-Pasco-Richland, WA
28420 28420 MSA
Killeen-Temple-Fort Hood, TX
28660 28660 MSA

Lsad density
10
factor
M1
81.22
M1
94.53
M1
80.74
M1
92.42
M1
101.75
M1
94.58

mix
centering
factor
factor
51.13
87.85
106.30
96.44
56.78
75.30
93.70
89.90
126.18
81.45
87.63
102.79

street composite
factor
index
61.44
61.06
100.59
115.15
86.02
64.84
73.85
86.67
82.42
92.24
99.45
111.36

M1
M1
M1

91.39
87.33
89.90

91.78
99.05
90.49

99.15
94.05
66.77

74.75
85.82
53.34

79.18
103.61
65.35

M1

88.22

80.57

84.94

70.70

63.50

M1
M1

86.69
86.03

72.89
69.80

81.15
80.53

71.40
97.52

58.98
87.61

M1
M1

84.10
93.54

74.10
102.14

112.54
99.29

78.51
119.17

94.13
111.40

M1

100.12

113.10

119.54

72.59

93.50

M1
M1

78.64
86.45

40.46
81.52

67.00
78.64

56.95
71.71

24.86
78.17

M1

83.73

75.47

106.77

86.11

100.13

M1

108.30

102.66

92.56

129.43

76.74

M1
M1
M1
M1
M1
M1
M1

84.25
86.18
98.11
87.35
96.81
85.55
96.84

67.73
58.29
99.65
64.41
82.50
75.00
109.49

142.77
89.43
98.42
105.46
90.17
85.58
80.45

108.91
99.31
102.31
73.80
111.76
64.97
103.52

118.43
78.02
83.89
72.30
80.85
70.32
77.60

M1

92.84

108.63

81.96

85.86

105.03

M1

89.16

79.86

78.17

94.80

83.12

187

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code

Geoid

28700
28940
29140
29180

28700
28940
29140
29180

29420

29420

29460
29540
29620
29700
29740
29820
30460
30700

29460
29540
29620
29700
29740
29820
30460
30700

30780
30980

30780
30980

31140
31180
31340
31420
31540
31700

31140
31180
31340
31420
31540
31700

32580
32780
32820
32900

32580
32780
32820
32900

33340

33340

33460
33660
33700
33860

33460
33660
33700
33860

Name
Kingsport-Bristol-Bristol, TN-VA
MSA
Knoxville, TN MSA
Lafayette, IN MSA
Lafayette, LA MSA
Lake Havasu City-Kingman, AZ
MSA
Lakeland-Winter Haven, FL
MSA
Lancaster, PA MSA
Lansing-East Lansing, MI MSA
Laredo, TX MSA
Las Cruces, NM MSA
Las Vegas-Paradise, NV MSA
Lexington-Fayette, KY MSA
Lincoln, NE MSA
Little Rock-North Little RockConway, AR MSA
Longview, TX MSA
Louisville/Jefferson County, KYIN MSA
Lubbock, TX MSA
Lynchburg, VA MSA
Macon, GA MSA
Madison, WI MSA
Manchester-Nashua, NH MSA
McAllen-Edinburg-Mission, TX
MSA
Medford, OR MSA
Memphis, TN-MS-AR MSA
Merced, CA MSA
Milwaukee-Waukesha-West
Allis, WI MSA
Minneapolis-St. PaulBloomington, MN-WI MSA
Mobile, AL MSA
Modesto, CA MSA
Montgomery, AL MSA

Lsad density
10
factor

mix
factor

centering
factor

street
factor

composite
index

M1
M1
M1
M1

78.73
88.10
95.46
90.03

40.53
60.62
90.63
87.35

89.67
100.77
94.82
115.90

82.87
82.53
83.10
92.72

60.00
68.22
106.55
111.44

M1

85.24

55.15

73.04

65.97

60.13

M1
M1
M1
M1
M1
M1
M1
M1

87.51
95.61
101.03
104.20
89.33
142.12
99.56
111.55

54.24
110.05
92.21
117.12
84.27
105.02
110.42
132.99

95.32
124.31
141.56
99.89
108.16
136.42
115.34
96.74

128.15
84.74
72.80
106.87
89.06
114.29
95.11
96.78

87.64
112.64
111.61
131.25
109.17
121.20
116.76
131.95

M1
M1

88.00
81.66

75.36
71.62

93.55
81.06

90.35
68.46

76.08
73.06

M1
M1
M1
M1
M1
M1

98.44
97.23
81.51
84.72
101.00
95.10

89.48
116.70
57.07
71.90
115.83
104.38

93.12
87.56
76.38
86.32
168.11
114.15

102.87
90.44
77.42
74.47
94.85
89.28

82.92
113.41
63.97
79.92
136.69
112.19

M1
M1
M1
M1

94.43
89.67
96.60
93.90

76.78
115.31
77.76
114.76

90.99
128.06
94.23
96.48

104.60
80.42
90.62
66.25

83.89
128.86
70.77
105.86

M1

113.31

126.73

153.40

130.35

134.18

M1
M1
M1
M1

105.92
92.43
109.91
90.01

110.34
88.23
140.69
85.97

111.41
78.79
62.32
98.71

108.60
112.30
102.89
80.50

88.69
97.48
113.28
91.20

188

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code

Geoid

34820
34940

34820
34940

34980
35300

34980
35300

35380

35380

35840
35980
36100
36260
36420
36500

35840
35980
36100
36260
36420
36500

36540

36540

36740

36740

37100

37100

37340

37340

37860
37900

37860
37900

38060
38300

38060
38300

38860

38860

38900
38940

38900
38940

39100
39140

39100
39140

39300
39340

39300
39340

Name
Myrtle Beach-North Myrtle
Beach-Conway, SC MSA
Naples-Marco Island, FL MSA
Nashville-Davidson-Murfreesboro--Franklin, TN
MSA
New Haven-Milford, CT MSA
New Orleans-Metairie-Kenner,
LA MSA
North Port-BradentonSarasota, FL MSA
Norwich-New London, CT MSA
Ocala, FL MSA
Ogden-Clearfield, UT MSA
Oklahoma City, OK MSA
Olympia, WA MSA
Omaha-Council Bluffs, NE-IA
MSA
Orlando-Kissimmee-Sanford, FL
MSA
Oxnard-Thousand OaksVentura, CA MSA
Palm Bay-Melbourne-Titusville,
FL MSA
Pensacola-Ferry Pass-Brent, FL
MSA
Peoria, IL MSA
Phoenix-Mesa-Glendale, AZ
MSA
Pittsburgh, PA MSA
Portland-South PortlandBiddeford, ME MSA
Portland-Vancouver-Hillsboro,
OR-WA MSA
Port St. Lucie, FL MSA
Poughkeepsie-NewburghMiddletown, NY MSA
Prescott, AZ MSA
Providence-New Bedford-Fall
River, RI-MA MSA
Provo-Orem, UT MSA

Lsad density
10
factor

mix
factor

centering
factor

street
factor

composite
index

M1
M1

83.43
91.57

54.95
81.95

104.88
55.19

95.40
90.69

88.70
75.23

M1
M1

91.54
106.86

63.92
127.52

96.17
113.51

77.00
97.82

51.74
116.29

M1

104.84

117.83

96.09

149.94

119.74

M1
M1
M1
M1
M1
M1

97.45
87.22
80.80
100.96
94.64
89.23

101.45
84.71
41.30
120.39
96.26
80.87

84.95
137.44
105.49
62.22
89.86
121.00

126.69
71.04
91.78
103.52
100.38
98.73

105.49
108.85
74.67
99.58
82.07
114.63

M1

102.64

120.53

99.67

103.54

108.42

M1

102.40

85.79

89.29

129.14

83.97

M1

107.91

133.35

78.01

118.31

113.87

M1

96.94

79.64

60.02

105.42

77.91

M1
M1

88.54
88.93

81.12
100.39

75.12
109.76

88.65
97.72

76.84
110.49

M1
M1

111.60
96.16

102.36
115.14

96.37
107.78

111.33
119.33

78.32
95.45

M1

86.06

79.09

157.47

80.24

107.72

M1
M1

111.14
92.74

136.12
77.05

100.81
62.73

124.98
106.43

109.85
80.75

M1
M1

89.38
82.33

95.38
53.19

97.49
58.15

70.30
69.96

79.51
48.96

M1
M1

105.40
104.53

83.28
123.55

112.77
77.37

141.95
100.08

104.34
108.45

189

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code Geoid


Name
39580 39580 Raleigh-Cary, NC MSA
39740 39740 Reading, PA MSA
39900 39900 Reno-Sparks, NV MSA
40060 40060 Richmond, VA MSA
Riverside-San Bernardino40140 40140 Ontario, CA MSA
40220 40220 Roanoke, VA MSA
40380 40380 Rochester, NY MSA
40420 40420 Rockford, IL MSA
Sacramento--Arden-Arcade-40900 40900 Roseville, CA MSA
Saginaw-Saginaw Township
40980 40980 North, MI MSA
41180 41180 St. Louis, MO-IL MSA
41420 41420 Salem, OR MSA
41500 41500 Salinas, CA MSA
41620 41620 Salt Lake City, UT MSA
San Antonio-New Braunfels, TX
41700 41700 MSA
San Diego-Carlsbad-San
41740 41740 Marcos, CA MSA
San Jose-Sunnyvale-Santa
41940 41940 Clara, CA MSA
San Luis Obispo-Paso Robles,
42020 42020 CA MSA
Santa Barbara-Santa Maria42060 42060 Goleta, CA MSA
Santa Cruz-Watsonville, CA
42100 42100 MSA
42220 42220 Santa Rosa-Petaluma, CA MSA
42340 42340 Savannah, GA MSA
Scranton--Wilkes-Barre, PA
42540 42540 MSA
Shreveport-Bossier City, LA
43340 43340 MSA
43620 43620 Sioux Falls, SD MSA
South Bend-Mishawaka, IN-MI
43780 43780 MSA
43900 43900 Spartanburg, SC MSA
44060 44060 Spokane, WA MSA

Lsad density
10
factor
M1
96.99
M1
102.22
M1
100.78
M1
96.36

mix
centering
factor
factor
87.30
109.43
121.83
129.72
93.69
137.29
78.08
101.95

street composite
factor
index
88.16
84.25
113.76
137.90
94.06
120.85
92.83
76.41

M1
M1
M1
M1

103.72
90.65
96.12
94.78

111.18
85.88
103.86
110.04

77.03
83.67
96.77
91.83

80.33
93.21
62.00
107.05

56.25
93.77
74.50
114.98

M1

111.65

119.11

104.19

108.92

99.27

M1
M1
M1
M1
M1

86.77
97.68
93.11
101.65
117.77

93.77
108.29
123.48
116.00
125.49

110.97
93.86
113.50
102.94
93.32

93.62
113.80
98.10
90.70
97.63

116.62
82.06
123.35
115.19
106.96

M1

100.67

93.56

95.15

102.43

77.37

M1

125.08

130.37

100.90

119.95

105.18

M1

149.50

148.76

86.80

131.45

128.76

M1

89.90

119.80

103.87

88.53

118.90

M1

112.28

148.85

109.48

122.05

146.59

M1
M1
M1

98.88
93.70
90.08

146.15
132.31
84.94

107.90
91.91
115.36

112.18
96.82
115.03

145.02
113.92
115.81

M1

91.28

116.46

95.07

123.01

115.84

M1
M1

87.79
97.68

76.94
104.85

72.39
95.96

84.53
60.16

72.63
101.75

M1
M1
M1

90.94
81.26
98.98

94.08
68.26
115.82

111.91
91.26
108.57

118.68
72.48
128.26

121.71
74.00
129.40

190

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code Geoid


Name
44100 44100 Springfield, IL MSA
44180 44180 Springfield, MO MSA
44700 44700 Stockton, CA MSA
45060 45060 Syracuse, NY MSA
45220 45220 Tallahassee, FL MSA
Tampa-St. Petersburg45300 45300 Clearwater, FL MSA
45780 45780 Toledo, OH MSA
45820 45820 Topeka, KS MSA
45940 45940 Trenton-Ewing, NJ MSA
46060 46060 Tucson, AZ MSA
46140 46140 Tulsa, OK MSA
46220 46220 Tuscaloosa, AL MSA
46340 46340 Tyler, TX MSA
46540 46540 Utica-Rome, NY MSA
46700 46700 Vallejo-Fairfield, CA MSA
Virginia Beach-Norfolk47260 47260 Newport News, VA-NC MSA
47300 47300 Visalia-Porterville, CA MSA
47380 47380 Waco, TX MSA
48620 48620 Wichita, KS MSA
48900 48900 Wilmington, NC MSA
49180 49180 Winston-Salem, NC MSA
49420 49420 Yakima, WA MSA
49620 49620 York-Hanover, PA MSA
Youngstown-Warren49660 49660 Boardman, OH-PA MSA
1698016974 16980 Chicago-Joliet-Naperville, IL MD
1698023844 16980 Gary, IN MD
Lake County-Kenosha County,
1698029404 16980 IL-WI MD
1910019124 19100 Dallas-Plano-Irving, TX MD
1910023104 19100 Fort Worth-Arlington, TX MD
Detroit-Livonia-Dearborn, MI
1982019804 19820 MD
Warren-Troy-Farmington Hills,
1982047644 19820 MI MD
Los Angeles-Long Beach3110031084 31100 Glendale, CA MD

Lsad density
10
factor
M1
90.39
M1
89.10
M1
106.54
M1
94.75
M1
91.64

mix
centering
factor
factor
100.51
160.03
89.25
75.99
135.75
82.11
100.93
122.57
68.25
130.77

street composite
factor
index
96.74
142.24
91.87
83.96
121.04
120.28
69.91
96.65
79.80
98.95

M1
M1
M1
M1
M1
M1
M1
M1
M1
M1

105.18
95.30
88.98
115.88
100.79
90.54
85.85
85.76
90.87
105.38

105.35
120.34
83.12
128.00
90.96
92.40
68.60
72.48
83.53
132.03

93.00
85.46
102.18
97.36
78.71
93.54
154.72
122.62
98.35
79.32

150.09
95.85
71.38
139.06
94.72
103.35
92.03
93.19
61.91
115.90

98.49
100.90
94.82
144.71
78.92
86.65
122.18
110.66
84.71
124.16

M1
M1
M1
M1
M1
M1
M1
M1

106.41
91.94
87.96
95.63
85.89
86.43
90.95
90.92

105.24
106.37
96.10
107.27
73.12
68.62
117.91
95.83

102.38
79.64
100.62
88.57
83.92
87.42
133.08
113.20

131.60
83.98
107.83
83.65
84.13
68.47
65.81
90.32

104.45
91.55
117.11
91.74
77.27
63.44
123.19
105.12

M1
M3
M3

87.36
145.50
94.53

100.76
140.09
107.73

74.10
143.24
82.31

81.52
160.21
106.33

78.08
125.90
96.70

M3
M3
M3

101.65
111.46
103.71

112.39
105.90
100.89

67.78
94.21
72.55

132.08
129.74
117.21

103.10
86.15
78.56

M3

125.20

124.65

107.48

183.98

137.17

M3

97.88

110.33

70.54

96.17

67.03

M3

187.39

160.18

115.66

154.40

130.33

191

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

MSAc Code

Geoid

3110042044

31100

3310022744

33100

3310033124

33100

3310048424
3562020764
3562035004
3562035084

33100
35620
35620
35620

3562035644
3798015804
3798037964
3798048864

35620
37980
37980
37980

4186036084

41860

4186041884

41860

4266042644
4266045104

42660
42660

4790013644

47900

4790047894

47900

Name
Santa Ana-Anaheim-Irvine, CA
MD
Fort Lauderdale-Pompano
Beach-Deerfield Beach, FL MD
Miami-Miami Beach-Kendall, FL
MD
West Palm Beach-Boca RatonBoynton Beach, FL MD
Edison-New Brunswick, NJ MD
Nassau-Suffolk, NY MD
Newark-Union, NJ-PA MD
New York-White Plains-Wayne,
NY-NJ MD
Camden, NJ MD
Philadelphia, PA MD
Wilmington, DE-MD-NJ MD
Oakland-Fremont-Hayward, CA
MD
San Francisco-San MateoRedwood City, CA MD
Seattle-Bellevue-Everett, WA
MD
Tacoma, WA MD
Bethesda-Rockville-Frederick,
MD MD
Washington-ArlingtonAlexandria, DC-VA-MD-WV MD

Lsad density
10
factor

mix
factor

centering
factor

street
factor

composite
index

M3

161.91

155.02

79.64

181.81

139.86

M3

140.93

136.53

61.79

153.66

121.41

M3

160.18

136.41

117.91

166.90

144.12

M3
M3
M3
M3

110.73
109.41
123.33
126.86

121.02
125.05
144.75
139.67

69.66
69.02
81.01
90.43

118.46
137.91
155.85
113.76

98.18
96.77
117.04
109.62

M3
M3
M3
M3

384.29
105.39
141.01
102.42

159.34
125.72
142.25
109.29

213.49
78.53
115.95
96.53

193.80
120.07
140.06
120.29

203.36
103.22
122.42
112.94

M3

136.28

145.75

88.11

159.44

127.24

M3

185.97

167.17

230.92

162.83

194.28

M3
M3

121.27
103.62

123.99
105.56

121.68
92.25

131.86
119.05

116.11
107.48

M3

115.08

123.84

98.97

118.94

114.66

M3

122.35

117.61

133.16

125.91

107.21

192

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix E. Urbanized Areas Compactness Indices 2010


UA10 UA00
code code
199
766
970
1171
1495
2602
2683
2764
3358
3817
3898
4222
4384
4681
4843
5680
7786
8785
8974
10972
11350
13375
13510
15508
15670
15832
16264
16885
17668
18964
19099
19234
19504
19558
19755

199
766
928
1171
1495
2602
2683
2764
3358
3817
3898
4222
4384
4681
4843
5680
7786
8785
8974
10972
11350
13375
13510
15508
15670
15832
16264
16885
17668
18964
19099
19234
19504
19558
87328

UZA name
Aberdeen--Bel Air South--Bel Air
North, MD
Akron, OH
Albany--Schenectady, NY
Albuquerque, NM
Allentown, PA--NJ
Ann Arbor, MI
Antioch, CA
Appleton, WI
Asheville, NC
Atlanta, GA
Atlantic City, NJ
Augusta-Richmond County, GA--SC
Austin, TX
Bakersfield, CA
Baltimore, MD
Baton Rouge, LA
Birmingham, AL
Boise City, ID
Bonita Springs, FL
Brownsville, TX
Buffalo, NY
Canton, OH
Cape Coral, FL
Charleston--North Charleston, SC
Charlotte, NC--SC
Chattanooga, TN--GA
Chicago, IL--IN
Cincinnati, OH--KY--IN
Cleveland, OH
Columbia, SC
Columbus, GA--AL
Columbus, OH
Concord, CA
Concord, NC
Conroe--The Woodlands, TX

density
mix
centering street composite
factor10 factor10 factor10 factor10 index10
85.49
81.39
97.77
116.44
97.96
97.99
114.41
95.15
60.41
84.64
93.87
72.48
113.28
125.2
129.32
81.92
73.46
108.78
77.33
104.71
108.69
78.14
71.37
89.82
82.95
68.92
138.66
96.17
98.46
77.26
83.9
109.73
117.87
61.76
84.06

120.77
116.43
118.87
78.03
143.39
79.81
159.21
115.28
95.23
75.63
91.07
77.69
81.33
121.55
121.02
75.3
86.42
117.41
82.83
69.61
129.82
120.31
48.77
87.42
64.56
54.18
115.95
108.85
119.6
72.43
81.28
111.69
127.75
92.03
74.6

76.74
93.24
112.62
93.5
104.15
147.32
55.47
129.16
103.73
107.29
157.06
94.35
134.13
76.44
123.1
77.21
105.98
75.99
62
60.4
93.58
79.59
102.22
117.43
115.94
97.03
146.41
108.51
95.01
117.99
109.53
106.51
88.56
63.73
90.9

77.96
89.2
89.2
122.76
137.07
63.7
116.28
109.29
77.43
36.84
143.86
84.62
86.92
116.2
122.24
77.61
112.13
117.27
76.13
113.57
79.29
119.98
108.16
97.96
53.01
70.33
132.57
70.43
56.7
80.39
85.81
101.67
108.39
68.79
55.55

96
92.2
106.98
101.29
131.35
102.94
126.73
131.07
83.12
37.45
144.25
76.28
96.11
116.85
122.49
64.38
88.06
113.63
66.52
90.72
98.81
107.69
73.12
97.6
57.41
60.96
121.64
81.34
74.58
79.72
93.81
101.64
116.23
66.05
72.27

193

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10 UA00
code code
20287
22042
22366
22528
23500
23527
23743
23824
25228
27253
28117
28333
29440

20287
22042
22366
22528
23500
23527
23743
23824
25228
27253
28117
28333
29440

29494
30628
31087
31843
34300
34813
35164
35461
35920
37081
37243
38647
40429
40753
40780
41212
41347
42211
42346
43723
43912
44479
44992
45451

29494
30628
31087
31843
34300
34813
35164
35461
35920
37081
37243
38647
40429
40753
40780
41212
41347
42211
42346
43723
43912
44479
44992
45451

UZA name
Corpus Christi, TX
Dallas--Fort Worth--Arlington, TX
Davenport, IA--IL
Dayton, OH
Denton--Lewisville, TX
Denver--Aurora, CO
Des Moines, IA
Detroit, MI
Durham, NC
El Paso, TX--NM
Eugene, OR
Evansville, IN--KY
Fayetteville, NC
Fayetteville--Springdale--Rogers,
AR--MO
Fort Collins, CO
Fort Wayne, IN
Fresno, CA
Grand Rapids, MI
Green Bay, WI
Greensboro, NC
Greenville, SC
Gulfport, MS
Harrisburg, PA
Hartford, CT
Hickory, NC
Houston, TX
Huntington, WV--KY--OH
Huntsville, AL
Indianapolis, IN
Indio--Cathedral City, CA
Jackson, MS
Jacksonville, FL
Kalamazoo, MI
Kansas City, MO--KS
Kennewick--Pasco, WA
Killeen, TX
Kissimmee, FL

density
mix
centering street composite
factor10 factor10 factor10 factor10 index10
106.9
115.92
90.86
87.21
104.86
128.15
99.26
106.01
94.32
118.51
114.84
94.15
79.4

117.65
90.22
140.14
126.45
111.25
94.52
110
112.41
67.57
78.44
134.37
101.62
73.65

86.49
101.95
73.06
89.21
66.35
118.79
92.79
91.65
96.34
78.15
134.15
94.07
67.16

119
117.33
128.02
95.11
100.65
127.61
100.45
109.31
68.93
123.97
123.07
105.58
64.43

118.91
84.43
121.31
96.47
98.54
110.96
103.87
85.73
76.75
95.69
152.54
108.97
61.05

81.99
101.36
85.44
128.7
91.17
93.27
87.97
67.92
68.81
90.63
93.32
46.92
114.84
78.77
73.08
94.06
96.72
75.8
96.67
76.6
98.85
89.99
95.9
87.75

95.85
112.13
100.41
131.47
108.19
91.28
98.75
75.26
73.89
110.89
106.67
78.41
88.59
114.67
63.42
90.56
112.31
70.34
84.48
86.21
105.06
107.69
100.91
49.1

95.95
97.17
93.76
85.2
107.46
74.78
92.59
89.88
85.65
104.73
129.53
72.2
100.16
141.6
81.92
95.71
71.8
112
99.64
104.69
92.38
89.25
69.3
60.42

69.57
104.12
86.58
114.97
74.4
92.29
69.88
57.88
104.8
119.9
45.2
44.94
121.05
119.14
96.52
88.12
107.44
70.53
97.33
69.49
103.91
89.51
101.84
104.82

85.16
115.05
93.59
122.62
92.57
92.67
86.85
60.57
85.14
113.49
84.27
48.64
84.54
133.96
74.11
76.17
101.29
77.22
83.97
86.63
88.64
102.36
98.59
67.9

194

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10 UA00
code code
45640
46045
46828
47530
47611
47719
47854
47995
49582
49933
50392

45640
46045
46828
47530
47611
47719
47854
47962
49582
49933
50392

51445
51755
51877
52390
53200
56116
56602
57466
57628

51445
51715
51877
52390
53200
56116
56602
57466
57628

57709
57925
58006
58600
60799
60895
61273
62407
62677
63217
64135
64945
65080
65269
65863
66673

57709
57925
58006
58600
87004
60895
61273
62407
62677
63217
64135
64945
65080
65269
65863
66673

UZA name
Knoxville, TN
Lafayette, LA
Lakeland, FL
Lancaster, PA
Lancaster--Palmdale, CA
Lansing, MI
Laredo, TX
Las Vegas--Henderson, NV
Lexington-Fayette, KY
Lincoln, NE
Little Rock, AR
Los Angeles--Long Beach-Anaheim, CA
Louisville/Jefferson County, KY--IN
Lubbock, TX
McAllen, TX
Madison, WI
Memphis, TN--MS--AR
Miami, FL
Milwaukee, WI
Minneapolis--St. Paul, MN--WI
Mission Viejo--Lake Forest--San
Clemente, CA
Mobile, AL
Modesto, CA
Montgomery, AL
Murrieta--Temecula--Menifee, CA
Myrtle Beach--Socastee, SC--NC
Nashville-Davidson, TN
New Haven, CT
New Orleans, LA
New York--Newark, NY--NJ--CT
Norwich--New London, CT--RI
Ogden--Layton, UT
Oklahoma City, OK
Omaha, NE--IA
Orlando, FL
Oxnard, CA

density
mix
centering street composite
factor10 factor10 factor10 factor10 index10
71.09
81.01
86.1
90.53
111.72
98.23
123.87
147.64
126.87
118.63
86.38

53.58
94.51
46.46
127.52
111.36
68.2
131.21
63.47
122.82
127.46
82.4

155.82
92.99
106.74
132.24
54.81
134.04
81.56
121.83
121.63
97.02
97.12

67.02
88.43
106.84
79.34
82.34
86.9
166.54
107.58
98.98
141.77
115.29

79.3
92.42
87.65
116.07
90.2
102.07
151.8
102.24
136.19
143.38
95.84

212.21
97.86
107.82
88.19
118.16
93.13
143.68
112.66
112.17

144.75
82.91
127.9
63.8
121.82
63.89
108.89
116.03
98.47

102.23
92.73
75.58
85.12
182.19
101.9
109.46
164.62
119.91

138.92
90.53
130.09
99.04
99.33
86.31
134.49
112.47
108.34

143.42
79.4
126.98
71.63
152.87
70.86
112.06
132.07
97.57

127.87
77.59
127.77
85.32
103.58
57
87.51
90.04
125.35
197.5
72.73
98.34
95.88
110.48
109.38
147.55

147.54
103.81
145.02
118.52
98.68
48.17
47.43
111.59
102.93
106.8
88.4
117.46
87.23
110.3
78.11
137.14

62.55
71.28
79.06
97.11
60.2
100.23
111.18
139.46
93.92
179.1
132.22
63.46
96.6
98.07
92.26
82.42

118.63
106.28
109.17
76.74
73.39
94.2
70.03
58.5
187.3
125.06
60.7
87.66
101.44
128.31
109.72
135.08

122.47
90.23
130
100.22
77.41
71.35
60.27
100.08
138.57
142.71
93.49
87.35
87.68
116.15
84.41
146.19

195

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10 UA00
code code

UZA name

67105 67105 Palm Bay--Melbourne, FL


Palm Coast--Daytona Beach--Port
67134 22636 Orange, FL
68482 68482 Pensacola, FL--AL
68509 68509 Peoria, IL
69076 69076 Philadelphia, PA--NJ--DE--MD
69184 69184 Phoenix--Mesa, AZ
69697 69697 Pittsburgh, PA
71263 71263 Portland, ME
71317 71317 Portland, OR--WA
71479 71479 Port St. Lucie, FL
71803 71803 Poughkeepsie--Newburgh, NY--NJ
72559 72559 Provo--Orem, UT
73261 73261 Raleigh, NC
73693 73693 Reading, PA
74179 74179 Reno, NV--CA
74746 74746 Richmond, VA
75340 75340 Riverside--San Bernardino, CA
75421 75421 Roanoke, VA
75664 75664 Rochester, NY
75718 75718 Rockford, IL
Round Lake Beach--McHenry-76474 76474 Grayslake, IL--WI
77068 77068 Sacramento, CA
77770 77770 St. Louis, MO--IL
78229 78229 Salem, OR
Salt Lake City--West Valley City,
78499 78499 UT
78580 78580 San Antonio, TX
78904 78904 San Francisco--Oakland, CA
79039 79039 San Jose, CA
79309 79309 Santa Clarita, CA
79606 79606 Sarasota--Bradenton, FL
79768 79768 Savannah, GA
80227 80227 Scranton, PA
80389 80389 Seattle, WA
81739 81739 Shreveport, LA
83116 83116 South Bend, IN--MI

density
mix
centering street composite
factor10 factor10 factor10 factor10 index10
88.7

78.17

60.31

88.15

68.9

84.24
73.9
85.82
127.16
119.2
93.7
88.48
127.64
78.65
75.26
113.81
90.27
127.71
101.13
94.09
119.16
83.87
103
86.89

82.84
71.76
104.21
124.32
79.12
119.21
123.43
129.26
57.1
112.65
130.08
77.3
150.87
59.47
83.03
112.94
108.45
101.93
103.05

66.42
74.11
125.17
131.46
99.99
125.55
148.13
107.58
76.99
121.96
77.33
112.47
124.45
123.12
111.23
81.81
81.82
103.14
96.03

108.04
111.04
113.45
105.73
106.59
117.1
85.12
135.17
103.04
43.25
100.13
54.9
147.46
94.24
109.31
82.29
110.57
61.44
119.57

82.45
78.47
120.49
109.05
80.27
109.25
130.27
126.14
71.26
84.82
110.6
68.86
169.32
95.67
93.1
84.2
105.72
85.12
109.98

80.35
125.63
100.04
115.59

83.85
104.85
114.26
125.71

79.57
109.31
104.27
112.5

90.59
107.52
110.7
103.81

81.75
106.02
96.18
133.51

130.73
113.62
205.69
181.13
120.29
84.4
82.3
95.52
118.83
80.89
79.42

116.64
85.4
129.92
136.26
129.69
94.94
96.52
145.27
89.41
74.13
91.31

84.13
92.02
164.34
86.67
81.81
93.16
110.11
102.54
142.43
70.96
105.6

99.9
104.86
153.38
127.03
93.51
115.72
111.85
133.3
110.09
102.49
129.25

105.81
85.2
180.94
139.98
119.53
93.95
109.61
135.5
104.65
79.07
110.77

196

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10 UA00
code code
83764
83953
85087
86302
86464
86599
87868
88462
88732
88948
90541
90946
92242
95077
95833
96670
96697
97750
97831

83764
83953
85087
86302
86464
86599
87868
88462
88732
88948
90541
90946
92242
95077
95833
96670
96697
97750
97831

UZA name
Spokane, WA
Springfield, MO
Stockton, CA
Syracuse, NY
Tallahassee, FL
Tampa--St. Petersburg, FL
Toledo, OH--MI
Trenton, NJ
Tucson, AZ
Tulsa, OK
Victorville--Hesperia, CA
Visalia, CA
Washington, DC--VA--MD
Wichita, KS
Wilmington, NC
Winston-Salem, NC
Winter Haven, FL
York, PA
Youngstown, OH--PA

density
mix
centering street composite
factor10 factor10 factor10 factor10 index10
97.43
86.57
126.41
100.81
97.65
103.05
94.96
123.71
100
90.85
82.38
118.08
142.28
96.94
81.25
66.31
67.51
91.8
76.37

109.36
110.76
131.34
110.01
71.33
92.69
127.77
121.54
70.98
97.81
67.79
126.58
96.36
92.64
102.01
68.97
52.97
129.86
134.31

103.98
68.49
98.73
133.54
144.71
93.15
90.58
106.59
90.13
96.64
57.01
92.94
136.7
94.44
89.16
88.15
77.78
121.89
77.47

141.33
115.03
117.22
85.98
84.14
122.73
100.46
108.84
94.08
99.33
61.88
127.07
104.92
110.31
96.91
54.29
110.62
103.78
90.73

125.49
101.06
134.67
116.05
109.39
87.63
106.97
132.08
77.54
92.29
54.15
137.22
107.69
100.02
99.02
55.56
75.86
129.62
96.18

197

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

Appendix F. Urbanized Areas Compactness Indices 2000


UA10
code

UA00
code

199
766
970
1171
1495
2602
2683
2764
3358
3817
3898
4222
4384
4681
4843
5680
7786
8785
8974
10972
11350
13375
13510
15508
15670
15832
16264
16885
17668
18964
19099
19234
19504
19558
19755

199
766
928
1171
1495
2602
2683
2764
3358
3817
3898
4222
4384
4681
4843
5680
7786
8785
8974
10972
11350
13375
13510
15508
15670
15832
16264
16885
17668
18964
19099
19234
19504
19558
87328

UZA name
Aberdeen--Bel Air South--Bel Air
North, MD
Akron, OH
Albany--Schenectady, NY
Albuquerque, NM
Allentown, PA--NJ
Ann Arbor, MI
Antioch, CA
Appleton, WI
Asheville, NC
Atlanta, GA
Atlantic City, NJ
Augusta-Richmond County, GA--SC
Austin, TX
Bakersfield, CA
Baltimore, MD
Baton Rouge, LA
Birmingham, AL
Boise City, ID
Bonita Springs, FL
Brownsville, TX
Buffalo, NY
Canton, OH
Cape Coral, FL
Charleston--North Charleston, SC
Charlotte, NC--SC
Chattanooga, TN--GA
Chicago, IL--IN
Cincinnati, OH--KY--IN
Cleveland, OH
Columbia, SC
Columbus, GA--AL
Columbus, OH
Concord, CA
Concord, NC
Conroe--The Woodlands, TX

density
factor00

mix
factor00

centering
factor00

street
factor00

composite
index00

85.05
83.56
97.62
115.4
98.6
109.06
112.73
109.95
56.51
88.54
93.52
74.42
121.77
121.27
129.5
83.46
81.96
104.83
76.78
108.24
114.62
83.07
81.59
87.45
85.5
65.83
148.83
100.48
111.72
82.21
85.1
114.53
123.25
55.14
94.53

130.29
124.2
126.41
83.47
154.6
89.82
162.91
136.48
112.01
90.28
86.06
87.91
113.65
134.54
120.15
72.66
94.9
131.24
77.85
107.19
131.15
135.07
79.16
95.25
99.95
55.21
120.65
116.34
122.33
84.52
69
124.36
126.99
99.79
92.36

74.74
91.55
111.23
88.92
98.74
125.09
50.78
103.75
102.96
106.29
158.52
89.02
133.13
78.88
128.93
85.07
100.51
71.93
61.38
65.79
103.32
79.23
95.48
127.26
108.27
92.3
131.04
116.18
112.13
125.63
117.06
102.14
82.93
79.86
86.68

47.27
81.39
76.74
111.01
137.75
58.53
108.2
115.15
63.14
19.9
140.57
70.65
82.98
118.03
96.29
64.16
99.74
104.89
46.22
106.8
75.28
123
90.72
96.79
35.77
53.9
122.64
61.77
56.21
77.91
77.88
96.89
82.12
68.46
54.45

85.29
89.78
102.59
96.53
133.48
98.12
122
135.96
81.59
39.5
139.23
71.97
113.25
122.72
113.51
61.39
86.68
110.79
52.49
106.23
101.45
114.04
82.2
103.51
66.06
49.7
117.76
84.83
86.01
88.92
86.41
105.27
104.03
76.14
88.85

198

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10
code
20287
22042
22366
22528
23500
23527
23743
23824
25228
27253
28117
28333
29440

UA00
code
20287
22042
22366
22528
23500
23527
23743
23824
25228
27253
28117
28333
29440

29494
30628
31087
31843
34300
34813
35164
35461
35920
37081
37243
38647
40429
40753
40780
41212
41347
42211
42346
43723
43912
44479
44992
45451

29494
30628
31087
31843
34300
34813
35164
35461
35920
37081
37243
38647
40429
40753
40780
41212
41347
42211
42346
43723
43912
44479
44992
45451

UZA name
Corpus Christi, TX
Dallas--Fort Worth--Arlington, TX
Davenport, IA--IL
Dayton, OH
Denton--Lewisville, TX
Denver--Aurora, CO
Des Moines, IA
Detroit, MI
Durham, NC
El Paso, TX--NM
Eugene, OR
Evansville, IN--KY
Fayetteville, NC
Fayetteville--Springdale--Rogers,
AR--MO
Fort Collins, CO
Fort Wayne, IN
Fresno, CA
Grand Rapids, MI
Green Bay, WI
Greensboro, NC
Greenville, SC
Gulfport, MS
Harrisburg, PA
Hartford, CT
Hickory, NC
Houston, TX
Huntington, WV--KY--OH
Huntsville, AL
Indianapolis, IN
Indio--Cathedral City, CA
Jackson, MS
Jacksonville, FL
Kalamazoo, MI
Kansas City, MO--KS
Kennewick--Pasco, WA
Killeen, TX
Kissimmee, FL

density
factor00
105.97
117.14
96.81
89.73
97.48
135.76
111.38
113.29
91.24
118.39
121.5
96.61
78.97

mix
factor00
125.48
102.37
152.84
123.77
116.34
108.29
121.97
112.29
77.48
84.76
141.47
124.11
98.97

centering
factor00
84.91
95.57
73.58
101.57
67.11
116.57
97.25
106.16
106.3
79.51
130.73
99.57
62.63

street
factor00
104.4
98.9
114.3
88.21
75.13
129.44
101.92
100.28
58.06
111.2
114.89
100.57
56.65

composite
index00
113.13
81.46
120.78
96.56
86.9
120.11
115.33
89.38
78.11
93.22
151.42
116.75
64.13

86.95
105.79
91.02
131.07
94.4
99.89
92.19
75.59
77.88
90.93
93.14
49.14
115.73
82.45
73.26
97.94
99.18
86.8
100.97
82.13
101.66
84.2
99.15
85.68

121.68
111.9
114.76
145.64
120
115.98
125.19
101.55
94.44
118.89
114.07
81.34
98.06
124.09
78.59
101.45
127.03
87.99
92.98
111.47
115.2
126.72
110.93
66.13

115.49
89.91
89.81
93.2
102.39
71.74
96.01
89.54
94.46
103.03
126.66
75.33
94.99
152.46
90.56
112.38
71.49
112.18
92.56
99.11
90.55
99.02
77.07
58.9

60.4
95.01
86.8
120.39
61.32
91.68
67.38
59.56
91.98
108.09
32.36
42.67
97.43
109.9
55.71
86.6
107.22
63.13
96.26
64.66
99.55
89.04
98.45
106.74

104.38
109.69
97.94
134.15
89.35
101.39
98.07
74.82
92.1
110.9
79.58
48.76
79.22
138
67.33
88.58
108.33
84.77
85.83
92.53
90.81
111.45
105.12
76.56

199

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10
code
45640
46045
46828
47530
47611
47719
47854
47995
49582
49933
50392

UA00
code
45640
46045
46828
47530
47611
47719
47854
47962
49582
49933
50392

51445
51755
51877
52390
53200
56116
56602
57466
57628

51445
51715
51877
52390
53200
56116
56602
57466
57628

57709
57925
58006
58600
60799
60895
61273
62407
62677
63217
64135
64945
65080
65269
65863
66673

57709
57925
58006
58600
87004
60895
61273
62407
62677
63217
64135
64945
65080
65269
65863
66673

UZA name
Knoxville, TN
Lafayette, LA
Lakeland, FL
Lancaster, PA
Lancaster--Palmdale, CA
Lansing, MI
Laredo, TX
Las Vegas--Henderson, NV
Lexington-Fayette, KY
Lincoln, NE
Little Rock, AR
Los Angeles--Long Beach-Anaheim, CA
Louisville/Jefferson County, KY--IN
Lubbock, TX
McAllen, TX
Madison, WI
Memphis, TN--MS--AR
Miami, FL
Milwaukee, WI
Minneapolis--St. Paul, MN--WI
Mission Viejo--Lake Forest--San
Clemente, CA
Mobile, AL
Modesto, CA
Montgomery, AL
Murrieta--Temecula--Menifee, CA
Myrtle Beach--Socastee, SC--NC
Nashville-Davidson, TN
New Haven, CT
New Orleans, LA
New York--Newark, NY--NJ--CT
Norwich--New London, CT--RI
Ogden--Layton, UT
Oklahoma City, OK
Omaha, NE--IA
Orlando, FL
Oxnard, CA

density
factor00
68.61
78.58
76.03
96.79
109.53
101.03
134.65
155.61
132.62
118.03
93

mix
factor00
70.74
104.29
86.84
132.28
123.6
96.55
148.02
69.07
121.28
133.12
95.64

centering
factor00
130.5
100.28
114.04
126.42
56.21
105.33
86.2
127.05
125.89
97.15
93.19

street
factor00
57.65
74.01
85.81
65.11
73.56
80.06
189.55
105.4
78.65
135.15
103.44

composite
index00
71.74
93.04
94.03
112.02
91.06
97.71
174.12
111.38
130.01
141.19
96.64

212.14
101.73
112.3
76.58
122.06
101.44
142.94
118.7
113.41

131.99
93.98
132.03
79.91
126.86
72.29
107.92
128.61
90.31

105.37
90.76
74.66
84.37
158.37
100.26
93.37
125.45
118.69

127.52
81.59
124.18
90.18
101.3
71.3
131.01
106.94
95.92

135.59
80.16
126.23
70.76
147.2
69.71
104.22
120.5
89.25

129.04
81.05
127.86
95.52
95.58
66.75
89.26
88.44
161.24
197.18
70.09
98.24
100.17
113.56
106.07
151.09

140.04
108.53
147.59
130.65
105.7
80.03
67.83
119.87
106.84
115.6
97.16
124.92
107.78
122.98
87.13
138.91

64.39
73.69
97.76
103.01
108.31
108.57
106.22
132.78
95.97
170.57
133.84
64.85
93.5
98.8
94.04
76.66

92.13
70.94
105.06
80.25
72.89
105.93
46.1
47.06
181.06
120.19
48.33
76.93
94.11
124.72
96.83
115.75

108.57
77.49
135.64
112.96
100.43
98.74
58.11
93.54
149.64
141.75
90.23
86.84
92.8
120.65
83.39
136.4

200

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10 UA00
UZA name
code
code
67105 67105 Palm Bay--Melbourne, FL
Palm Coast--Daytona Beach--Port
67134 22636 Orange, FL
68482 68482 Pensacola, FL--AL
68509 68509 Peoria, IL
69076 69076 Philadelphia, PA--NJ--DE--MD
69184 69184 Phoenix--Mesa, AZ
69697 69697 Pittsburgh, PA
71263 71263 Portland, ME
71317 71317 Portland, OR--WA
71479 71479 Port St. Lucie, FL
71803 71803 Poughkeepsie--Newburgh, NY--NJ
72559 72559 Provo--Orem, UT
73261 73261 Raleigh, NC
73693 73693 Reading, PA
74179 74179 Reno, NV--CA
74746 74746 Richmond, VA
75340 75340 Riverside--San Bernardino, CA
75421 75421 Roanoke, VA
75664 75664 Rochester, NY
75718 75718 Rockford, IL
Round Lake Beach--McHenry-76474 76474 Grayslake, IL--WI
77068 77068 Sacramento, CA
77770 77770 St. Louis, MO--IL
78229 78229 Salem, OR
Salt Lake City--West Valley City,
78499 78499 UT
78580 78580 San Antonio, TX
78904 78904 San Francisco--Oakland, CA
79039 79039 San Jose, CA
79309 79309 Santa Clarita, CA
79606 79606 Sarasota--Bradenton, FL
79768 79768 Savannah, GA
80227 80227 Scranton, PA
80389 80389 Seattle, WA
81739 81739 Shreveport, LA
83116 83116 South Bend, IN--MI

density
factor00
76.29

mix
factor00
75.93

centering
factor00
62.16

street
factor00
77.64

composite
index00
58.18

95.16
73.65
91.09
131.05
130.95
96.98
89.38
124.3
71.15
78.28
127.45
85.53
119.44
113.14
92.38
116.92
86.36
108.58
89.96

98.84
74.45
126.26
121.96
100.13
127.23
134.9
134.07
72.97
112.78
156.85
110.35
157.15
72
90.82
113.99
111.67
97.82
114.18

68.39
72.66
99.4
126.98
97.92
118.72
155.27
102.05
75.25
115.34
75.63
75.3
126.12
126.28
104.41
91.49
76.59
103.61
95.46

109.08
94.85
105.88
101.25
103.57
106.44
66.41
128.07
94.03
22.28
94.04
52.22
118.53
92.04
88.19
83.23
86.48
50.71
103.81

94.85
69.79
115.78
106.14
92.82
105.11
128.14
121.95
70.87
74.14
126.13
67.3
155.74
105.82
83.85
89.17
93.5
79.59
107.07

76.99
124.59
103.99
115.57

115.86
120.43
123.85
137.9

79.25
124.87
101.53
112.49

75.95
98.6
96.97
99.24

86.73
115.3
93.99
134.82

133.2
117.87
219.66
178.91
118.24
90.69
99.99
101.5
113.58
86.6
83

130.41
96.14
128.39
134.54
137.68
100.26
89.55
155.32
93.37
82.72
111.98

88.53
99.08
162.41
82.37
79.83
114.14
108.58
100.64
135.64
74.56
104.07

99.2
106.77
149.84
116.63
67.1
110.88
123.77
129.53
97.4
93.39
99.66

113.34
96.28
184.06
131.9
111.89
104.18
117.33
136.8
96.57
80.28
104.93

201

MEASURING URBAN SPRAWL AND VALIDATING SPRAWL MEASURES

UA10
code
83764
83953
85087
86302
86464
86599
87868
88462
88732
88948
90541
90946
92242
95077
95833
96670
96697
97750

UA00
code
83764
83953
85087
86302
86464
86599
87868
88462
88732
88948
90541
90946
92242
95077
95833
96670
96697
97750

UZA name
Spokane, WA
Springfield, MO
Stockton, CA
Syracuse, NY
Tallahassee, FL
Tampa--St. Petersburg, FL
Toledo, OH--MI
Trenton, NJ
Tucson, AZ
Tulsa, OK
Victorville--Hesperia, CA
Visalia, CA
Washington, DC--VA--MD
Wichita, KS
Wilmington, NC
Winston-Salem, NC
Winter Haven, FL
York, PA

density
factor00
99.69
89.76
134.42
104.93
93.87
106.59
102.27
130.68
103.97
96.26
74.79
116.84
133.79
101.36
74.31
66.67
72.57
84.59

mix
factor00
110.75
138.09
145.18
115.92
68.08
94.04
129.37
138.84
93.06
101.83
84.24
142.48
104.48
107.4
109.17
68.56
72.7
139.34

centering
factor00
102.36
66.86
104.41
130.6
112.12
89.98
93.5
103.24
82.2
93.07
56.75
107.53
112.04
97.06
91.78
93.67
75.82
129.88

street
factor00
140.36
87.37
124.09
76.08
61.01
122.04
92.92
106.32
91.29
96.58
51.04
108.93
85.55
112.11
77.06
44.02
100.19
93.59

composite
index00
124.73
101
147.55
112.42
82.66
88.57
106.17
137.57
83.13
92.84
55.43
145.05
90.84
108.03
92.18
53.49
80.21
128.45

202