Author: Brenda Gunderson, Ph.D.
, 2013
License: Unless otherwise noted, this material is made available under the terms of the Creative
Commons AttributionNonCommercialShare Alike 3.0 Unported License:
http://creativecommons.org/licenses/byncsa/3.0/
The University of Michigan Open.Michigan initiative has reviewed this material in accordance with U.S.
Copyright Law and have tried to maximize your ability to use, share, and adapt it. The attribution key
provides information about how you may share and adapt this material.
Copyright holders of content included in this material should contact open.michigan@umich.edu with any
questions, corrections, or clarification regarding the use of content.
For more information about how to attribute these materials visit:
http://open.umich.edu/education/about/termsofuse. Some materials are used with permission from the
copyright holders. You may need to obtain new permission to use those materials for other uses. This
includes all content from:
Mind on Statistics
Utts/Heckard, 4th Edition, Cengage L, 2012
Text Only: ISBN 9781285135984
Bundled version: ISBN 9780538733489
SPSS and its associated programs are trademarks of SPSS Inc. for its proprietary
computer software. Other product names mentioned in this resource are used for identification purposes
only and may be trademarks of their respective companies.
Attribution Key
For more information see: http:://open.umich.edu/wiki/AttributionPolicy
Content the copyright holder, author, or law permits you to use, share and adapt:
Creative Commons AttributionNonCommercialShare Alike License
Public Domain Self Dedicated: Works that a copyright holder
has dedicated to the public domain.
Make Your Own Assessment
Content Open.Michigan believes can be used, shared, and adapted because it is ineligible for copyright.
Public Domain Ineligible. Works that are ineligible for
copyright protection in the U.S. (17 USC 102(b)) *laws in your
jurisdiction may differ.
Content Open.Michigan has used under a Fair Use determination
Fair Use: Use of works that is determined to be Fair consistent
with the U.S. Copyright Act (17 USC 107) *laws in your jurisdiction may differ.
Our determination DOES NOT mean that all uses of this thirdparty content are Fair Uses and we DO
NOT guarantee that your use of the content is Fair. To use this content you should conduct your own
independent analysis to determine whether or not your use will be Fair.
Statistics 250
1
Lab Workbook
Fall 2013  Winter
2014
Weekly Labs, InLab Projects,
Supplements,
and Old Exams for Review
Used in all lab sections of Stat
250
Dr. Brenda Gunderson
Department of Statistics
University of Michigan
Table of Contents
2
Material
Page
Note to Students and Supplements
Supplement 1: SPSS Commands
Summary
Supplement 2: Notation Sheet
Supplement 3: Name That Scenario
Supplement 4: Editing Charts in SPSS
Supplement 5: Notes about SPSS t
Procedures
Supplement 6: Interpretation Examples
Supplement 7: Summary of the Main tTests
Supplement 8: Regression Output in
SPSS
Lab 1:
Scavenger Hunt; Mean and
Median
Lab 2:
Describing Data with Graphs
and Numbers
Lab 3:
Time Plots and QQ Plots
Lab 4:
Probability and Random
Variables
Lab 5:
Confidence Intervals for
Population Proportions
Lab 6:
Hypothesis Testing for
Population Proportions
Lab 7:
Sampling Distributions and the
CLT
Lab 8:
OneSample tTest Procedures
Lab 9:
Paired tTest Procedures
Lab 10:
Two Independent Samples tTest Procedures
Lab 11:
OneWay Analysis of Variance
(ANOVA)
Lab 12:
Simple Linear Regression
Lab 13:
ChiSquare Tests
1
2
4
6
8
10
12
14
17
22
29
36
46
56
62
70
80
89
96
105
115
127
Old Exams for Review
Exam 1 Questions
Exam 2 Questions
Final Exam Questions
135
169
203
Note to Students
Welcome to Statistics 250 at the University of Michigan! This lab
workbook is designed for you to use in lab and as extra
preparation for exams. In the workbook, you will find the following
materials:
Supplemental Material great summaries for reference
throughout the term:
1. SPSS Commands Reference
2. Notation Sheet
3. Name That Scenario
4. Editing Charts in SPSS
5. Important Notes for Hypothesis Testing
6. Interpretation Examples
7. Summary of Ttests and Name That Scenario Practice for
Means
8. Regression Output in SPSS
Weekly Labs (numbered 1 to 13) each lab contains the follow
parts:
o Lab Background objective and brief overview material,
which is good to take a couple minutes to read before you come
to lab each week.
o WarmUp Activity quick questions for you to do before the
InLab Project, usually a quick review of concepts you have
seen in lecture.
o ILP (InLab Project) one or more activities that you will work
on in lab, in groups. A copy of the ILP will be provided to each
group for turning in at the end of the lab period.
o CoolDown Activity questions for you to do after the ILP for
further reflection and application of the concepts covered in the
ILP.
o Example Exam Questions old exam questions on the lab
topic(s) for additional practice.
Old Exams complete sets of actual old exam for studying. Be
sure to refer to CTools to see if any problems on these old exams
are not relevant for your particular upcoming exam (due to
differences in the semester schedule). This information, in addition
to solutions, will be posted on CTools in the Review Info folder
under the Resources tab closer to each exam date.
The Labs are designed to be interactive and to provide you with a
complete example for each concept.
Completing the
corresponding PreLab assignment (a link to video instructions for
PreLabs will be on CTools and the Stat 250 YouTube channel) and
reading the upcoming lab background overview before lab each
week is a good way to prepare for the various lab activities.
Good luck in Statistics 250!
 The Stat 250 Instructors and GSIs
Supplement 1: SPSS Commands
Summary
By Lab For Quick Reference
Lab 2 Descriptive Statistics
Open a data file after having SPSS already open: File> Open>
Data
To produce a Histogram: Graphs> Legacy Dialogs> Histogram
To produce a Boxplot for a single variable with no groups:
Graphs> Legacy Dialogs> Boxplot> Simple> Summaries for
separate variables
To use Data Label Mode: From inside the Chart Editor,
Elements> Data Label Mode
To generate Descriptive Statistics I: Analyze> Descriptive
Statistics> Descriptives
To generate Descriptive Statistics II: Analyze> Descriptive
Statistics> Frequencies
To produce a Bar Chart: Graphs> Legacy Dialogs> Bar> Simple>
Summaries of
separate variables
To Split (or unsplit) the data (get charts and statistics by
group): Data> Split File
To produce SidebySide Boxplots: Graphs> Legacy Dialogs>
Boxplot> Simple> Summaries for groups of cases
Lab 3 Time Plots and QQ Plots
To produce a Sequence (Time) Plot: Analyze> Forecasting>
Sequence Charts
To produce a QQ Plot: Analyze> Descriptive Statistics> QQ Plots
Lab 8 OneSample Confidence Intervals for a
Population Mean
To produce a Confidence Interval for a population mean
(method I):
Analyze> Descriptive Statistics> Explore> Statistics option
To produce a Confidence Interval for a population mean
(method II):
Analyze> Compare Means> OneSample T Test
Lab 8 OneSample t Procedures for a Population
Mean
To perform a OneSample T Test for a population mean:
Analyze> Compare Means> OneSample T Test
Lab 9 Paired t Procedures
To calculate a confidence interval for D: Analyze> Compare
Means> PairedSamples T Test
To perform a Paired T Test: Analyze> Compare Means> PairedSamples T Test
To compute Differences: Transform> Compute
Lab 10 Independent Samples t Procedures
To construct a confidence interval for 1  2:
Analyze> Compare Means> IndependentSamples T Test
To perform a TwoSamples T Test: Analyze> Compare Means>
IndependentSamples T Test
Lab 11 Oneway Analysis of Variance (ANOVA)
To perform an ANOVA: Analyze> Compare Means> OneWay
ANOVA
Lab 12 Simple Linear Regression
To produce a Scatterplot: Graphs> Legacy Dialogs> Scatter/Dot
To perform a Linear Regression: Analyze> Regression> Linear
To produce a Residual plot: Graphs> > Legacy Dialogs>
Scatter/Dot
Lab 13 ChiSquare Tests
To weight cases by Counts: Data> Weight Cases
To perform a Goodness of Fit Test: Analyze> Nonparametric
Tests> ChiSquare
To perform a Test of Independence: Analyze> Descriptive
Statistics> Crosstabs
To perform a Test of Homogeneity: Analyze> Descriptive
Statistics> Crosstabs
Supplement 2: Notation Sheet
The table below defines important notations, including that used
by SPSS, which you will come across in the course. This is not an
exhaustive list, but it is a fairly comprehensive overview of the
strange letters used in the course.
Note: Blank cells mean there is no corresponding notation.
Name
Mean
Proportion
Standard
deviation
Variance
Sample size
Multipliers
Population
Notation
Sample
Notation
Summary Measures
(read as
x (xbar)
mu)
p
p (phat)
(sigma)
s2
n
Confidence Intervals
z* (zstar)
t* (tstar)
Margin of
error
Test
statistics
Note: t, F, and
2 statistics
Notation
used in
SPSS
Mean
Std. Deviation
Variance
N
m, m.e.
Hypothesis Testing
z
t
F
have degrees of
freedom
(abbreviated df)
associated with
them. Look for
these on your
Formula Card.
2 (chisquare)
Significance
level
pvalue
t
F
Chisquare
(alpha)
pvalue
Sig.
Notation
used in
SPSS
Analysis of Variance (abbreviated ANOVA)
Between
Sum of
groups (look
squares for
SSG
in the column
groups
labeled Sum
of Squares)
Within groups
Sum of
(look in the
squares for
SSE
column
error
labeled Sum
of Squares)
Between
Mean
groups (look
square for
MSG
in the column
groups
labeled Mean
Square)
Within groups
(look in the
Mean
MSE
column
square error
labeled Mean
Square)
Regression
Response
(given by
(dependent)
y
y
name of yvariable
variable)
Predicted
E(y)
y (yhat)
(estimated)
(expected
response
value of y)
Explanatory
(given by
(independen
x
x
name of xt) variable
variable)
B (look in the
o (betanot)
yintercept
bo
row labeled
(Constant))
B (look in the
row labeled
Slope
b1
with the
1 (betaone)
name of the
xvariable)
Coefficient
of
r
R, Beta
correlation
Name
Population
Notation
Sample
Notation
Coefficient
of
determinati
on
Error terms
vs Residuals
(error
terms)
r2
R Square
e (residuals)
Unstandardiz
ed residuals
Supplement 3: Name That Scenario
The first thing to do in any research inference problem is
determine what type of inference problem it is. This will help in
deciding what procedure/formulas are appropriate to use. The
following questions can help you determine the data scenario you
are working with.
Please note, when answering, How many variables are there? do
not count the variable which defines the populations (if there is
more than one population).
How many populations are there?
One
Two
How many variables are there?
One
Two
What type of variable(s)?
Categorical
More than two
Quantitative
Then use the following table to determine which type of
inference would be appropriate for this scenario.
Note the corresponding parameter is in parentheses, where
appropriate.
Number of Populations
Number of
Variables and
Type
Categori
cal
One
Two
1sample
inference
for
population
proportion
(p)
(Labs 5 an
6)
2
independe
nt samples
inference
for the
difference
between 2
population
proportions
Chisquare:
Goodness
of Fit (Lab
13)
On
e
Tw
o
Categori
cal
(relationshi
p)
Quantita
tive
(relationshi
Chisquare:
Homogenei
ty
(Lab 13)
(p1 p2)
Chisquare:
Homogenei
ty
(Lab 13)
2
independe
nt samples
inference
for the
difference
between
2
population
means
(1  2)
(Lab 10)
1sample
inference
for
population
mean ()
(Lab 8)
Quantita
tive
More Than
Two
Paired
samples
inference
for a
population
mean
difference
(D)
(Lab 9)
Chisquare:
Independen
ce
(Lab 13)
Regression
(1)
(Lab 12)
ANOVA
(i where
there is
one i for
each
population)
(Lab 11)
p)
Supplement 4: Editing Charts in
SPSS
Once we have a histogram (or any chart) made, we may wish to
edit the chart (perhaps to change the color of the bars or change
the number of class intervals). To do this, double click on the chart
displayed in the output viewer window. This will open the chart in
the SPSS Chart Editor window.
Suppose we want to change the color of the histogram bars from
tan to a light green. To do this, double click on one of the tan bars
and click on the Fill & Border tab in the properties window.
Change the fill color to light green (click on the light green box
color) and click on apply and then close.
To change some aspect of the xaxis, such as the scaling,
double click on the xaxis and its corresponding properties box will
appear with many tab options. The scale tab can be used to
change the endpoints and major increments of the xaxis value
labels. You could adjust the minimum to 10,000 and the maximum
to 150,000. Leaving the Auto box checked for the Major Increment
option will let SPSS create the increment size. The Histogram
Options tab would allow you to add a normal curve to a
histogram, as well as change the starting position and size of the
bins (classes or intervals) in the histogram. In general, SPSS uses
algorithms to produce a nice display of the data. These options
are helpful if you have multiple plots that you would like to display
using the same xaxis values so comparisons can be more easily
made. You can also change the gray background color to white
under the Properties window with the background highlighted.
The help button in the lower right corner of the Properties box can
be selected to provide more details about any of the various
options for that tab. Once you have finished customizing your
chart, you can close out the chart editor.
There are alternate ways to get to the Properties box in order to
customize your chart. Once you have doubleclicked on your chart
to open the Chart Editor, click once on the part of the chart that
you wish to customize (so that is it highlighted). Then, click on the
Show Properties Window tab (it looks like a paint palette) in
your menu. This will open the Properties box. Another alternative
is to simply select Properties under the Edit menu in the Chart
Editor. Also, note that if you do not close the Properties box, and
you continue highlighting different parts of your chart, the
Properties box updates so that you can customize those parts as
you go.
10
For boxplots, if there are any points denoted as outliers, you can
identify them by looking at their case label number in the default
output. The Chart Editor provides a special mode for identifying
individual cases whose data labels you want to display. This is the
data label mode, and when you are in data label mode, you can't
change anything else in the chart. From the menus, choose
Elements> Data Label Mode. The cursor changes shape to
indicate that you are in data label mode. Click the data element
for which you want to display the case label.
If there are
overlapping data elements in the spot that you click, the Chart
Editor displays the Select Data Element to Label dialog box.
This dialog allows you to select the specific data element or
elements for which you want to display data labels. The Chart
Editor displays the data label in a default position related to the
data element. When you are finished choosing data elements,
from the menus choose Elements> Data Label Mode again, and
the cursor changes back to the arrow to indicate that you are no
longer in Data Label Mode.
The Options menu lets you customize your chart further. You may
add a title or text box from this menu. Text boxes can appear
anywhere in a chart.
From the Chart Editor menus, select
Options> Text Box or Options> Title depending on which you
want.
For titles, the Chart Editor creates the title box and
automatically positions it in the top center of the chart. Type the
text and press enter when you are finished typing. To enter line
breaks, press Shift+Enter. If necessary, use the Text tab to format
the text. For text boxes, you can drag and drop to reposition
them. You may need to resize the graph so the text box will not
cover up part of the graph. You can also copy the plots onto
MSWord or another text editor and then type in your name and
title within the document.
Saving Output Boxes and Graphs
Images and other output from SPSS can often be copied and then
pasted into a document by selecting the desired output, rightclicking, and choosing Copy.
1. To save an output box, such as a table of descriptive
statistics, first have the location where you would like to store
the output open. Then rightclick on the output table and
select Copy. The table can then be pasted into a document or
textfield (such as those in your PreLab assignments or
homework). If you are pasting into a Word document and if the
output does not appear to format correctly, it may be a good
idea to choose Paste Special and paste as an image. Your
11
stats 250 GSIs prefer not to receive Word documents.
2. To save a graph, such as a boxplot or histogram, you will
want to Export the graph as an image. Select the graph you
wish to export then select Export from the File menu. At the
top click the Selection button, select None (graphics only)
under Document, and then choose the file type at the bottom.
For uploading to www.lecturebook.com, the extension .jpg or
.png is required. Before completing the export command, be
sure to give your file an informative name and note the location
where the file will be saved.
3. To save an entire SPSS session, you can export output (all
output in the viewer, charts only or text only) in many possible
formats (html, jpeg, bitmap, etc.). You first make the Viewer
the active window and select the Export command is under
the File menu.
You can also print the contents of your output viewer window (all
output, text as well as charts) or any selected portion. Click on
File from the menu bar and then choose Print. You could also just
save your output file within SPSS by selecting the Save as option
and giving a name for your file and click OK.
Supplement 5: Notes about SPSS t
Procedures
1. The reported pvalue under the column heading of Sig. (2tailed)
is
for
a
2sided test. For a onesided test, you first divide the reported
2tailed pvalue in half (p/2). If the tstatistic is positive and the
alternative hypothesis was uppertailed (>), then p/2 is the pvalue.
If the tstatistic is negative and the alternative
hypothesis was lowertailed (<), then p/2 is the pvalue.
However, if the tstatistic is positive and the alternative
hypothesis was lowertailed (<) or if the tstatistic is negative
and the alternative hypothesis was uppertailed (>), then the pvalue is 1  p/2.
Alternative is >, then pvalue is sig/2
positive
Alternative is <, then pvalue is 1 (sig/2)
tstatistic
negative
Alternative is >, then pvalue is 1 (sig/2)
Alternative is <, then pvalue is sig/2
12
For example, consider a 2sided test with an observed t
statistic value of 0.948 and a pvalue of 0.364. This 0.364 is
actually the sum of two equal areas: one being the area to the
right of 0.948 and the other being the area to the left of
0.948 under the tdistribution with df = 11 curve (see Figure 1
below).
If the alternative hypothesis had been uppertailed (>), then
the pvalue would be only the area to the right (in the direction
of
extreme)
of
the
observed
tstatistic of 0.948, which is half of the twosided pvalue, or
0.182. But if the alternative hypothesis had been lowertailed
(<), the pvalue would be the area to the left (in the direction of
extreme) of the observed tstatistic of 0.948. So the pvalue
would be 1 (0.364/2) = 0.818.
2. Sometimes SPSS will display a pvalue of 0.000. Clearly, the
probability is not exactly zero. Rather, it is zero to 3 significant
digits. Thus, it is correct to say that the pvalue is less than
0.0005, since anything greater when rounded would have
resulted in Sig. 2tailed pvalue of 0.001 or more.
3. One can use a onesample t (1  )% confidence interval to test
a twosided hypothesis at the % level by checking whether
the hypothesized or null test value is contained in the interval.
If it is, we cannot reject H 0, but if not, we would reject H 0.
Recall that if you want to produce a confidence interval for the
population mean, , you must specify the test value to be 0 in
the onesample ttest dialog box. For example, a test value of
3.0 that would result in a 95% confidence interval of the
difference given as (0.265, 0.665) is actually a confidence
interval for 3.0. See Lab 5 for more details.
13
4. Supplement 6 provides an overview of the assumptions for all
three ttests that are presented in Labs 8, 9, and 10.
14
Supplement 6: Interpretation
Examples
In 1980, Bausch and Lomb Corporation developed a new type of
extendedlife contact lens made of silicone, which it claimed had a
useful
life
of
more
than
4 years. During the research and development period, a random
sample of 6 contact wearers was asked to wear the new contact
lenses and record how long they lasted. The average useful life of
the six pairs of lenses was 4.6 years, with a standard deviation of
0.49 years.
a. Interpretation of the Standard Deviation:
The average distance of the observed useful lives of these
lenses from their mean useful life of 4.6 years is about 0.49
years.
b.
Calculate the value of the standard error of the mean.
SE ( X )
s
n
0.49
6
0.200
Interpretation: We would estimate that the average
distance of the possible sample mean useful life values
(obtained from repeated random samples of size n = 6 pairs of
such lenses) from the population mean useful life, , to be
roughly 0.20 years.
c. Construct a 90% confidence interval for the population mean
life of all such siliconebased lenses:
4.6 (2.015)(0.200) (4.197, 5.003)
Interpretation of the Interval: This interval provides a range
of reasonable values for the population mean useful life, . We
would estimate the population mean useful life, , to be
between 4.197 years and 5.003 years, with 90% confidence.
Interpretation of the 90% Confidence Level: This interval
was made with a method, which if repeated, would generate
many 90% confidence intervals. We would expect 90% of
these resulting intervals to contain the population mean life,
d. State the hypotheses to test the claim made by Bausch and
Lomb about their new contact lens; that is, test if the
15
population
4 years.
mean
useful
life
is
more
than
H 0 : 4 , H a : 4 , with an observed ttest statistic of
x 0 4.6 4
t
0.200
3.00
.
The pvalue for this test is the probability of getting a ttest
statistic at least as extreme as the observed test statistic,
assuming
the
null
hypothesis
is
true.
n
So we have the pvalue is P T 3.00 found under the t(5)
distribution. This pvalue turns out to be equal to 0.015.
Interpretation of the value of the test statistic t = 3.00
in terms of a distance: The observed sample mean was 3
average distances (i.e. 3 standard errors) above the
hypothesized mean of 4.
Interpretation of the resulting pvalue of 0.015: If the null
hypothesis was true (the population mean useful life is just 4
years) and this procedure (study) was repeated many times, we
would expect to see a ttest statistic value of 3.00 or larger in
only 1.5% of the repetitions. Thus are data are somewhat
unusual under the null hypothesis theory, providing evidence
for the alternative theory that the population mean useful life is
greater than 4 years.
e. At a 10% significance level, what is the decision?
Reject H0 since the pvalue is less than 0.10.
f.
What is the conclusion? There is sufficient evidence to
conclude that the population mean useful life of the new
lenses is greater than 4 years.
NOTE:
These interpretations can be extended to the any test and
confidence
interval,
adjusting for the different parameters, different directions of
extreme, different test statistics, etc.
16
Supplement 7: Summary of the
Main tTests
The three inference scenarios presented in Labs 8, 9, and 10 are:
onesample t procedures, paired t procedures, and two
independent samples t procedures. It is important to look at
the data to determine if doing a particular t procedure is
appropriate. That is, we need to check the assumptions. (Recall
that checking assumptions is the second step in performing a
hypothesis test.)
The t procedures have the following
assumptions:
1. Each sample is a random sample (the observations
can be viewed as realizations of independent and
identically distributed random variables). In the paired t
procedures, the differences are assumed a random
sample.
2. Each sample is drawn from a normal population, that
is, the response variable has a normal distribution for
each population.
In the paired t procedures, the
population of differences is assumed to have a normal
distribution. In the twosample case, both populations of
responses are assumed to have normal distributions.
You need normality of the underlying population for the
response in order to have normality for the sample
mean. In the case where you do not have a normal
population, you can still have normality of the sample
mean if you have a large enough sample size (most texts
state at least 30). In Lab 7 (Sampling Distributions and
the CLT) you worked with an applet that demonstrated
the CLT using a large sample size of 25. Thus we will
accept at least 25 as large enough for inference about
means.
3. For the two independent samples t procedures, we also
assume that the two samples are independent. We
also need to assess whether the two population
variances can be assumed equal in order to decide
between the pooled and the unpooled t tests.
Graphical tools can be used to check these assumptions (see
Labs 2 and 3 for more details about these various graphs).
Time Plots (or Sequence Plots): If your quantitative data have
been gathered over time, then a time plot can be used to
determine if the underlying process that generated that time
dependent data appears to be stable. So these plots can help us
17
assess if the random sample assumption is reasonable. For the
paired design problems, we assume our set of differences
calculated from the paired observations (D 1, D2,..., Dn) are a
random sample. So if these differences were obtained over time,
they should be plotted against their order to see if they look like
they came from one population of all differences (no changing
mean or variability over time). Remember: Time or Sequence
plots are useful for checking stability only when the data
are ordered in some sense. If there is no inherent order to
the data, a sequence plot should not be made.
Histograms: Histograms are especially useful for displaying the
distribution of a quantitative response variable. You could make a
histogram of the observations in a onesample problem, of the
differences in a matched pairs design, and of each of the two
samples separately in the independent samples design. Examine
the histogram for evidence of strong departures from normality,
such as bimodality or extreme outliers. Since you are just plotting
data (just a sample and not the entire population of responses),
your histogram may not look perfectly bellshaped or normal.
QQ plots: QQ plots (or quantile plots or normal probability plots)
are generally better than histograms for assessing if a normal
model is appropriate. If the points in a QQ plot fall approximately
in a straight line (with a positive slope) then the normal model
assumption is reasonable.
Boxplots: Boxplots are most useful for assessing the validity of
the assumption of equality of population variances in the two
independent samples design. We would see if the IQRs (shown
graphically by the length of the boxes) are comparable, and also
compare the overall ranges. If they do have comparable lengths
or sizes (they do not need to be lined up), then we have support
that the equality of population variances assumption is
reasonable. We would also want to compare the two sample
standard deviations themselves, and Levenes test of equality of
the two population variances may also be available.
18
Name that Scenario Practice for the Three
T Tests:
Having just reviewed the three main ttest inference scenarios, you
should understand the testing procedures and be able to interpret the
results of a test. However, it is important to know when each scenario
applies. Read each of the following inference scenarios and determine
which of the three ttest procedures would be most appropriate: the onesample ttest, the paired ttest, or the twoindependent samples ttest.
1. A researcher is studying the effect of a new teaching technique for
middle school students. One class of 30 students is taught using the
new technique and their mean score on a standardized test is
compared to the mean score of another class of 27 students who
were taught using the old technique.
2. A company claims that the economy size version of their product
contains
32
ounces.
A consumer group decides to test the claim by examining a random
sample of 100 economy size boxes of the product, since they have
received reports that the boxes contain less than the 32 ounces
claimed.
3. At some universities, athletic departments have come under fire for
low academic achievement among their athletes. An athletic director
decides to test whether or not athletes do in fact have lower GPAs. A
random sample of 200 student athletes and a random sample of 500
nonathlete students are taken and their GPAs are recorded.
4. As part of a biology project, some high school students compare heart
rates of 40 of their classmates before and after running a mile. They
want to see if the heart rate of students their age is faster after
running a mile than before, on average.
5. A hospital is studying patient costs; they decide to follow 500 surgery
patients hospital and medical bills for a year after surgery, and
compare them to the estimated costs provided to the patients before
surgery. They want to see if the estimated and actual costs are
comparable on average.
6. A chemical process requires that no more than 23 grams of an
ingredient be added to a batch before the first hour of the process is
complete. An analyst feels that due to current settings more than 23
grams may actually be added. If the analyst is correct, the settings
19
need to be altered and recent batches recalled. A random sample of
25 batches is obtained from the machine that is supposed to add the
ingredient. The measurements are used to test the analysts claim.
Supplement 8: Regression Output in
SPSS
There are four parts to the default regression output. Use the
scroll bar at the right edge of the Output Window to scroll up to
the top of the regression output. The first section just reminds you
which
variable
was
entered
as
the
explanatory
x variable; for this example, the explanatory variable is DNA.
The second section has the heading Model Summary. The Model
Summary starts with the correlation between the two variables, R,
which is the absolute value of the correlation coefficient, r.
You need to look at the sign of the slope of the regression line to
determine if you need to put a minus sign in front of this value to
correctly report the correlation coefficient. (The actual value of
the correlation coefficient is also reported in the last section of
regression output, under the column heading Beta.)
The
correlation coefficient measures the strength of the linear
association between the two variables. The closer it is to +1 or 1,
the stronger the linear association. The square of the correlation,
the R Square quantity, has a useful interpretation in regression.
It is often called the coefficient of determination and measures
the proportion of the variation in the response that can be
explained by the linear regression of y on x. Thus, it is a measure
of how well the linear regression model fits the data. The Std.
Error of the Estimate gives the value of s, the estimate of the
population standard deviation .
Model Summary
Model
1
R
R Square
a
.856
.732
Adjusted
R Square
.699
Std. Error of
the Estimate
4.851
a. Predictors: (Constant), DNA
The third part of the output contains the ANOVA table for
regression, used for assessing if the slope is significantly
20
different from 0 via an F test. The corresponding ttest will be
discussed first and we return to this ANOVA part later.
ANOVAb
Model
1
Regression
Residual
Total
Sum of
Squares
515.141
188.228
703.369
df
1
8
9
a. Predictors: (Constant), DNA
b. Dependent Variable: PLAQUE
21
Mean Square
515.141
23.528
F
21.894
Sig.
.002a
The last portion of the output falls under the heading
Coefficients. In this section, the least square estimates
for the regression line are given. These estimated regression
coefficients are found under the column labeled B. The estimated
slope is next to the independent variable name (in this example it
is DNA), and the estimated intercept is next to (Constant). So, b0
is the coefficient for the variable (Constant), and b1 is the
coefficient for the independent variable x in the model. The next
column heading is Std. Error, which provides the corresponding
standard error of each of the least squares estimates. Also
produced
in
this
table,
are
the
ttest statistics in the column labeled t and Sig., which reports
the
twosided
pvalues for these ttest statistics.
Coefficientsa
Model
1
(Constant)
DNA
Unstandardized
Coefficients
B
Std. Error
.548
8.193
.167
.036
Standardi
zed
Coefficien
ts
Beta
.856
t
.067
4.679
Sig.
.948
.002
a. Dependent Variable: PLAQUE
The tstatistic for the slope, in the second row, is a test of the
significance of the model with x versus the model without x, that
is,
for
testing
H0:
1
=
0
versus
Ha: 1 0. The tstatistic for the yintercept, in the first row, is a
test of whether the yintercept (o) is different from zero. This test
is not often of interest unless a value of 0 for the yintercept is
meaningful
and
of
interest.
For
example,
if
x = amount of soap used and y = height of the suds, then an
intercept value of 0 is meaningful as no soap would lead to no
suds. The column labeled Sig. gives the twosided pvalue for
the corresponding hypothesis test.
SPSS also provides the information to calculate confidence
intervals for the parameter estimates. The column labeled Std.
Error provides standard errors (estimated standard deviations)
of the parameter estimates and is the quantity that is multiplied
by the appropriate t* value in computing the halfwidth of the
confidence interval. Recall that you can request SPSS to produce
22
these confidence intervals for you using the Statistics button in
the Regression dialog box.
23
Interpretation of estimated slope b1:
According to our regression model, we estimate that increasing
DNA by one unit has the effect of increasing the predicted plaque
by .167 units.
Interpretation of r2:
According to our model, 73% of variation in plaque levels can be
accounted for by its linear relationship with DNA.
Decision for test of a significant linear relationship:
Since the pvalue = .002 is less than the significance level = .05,
we can reject the null hypothesis that the population slope, 1,
equals 0.
Conclusion: There is sufficient evidence to conclude that in the
linear model for plaque based on DNA the population slope, 1,
does not equal zero. Hence, it appears that DNA is a significant
linear predictor of plaque.
Lets return to the ANOVA table in the middle of the regression
output.
ANOVAb
Model
1
Regression
Residual
Total
Sum of
Squares
515.141
188.228
703.369
df
1
8
9
Mean Square
515.141
23.528
F
21.894
Sig.
.002a
a. Predictors: (Constant), DNA
b. Dependent Variable: PLAQUE
The Regression Sum of Squares corresponds to the portion of
the total variation in the data that is accounted for by the
regression line. Everything that is left over and not accounted for
by the regression line is placed in the Residual Sum of Squares
category. Then, dividing the sum of squares by their respective df
(degrees of freedom) yields the Mean Squares.
Finally, the ratio of the Mean Squares provides the F statistic which
tests if the slope is significantly different from zero (i.e. if there is a
significant nonzero linear relationship between the two variables
H0: 1 = 0 versus Ha: 1 0.) The Sig. is the corresponding pvalue
for the F test of these hypotheses.
24
In simple linear regression, the ttest in the Coefficients output
for the slope is equivalent to the ANOVA Ftest. Notice that the
square of the tstatistic for testing about the slope is equal to the
Fstatistic in the ANOVA table, and the corresponding pvalues are
the same
25
Checking the Simple Linear Regression Assumptions
Here is a summary of some graphical procedures that are useful in
detecting departures from the assumptions underlying the simple
linear regression model.
1. LINEARITY: Do a scatter plot of y versus x.
The plot should appear to be roughly linear.
2. STABILITY: Do a sequence plot of the residuals.
The plot should show no pattern indicating any trend in the
mean or in the variance of the residuals. An example series
plot is shown below. Remember that it is only appropriate to
make sequence plots when there is some ordering present in
the data.
26
3. NORMALITY: Examine a QQ plot of the residuals to check on
the assumption of normality for the population (true) error
terms. An example QQ plot is shown below.
4. CONSTANT STANDARD DEVIATION of the population (true)
error terms: Make a plot of the residuals versus x. This plot is
called a residual plot. The residuals represent what is left
over after the linear model has been fit. The residual plot
should be a random scatter of points in roughly a horizontal
band, with no apparent pattern. An example residual plot is
shown at the right. Sometimes this plot can also reveal
departures from linearity (i.e. that the regression analysis is not
appropriate due to lack of a linear relationship).
27
Lab 1: Scavenger Hunt; Mean and
Median
Objective: In this lab you will visit several of the sites that will be
used throughout the course of the semester and learn the
locations of the important resources. In the second part of the lab,
you will make use of an applet to explore ideas related to
measures of center for a data set.
Overview: In this course you will be using various forms of
technology including applets that will be useful for exploring
statistical concepts. There will be a lot of places to visit and the
links for all these sites are available on CTools. In the first portion
of the InLab Project, you will complete a Scavenger Hunt that will
allow you to become familiar with the resources available on
CTools. In the second part of the InLab Project you will use the first
applet to enhance your understanding of the mean and median.
Measures of Center: Measures of center are numerical values
that tend to report the middle of a set of data. The two that we
will focus on are the mean and the median.
1. Mean: The mean of a set of n observations is simply the
sum of the observations divided by the number of
observations, n.
2. Median: The median of a set of observations, ordered from
smallest to largest, is the middle value. It will be the value
such that (at least) half of the observations are less than or
equal to that value and (at least) half the observations are
28
greater than or equal to that value.
Measures of Variation or Spread: Measures of variation include
the Interquartile Range (IQR) and standard deviation. These
numerical summaries describe the amount of spread that is found
among the data, with larger values indicating more variability or
more spread.
1. Standard Deviation: Standard deviation is a measure of
the spread of the observations from the mean. It is actually
the square root of an average of the squared deviations of
the observations from the mean. We can think of the
standard deviation as approximately an average distance of
the observations from the mean.
2. IQR: The IQR measures the spread of the middle 50% of the
data. It is defined as the difference between the 3 rd quartile
(Q3) and the 1st quartile (Q1). These quartiles are also
called the 75th and 25th percentiles, respectively. IQR = Q3
Q1.
29
WarmUp:
Variables
Categorical
and
Quantitative
Take a few minutes to recall the two types of variables that have
been introduced in class: categorical and quantitative. For each of
the following variables, determine whether it is a categorical or
quantitative variable. Recall that numerical summaries such as
mean and median can only be computed for quantitative
variables.
Cell Phone Model (iPhone, Android)
Categorical
Quantitative
Number of countries visited in a year
Categorical
Quantitative
Credit Hours enrolled per semester
Categorical
Quantitative
Eye color
Categorical Quantitative
Points scored in a basketball game
Categorical
Quantitative
ILP: Scavenger Hunt
The first activity in this InLab Project is a Scavenger Hunt that will
allow you to become familiar with the locations for all your
important resources on CTools.
Task
Answer
30
1. Go to the CTools Site and find the
formula card page in Resources:
What is the topic header at the top of
page 3
of formula card?
2. Go to the CTools Site and find the
Data Sets folder in Resources:
What is the name of the last data set
(when sorted alphabetically a z)?
3. Go to the CTools Site and find the Lab
Info folder
in Resources:
Go to the folder for your lab section
and find your lab syllabus.
Look at the Grading Policy section The InLab Projects will count what % toward your
final course grade?
31
Task
Answer
4. Go to the CTools Site and find the
Stats 250 Prelab link:
Click on the link that will take you to the
Prelab Sitemaker Site.
Click on LESSON 1.
How many short videos do you have to
watch for your Lesson 1?
5. Go to the CTools Site and find the
Assignments link in the left menu:
Find the PRELAB LESSON 1 assignment.
When is your PRELAB LESSON 1
assignment due?
6. Go to the CTools Site and find the
Link to the Online HW site (called
Lecturebook):
If you have your HW tool subscription,
log in and select your GSI.
What is our lab section number?
Note: If you dont have your GSI
selected (with correct section number),
your GSI cannot see your homework and
thus you will receive 0 points!
7. Go to the CTools Site and find the
Lab Info Folder in Resources:
Go to the folder for Links to Lab Applets.
Click on LAB 1.
What is the title of the Applet?
For the second part of the ILP
you will work with this applet.
32
ILP: The Mean and the Median
The applet that you opened in the last step of the Scavenger Hunt
portion of the ILP will now be used to help you discover how the
shape of the distribution for a set of data can provide important
details regarding the relationship between the mean and the
median for that data. In this activity you will observe the mean
and the median for a variety of shapes of distributions.
Open the Lab 1 Descriptives Applet from the Links to Lab
Applets folder on the Stat 250 CTools site (in the Lab Info folder
under Resources).
Alternatively, the original applet can be found at:
http://onlinestatbook.com/stat_sim/descriptive/index.
html
33
This web site contains a Java applet that will help you
understand the relationship between the mean and the
median.
1. Read the applet instructions.
2. Click Begin and you will see a histogram of nine numbers:
3, 4, 4, 5, 5, 5, 6, 6 and 7.
This histogram shows a symmetric distribution. The summary
in the upper left corner shows that the mean and the median
are both equal to 5, the standard deviation is 1.15 and there is
no skewness (note the skewness measure is 0).
3. Change the distribution so that it now has a positive skew by
painting the histogram with the mouse. Does this correspond
to a right or left skewed distribution? Which is bigger, the
mean or the median?
4. Change the distribution so that it has a negative skew. Which
direction is this distribution skewed? Now which is bigger, the
mean or the median?
5. Try a few other distributions (uniform, ushaped, etc.) and see
how the mean and median compare.
Comment on your
34
findings here.
6. Summarize what you have learned about the relationship
between the shape of a distribution and the mean and median.
CoolDown #1: Which to Report?
You have seen that the mean is more sensitive to outliers than the
median. For a data set that contains several outliers, which
measure of center would you choose to report? What measure of
spread? Explain.
CoolDown #2: Real World Example
You are the manager of a local grocery store who is put in charge
of setting the prices for your stock. You will determine the prices
for each product by examining the prices of your competitors in
the neighborhood. Suppose your neighborhood consists mainly of
chain store supermarkets along with 2 highend grocery stores.
You want to set your prices low enough to attract customers but
high enough so you will make a profit. How would you use these
measures of center to help you determine the prices?
35
Example
Median
Exam
Question
on
Mean
and
Tattoos The Pew Research Center took a survey of Millenials,
that is, young adults between the ages of 18 and 29. The survey
looked at characteristics that researchers thought described
Millenials social networking, life priorities, and aspects of the
respondents' appearances. One question asked how many tattoos
respondents had, and there were 486 respondents who had at
least one tattoo. Data for these 486 respondents are shown in the
histogram below.
a. How could this histogram be described?
Choose all that apply.
RIGHT SKEWED
LEFT SKEWED
SYMMETRIC
STABLE
BIMODAL
UNIMODAL
DECREASING TREND
Note: the first interval is
[1, 3).
b.
If the mean number of tattoos for these Millenials is 3.01
tattoos, which of the
following is a reasonable value for the
median?
2
3.01
c. Is the following statement true or false?
Based on the histogram,
we can be sure the range of the tattoo data is exactly 20.
TRUE
FALSE
36
d. The standard deviation is computed exactly to be 2.91. Write a
sentence to interpret this standard deviation (in terms of an
average distance) in the context of the problem.
37
Lab 2: Describing Data with Graphs
and Numbers
Objective: In this lab, you will use some graphical and numerical
tools to summarize the distribution for a quantitative variable or
response a histogram, a boxplot, mean, median, standard
deviation, and interquartile range (IQR).
You will also be
introduced to sidebyside boxplots for comparing two or more
distributions and bar charts for summarizing categorical data.
These techniques can be very useful at the start of data analysis
to get a feel for the data.
Overview: Two graphs that can be used to summarize the
distribution for a single quantitative variable or response are a
histogram and a boxplot.
Each graph provides different
information about the distribution. When used properly, graphs
can be a very effective way to summarize data. Data on a single
quantitative variable should first be examined graphically. The
overall shape of the distribution and existence of outliers can
generally be used to assess if the data appear to be coming from a
relatively homogenous population. If so, then various numerical
summaries may be used to characterize the center of the
distribution (such as mean and median) and the spread of the
distribution (such as the standard deviation and the IQR). For
categorical variables, a bar chart can be used to display the
number falling in each category (frequency distribution).
Histograms: A histogram displays the distribution of a
quantitative variable by showing the frequency (count) or percent
of the values that are in various classes. The classes are typically
intervals of numbers that cover the full range of the variable.
Histograms can be used to assess the symmetry and modality of
a single distribution or for comparing the relative locations and
shapes of several distributions.
Boxplots: One plot that can detect extreme observations or
outliers is the boxplot. A boxplot is a graphical representation of
the fivenumber summary, namely, the minimum, first quartile,
median, third quartile, and maximum of the data. The centerline
of the box marks the median or the 50 th percentile. The sides of
the box show the first (lower) quartile, Q1, and the third (upper)
quartile, Q3. Thus a boxplot shows the overall range (maximum
minimum) and the interquartile range (IQR = Q3 Q1). A
modified boxplot uses a rule for identifying values that are
38
extraordinary compared to the others (outliers or outside
values). Circles (o) are used to denote outliers and asterisks (*) to
denote extreme outliers if any are present. Any point below Q1
(1.5 x IQR) or above Q3 + (1.5 x IQR) is considered an outlier.
Extreme outliers are points below Q1 (2 x IQR) or above Q3 + (2
x IQR). Box plots cannot tell you the shape of the distribution.
39
Sidebyside Boxplots: These plots are helpful for comparing
two or more distributions with respect to the fivenumber
summary. For example, suppose you are interested in comparing
the distribution of a variable, such as the salary of the employees
of a certain company. If you have information on sex for the
group, you might be interested in comparing the distribution of
salary of females with respect to males. In this case, the sidebyside boxplot will be an important part of the descriptive analysis of
the data set involved.
Bar Charts: One way to display the number or frequency
distribution for a categorical variable is with a bar chart. A bar
chart shows the percentage of items that fall into each category
or value of a categorical variable. It displays a bar for each
category with the height of each bar equal to the number, the
proportion, or the percentage of items in that category. If the
categories have no inherent order, we could rearrange the bars in
the graph in any way we like. In such cases, the shape of the bar
graph would have no bearing on its interpretation.
WarmUp: Matching
Match the graph or descriptive statistic to one of its primary uses
(some may have more than one and you may use an answer more
than once).
____ i. Histogram
outliers
A. Measure of center, not sensitive to
____ ii.Bar Chart
shapes)
B. Compare distributions (but not their
____ iii.
Mean
categorical variable
C.
____ iv.
Median
quantitative variable
Examine
distribution
of
D. Examine distribution of a
____ v. Sidebyside Boxplots
E. Measure of spread
____ vi.
outliers
F. Measure of center, sensitive to
IQR
40
ILP: Visualizing and Exploring Data
In this InLab Project, you will learn how to create graphs and
obtain descriptive statistics for a data set using SPSS.
Task: The data set employee data.sav contains information on
employees at a company. Explore possible questions this data
could be used to address. Create appropriate graphs and obtain
descriptive statistics for current salary, and discuss the results.
1. Log onto your computer. To obtain the data set, go to CTools,
and find Datasets for Labs and HW in the Resources folder.
Select employee data.sav and save it to a directory of your
choice (alternatively you may open the data set directly, in
which case you do not need to open SPSS after). Once you
have saved the data set, go to Programs, followed by
Statistics Packages & Math Programs and then select
SPSS.
2. To open the employee data.sav data set from within SPSS,
select the option Open an existing data source from the
dialog box with the More Files line highlighted and click on
OK. Change the directory to where you saved the data set,
select employee data.sav and click on the Open button. The
data set will open, and you can view it (it works like an Excel
spreadsheet).
3. The starting view of the data is the Data Editor window. Here,
you can see the variables in the data set and their values. The
first variable you should see is ID.
What
What
What
What
is the second variable present in the data set?
type of variable is it?
is the eighth variable present in the data set?
type of variable is it?
4. Brainstorm possible questions this data set may have been
collected to address.
5. Focus on the variable current salary.
41
What are some graphs
that would be appropriate to make for this variable?
6. Create a histogram for current salary. Use the graphs menu Graphs> Legacy Dialogs> Histogram, select (current)
salary, and move it to the variable box. Editing details can be
found in the Editing Charts in SPSS section (Supplement 4)
of this workbook.
Note: All Statistics 250 homework and labwork will require that
students provide an appropriate title and their name on each SPSS
chart or output. For histograms, click on the Titles button and
enter the information and click on Continue.
Describe what the histogram shows about the distribution of
current salary.
A good description includes information
regarding the shape, modality and range of the data along with
possible outliers.
7. Obtain a boxplot for current salary. Use: Graphs> Legacy
Dialogs> Boxplot> Simple> Summaries of separate
variables. This is appropriate for one variable with no groups.
Click on the button Define to open another dialog box that
defines the variables for our analysis. Click once on salary to
highlight it and then on the Boxes Represent arrow to select
it.
Note: Boxplots do not have a Titles option. However, you may
add a title via the Chart Editor. Double click the graph, and from
the Chart Editor menus select Options> Title. The Chart Editor
creates the text box and automatically positions it in the top
center of the chart. Type the text and press enter when you are
finished typing. To enter line breaks, press Shift+Enter.
Describe what the boxplot shows about the distribution of
current salary. What do the various lines on the boxplot
represent?
42
8. Numerical summaries may also be obtained for any
quantitative variable. To obtain the fivenumber summary do
Analyze> Descriptive Statistics> Frequencies and then
choose the summary measures you want under the Statistics
button.
Fill in the basic summary measures for current salary
(some require hand calculation).
Mean:
Standard Deviation:
Median:
Q1:
Q3:
IQR: Q3Q1 =
Min:
Max:
Range:
MaxMin =
9. To obtain numerical summaries or any graph (except boxplots)
for current salary by minority status, we need to split the data
file. Use Data> Split File and choose Organize output by
groups.
Doing this successfully does not produce any
noticeable changes in the SPSS windows; there will only be two
short lines of output in the Output window confirming the split.
The grouping variable is minority classification.
Obtain
descriptive statistics for current salary by minority status (once
the data is split, just generate descriptive statistics). List some
of your findings below.
Minority:
NonMinority:
10.Create histograms for both minorities and nonminorities.
(Leave the data split and create histograms as before). For
each category, would it be appropriate to summarize the shape
of the distribution of the current salary using descriptors such
as skewed or symmetric? Why?
Important Note: When you are finished conducting analyses by
43
group, you need to go back to the Split File dialog box and
choose Analyze all cases, do not create groups. Again, the
only change will be one line of output confirming the data has
been unsplit.
11.Create sidebyside boxplots for current salary. The data file
should NOT be split to create these. Use Graphs> Legacy
Dialogs> Boxplot with Simple and Summaries for groups
of cases. Minority Status is the variable for the category axis,
and current salary is the variable.
How does the distribution for current salary compare for
minorities versus nonminorities (based on the sidebyside
boxplots, histograms, and descriptives)?
CoolDown: Check
about Std Dev
Your
Understanding
Recall the definition of Standard Deviation from Lab 1 the
standard deviation is a measure of the spread of the observations
from the mean. It is actually the square root of an average of the
squared deviations of the observations from the mean. We can
think of the standard deviation as approximately an average
distance of the observations from the mean.
a. Suppose we are interested in learning about heights of
Michigan students. We take a simple random sample of 100
students and find that the average height for this sample is 66
inches, with a standard deviation of 2 inches. Below are some
interpretations of this standard deviation.
For each one,
evaluate if it is a correct interpretation or say why it is
incorrect.
1. The average distance between the height values and the
mean height is roughly 2 inches.
Correct
Incorrect
because
___________________________
_____________________________________________________________
44
2. The height values differ from the
approximately 2 inches, on average.
Correct
Incorrect
mean
height
by
because
___________________________
_____________________________________________________________
3. The average distance between the height values is roughly
2 inches.
Correct
Incorrect
because
___________________________
_____________________________________________________________
b. A student provided the following incorrect interpretation of
standard deviation.
68% of the height values are within 2 inches of the mean
height
Why is this interpretation incorrect in general? What graph of
the height data would you make to check if the statement
could be correct? What would you look for in the graph?
45
Example Exam Question on Boxplots
Fiftyfive parents of gradeschool children were recently
interviewed regarding the breakfast habits in their family. One
question asked was if their children take the time to eat a
breakfast (recorded as breakfast status Yes or No). The grades of
the children in some core classes (e.g. reading, writing, math)
were also recorded and a standardized grade score (on a 10point
scale) was computed for each child. At the end of the study it was
discovered that the children who do take time to eat breakfast get
higher grade scores than those who dont.
a. What type of study is this?
Observational study
Experiment
b. What is the response variable in this study?
_____________________________
c. What is the explanatory variable in this study?
___________________________
d. What type of variable is the explanatory variable? Categorical
Quantitative
Sidebyside boxplots of the childrens standardized grade scores
are provided.
11
10
9
8
7
6
Grades
5
4
3
No
Yes
Do you have breakfast?
e. What is (approx.) the lowest grade scored by a child who does
have breakfast?
___________ points
f.
What is (approx.) the IQR for the grade scores of children who
do eat breakfast?
___________ points
g. Using one of the measures displayed in the boxplot, complete
46
this sentence.
The highest grade scored by one of the children not eating
breakfast is (approx) equal to the ______________________ for the
children who do eat breakfast.
h. True or false: The symmetry in the b for the children not eating
breakfast implies that the histogram of the same data is also
symmetric.
Circle one: True
False
Explain:
Lab 3: Time Plots and QQ Plots
Objective: In this lab, you will add to your set of graphical tools
for examining data.
The graphs you will examine include
sequence (time) plots for data collected over time and QQ plots
for checking whether a normal model is a reasonable distribution
for a quantitative variable.
Overview: Lab 2 provided a summary of some graphical and
numerical tools that can be used to summarize the distribution for
quantitative and qualitative variables or responses. We may use
those tools for the activities in this InLab Project, but we will also
need to utilize the new tools described below. Note that these
graphical tools are introduced solely in lab, not in lecture, so it will
benefit you to read this overview thoroughly.
Time (Sequence) Plots: Data is often gathered over time.
Employment rate, stock prices, and sales figures are just a few
examples. When data is gathered over time, it is generally wise to
examine the data plotted against time. Plots against time can
reveal the main features of a time series, overall patterns and
striking deviation from those patterns. Some overall patterns that
may arise are:
A persistent, longterm rise or fall called a trend
(either increasing or decreasing).
A pattern that repeats itself at regular intervals of
time called seasonal variation.
A persistent, longterm increase or decrease in the
variation of the observations called a pattern in
variation.
If data is collected over time, a time plot can be used to check the
assumption of a random sample, which will be needed for
inference procedures. As you have learned in your lecture notes
47
on sampling, a random sample consists of independent and
identically distributed (i.i.d.) observations. This means that the
observations can be considered as all coming from the same
parent population (with the same or identical distribution) and are
independent of one other. With a sequence plot, you can check
the identically distributed aspect of a random sample by looking
for evidence of stability in the plot. Stability is supported when
both the mean of the observations and the amount of variation
among observations appear to be constant over time and there
does not appear to be any pattern in the resulting plot.
48
Below are two sequence plots; in the first plot the observations
appear to support that the underlying process that generated the
observations is stable, but that is not the case for the observations
in the second plot on the right. In this case, there appears to be
an increasing trend, thus the underlying process does not appear
to be stable; the observations should not be considered a random
sample.
QQ Plots: Later in this class, we will see that the assumption of
a normal model for a population of responses will be
needed in order to perform certain inference procedures.
Previously, we have seen that a histogram can be used to get an
idea of the shape of a distribution. However, there are more
sensitive tools for checking whether the shape is close to a normal
(bellshaped) model.
The best plot that can be used to check for normality is called a QQ Plot, which is a plot of the percentiles (or quantiles) of a
standard normal distribution against the corresponding percentiles
of the observed data. If the observations follow an approximately
normal distribution, the resulting plot should be roughly a straight
line with a positive slope. Deviations from this indicate possible
departures from a normal distribution.
At the right is an example of a QQ Plot showing strong support to
say the data that does seem to come from a population with an
approximately normal distribution.
49
The three graphs below are
examples for which a normal
model for the response is not
reasonable.
The QQ plot on the left indicates the existence of two clusters of
observations. The QQ plot in the center shows an example where
the shape of the distribution appears to be skewed right. The QQ
plot on the right shows evidence of an underlying distribution that
has shorter tails compared to those of a normal distribution.
Note: It is only important that you can see the departures in the
above graphs and not as important to know if the departure
implies skewed left versus skewed right and so on. A histogram
would allow you to see the shape and type of departure from
normality.
Finally, we consider
an example QQ plot
(shown at the right)
that appears normal
with the exception of
one data point.
50
In this case, we
would say the QQ
plot shows evidence
of
an
underlying
distribution which is
approximately
normal except for
one large outlier that
should be further
investigated.
Note that outliers
could
appear
in
either the upper or
lower tail.
WarmUp: Check Your Understanding
Before beginning the InLab Project, review your understanding of
the key concepts related to time plots and QQ plots.
1. For each word pair, select the appropriate word(s) to complete
the sentences.
We use sequence (or time) plots to check the
independent
identically distributed
part of the random sample assumption
by looking to see if the data appear to be
stable
normal ,
that is, have a constant mean and constant variation over
time.
If there is any pattern in the observations over time, we
should
should not
make a histogram of the observations for further analysis.
2. If the time plot supports that we can consider our observations
to be a random sample (that is, shows our underlying process
appears to be stable), we could make a histogram to help
51
assess if the model for our response in the underlying
population is normally distributed.
What other graph could we make to help assess this normality
assumption?
What would we hope to see in that graph to support normality
is reasonable?
ILP: Time Plot and QQ Plot Examples
In this first part of the InLab Project, you will look at more
examples of time plots to help learn how to better read such
graphs for assessing whether our data appear to be stable and
support the random sample condition.
Task 1: Go to the Stat 250 Prelab Site and find the Time Series
tab along the top. Download the timeseries.rdata file script you will
be using in this part of the ILP. When you double click on this
script file it should open up the R program (on all campus
machines, and you can download R to your computer free too).
1. Begin the program by entering the following command.
timeseries()
2. Select your sample size by entering a number between 1 and
10000.
3. Select if you would like to see an example of a stable or
unstable time plots.
4. If unstable time plots are selected, you will be asked questions
to determine the type of unstable pattern youd like to see.
5. Once your time plots have been created, you will be asked if
you want to save your plots to the desktop as an image.
Answer and then you will again be asked to select a next
sample size. Try out the various options and explore the various
patterns of time plots that were mentioned in the introduction.
6. Sketch below a time plot which indicates both an increasing
mean and a decreasing variance.
52
Task 2: Go to the Stat 250 Prelab Site and find the QQ Plot tab
along the top. Download the qqplot.rdata file script you will be
using in this part of the ILP. When you double click on this script
file it should open up the R program (on all campus machines, and
you can download R to your computer free too).
1. Begin the program by entering the following command.
qqplot()
2. Select your sample size by entering a number between 1 and
10000.
3. Select the type of distribution you would like an example a QQ
plot be generated from.
4. Once your QQ plots and the corresponding histograms have
been created, you will be asked if you want to save your plots
to the desktop as an image. Answer and then you will again be
asked to select a next sample size. Try creating QQ plots for
many different distributions and sample sizes.
5. Sketch below one of the resulting QQ plots and Histograms for
a sample of 1000 observations from a skewed right distribution.
QQ plot:
Histogram:
53
ILP: TimeDependent Data
Background 1: The data set deathrate.sav contains the
death rate (number of deaths per 100 million miles driven) taken
at twoyear intervals from 1960 to 2004.
Display and summarize this data in an appropriate and useful way.
What do you see? Would it make sense to make a histogram of
the death rates? The following steps will guide your thinking as
you complete this task.
1. Once the data set is open, swap from Data View to Variable
View. Here, you can set the number of decimal places for each
variable, give them longer labels, and input coding for
categorical variables. What is the extended label description of
the variable Rate?
2. Why should a sequence plot be made to display this data?
3. Make a sequence plot for the data using Analyze>
Forecasting> Sequence Charts.
What does the graph
show? Comment on if you see any trend, seasonal variation, or
pattern in variation in this graph.
4. Does the plot appear to be stable? What would you conclude if
asked if the data were a random sample of death rates?
5. Would it make sense to make a histogram of the death rates?
Why or why not?
54
Background 2: The data set oldfaithful.sav contains the
date and duration of eruptions (in minutes) of the Old Faithful
geyser. The data was collected several times per day over 23
consecutive days.
Display and summarize the data in an
appropriate and useful way. What do you see? Does there appear
to be any pattern to this process? The following steps will guide
your thinking as you complete this task.
1. Make a time plot for the data using Analyze> Forecasting>
Sequence Charts. What does the graph show? Are there any
patterns to the process?
2. Does the plot appear to be stable? What would you conclude if
asked if the data were a random sample of eruptions?
ILP: Assessing Normality
You have discussed using a histogram to examine the shape of the
distribution of a quantitative variable. If the histogram shows a
fairly homogeneous, unimodal set of observations, we might like to
assess whether a normal distribution is a reasonable model for the
response. A better graph for assessing normality is a QQ plot. In
this problem, we will examine a few distributions and see what
each corresponding QQ plot looks like.
Background 1: Suppose a study examined high school
students and the relationship between IQ and GPA. Use the iq.sav
dataset and examine the distribution of IQ.
1. Create a histogram and a QQ plot for the IQ values. QQ plots
are created via Analyze> Descriptive Statistics> QQ plots.
Provide rough sketches.
Histogram:
QQ Plot:
2. Describe the shape of the resulting histogram.
55
3. Is a normal distribution a reasonable model for IQ scores in the
population based on this QQ plot?
Background 2: We have previously explored the employee
data.sav dataset variable of salary. Now, lets check to see if the
overall distribution of salary can be considered normal, and then to
see if the distribution of salary might be normal depending on
minority status.
1. Create a histogram for the variable salary, and describe its
shape.
2. Create a QQ plot for salary using Analyze> Descriptive
Statistics> QQ Plots. Based on the evidence of these
graphs, is a normal distribution an appropriate model for
current salary? Why or why not?
3. Create histograms and QQ plots separately for minorities and
nonminorities (recall the Data> Split File command). Does
the distribution of salary appear to be different for either
group? Comment on both histograms and QQ plots.
4. For salary, is a normal model reasonable for either minorities or
nonminorities?
56
57
CoolDown: Matching Graphs
Match the corresponding histograms, boxplots, and QQ plots.
C
Histogram A
Boxplot A
Boxplot C
QQ Plot A
Histogram B
Histogram
Boxplot B
QQ Plot B
58
QQ Plot
Which type of graph best shows the shape of the underlying
distribution?
Which type of graph best shows if the underlying distribution
appears to be normal (bellcurve)?
Example Exam Question on Sequence Plots and QQ Plots
A new method of measuring phosphorus levels in soil is under
consideration. A sample of 11 soil specimens is analyzed using
the new method. The time series (sequence plot) for the 11
observations is presented below.
a. Comment on the overall stability of these data based on this
plot.
660
640
620
600
580
PHSLEV
560
540
520
1
10
11
Sequence number
b. An assumption of many statistical inference methods is that the
data follow a normal distribution. In the space provided below,
sketch how the QQ plot would appear if a normal distribution
was a good model for phosphorus levels.
59
Lab 4: Probability and Random
Variables
Objective: The objective of this lab is to become familiar with
using the models for random variables and to find the probabilities
associated with the models you have learned. The probabilities we
compute from these models (for example, pvalues in testing
theories) will help us make reasonable decisions. You will work with
three random variables and the methods used to calculate
probability for each variable. You will also become familiar with
several concepts that allow for easier calculation of probabilities.
Overview: In this lab you will be introduced to several random
variables and their models. These variables can be classified as
one of two types: a discrete random variable, which has a finite
number of outcomes, and a continuous random variable which
has an infinite number of outcomes.
You will practice computing probabilities for each of these
variables, using concepts such as a standardized score and a
Normal Approximation. There are several situations that can
make computing probabilities easier, such as independent
events and mutually exclusive events that will also be
explored in this lab.
60
Independent Events: Two events A, B are said to be
independent if knowing that one will occur (or has
occurred) does not change the probability that the
other occurs. In probability notation this can be
expressed as P(AB) = P(A).
Mutually Exclusive: Two events A, B are mutually
exclusive (or disjoint) if they do not contain any of the
same outcomes. So their intersection is empty.
Random Variables: A random variable assigns a number to each
outcome of a random circumstance, or, equivalently, a random
variable assigns a number to each unit in a population. The
distribution of a random variable is a model that shows us what
values are possible for that particular random variable and how
often those values are expected to occur (i.e. their probabilities).
The model can be expressed as a function or table or picture,
depending on the type of variable it is.
We will consider two broad classes of random variables: discrete
random variables and continuous random variables.
61
Discrete Random Variable: A discrete random variable, X , is a
random variable with a finite or countable number of possible
outcomes. The probability distribution function (pdf) for a discrete
random variable X is a table or rule that assigns probabilities to
the possible values of the X.
Two conditions that must always apply to the probabilities for a
discrete random variable are:
Condition 1: The sum of all of the individual probabilities must
equal 1.
Condition 2: The individual probabilities must be between 0 and
1.
Binomial Random Variable: One type of a discrete random
variable is the binomial random variable, which counts the number
of times a certain event occurs out of a particular number of
observations or trials of a random experiment.
A binomial experiment is defined by the following conditions:
1. There are n trials where n is determined in advance, not a
random value.
2. There are two possible outcomes on each trial, called a
success (S) and a failure (F).
3. The outcomes are independent from one trial to the next.
4. The probability of a success remains the same from one
trial to the next, and this probability is denoted by p. The
probability of a failure is 1 p for every trial.
A binomial random variable is defined as: X = number of
successes in the n trials of a binomial experiment.
Continuous Random Variable: A continuous random variable, X
, takes on all possible values in an interval (or a collection of
intervals). The way that we determine probabilities for continuous
random variables differs in one important respect from how we
determine probabilities for discrete random variables. For a
discrete random variable, we can find the probability that the
variable X exactly equals a specified value. We cant do this for a
continuous random variable. Instead, we are only able to find the
probability that X could take on values in an interval. We do this by
determining the corresponding area under a curve called the
probability density function of the random variable.
So the probability distribution of a continuous random variable is
described by a density curve. The probability of an event is the
area under the curve for the values of X that make up the event.
62
The probability model for a continuous random variable assigns
probabilities to intervals.
Definition: A curve (or function) is called a Probability Density
Curve if:
1. It lies on or above the horizontal axis.
2. Total area under the curve is equal to 1.
Normal Random Variable: The family of normal distributions is
very important because many variables have this shape and form
approximately and many statistics that we use in our inference
methods are based on sums or averages which generally have
(approximately) a normal distribution.
A normal curve is symmetric, bellshaped, centered at the mean
and its spread is determined by the standard deviation. In fact, the
points of inflection on each side of the mean mark the values
which are one standard deviation away from the mean.
Standardized Scores: A normal distribution is indexed by its
population mean, and its population standard deviation. Recall
that the standard deviation is a useful yardstick for measuring
how far an individual value falls from the mean. The standardized
score or zscore is the distance between the observed value and
the mean, measured in terms of number of standard deviations.
Values that are above the mean have positive zscores, and values
that are below the mean have negative zscores.
Normal Approximation to the Binomial Distribution: The
easier way involves using a normal distribution. The normal
distribution can be used to approximate probabilities for other
types of random variables, one being binomial random variables
when the sample size n is large.
Expected Value: The expected value of a random variable is the
mean value of the variable X in the sample space, or population, of
possible outcomes. Expected value, denoted by E(X), can also be
interpreted as the mean value that would be obtained from an
infinite number of observations on the random variable.
Standard Deviation: The standard deviation can be viewed as
approximately the average distance of the possible values of X
from its mean.
63
64
Formula Card
65
WarmUp: Types of Variables
Todays typical undergraduate student is often characterized as
preferring teamwork, experiential activities, and the use of
technology. An ECAR (Educause Center for Applied Research)
study was published on technology use among undergraduate
students. The study used survey and interviewer data to create a
portrait of todays students experiences with and skill using
information technology.
1. Listed below are some of the response variables that were
measured in this study. For each of these determine whether it
is categorical, quantitative discrete, or quantitative continuous.
a. Technology ownership: Do you own a computer?
categorical
quantitative discrete
quantitative continuous
b. Time (per week in minutes) spent using a computer for writing
documents (word processing).
categorical
quantitative discrete
quantitative continuous
c. Which social networking site(s) are you a member? (facebook,
myspace, friendster, etc.)
categorical
quantitative discrete
quantitative continuous
2. Identify appropriate model for each variable. (Be complete).
a. Let X represent waiting time to be served at a deli.
The model for X is shown below.
66
X has a
___________________________ distribution
b. Suppose that 45% of Michigan residents own dogs.
Let X represent the number of Michigan residents with a dog in
a random sample of 10 Michigan residents.
X has a ___________________ ________________ distribution
c. Let X represent the score (in points out of 100) on a
standardized exam.
The model for X is shown below.
X has a
____________________________ distribution
ILP: Probability and Random Variable
Review
67
Problem 1: Study on Smiling
In a recent study people were observed for about 10 seconds in
public places (e.g. malls and restaurants) to determine whether
they smiled during the randomly chosen 10second interval. The
table shows the results for comparing males (group 1) and females
(group 2).
a. What is the probability that a randomly selected person smiled?
b. The researcher would like to assess if smiling status is
independent of gender.
i. To check for independence, the probability found in part (a)
should be compared to which of the following probabilities?
P(smiled and male)
P(smiled
given
male)
P(male given smiled)
ii. Find the probability selected
appropriate conclusion.
P(male)
above
and
circle
the
The probability = __________________________________
Thus it appears that smiling status is is not
independent of gender.
Problem 2: Summer Trip Length
Did high gas prices keep Americans from hitting the road this past
summer? In a nationwide survey of adults, one variable measured
was how many days vacationers spent driving on the road on their
longest trip.
Consider the following (partial) probability
distribution for the random variable X = the number of days for
the longest car trip.
X
4
5
6
7
8
Probab
0.10
0.20
0.25
ility
68
a. Suppose the probability of 7 days is twice as likely as the
probability of 8 days. Complete the probability distribution for
X. Show your work.
b. What is the expected number of days for the longest trip?
Include symbol, value, and units.
Problem 3: Surviving Trees A landscaping company claims
that 90% of the trees they plant survive (defined as being still
alive one year from planting). If a tree does not survive, the
company will replace the tree with a new one.
a A homeowner will have 5 trees planted in his yard by this
landscaping company. You can consider these 5 trees to be a
random sample of all trees planted by this company.
If
companys claim is correct, what is the probability at least 4
trees will survive?
An Office Space Developer will have 200 trees planted all
around a new office space building complex by this landscaping
company. You can consider these 200 trees to be a random
sample of all trees planted by this company. If the companys
claim is correct, what is the probability that 30 or more trees
will need to be replaced (i.e. not survive)?
Problem 4: How Much Time to You Spend Studying? A
Washington Post article Is college too easy? As study time falls,
debate rises (May 21, 2012) stated that the amount of time
college students actually study has dwindled from an average of
24 hours per week to about 15 hours (based on a survey).
A professor of statistics decided to ask all of his current semester
students to report the number of hours per week they spend
69
studying his course material (on a regular, nonexam week). The
mean for the female students was 10 hours and the standard
deviation was 3.5 hours.
a. Consider the following interpretations of the standard deviation
and clearly circle those that are correct.
On average, the number of hours spent studying statistics
varied from the mean by about 3.5 hours.
The average distance between the number of hours spent
studying statistics is roughly 3.5 hours.
The average number of hours spent studying statistics is
about 3.5 hours away from the mean.
b. Julie is one of his current students and she studies statistics for
about 6 hours per week. What is her corresponding zscore?
c. The male students had a lower mean and a larger standard
deviation. Suppose Jakes response corresponds to a zscore of
2.1. Complete the sentence to explain what this tells us about
the number of hours that Jake studies statistics per week. (Be
as specific as you can).
Jakes number of hours he studies statistics per week is
70
d. The professor was interested in how the results of his students
compare to those taking a Chemistry class. The distribution of
hours spent studying (per week) for students in the Chemistry
class were reported as being approximately normal with a
mean of 12 and a standard deviation of 3.
i
What is the probability a randomly selected Chemistry
student studies between 16 and 20 hours per week?
First find the z scores, and use the Ztable to estimate the
probability. Next, use the prob() R script to get a more exact
probability and to produce a nice picture and solution.
ii
Jing learns that she is in the top 30% of this distribution.
Based on the distribution, Jing must study at least how
many hours per week? Make a hand sketch of what you are
trying to find to help show your work.
CoolDown: True or False
Decide whether the following questions regarding probability and
random variables are true or false.
1. If the time to wait for pharmacy help has a uniform distribution
from
0 minutes to 30 minutes, then 33% of the customers are
expected to wait more than 20 minutes.
True
False
2.
If X has a Binomial (50, 0.7) distribution, then the criteria to
use the normal approximation are met.
True
False
3. 68% of test scores are always within one standard deviation of
the mean test score.
71
True
False
Example Exam Question on
Probability and Random Variables
Suppose that the amount of time spent waiting for your bus to our
campus each day is a uniform random variable between 0 to 20
minutes.
a. Sketch a picture of the model for waiting time for the bus.
Provide labels for each axis and some values along each axis.
Lets define the following events:
A is the event that you wait at least 10 minutes, that is your
waiting time is in the interval [10,20].
B is the event that you wait at most 15 minutes, that is your
waiting time is in the interval [0,15].
C is the event that you wait at most 10 minutes, that is your
waiting time is in the interval [0,10].
Answer the following questions based on the information given.
Show all work.
b. What is P(A)?
Final answer:
________________
c. What is P(A and B)?
Final answer:
________________
d. What is P(BA)?
Final answer:
________________
e. What is P(A and C)?
72
Final answer:
________________
f.
Are the events A and C mutually exclusive?
Yes
No
Explain briefly.
Circle one:
Lab 5: Confidence Intervals (CI)
for a Population Proportion
Objective: This module will help you better understand the ideas
involved in confidence interval estimation as well as how to
interpret both the confidence level and confidence interval for a
population proportion. You will construct onesample confidence
intervals for a population proportion, and to check that the
conditions necessary for the interval are valid.
Overview: Since, generally, a population proportion is an
unknown number, we are interested in seeing how close a sample
proportion (computed from a random sample) will be to its
corresponding true population value. Hence, it is important to
provide both the sample proportion (the estimate) in addition to a
statement that describes the precision of the estimation process.
Confidence intervals provide a method of stating both how close
the value of a sample proportion is likely to be to population
proportion and the accuracy of it being that close. The basic
structure for any confidence interval:estimate (multiplier x
standard error). The multiplier x standard error portion is also
called the margin of error (or error margin). The multiplier used
will depend on the confidence level and will be a z multiplier when
we are constructing an interval for a population proportion. In
contrast, the confidence level is the proportion of times the
method will produce an interval that contains the true
parameter in repeated random sampling.
Confidence Interval (CI) Summary
1. Principles for using Confidence Intervals to Guide
Decision Making:
Principle 1:
A value in a CI is an acceptable or
reasonable possibility for the value of a population
parameter. A value not in a CI can be rejected as a likely value
of the population parameter.
Principle 2:
When the CIs for parameters for two different
populations do not overlap, it is reasonable to conclude that the
73
parameters for the two populations are different.
2. The probability that the true parameter is a particular, already
computed confidence interval is either 0 or 1. Both the interval
and the parameter are fixed entities, so either the parameter is
in that particular interval, or it is not.
3. Interpreting a 95% Confidence Interval: We are 95%
confident that the true parameter value lies inside the CI. The
interval provides a range of reasonable values for the
population parameter.
4. Interpreting a 95% Confidence Level: If the procedure were
repeated many times (that is, if we repeatedly took a random
sample of the same size, and computed a 95% CI based on
each sample), we would expect 95% of the resulting CIs to
contain the true population parameter.
The interpretations in the previous CI summary can be used for
any problem by adjusting the confidence level and parameter
under study, given the context of the problem. Additionally,
phrases such as true parameter value and population should
be stated in context of the problem.
WarmUp: Confidence Level Interpretation
Background: A study was conducted to learn about eating habits
for American families. A random sample of 200 families was
selected, and an adult head of household was asked to complete a
survey. One question asked was, Did your family eat dinner
together last Sunday night yes, no?
Based on the results, a 95% (conservative) confidence interval for
the proportion of all such families that ate dinner together last
Sunday night is given by (0.56, 0.70).
Task: Here are two possible interpretations of the confidence level
given by a previous student in statistics.
Are the following
interpretations of the confidence level correct or incorrect?
1.
If this survey was repeated many times, we are confident the
true population proportion of families that ate dinner together
last Sunday night is represented in the data 95% of the time.
Correct
Not Correct
2. The 95% confidence level means that with this method, and for
74
similar samples, we would construct many confidence intervals,
and of these, 95% would contain the true population proportion
of families that ate dinner together last Sunday night.
Correct
ILP: Calculating
Proportion
Not Correct
CI
for
Population
In this ILP, we are going to investigate using the sample proportion
as a way of estimating the population proportion. For the purpose
of this activity, we are going to assume that the lab section is a
representative random sample of the UM student body. Follow the
steps below to estimate the population proportion.
1. Determine a question to investigate by filling in the following
question.
What
proportion
of
UM
students__________________________________?
2. Once the question has been determined, the instructor will
pose the question to the class using clickers.
Record the results below and provide an estimate of the
population proportion of all UM students that ___________ (that
is, report the sample proportion).
# of students responding:
# of student who responded yes:
Sample proportion of student who
responded yes:
(symbol and value)
3. (a)
The estimate in part (2) is a sample proportion.
About how far away from the population proportion would you
expect such estimates to be, on average? (i.e. report the
standard error of the sample proportion)
(b) Construct a ____ % confidence interval for the population
proportion
of
all
UM
students
that
_____
75
__.
(c) Interpret the interval you constructed in part (b).
(d) Explain what the corresponding confidence level means.
Note: There are certain conditions required for this inference
procedure to be valid. Are they met here? If not, how would
you update this study so the conditions would be met?
CoolDown:
More
Confidence Levels
Interpretations
of
Consider once again the survey of American families, and the
question of whether they ate dinner together last Sunday night.
Here are a few more possible interpretations of the 95%
confidence interval for practice. Can you tell which are correct and
which are not?
1. If the procedure is repeated many times, about 95% of the
confidence intervals will contain the true population proportion
of families that ate dinner together last Sunday night.
Correct
2.
Not Correct
In the long run, the population proportion of families that ate
dinner together last Sunday night will be found in the interval
0.56 to 0.70, 95% of the time.
Correct
Not Correct
76
3.
In data collected from a normal distribution, 95% of the
observed values will lie within two standard deviations of the
mean.
Correct
4.
Not Correct
If we take a large number of independent samples from the
population of all American families, 95% of the intervals we
calculate based on these samples will contain the true
population proportion of families that dinner together last
Sunday night.
Correct
5.
Not Correct
The data has a 95% confidence level that the true proportion
of families that ate dinner together last Sunday night is within
the calculated confidence interval of (0.56, 0.70).
Correct
6.
Not Correct
95% of all confidence intervals for different samples taken will
contain the true population proportion of families that ate
dinner together last Sunday night.
Correct
Not Correct
Example Exam Question on
Confidence Intervals for
Proportion
Population
1. Point Estimate and Standard Error
After once again losing a basketball game to their archrival, a
colleges alumni association conducted a survey to see if
alumni were in favor of firing the coach. A random sample of
100 alumni was taken and 70 were in favor of firing the coach.
a. Report the sample proportion of alumni in favor of firing the
coach and the corresponding standard error.
b. Which of the following is/are correct interpretation(s) of this
77
standard error?
i. If repeated samples of 100 alumni were obtained, we would
estimate the resulting sample proportions to be about 0.046
from the true population proportion p on average.
ii. If repeated samples of 100 alumni were obtained, we would
estimate the sample proportion of 0.70 to differ from the
true population proportion p by about 0.046.
iii. We would estimate that our sample proportion of 0.70 is
about 0.046 away from the true population proportion p on
average.
2. Construct a CI for population proportion p and Find the
Sample Size.
a.
Give a (general) 95% confidence interval estimate for the
population proportion of alumni in favor of firing the coach.
b.
Interpret the 95% confidence interval.
c.
Interpret the 95% confidence level.
78
d. Give a conservative 95% confidence interval estimate for the
population proportion of alumni in favor of firing the coach.
e. Is the conservative interval in part (d) wider or narrower than
the interval in part (a)?
f.
What sample size would be needed for a 95% (conservative)
confidence interval with an error margin of 0.05?
3. Point Estimate and Interpretation of the Confidence
Level
The 99% confidence interval for the population proportion of
members who participated in the fundraiser is given by (0.30,
0.40).
a. The above interval was constructed from the results of a
random
sample
of
n = 200 members.
i.
What was the proportion of members in the sample
that participated in the fundraiser?
ii.
How many members in the sample participated in the
fundraiser?
b. Does the 99% confidence level imply that P(0.30 p 0.40) =
0.99?
Yes No
Explain.
79
Lab 6: Hypothesis Testing (HT)
For Population Proportions
Objective: In this lab, you will learn an important statistical
technique that will allow you to answer the question, Was our
observation due to chance, or is it more significant?
The
objective is to guide you through the ideas behind tests of
statistical significance and the statistical language involved. This
lab first presents a general overview of testing. Then, the InLab
Project discusses the large sample z test for a population
proportion, as well as providing practice in answering the question
of whether an event can be attributed to chance.
Overview:
Hypothesis Tests: A test of hypotheses (or significance test)
is a procedure designed to assess what the evidence provided by
the data says about some statement about a population
parameter. The key elements of a statistical test include: a null
and alternative hypothesis, assumptions that must be checked, a
test statistic and its pvalue, a decision, and a conclusion. These
components depend on the distribution of the population from
which the data come as well as the question being studied.
Hypothesis Testing Steps:
1. Determine appropriate null and alternative hypotheses.
2. Check assumptions for performing the test and calculate the
test statistic.
3. Calculate the pvalue under the assumption the null
hypothesis is true.
4. Determine if the result is statistically significant.
5. Report a conclusion in the context of the problem.
The first step of a hypothesis test is to identify the hypotheses;
this step is crucial, as it dictates the procedures for the remainder
of the test. The null hypothesis, H0, represents the status quo or
statement of no effect. That is, it is the model from which the
experimenter would like to show a deviation. The alternative
hypothesis, Ha (or sometimes H1), represents the experimenter's
new model, or what the experimenter would like to show. In both
instances, the hypotheses are postulated about the population
parameter of interest. The alternative hypothesis can take three
different forms it may be a denial of the null hypothesis (uses ;
called a twosided test), or it may specify a direction of interest
80
(uses > or <; called a onesided test). Notice that we never test
for equality in Ha.
The purpose of a significance test is to assess whether or not the
observed data are consistent with the null hypothesis (within the
reasonable bounds of sampling variability). If the data seem
unlikely to occur if the null hypothesis is assumed true, then we
would reject the statement made in the null hypothesis. To help us
make this decision, we use a test statistic, which represents a
summary of the data. More specifically, a test statistic is a
random variable related to the hypotheses of interest that has a
known probability distribution (under the null hypothesis), and will
be examined for evidence in favor of or against H 0.
In hypothesis testing, another frequently reported value is the pvalue, a number that is used to indicate the degree of significance
of the data. The pvalue is the probability of getting a test
statistic as extreme or more extreme than the observed value of
the test statistic, assuming the null hypothesis is true. Here
extreme means in the direction of Ha, or providing more
evidence against H0 .
We must decide in advance how much evidence against H 0 we will
require for rejection. This designated amount of evidence is called
the level of significance, denoted by (alpha). Common values
of are 0.01, 0.05, and 0.10. If the pvalue is less than or equal to
, we make the decision to reject H0. If we reject the null
hypothesis, the results of the test are said to be statistically
significant at level . A significant result in the statistical sense
does not necessarily imply an important result in the practical
sense. It simply means that such a difference from the null
hypothesis is not likely to happen just by chance.
Of course, no procedure is perfect, and as such, there are two
types of errors possible during hypothesis testing. If the null
hypothesis is true but the decision is to reject H 0, then we say that
a Type I error has occurred. A Type II error occurs when the
alternative hypothesis is true, but we fail to reject H 0. Each type of
error has a probability of occurring. If the null hypothesis is true,
the level of significance, , is also the probability of a Type I error,
while the probability of a Type II error is denoted by .
Another important component of a hypothesis test is its power.
The power of a test measures its ability to detect an alternative
hypothesis when it is true. Power of a particular test is calculated
81
as the probability that the test will reject H 0 when the alternative
hypothesis is true. Since we just learned that is the probability
that we did NOT reject H 0 when the Ha is true, we can see that
power is represented by 1 .
Truth
H0
True
Ha
True
Decision
Made
Result
Associated
Probability
Reject H0
Type I Error
Do Not Reject
H0
Correct
Decision
Reject H0
Correct
Decision
1 = power
Do Not Reject
H0
Type II Error
A Few Notes:
1. In practice we want to protect the status quo, so we are usually
most concerned with Type I error. We set Type I error at a
constant value by fixing the significance level .
2. Most tests we describe have the smallest for given .
3. For a fixed sample size n, there is a tradeoff between and
decreasing one type of error increases the other.
Ideally, we want the probabilities of making a mistake to be small.
However, once a decision is made, it is either right or
wrong only prior to looking at the data can we talk about
probabilities of making these two types of errors.
Large Sample ZTest: A hypothesis test for a proportion is used
to test whether the proportion of a variable is significantly different
from some value. This value, the test value or p0 , is given in
the null hypothesis (H0: p = p0).
Conditions:
The test relies on one key assumption: that our sample size is
large enough. To achieve a large enough sample size, we need at
least 10 hypothesized yes answers and 10 hypothesized no
answers; namely
np0 10
and
1p
n( 0) 10.
We also must
assume are sample can be considered a random sample from the
larger population of interest.
82
Once the hypotheses are stated, we
carry out the test by first calculating
the
test
statistic,
z=
^p p 0
( p ) (1 p )/n
0
The test statistic tells us how many
standard
errors
the
sample
proportion is from the null test value,
p0 .
To summarize the statistical
importance of this distance, we
calculate the pvalue using the
distribution of the test statistic. Our
test statistic, z, has a standard
normal distribution.
Formula Card:
The relevant parts of your Stats 250
formula card are shown at the right.
83
WarmUp:
Stating the Hypotheses
and Defining the Parameter
Interest
of
For each example below fill in the hypotheses and define the
parameter of interest.
1. The Detroit Tigers advertising team believes that about 70 % of
Ann Arbor residents will attend a Tigers game this season.
Suppose the general manager wants to cut back on advertising
in the area and speculates that Ann Arbor residents are less
likely to attend a Detroit Tigers game this season than
previously assumed.
Ho:__________________________ Ha:___________________________
Let _
represent
2. The nationwide college graduation rate is about 60%. Suppose
that a researcher speculates that the University of Michigan
has a higher graduation rate than the nation.
Ho:__________________________ Ha:___________________________
Let _
represent
ILP: Hypothesis Testing for a Population
Proportion
In this ILP, we are going to use a sample proportion to test a
theory about the value of the population proportion. Again, we
assume the lab section is a representative random sample of the
UM student body.
1. Determine a question to investigate by filling in the following
84
question.
What proportion of UM students____________________________?
85
2.
Once the question has been determined, the instructor will
pose the question to the class using clickers.
Record the results below and provide an estimate of the
population proportion of all UM students that ___________ (that
is, report the sample proportion).
# of students responding:
# of student who responded yes:
Sample proportion of student who
responded yes:
(symbol and value)
3. (a)
The estimate in part (2) is a sample proportion.
About how far away from the population proportion would you
expect such estimates to be, on average? (i.e. report the
standard error of the sample proportion)
4. We wish to use the data collected to assess:
Do a (select one) minority
majority
of UM
students___________?
Express this as an appropriate null and alternative hypothesis:
H0:_____________________________________
Ha:_______________________________________
where p represents
86
5. A large sample Z test can be performed using a normal
approximation for the binomial if the sample size is large
enough.
Provide the checks necessary to see if this
approximation may be made.
6. Assume the sample is large
corresponding z test statistic.
enough,
calculate
the
What is the distribution of the test statistic under the null
hypothesis?
7. Use your test statistic to find the corresponding pvalue. First
make a quick sketch by hand and use Table A.1 to find your pvalue; then try the new pval() script in R to check your answer.
8. What is your decision at the 5% significance level?
Reject H0
Fail to Reject H0
Also write out your real world conclusion in the context of the
problem.
87
CoolDown: True or False
Determine whether the following statements are True or False.
1. If the pvalue for the largesample ztest turned out to be
0.028, then we would reject the null hypothesis at a 1% level.
True
False
2. The significance level, , is determined prior to the study by
considering the consequence of incorrectly rejecting H 0.
True
False
3. If the test of H0: p = 0.75 versus Ha: p > 0.75 resulted in a pvalue of 0.083, then the probability that H 0 is true is 0.083.
True
False
4. One condition for using a Z statistic for testing hypotheses
about a population proportion p is that the population has a
normal distribution.
True
5.
False
If the null hypothesis is actually true, then it is not possible to
make a Type 1 error.
True
False
6. A statistically significant finding does not necessarily have
practical significance.
True
False
88
89
Example
Testing
Exam
Question
for
Proportion
on
Hypothesis
Population
Obama versus McCain: A mock presidential election is planned
on the UM college campus. An enterprising Statistics 350 student
plans to use the results of the mock election to test the hypothesis
that Obama would have a majority of votes among all UM college
students. The significance level of the test is set at 5%. A random
sample of 3000 college students at the UM was selected to take
part in this mock election which resulted in 1578 voting for
Obama.
a. State the hypotheses to be tested and completely define the
parameter.
Ho:__________________________ Ha:___________________________
where the parameter _______ is defined (in the context of this
problem) as
b. Provide the value of the test statistic (including its symbol) and
the corresponding pvalue. Include a sketch to show how the pvalue is found.
Sketch (include labels):
Symbol
Test Statistic =
=
_______________
pvalue = _________________________
c. Give the decision and corresponding conclusion by circling the
appropriate statements:
The statistical decision at a 5% significance level is:
90
Reject Ho
Fail to reject Ho
Therefore, there
is
is not
sufficient
evidence to say that,
a majority of all students (represented by this sample) would
vote for Obama.
d. If this decision in part (c) is incorrect,
it would be a (circle one) Type 1
error.
Type
Lab 7: Sampling Distributions and
the CLT
Objective: In this lab, you will use some graphical and numerical
tools to summarize the distribution for a quantitative variable or
response a histogram, a boxplot, mean, median, standard
deviation, and interquartile range (IQR). You will
Objective: The objective of this lab is to give you a handson
discussion and understanding of sampling distributions and the
Central Limit Theorem (CLT), a theorem that plays an important
role in statistics. The sampling distribution of a statistic can be
obtained mathematically, but we will simulate the sampling
process and will observe the empirical sampling distribution of
various statistics. In this lab, you will simulate random samples
from a known population distribution and compute a sample
statistic for each of the generated samples.
The generated
sample statistics can be examined to learn about properties of the
sampling distribution of the statistic.
Overview: Statistical inference is the process of drawing
conclusions about a population parameter based on data. When a
sample is selected from a population, a summary number can be
computed from the observations resulting in the value of a
statistic. A statistic is used to estimate the corresponding value
for a population (that is, a sample statistic estimates a population
parameter).
However, a sample chosen at random will not
necessarily yield an estimate (or statistic) that is exactly equal to
the corresponding parameter for the population; the next selected
sample of the same size will probably give a different estimate
from the first one. If additional samples of the same size were
taken, you would begin to see how the possible estimates
91
(possible values of the statistic) vary and how close they tend to
be to the parameter value.
With a large number of samples, you can assess whether the value
of the statistic (e.g., sample mean, X ) will frequently be close to
the true value of the population parameter (e.g., population mean,
), and if so, how close on average. This can be seen more easily
through some pictures:
When data are gathered by random sampling, the statistic will be
a random variable and as such, it will have a probability
distribution. The probability distribution of the sample statistic is
called its sampling distribution.
Generally, if we use a statistic to make an inference about a
population parameter, we want its sampling distribution to be
centered at the true parameter (a characteristic which allows us to
call that statistic unbiased), and we would like variability in the
estimates to be as small as possible.
Below, we have two estimators that are both unbiased, but
Estimator I has less variability (is more precise). Thus, we would
prefer Estimator I to Estimator II.
We will next examine the sampling distribution of the sample
statistic most commonly used for measuring the center of a
distribution  the sample mean.
92
Formula Card:
93
WarmUp: Check Your Understanding
Suppose the time to wait for placing an order at a drivethrough
window has a uniform distribution between 0 and 8 minutes. The
mean waiting time is 4 minutes and the standard deviation is 2.3
minutes.
Suppose a random sample of 100 drivethrough window orders will
be selected from over the past week and the average waiting time
for placing the orders will be calculated. Which graph below best
represents the model that could be used to approximate the
probability that the average waiting time is at most 5 minutes?
ILP: How Do Sample Size and Distribution
of the Parent Population Affect the
Sampling Distribution of the Sample
Mean?
In this activity, you will observe the effects that sample size and
the distribution of the population you are sampling from have on
the sampling distribution of the sample mean. The sampling
distribution of the sample mean, X , is the distribution of the
sample mean values for all possible samples of the same size from
the same population.
Open the Lab 7 Sampling Distribution applet from the Links
to Lab Applets folder on the Stat 250 CTools site (Resources
Lab Info folder).
Alternatively, the original applet can be found at:
http://onlinestatbook.com/stat_sim/sampling_dist/inde
x.html
94
This applet will help you simulate sampling distributions for a
variety of statistics and allows you to vary the sample size and the
population from which the samples are taken.
Read the Applet instructions.
Press Begin to open the applet; you will see the screen pictured
below.
Notice that when the applet begins, a histogram of the normal
distribution with mean 16 and standard deviation 5 is displayed for
the default parent distribution.
The Sampling Distribution Applet has several options from which
you can choose:
The 1st histogram, the Parent Population histogram, is the
population from which the sample will be drawn. You can select
from Normal, Uniform, Skewed, or even customize the
distribution by selecting Custom and dragging the mouse over
the plot. For now, keep the default N(16, 5) distribution as the
parent population.
When you are done with a particular
simulation, you can click on Clear lower 3 button to clear the
remaining histograms, and select new settings for your next
simulation.
The 2nd plot, the Sample Data histogram, displays a
histogram of the sampled data. This histogram is initially
blank. You can select to draw Animated Sample, 5 Samples,
95
1,000 Samples,
population.
or 10,000
Samples
from the parent
The 3rd and 4th histograms show the distribution of
statistics computed from the sampled data. The number of
samples (replications) on which the 3rd and 4th histograms are
based is indicated by the label "Reps=," which is displayed
once the simulation is started. You can also control which
statistic to examine, as well as the sample size by using the
dropdown menu options to the right of each plot. (Note that
the applet uses N to denote sample size, whereas we generally
use n.)
96
The statistic options include:
Mean
Median
sd = standard deviation (uses N in the denominator)
Variance = variance of the sample (uses N in the
denominator)
Variance (U) = unbiased estimate of variance
(uses N1 in the denominator)
MAD = mean absolute value of the deviation from the
mean
Range
Select Mean as the statistic in the 3 rd histogram and a sample size
of 5 (default), then click on Animated Sample to draw one
sample of size n = 5 from the normal parent population. You will
see five observations appear in the 2 nd histogram, and the sample
mean of the five numbers will appear in the 3 rd histogram as a blue
rectangle. This graphically shows the process of attaining the
sample mean from one sample of size 5. Repeat this several times
and you will see how the sampling distribution of the sample mean
starts to form in the 3 rd histogram. Once you have a feeling of this
works, you can speed things up by choosing the larger sampling
options, such as 5, 1,000, or 10,000 samples.
3.
Select the Normal distribution as a parent population.
a. What are the mean and standard deviation of
population?
this
b. Select Mean (sample mean) as the statistic of interest in
both the 3rd and 4th histograms, sample size n = 5 for the 3rd
histogram, and n = 25 for the 4th. Do about 5 animated
samples, and then take 10,000 samples at once. Draw
rough sketches of each of the distributions of the sample
means. Make sure to label both axes.
n = 5:
n = 25:
c. How do the distributions of each sample mean in the 3 rd and
4th histograms compare with the parent population in the 1 st
histogram? Comment on shape, mean, standard deviation,
etc.
97
n = 5:
n = 25:
d.
Looking at the properties of the population and sample
distributions (displayed to the left of their respective
histograms), what can you say about the relationship
between the standard deviation of the sample mean and the
population standard deviation?
e.
What can you say about the relationship between the
sample size and the standard deviation of the sample
mean?
f.
Does the number of replications influence the shape of
the sampling distribution?
That is, as you take more
samples, does the shape of the sampling distribution
change significantly?
2. Clear the lower three graphs and then select the Skewed
distribution as a parent population.
a. Select Mean (sample mean) as the statistic of interest in
both the 3rd and 4th histograms, sample size n = 5 for the 3rd
histogram, and n = 25 for the 4th. Do about 5 animated
samples, and then take 10,000 samples at once. Draw
rough sketches of each of the distributions of the sample
means. Make sure to label both axes.
n = 5:
n = 25:
b. How do the distributions of each sample mean in the 3 rd and
4th histograms compare with the parent population in the 1 st
98
histogram? Comment on shape, mean, standard deviation,
etc.
n = 5:
n = 25:
99
c. How do the distributions of each sample mean in the 3 rd and
4th histograms compare to each other? Comment on shape,
mean, standard deviation, etc.
n = 5:
n = 25:
d. How do the distributions of each sample mean in the 3 rd and
4th histograms compare with those created of the sample
mean when the parent population was normal (in question
1)? Comment on shape, mean, standard deviation, etc.
n = 5:
n = 25:
e. What should be the value of the standard deviation of the
sample mean if the population standard deviation is 6.22
and the sample size is n = 25? (Show the calculation.) How
does this value compare to the standard deviation displayed
to the left of the 4th histogram created above?
3. Clear the lower three graphs and select the Custom
distribution as a parent population. The parent population plot
should be empty. To create a distribution, you will need to use
the mouse to click and drag on different parts of the parent
population graph until youve drawn a distribution that you like.
a. Provide a rough sketch of your custom population. Be sure
to note the mean and standard deviation.
100
b. Select Mean (sample mean) as the statistic of interest in
both the 3rd and 4th histograms, sample size n = 5 for the 3rd
histogram, and n = 25 for the 4th. Do a few animated
samples, and then take 10,000 samples at once. Draw
rough sketches of each of the distributions of the sample
means. Make sure to label both axes.
n = 5:
n = 25:
c. How do the distributions of each sample mean in the 3 rd and
4th histograms compare with the parent population in the 1 st
histogram? Comment on shape, mean, standard deviation,
etc.
n = 5:
n = 25:
d. How do the distributions of each sample mean in the 3 rd and
4th histograms compare to each other? Comment on shape,
mean, standard deviation, etc.
n = 5:
n = 25:
e. Considering the changes observed from n = 5 to n = 25 in
questions 2 and 3, what can you say about the shape of the
distribution of the sample mean with respect to the sample
size n?
f.
What should be the standard deviation of the sample mean
for samples of size n = 25 from your custom population?
(Show the calculation.) How does this value compare to the
standard deviation displayed to the left of the 4th histogram
created above?
101
102
CoolDown: Check Your Understanding
1. Fill in the blanks to summarize your findings in the ILP:
a. If the parent population is a normal distribution with a
mean and a standard deviation , then for any sample
size, the sample mean will have a __________________
distribution with a mean of __________ and a standard
deviation of _____________.
b.
If
the
parent
population
is
NOT
normal
distribution, but with a mean and a standard deviation ,
then for a large sample size, the sample mean will have
approximately a __________________ distribution with a
mean
of
__________
and
standard
deviation
of
_____________.
The result in 1(a) is known as the Sampling Distribution of
the Sample Mean. The result in 1(b) is known as the Central
Limit Theorem. While you should note that there are several
similarities between them, make sure you can see and
understand the difference between the two results.
2. Fill out the chart below to further summarize your findings
regarding the sampling distribution of the sample mean based
on the CLT.
Will the sampling
distribution of
sample mean be approx
normal?
Sample Settings
n = 10, Parent Population
Normal
n = 10, Parent Population NOT
Normal
103
n = 50, Parent Population
Normal
n = 70, Parent Population NOT
Normal
104
Example Exam
Distribution
Question
of
Mean
on
the
Sampling
Sample
For a particular community it is known that the mean amount of
water used per home during October is 1250 gallons and the
standard deviation is 325 gallons.
a. The distribution for amount of water used is skewed to the
right. Sketch a skewed right distribution below and label both
axes.
b. For a promotional campaign, a radio station plans to randomly
select
50 homes and pay their water bills for the month of October.
Describe the approximate sampling distribution of the
sample mean amount of water used for a random
sample of 50 homes. Provide all features of the distribution.
c. The radio station can afford to pay for a total of 67,000 gallons.
What is the probability that the total number of gallons for a
random sample of 50 homes will exceed 67,000 gallons?
Hint: Think about how a total and an average are related.
105
Lab 8: OneSample tTest
Procedures
Objective: In this lab, you will learn an important statistical
technique that will allow you to answer the question, Was our
observation due to chance, or is it more significant? The
objective is to guide you through the ideas behind tests of
statistical significance and the statistical language involved. This
InLab Project discusses the onesample ttest for a population
mean, as well as provide practice in answering the question of
whether an event can be attributed to chance.
Overview:
Hypothesis Tests: A onesample t test is a procedure designed
to assess what the evidence provided by the data says about
some statement about the population mean. The key elements of
the onesample t test include: a null and alternative hypothesis,
assumptions that must be checked, a test statistic t and its pvalue, a decision, and a conclusion.
Hypothesis Testing Steps:
1. Determine appropriate null and alternative hypotheses about
2. Check assumptions for performing the test and calculate the
test statistic t.
3. Calculate the pvalue under the assumption the null
hypothesis is true.
4. Determine if the result is statistically significant.
5. Report a conclusion in the context of the problem.
The first step of a hypothesis test is to identify the hypotheses;
this step is crucial, as it dictates the procedures for the remainder
of the test. A onesample ttest is used to test whether the mean
of a quantitative variable is significantly different from some value.
This value, the test value or 0 , is given in the null hypothesis
(H0: = 0), and is often taken to be zero (i.e., H 0: = 0). The
alternative hypothesis, Ha (or sometimes H1), represents the
experimenter's new model, or what the experimenter would like to
show. The alternative hypothesis can take three different forms
it may be a denial of the null hypothesis (uses ; called a twosided test), or it may specify a direction of interest (uses > or <;
106
called a onesided test). Notice that we never test for equality in
H a.
The test relies on two key assumptions: (1) the data is a random
sample and (2) the data are observations from a normally
distributed population. The Central Limit Theorem (see Lab 7)
allows the assumption of normality to be more relaxed if our
sample size, n, is large. (Generally, large means more than 30
observations, although it also depends somewhat on how serious
the data depart from normality.) In Lab 7 (CLT) you worked with an
applet that demonstrated the CLT using a large sample size of 25.
Thus we will accept at least 25 as large enough for inference about
means.
The purpose of a significance test is to assess whether or not the
observed data are consistent with the null hypothesis (within the
reasonable bounds of sampling variability). If the data seem
unlikely to occur if the null hypothesis is assumed true, then we
would reject the statement made in the null hypothesis. To help us
make this decision, we use a test statistic, which represents a
summary of the data.
For the onesample t test, our test statistic takes the form: t
x 0
s/n
The test statistic tells us how many standard errors the sample
mean, x , is from the test value, 0.
In hypothesis testing, another frequently reported value is the pvalue, a number that is used to indicate the degree of significance
of the data. The pvalue is the probability of getting a test statistic
as extreme or more extreme than the observed value of the test
statistic, assuming the null hypothesis is true. Here extreme
means in the direction of Ha, or providing more evidence against
H0. Our test statistic, t, has a tdistribution with n 1 degrees of
freedom (df), which is written as t(n 1).
We must decide in advance how much evidence against H 0 we will
require for rejection. This designated amount of evidence is called
the level of significance, denoted by (alpha). Common values of
are 0.01, 0.05, and 0.10. If the pvalue is less than or equal to ,
we make the decision to reject H0. If we reject the null hypothesis,
the results of the test are said to be statistically significant at level
.
A significant result in the statistical sense does not
necessarily imply an important result in the practical sense. It
107
simply means that such a difference from the null hypothesis is not
likely to happen just by chance.
Of course, no procedure is perfect, and as such, there are two
types of errors possible during hypothesis testing. If the null
hypothesis is true but the decision is to reject H 0, then we say that
a Type I error has occurred. A Type II error occurs when the
alternative hypothesis is true, but we fail to reject H 0. Each type of
error has a probability of occurring. If the null hypothesis is true,
the level of significance, , is also the probability of a Type I error,
while the probability of a Type II error is denoted by .
108
Formula Card:
WarmUp: Check Your Understanding
A researcher wants to determine if a filling machine at a cereal
plant fills the boxes well on average or if the machine needs to be
replaced. The machine is intended to fill the packages with 28
grams of cereal. He wants to test the machine to see if the
average weight of cereal is less than the intended 28 grams. He
performs a one sample t test and calculates his pvalue for the test
to be 0.027.
For which of the following levels of significance would the result be
statistically significant at?
1%
5%
10%
If the significance level was 5%, for which of the following pvalues
would your result be statistically significant?
0.0067
0.043
0.22
109
0.14
110
ILP: Is There Salmonella enteritidis in the
Ice Cream?
Background: A massive multistate outbreak of foodborne
illness was attributed to Salmonella enteritidis. Epidemiologists
determined that the source of the illness was ice cream. To
determine the level of Salmonella enteritidis in the ice cream, they
sampled nine production runs from the company that produced
the ice cream.
The levels of Salmonella enteritidis in Most
Probable Number/Gram (MPN/g) are given in the salmonella.sav
data set (Source Lyman Ott et.al., pg 232). The researchers would
like to use these data to determine whether the average level of
Salmonella enteritidis in the ice cream is greater than 0.3 MPN/g,
since such levels are considered very dangerous using a 5%
significance level. Time ordering of the data is present since the
sample is collected from production runs.
Task: Perform a test to assess if the average level of Salmonella
enteritidis in the ice cream is significantly larger than 0.3 MPN/g.
Recall: Complete the blanks in the Five Steps for conducting a
test of hypotheses.
1. Determine appropriate ____________ and __________________
hypotheses.
2. Check __________ for performing the test and calculate the
______________.
3. Calculate the
_______________is true.
under the assumption the
4. Determine if the result is
.
5. Report a
of the problem.
in the
Clarify: Before conducting any test, here are a set of questions to
ask yourself:
111
How many populations are there?
One
Two
More
than two
How many variables are there? One
What is the response variable?
What type of variable is the response? Categorical
Two
Quantitative
What type of parameter would be useful for summarizing
this response?
(See Supplement 3)
Proportion Mean
Other
Procedure: The appropriate inference procedure for this scenario
is the onesample t test and the specific parameter of interest is
___________________ .
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Hypotheses:
H0: ___________ = ___________ and Ha: ___________
___________ , where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
2. Checking the Assumptions and Computing the Test
Statistic
Assumptions:
a. For this scenario, we need to assume the data are a
___________ sample. To check this assumption, we would
make a ________ plot (if there was time order) of the
observations, and look for ____________________ .
b. We also need to assume that the responses come from a
__________________ distributed ___________________ . To check
this assumption, we would make a __________ plot.
112
c. Comment on each assumption, using graphs and output
when appropriate.
Normal QQ Plot of SAL
1.0
.8
.7
.8
.6
.6
.4
.4
.3
.2
.2
SAL
Expected Normal Value
.5
.1
.1
.2
.3
.4
.5
.6
.7
.8
0.0
1
Observed Value
Sequence number
d. Is the assumption of normality really that important with a
sample of size of 9? Why?
If the sample size were much larger, would this assumption
be as important? Why?
TestStatistic:
e. Generate the ttest output using Analyze> Compare
Means> OneSample T Test. Enter a test value of
_______ (this is the value you are testing against from the
null hypothesis).
f. What is the value of the test statistic? _____ =
_______________
g. Provide an interpretation of the test statistic value. (See
Supplement 6.)
h. What is the distribution of the test statistic
if the null hypothesis is true?
Note:
This is not the same as the distribution of the
population that the data were drawn from, and will be the
model used to find the pvalue.
3. Calculate the pvalue:
a What is the SPSS reported pvalue? _______________.
Is this the pvalue we want? __________.
Draw a picture of the desired pvalue.
113
Include labels for both the distribution and xaxis.
c. Ultimately, the pvalue we want is _____________
d. Provide an interpretation of the pvalue (see Supplement 6).
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
Note: Conclusions should include a reference to the population
parameter of interest. Your conclusion should not be too strong;
you can say you have sufficient evidence or something equivalent,
but NOT that you have proven anything.
ILP: Testing the Population Mean with pValue Practice
For each of the following sets of hypotheses and test statistic
values, provide a sketch of the pvalue. Assume the degrees of
freedom are 20 throughout. Then use your computer with the
pval() function in R (accessible under Lab Info folder, in the
Resources folder on the Stats 250 CTools site), your calculator (if
it has the capabilities), or the partial t distribution table provided
below, and compute the pvalue (or bounds for the pvalue).
Finally consider a significance level of 5%, and circle the
corresponding statistical decision.
df
1.28
Absolute Value of tStatistic
1.50
1.65
1.80
2.00
2.33
114
2.58
3.00
20
0.108
0.075
0.057
0.043
1. H0: = 20,
Ha: > 20,
Do Not Reject H0
Sketch:
0.030
0.015
t = 2.58
0.009
0.004
Reject H0
2. H0: = 6,Ha: > 6,
Reject H0
Sketch:
t = 2.12
3. H0: = 0,Ha: < 0,
Reject H0
Sketch:
t = 3.20
Reject H0
Do Not
4. H0: = 6,Ha: 6,
Reject H0
Sketch:
t = 1.87
Reject H0
Do Not
115
Reject H0
Do Not
CoolDown: Check Your Understanding
Note that while the alternative hypothesis and significance level
must be set in advance, we can think about what would have
happened if the alternative had originally been twosided.
Complete the following to see how the analysis of the ice cream
data in the InLab Project would change assuming the alternative
hypothesis was twosided.
How would the following change?
a.
ttest statistic value:
2.205
2.205
b.
Distribution of the ttest statistic assuming the null
hypothesis is true:
t(8)
c.
t(9)
t(16)
Interpretation of the value of the ttest statistic:
d.
pvalue:
0.0259
10.059
0.0295
e.
Decision at a 5% significance level:
Reject H0
CoolDown: Think About It
0.059
1
Reject H0
Fail to
If you were going to repeat the Salmonella study and wanted to
increase the power of the test, which of the following could you
do?
Increase alpha
Increase beta
116
Increase the sample size, n
Decrease the sample size, n
Example Exam Question on OneSample tTests
The Healthy Life
Company
produces
canned
whole
pineapples. The production line is supposed to yield cans
with an average weight of 1 kg. In order to assess if the
production line is operating properly, the quality control
statistician at Healthy Life selects a random sample of 36
cans produced over the course of the shift (keeping track
of the order number). The latest sample gave a mean
weight of 980 grams (1 kg=1000 grams). Based on this
sample, an estimate of the average distance that possible
sample means are from the true population mean is about
8.33 grams.
a. The value of 8.33 grams described above corresponds to what
statistical term or notation? (Be specific.)
WEIGHT
b. The hypotheses to be tested are H 0: = 1000 grams versus Ha:
1000 grams. Consider the
following graph of the data.
State what assumption can be
checked using the graph and
comment on the validity of the
assumption.
Assumption:
1100
1050
1000
950
900
Based on this graph, this
assumption appears:
valid
not valid because:
850
1
9 11 13 15 17 19 21 23 25 27 29 31 33 35
Order number
c. The test is to be performed using a 5% significance level. The
observed ttest statistic is t = 2.4 and the 2sided pvalue is
0.022. Consider the statements below and circle all that are
correct statements for this statistical analysis.
o The null hypothesis is rejected at the 5% level.
o The results are not statistically significant at the 5% level.
o The probability that the null hypothesis is true is 0.022.
o If this process were operating as it should and repeated
random samples of 36 cans were obtained, we would
117
observe a t statistic of 2.4 or smaller or a t statistic of 2.4
or larger in about 2.2% of the repetitions.
d. What would have been the pvalue
i. if Ha: < 1000 grams?
ii. if Ha: > 1000 grams?
e.
Compute a 95% confidence interval for the population mean
weight. Do you expect your interval to contain 1000? Why?
Lab 9: Paired tTest Procedures
Objective: In this lab, you will learn how to perform a hypothesis
test in the case when we have two quantitative variables collected
in pairs, called a paired ttest. You will make a confidence
interval for and test hypotheses about the population mean
difference, d.
Using these, you will be able to provide a
statement about how confident you are regarding your interval
estimate or in your decision.
Overview: Matched or paired data results from a deliberate
experimental design scheme.
For example, suppose we are
examining the effect of a drug on a certain type of response. The
drug is administered to a group of people, and responses for each
individual can be measured both before and after the drug is
given. Or consider an experiment where rats are matched by
weight, where one rat in each match receives a new diet and the
other rat in the match receives a control diet. These types of
design are called paired data designs. Note that paired designs
can occur when you have two measurements on the same
individual OR when you have two individuals that have been
matched or paired prior to administering a treatment.
The inference procedures for a paired data design are based on
the onesample ttest procedures from Lab 8. The change is that
we are interested in estimating or testing hypotheses about the
population mean difference d, which is generally compared to a
hypothesized value of zero, indicating no difference on average.
Formula Card:
118
WarmUp: Check Your Understanding
Eight cars were run to determine their mileage, in miles per gallon
(MPG). Then each car was given a tuneup, and run again to
measure the mileage a second time. The difference in mileage was
computed (diff = After MPG minus Before MPG). Assume that the
selection of 8 cars represents a random sample of cars.
1. Circle the appropriate alternative hypothesis to assess if on
average mileage significantly improves after a tune up.
a. Ha: d > 0
b. Ha: d < 0
c. Ha: d
2. The researcher of course is hoping that the results of the
experiment are statistically significant. What type of pvalue
would the researcher want to obtain? Circle your answer.
a. A large pvalue.
b. A small pvalue.
c. The magnitude of a pvalue has no impact on statistical
significance
ILP: Do Books Purchased from Barnes and
Noble (Instore) Cost More on Average than If
Purchased Online at Amazon.com?
119
Background: In recent years, the popularity of purchasing
books via the Internet has increased dramatically, and the
conventional bookstore no longer dominates the sales of books.
The most influential factor that sways customers into purchasing
books online is lower prices when compared to local bookstores. A
group of Statistics 250 students decided to perform a comparison
of the Amazon.com prices versus Barnes and Noble bookstore
(Ann Arbor) prices based on a sample of 40 books, selected from a
wide range of categories. For Amazon, a standard ground shipping
of $4.29 and local state tax of 6% were included in the cost. The
corresponding costs are available in the SPSS data set called
books.sav. Do the data provide sufficient evidence to conclude
that, on average, Barnes and Noble (instore) books are more
expensive than Amazon.com books?
Task: Perform a test regarding the mean difference in book price,
D , where the differences are computed as Barnes and Noble less
Amazon (i.e. price at Barnes and Noble minus price on
Amazon).
Clarify: Before conducting any test, here are a set of questions to
ask yourself:
How many populations are there?
One
Two
More
than two
How many variables are there? One
What is the response variable?
What type of variable is the response?
Two
Categorical
Quantitative
What type of parameter would be useful for summarizing
this response?
(See Supplement 3)
Proportion
Mean
Other
Procedure: The appropriate inference procedure for this scenario
is
a
paired
ttest,
and
the
specific
parameter
of
interest
is
___________________ .
NOTE: Why is this a paired procedure?
120
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Hypotheses:
H0: ___________ = ___________ and Ha: ___________
___________ , where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
2. Checking the Assumptions and Computing the Test
Statistic
Assumptions:
a. For this scenario, we need to assume that the sampled
differences are a ________________ sample. To check this
assumption, we would make a ________ plot (if there was
time order) of the ___________________ , and look for
____________________ .
b. We also need to assume that the ___________________ of
differences is normally distributed. To check this
assumption, we would make a __________ plot of the
_________________________________.
We will assume the assumptions are reasonable for this
example.
121
TestStatistic:
d. Generate the paired ttest output using Analyze>
Compare Means> PairedSamples T Test.
What is the value of the test statistic? ____ = ____________
e. What is the distribution of the test statistic if the null
hypothesis is true?
Note:
This is not the same as the distribution of the
population that the data were drawn from, and will be the
model used to find the pvalue.
3. Calculate the pValue:
a. What is the SPSS reported pvalue? _____________ .
Is this the pvalue we want? ________
b. Draw a picture of the desired pvalue.
Include labels for both the distribution and xaxis.
c. Ultimately, the pvalue we want is _____________
d. Provide an interpretation of the pvalue (see Supplement 6).
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
122
Note: Conclusions should include a reference to the population
parameter of interest. Your conclusion should not be too strong;
you can say you have sufficient evidence or something equivalent,
but NOT that you have proven anything.
CoolDown: Setting Up Appropriate
Hypotheses
and Writing
Conclusions
1. The Major League Baseball Organization would like to assess
whether or not players have the same batting averages during
the nighttime games and the daytime games at a 5%
significance level. Eight players are selected at random and
their nighttime and daytime batting averages are collected for
a given period of time. Let d represent the population mean
difference between the nighttime minus the daytime batting
average.
Write out the corresponding hypothesis using the appropriate
notation.
H0:_______________________
Ha:_______________________
Why is this a paired ttest?
The pvalue for this paired ttest was found to be 0.028. Make a
decision for this test and write a conclusion in context of the
problem.
2. Data was collected on the height of male college students and
their fathers.
Suppose we are interested in testing the
hypothesis that sons are taller than their fathers, on average
using a 5% significance level. Let d represent the population
mean difference between the sons height minus the fathers
height.
Write out the corresponding hypothesis using the appropriate
notation.
123
H0:_______________________
Ha:_______________________
Why is this a paired ttest?
The pvalue for this paired ttest was found to be 0.30. Make a
decision for this test and write a conclusion in context of the
problem.
124
3. In a recent survey, college students were asked the amount of
time they spend watching television on a typical day and the
amount of time they spend surfing on the Internet on a typical
day. The researchers were interested in determining whether
the time spent watching television was lower than the time
spent surfing the internet, on average at a 5% significance
level. Let d represent the population mean difference between
the amount of time spent watching TV minus the amount of
time spent surfing the Internet.
Write out the corresponding hypothesis using the appropriate
notation.
H0:_______________________
Ha:_______________________
Why is this a paired ttest?
The two sided pvalue for this paired ttest was found to be
0.05. Make a decision for this test and write a conclusion in
context of the problem.
Example Exam Question on Paired tTests
A study was conducted to see how often two rooms of a sports
facility were being used during the lunch hour. The number of
people in each room was counted at 12:30 PM each Monday for 10
weeks. The results are summarized below.
Suppose you want to test the hypothesis of no difference between
the utilization of the two rooms against the alternative that the
dance studio is used by more people during the lunch hour on
average.
125
You enter the above data into SPSS, and conduct a paired ttest to
obtain the following output.
P
ire
a
9
P
1
S
# a
3
#
ire
M td
e .D
u ri
s
.
6
5
0
u
s
d
5
a
D
%
In
ife
C
te
D td
S
a e ML n v e o
in
d
g
. 1
3
4 .6
in
w
g
le
re
o
vr
n
n
a
i E
fe
re ro
.
ia a wU
tio n p
a
n
9.2 7
2 .4 2
c
f
id
lo
n r
n
c
2 .1
e
e
fth
n
e
s
c
eS
et
e
3
h
ig
.0
d r ig
.(2r
s
9
6
tro
ta
t
u
o
ile
2
d
m
a. The observed test statistic is given as t = 2.13. State what this
value tells you about the location of the sample mean
difference of 3.6.
b. State the appropriate null and alternative hypotheses.
H0: ___________ = ___________ and Ha: ___________
___________ ,
where __________ represents:
c. Report the pvalue for the test in (b) and decision at a 10%
significance level.
pvalue: _______________ Decision: (circle)
Fail to Reject H0
Reject H0
d. A decision was made in (c). Which type of error could have
been made?
Use the appropriate statistical name to identify the mistake:
__
e. Which statements are assumptions required for performing
the paired ttest?
i.
The population standard deviation for the difference in room
use is known.
126
d
io
)

ii. The numbers using the dance studio are normally
distributed.
iii. The difference in room use is normally distributed.
iv. The standard deviation for the number using the dance
studio is equal to the standard deviation for the number
using the weight room.
v. The numbers using the dance studio are independent of the
numbers using the weight room.
Lab 10: Two Independent Samples
tTest Procedures
Objective: In this lab, you will learn an important statistical
technique that will allow you to compare two populations with
respect to their means by looking at 12. The aim is to help you
understand the ideas behind confidence intervals, tests of
significance, and the statistical language involved in the
comparison of two population means.
Overview: The two independent samples t procedures
(sometimes called Student's t procedures) are used when you
want to compare the means of two populations that are not related
or matched in any way. For example, you might want to test if the
GPA of English majors is different than that of Math majors. The
idea is that we use the two sample means to estimate the
corresponding population means. We may want to construct a
confidence interval to estimate the difference between the two
population means. Or, we may wish to test if the difference in the
two population means equals a value specified in a null
hypothesis; generally, the hypothesized test value is 0.
We will assume that we have a random sample from each of the
two populations and that the variable being measured has a
normal model for each population, although possibly with different
means.
In addition, the two random samples are assumed
independent of each other. However, there are two versions of
independent samples t procedures: pooled and unpooled (also
known as general). In a pooled independent t procedure, we
assume equal population variances for the two populations.
Whether or not we can assume the two populations have equal
variances will dictate which of the two independent samples t
127
procedures we use. Details on determining which version to use
are below
Checking the Equal Population Variance Assumption: There
are several ways to check the equal population variances
assumption required for a pooled ttest. Which one you will use
will often depend on the information provided.
Sidebyside Boxplots: Here, examine the IQRs of the
sample data. If they are similar, then the assumption is valid.
(NOTE: Similar in this sense does not mean that the boxes
need to line up right next to each other. It means that the
lengths or sizes of the boxes should be similar.)
Sample Standard Deviations: Since variance is standard
deviation squared, if the sample standard deviations are
similar, then the assumption is valid. (NOTE: Here, similar
means roughly that one should not be more than two times the
other.)
Levenes Test for Homogeneity of Variances: Levenes
test is a hypothesis test, but it is a bit different from hypothesis
tests you have seen previously since it is a test about variances
rather than means. Really, it is a hypothesis test conducted
before another hypothesis test. The null hypothesis is that the
populations of interest do have equal variances (H 0: 12 = 22)
and the alternative is that they do not have equal variances
(Ha: 12 22). Thus, when the null hypothesis from the
Levenes test is NOT rejected, it is reasonable to perform the
pooled ttest for two population means.
And if the null
hypothesis from Levenes test IS rejected, then the assumption
of equal population variances is unreasonable, and the
unpooled (general) version should be used instead of the
pooled. To make this determination, we will use an alpha of
10% for Levenes tests.
Formula Card:
128
WarmUp: Check Your Understanding
A student researcher, Jackie, is interested in learning about how
much time students spend studying per week. She is curious to
find out if study time is higher for female students (group 1)
versus male students (group 2). Data from a survey of 92
undergraduate students is collected. One question asked students
to report the number of hours that they study in a typical week.
An independent tTest was run and the difference in the two
sample means, x 1x 2 has been computed. This difference is 3
standard deviations above the presumed difference of 0 for
12 .
H0: ___________ = ___________
___________
and
Test Statistic Value: ___________________
Which of the following
sketches correctly
displays the pvalue?
If the pvalue for this
test was 0.076, are the
results significant at a
5% level?
Yes
No
129
Ha: ___________ =
ILP: Do Men and Women Differ in their Mean
SSHA Score?
Background: A total of 38 college freshmen at a private
college were administered the Survey of Study Habits and
Attitudes (SSHA), a psychological test designed to measure
motivation and attitude towards study habits in college students.
The sampled students were a simple random sample consisting of
18 females and 20 males. Scores on the test range from a low of 0
to a high of 200 and it is known that they may explain collegiate
success. School administrators are interested in whether or not
there is a difference between the mean scores for males and
females. The scores for females (Group 1) and males (Group 2)
are listed in the SSHA.sav data set. (Source: Moore and McCabe
(1999), pg 563.
Task: Perform a test to assess if there is a difference between the
mean score for women on the SSHA and the mean score for men
on the SSHA.
Clarify: Before conducting any test, here are a set of questions to
ask yourself:
How many populations are there?
One
Two
More
than two
How many variables are there? One
What is the response variable?
What type of variable is the response?
Two
Categorical
Quantitative
What type of parameter would be useful for summarizing
this response? (Supplement 3)
Proportion Mean
Other
Procedure: The appropriate inference procedure for this scenario
is the two independent samples ttest, and the specific parameter
of interest is ___________.
NOTE: Why is this not a paired procedure?
130
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Hypotheses:
H0: ___________ = ___________ and Ha: ___________
___________ , where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
2. Checking the Assumptions and Computing the Test
Statistic
Assumptions:
a. For this scenario, we need to assume that the two samples
are ____________________ from each other.
b. We need to assume that each sample is a ________________
sample.
To check this assumption, we would make a ________ plot (if
there was time order) for each sample, and look for
____________________ .
c. Each sample needs to come from a normally distributed
______________ . To check this assumption, we would make a
_____________ plot
for each ______________ .
d. Finally, in order to determine whether to use a pooled or
unpooled test, we need to check the assumption that both
populations have equal ______________________________ .
e. Comment on each assumption, using graphs and output
when appropriate.
i. Are the two samples independent?
ii. Are the samples random samples?
131
iii. Note there is no time order for this data. If there were,
since you need EACH sample to be a random sample,
how many time plots would you need to make to check
this assumption? _______ time plot(s)
132
iv. Does it appear that the assumption that each sample
comes from a normally distributed population is met?
Why?
Note: The equal population variances assumption will be
considered
after
the
ttest output is generated. The ttest output will include the
results from Levenes test that allows us to decide between the
pooled and general two independent samples t procedure.
TestStatistic:
e. Generate the ttest output using Analyze> Compare
Means> IndependentSamples T Test.
f.
The output provides both the pooled (equal variances
assumed) and unpooled (equal variances not assumed)
versions of the test. Use the output as well as the provided
sidebyside boxplots to determine which version of the test
to use utilizing all three methods discussed earlier.
i.
Sidebyside Boxplots:
Sidebyside boxplots show that the IQRs are: similar
not similar
ii. Sample Standard Deviations: The standard deviation for
females is ___________ and ___________ for males. This
indicates that the standard deviations are:
similar
not similar
iii. Levenes Test: Write the appropriate hypotheses for
Levenes test using correct notation:
H0: ______________________
133
Ha: ______________________
The Levenes test statistic is ____ = ____________ , with a
pvalue of ____________ . Therefore, we
can
cannot
reject the null hypothesis that the population
variances are equal.
As indicated by the previous Levenes test decision, we
can say that the assumption of equal population
variances is
is not
valid.
Thus, the symbol for the estimate of the common
population standard deviation is _______. Calculate its
value:
A Note on Output: Be careful not to confuse the Levenes test pvalue with the pvalue from the two independent samples ttest,
since your output will contain both. Levenes tests pvalue is
labeled Sig. in the Levenes section of the output. Further, note
that the first line of output corresponds to the pooled procedure,
while the second line corresponds to using an unpooled procedure.
g. Based on the version of the test you decided to conduct
(pooled or general) in part (f), what is the value of the test
statistic? Make sure you are looking at the correct line of
output (see note).
h. What is the distribution of the test statistic if the null
hypothesis is true?
Note:
This is not the same as the distribution of the population
that the data were drawn from, and will be the model used to find
the pvalue.
3. Calculate the pValue:
c. What is the SPSS reported pvalue? _____________ .
Is this the pvalue we want? ________
d. Draw a picture of the desired pvalue.
Include labels for both the distribution and xaxis.
134
c. Ultimately, the pvalue we want is _____________
d. Provide an interpretation of the pvalue (see Supplement 6).
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
Note: Conclusions should include a reference to the population
parameter of interest. Your conclusion should not be too strong;
you can say you have sufficient evidence or something equivalent,
but NOT that you have proven anything.
6. Confidence Intervals (CI):
a. What is the 95% confidence interval given in the output for
the difference in the two population mean scores?
a. Based on the confidence interval, would you reject the null
hypothesis at a 5% significance level?
Circle one: Yes
No
Explain.
b. Did your conclusion here match the one you made in part 4?
135
Circle one: Yes
No
c. Provide an interpretation of the 95% confidence interval.
d. Provide an interpretation of the 95% confidence level.
136
CoolDown: Check Your Understanding
1. T F
The test statistic of 2.032 in the activity implies that
the sample means differed by 2.032 pooled standard
errors.
2. T F
The assumption required by the pooled test is that
the samples
have equal variances.
3. Consider the following sets of boxplots of scores between two
age groups.
Set 1
Set 2
Set
a. Which sets that indicate a pooled test is appropriate:
Set 1
Set 2
Set 3
b. For which sets are you most likely to reject the null
hypothesis that the population mean scores are equal?
Set 1
Set 2
Set 3
Example Exam Question on TwoSample tTest
In an animallearning experiment, a researcher wanted to assess if
a particular drug speeds learning. One group of 5 rats (Group 1 =
control group) is required to learn to run a maze without use of the
drug, and a second, independent group of 8 rats (Group 2 =
experimental group) is administered the drug. The running times
(time to complete the maze) for the rats in each group were
entered into SPSS.
137
Group Statistics
Time to complete maze
Group
Control
Experimental
Mean
46.80
38.38
5
8
Std.
Deviation
3.42
4.78
Std. Error
Mean
1.53
1.69
Independent Samples Test
Levene's Test for
Equality of Variances
Run Time
Equal variances assumed
Equal variances not assumed
F
1.09
Sig.
.32
ttest for Equality of Means
t
3.41
3.70
df
11
10.653
Sig.
(2tailed)
.006
.004
Mean
Difference
8.42
8.42
Std. Error
Difference
2.47
2.28
Conduct the two independent samples ttest to address
the theory of the researcher.
State the null and alternate
hypotheses, report the appropriate test statistic and pvalue, and
state your decision and conclusion (in context) at the 5% level of
significance.
H0: _______________________
Ha: _______________________
Based on the above output, which results should you report?
Pooled or unpooled? Explain.
Report the results for testing about the population mean times:
Test statistic: ______________
pvalue: __________________
__
Decision: (circle one)
Fail to reject H0
Reject H0
Thus
Would the corresponding 95% confidence interval for the
difference in population mean times contain the value of 0?
138
Is it appropriate to use a 95% confidence interval to perform the
hypothesis test above? Why or why not?
139
Lab 11: OneWay Analysis of
Variance (ANOVA)
Objective: In this lab you will perform a oneway analysis of
variance, often abbreviated ANOVA. We have already seen that
the two independent samples ttest can be used to compare the
means of two populations (when the samples are independent).
However, when we want to compare the means of three or more
populations, we turn to ANOVA. You can think of ANOVA as an
extension of the two independent sample pooled ttest since it
compares several population means and requires the assumption
that the populations have equal variances.
Overview: ANOVA is a statistical tool for analyzing how the mean
value of a quantitative response (or dependent) variable is
affected by one or more categorical variables, known as treatment
variables or factors. For example, we might administer a new
antibiotic drug to a random sample of people. The response
variable is white blood cell count and the grouping variable or
factor could be age with 4 levels: 1 = 0 to 19 years, 2 = 20 to 29
years, 3 = 31 to 40 years, and 4 = 40 years and older. We would
then use the ANOVA method to see if the mean white blood cell
count for the four age group populations are all the same. In this
example, the number of populations under study is k = 4. Another
example would be a study of painkillers for relief of headache pain.
The response variable might be the time to relief and the factor or
treatment variable might be the type of painkiller, using different
levels: new drug, standard drug, and a placebo. In this example
the number of treatment groups is k = 3.
While ANOVA allows us to compare the means of more than two
populations, it can only tell us whether differences exist, not
specifically which means vary. Consequently, the appropriate
hypotheses for ANOVA are H0: 1 = 2 = = k (that the
population means are equal, where k is the number of treatment
groups) and Ha: at least one of the population means, i, is
different.
As with our other hypothesis tests, several
assumptions are required for ANOVA. The response for each
population is assumed to be normally distributed with equal
variance across the populations. The equal population variance
assumption here can be checked the same way as the two
independent samples ttest sidebyside boxplots, comparing
sample standard deviations, and Levenes test. Further, the data
are assumed to consist of independent random samples.
140
ANOVA involves decomposing the total variation of the responses
into two parts:
(1) that due to the variation among sample means (between
groups variation), and (2) that due to natural variation within
groups (variation due to error, or within groups variation).
SS Total = SS Groups + SS Error
If the sum of squares between groups (SS Groups) is large relative
to the sum of squares within groups (SS Error), it implies that the
model of different treatment means explains a significant portion
of the observed variability.
In order to determine what is
"relatively large," the sum of squares values are divided by their
respective degrees of freedom, creating what are called mean
square terms. The degrees of freedom for SS Groups is the
number of treatment groups, k, minus one (k  1), while for SS
Error, the degrees of freedom is the total number of observations,
N, minus the number of treatment groups (N k).
MS Groups = SS Groups/(k 1)
MS Error = SS Error/(N k)
The ratio of these two mean squares forms the Fstatistic, which
has numerator degrees of freedom (k  1) and denominator
degrees of freedom (N  k). Note in the following equation that
MSE stands for MS Error.
Variation among sample means MS Groups
Natural variation within groups
MSE
We can view this Fstatistic as the ratio of two estimators of the
common population variance, 2, where the denominator (MSE) is
a good (unbiased) estimator, and the numerator (MS Groups) is
only good when H0 is true (otherwise, it tends to overestimate 2).
Thus, large F values are evidence against the null hypothesis of
equal population means.
If we reject the null hypothesis, indicating at least one population
mean is different, then we turn to a multiple comparisons
procedure for determining which population mean(s) appear to
be different and how they differ. A common method is to look at
all pairwise comparisons. Two equivalent techniques can be used
for each pair of means: either perform a hypothesis test to see if
two population means are significantly different, or construct a
confidence interval for the difference in population means to see
whether the value of 0 is contained in the interval. Specifically, a
multiple comparisons procedure called Tukeys procedure, which
141
is available in SPSS, controls for the overall Type I error rate
(overall significance level) or the overall confidence level.
Formula Card:
WarmUp: Content of TV Shows
A study examined whether the content of TV shows influence the
ability of viewers to recall brand names of items featured in the
commercials. The researchers randomly assigned 90 adults to
watch one of three programs (30 to each). One program had
violent content, another sexual content, and the third neutral
content. Each program contained the same nine commercials.
After the shows ended, the subjects were asked to recall the
brands of products that were advertised. The ANOVA table based
on the data is provided below.
a. The researcher would like to assess if there is any effect of the
program type on the average number of brands recalled,
namely, test H0: 1= 2= 3. Provide the appropriate alternative
hypothesis in the context of the problem.
Ha: _____________________________________
b. For the inference to be valid certain conditions should be met.
We assume:
#1
Each sample is a random sample (the viewers are
representative of the populations of interest).
#2
The three samples are independent.
#3
The model for number of brands recalled for each
population is normal.
142
#4
i.
The three populations have equal standard
deviation for number of brands recalled.
Clearly state what graph(s) you would make to assess
assumption #3.
ii. Provide an estimate of that common population standard
deviation referred to in assumption #4. Include the units.
c. The ANOVA table produced by SPSS gave a test statistic value
of
7.47
and
a corresponding pvalue of 0.001. What distribution was used to
find
the
pvalue? Be specific.
ILP: Is There a Difference Among the Mean
Freshman GPAs for Three Different
Socioeconomic Classes?
Background: Sociologists often conduct experiments to
investigate the relationship between socioeconomic status and
college performance. Socioeconomic status is generally partitioned
into 3 classes: lower, middle, upper. Consider comparing the
mean grade point average (GPA) of college freshmen across the
three socioeconomic populations. The GPAs for random samples
of seven college freshmen from each of the three socioeconomic
classes were selected from a universitys files at the end of the
first academic year. The data are in the gpa.sav data set.
(Source: Mendenhall and Sincich, 1996, pg 589)
Task: Perform a test to assess whether the mean freshman GPAs
among the three socioeconomic classes differ. If there is sufficient
evidence to indicate significant differences, determine which
groups differ and how.
Clarify: Before conducting any test, here are a set of questions to
ask yourself:
143
How many populations are there?
One
Two
More
than two
How many variables are there? One
What is the response variable?
What type of variable is the response?
Categorical
Two
Quantitative
What is the explanatory variable?
What type of variable is the explanatory variable?
Categorical
Quantitative
What type of parameter would be useful for summarizing
this response, considering the explanatory variable? (See
Supplement 3)
Proportion Mean
Other
Procedure: Use your answers to these questions, to guide you to
the appropriate inference procedure. Refer to Supplement 3: Name
that Scenario for assistance.
The appropriate inference procedure for this scenario is ANOVA
and the value of k for this problem is ___________________ .
144
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Hypotheses:
H0: ___________ = ___________ and Ha: ___________
___________ , where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
2. Checking the Assumptions and Computing the Test
Statistic
Assumptions:
a. For this scenario, we need to assume that the k samples are
____________________ from each other.
b. We need to assume that each sample is a ________________
sample.
To check this assumption, we would make a ________ plot (if
there was time order) for each sample, and look for
____________________ .
c. Each sample needs to come from a normally distributed
___________________ . To check this assumption, we would
make a __________ plot for each ______________ .
d. Finally, for ANOVA, we need to check the assumption that all
k populations have equal ___________________ .
e. Comment on each assumption below, using graphs and
output when appropriate.
i. Are the three samples independent?
ii. Are the samples random samples?
145
146
iii. Note there is no time order for this data. If there were,
since you need EACH sample to be a random sample,
how many time plots would you need to make to check
this assumption? _______ time plot(s)
iv. Construct the QQ plots necessary to check the
assumption about normally distributed populations.
(Recall that in order to split a data file, the command is:
Data> Split File.) Does it appear that the assumption
that each sample comes from a normally distributed
population is met? Why?
Note: The equal population variances assumption
considered after the ANOVA output is generated.
will
be
TestStatistic:
f. Generate the ANOVA output using Analyze> Compare
Means> OneWay ANOVA. (Make sure you have turned
off any split file features you may have been using in your
assumptions checks.)
Under Options, select the
Descriptive (which gives you sample means and standard
deviations) and the Homogeneity of variance test
(Levenes test) options.
i.
Choose a way to determine if the assumption of equal
population variances is valid. Check the assumption and
comment.
ii. The assumption of equal population variances appears to
be:
valid
not valid
iii. The symbol for the estimate of the common population
standard deviation is ____ . For ANOVA, this value can
be obtained by computing
__________ , which is equal
to ____________ .
g. What is the value of the test statistic? ____ = ____________
147
h. What is the distribution of the test statistic if the null
hypothesis is true?
Note:
This is not the same as the distribution of the
population that the data were drawn from, and will be the
model used to find the pvalue.
148
3. Calculate the pValue:
e. What is the SPSS reported pvalue? _____________
b. Draw a picture of the desired pvalue. Include labels for
both the distribution and xaxis. Use the pval() function in R
to check your work.
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
Note: Conclusions should include a reference to the population
parameter of interest. Your conclusion should not be too strong;
you can say you have sufficient evidence or something equivalent,
but NOT that you have proven anything.
149
150
6. Followup Analyses: If ANOVA has indicated there appears to
be significant differences between two or more of groups, we
can use a multiple comparison test to tell us which groups
appear to be different and by how much.
a. Obtain the multiple comparisons output using Analyze>
Compare Means> OneWay ANOVA. Under Post Hoc,
choose Tukey, which has a default significance level of
0.05. The multiple comparisons output contains both
pvalues and confidence intervals for every possible
pairwise comparison of groups, and either can be used to
determine where differences exist. The pvalues that are
equal to or smaller than 0.05 or the confidence intervals
that do NOT contain 0 indicate a difference between those
two population means.
b. Summarize the findings about the differences in population
means for the GPAs of freshmen in the different
socioeconomic classes.
Which pairs are significantly
different?
c. Calculate a 95% confidence interval for the mean GPA for
the middle class group, where the sample mean based on
the 7 subjects involved in the group was 3.25.
151
CoolDown: Check Your Understanding
Complete the following sentences by circling words or filling in
blanks as necessary.
The pvalue of 0.025 from this activity implies that if this study
were repeated many times, we would see an F test statistic of
4.579 or greater
less in about _____ % of
repetitions if the population means were really all
equal not
equal.
ANOVA procedures can be thought of as an extension of the two
independent sample pooled
unpooled
ttest, and
hence requires the assumption of
equal
population
sample
variances.
One way to check this assumption is to use Levenes test and see
if the pvalue is greater than less than or equal to 0.10 (or
any reasonable significance level).
CoolDown: Think About It
For an ANOVA test, would there ever be a situation in which we
would need to divide the SPSS output pvalue by 2? Why or why
not?
152
153
Example Exam Question on ANOVA
A study was conducted to compare the effects of two different
therapy treatments and a control condition on weight gain in
anorexic girls. Group 1 was the control condition subjects that
received no intervention, Group 2 subjects received a cognitivebehavioral treatment condition, and Group 3 subjects received a
family therapy condition. The response was weight gain over a
fixed time period.
a. The ANOVA output provided below is used to test a set of
hypotheses.
ANOVA
Gain in Weight
Between Groups
Within Groups
Total
Sum of
Squares
601.916
3331.037
3932.953
df
2
60
62
Mean Square
300.958
55.517
F
5.421
Sig.
.007
i. State the null and alternative
hypotheses.
H0: ______________________________________________________
Ha: ______________________________________________________
ii. The pvalue is 0.007. Draw a sketch of the appropriate
distribution showing how the pvalue was determined for
this ANOVA study. Provide all details.
b. Multiple comparisons were
performed on the weight
gain data (using Tukeys
method). Use the results
to circle all pairs that are
significantly different
(using a 5% significance
level).
Multiple Comparisons
Dependent Variable: Gain in Weight
Tukey HSD
(I) Condition
Control
Cog Behav
Family
control versus cognitive
behavior
control versus family therapy
154
(J) Condition
Cog Behav
Family
Control
Family
Control
Cog Behav
Mean
Difference
(IJ)
3.65
8.29
3.65
4.64
8.29
4.64
95% Confidence Interval
Lower Bound Upper Bound
8.77
1.48
14.36
2.22
1.48
8.77
10.58
1.30
2.22
14.36
1.30
10.58
cognitive behavior versus family therapy
c. Calculate a 95% confidence interval for the mean weight gain
for the family therapy group, where the sample mean based on
the 23 subjects involved in group 3 was 7.4 pounds.
Final answer:
____________________________________
Lab 12: Simple Linear Regression
Objective: In this lab, you will examine relationships between two
quantitative variables using a graphical tool called a scatterplot.
You will interpret scatterplots in terms of form, direction, and
strength of the relationship, and use it to assess the
appropriateness of using a simple linear regression model to
describe the relationship between the two variables.
If
appropriate, you can then perform a simple linear regression
analysis to produce an estimated model that can be used to
predict the value of the response y for a given value of the
predictor x. You will learn how to perform hypothesis tests and
compute confidence intervals in regression; including learning how
to check the assumptions needed for these inference procedures
to be valid (see Supplement 8 for further details about these
assumptions).
Overview: A regression model describes how the mean of one
variable is thought to depend on the value of one or more other
variables. If we think a variable x may explain changes in another
variable y, we call x an explanatory variable (or predictor
variable or independent variable) and the variable y is called the
response variable (or dependent variable).
To start, we use a scatterplot to display the relationship between
two quantitative variables, plotting the explanatory variable x on
the xaxis and the response variable y on the yaxis. In examining
the relationship, look for an overall pattern showing the form
(linear, curved, clusters), direction (positive or negative), and
strength (how close the points are to the underlying form weak,
moderate, strong, etc.) of the relationship. Finally, check for
outliers or other deviations from the overall pattern.
One of the many misconceptions about regression arises from the
concept of association. Scatterplots can show the association
between variables, but you need to remember that correlation
155
does not imply causation. This can be seen in a very simple
example: weekly flu medication sales and weekly sweater sales for
an area with extreme seasons would exhibit a positive association
because both tend to go up in the winter and down in the summer.
However, it should be clear that neither of them causes the other.
156
There are four interpretations of an observed association. (Source:
Mind on Statistics, 3rd edition, by J. M. Utts, R. F. Heckard. 2007,
page 176)
1.
There is causation the explanatory variable is indeed
causing a change in the response variable.
2.
There may be causation, but confounding factors
contribute as well and make this causation difficult to
prove.
3.
There is no causation the association is explained by
how the explanatory and response variables are both
affected by other variables.
4.
The response variable is causing a change in the
explanatory variable.
When a scatterplot suggests that the dependence of y on x can be
summarized by a straight line, the least squares regression
line of y on x can be calculated. The least squares regression line
is the line that minimizes the sum of the squared vertical distances
of the data points to the line hence the name least squares.
This fitted line can be used to describe the linear relationship
between y and x and to predict y for a given value of x.
In this lab we will focus on simple linear regression, which is
based on a linear model relating a single explanatory variable x to
the mean response y as follows: E(Y) = 0 + 1(x). Here 0 and 1
are parameters fixed but unknown constants. Specifically, 0 is
the population yintercept (where the true regression line crosses
the yaxis when x = 0) and 1 is the population slope (the change
in the mean response y for a oneunit increase in x). These two
values are unknown, but can be estimated using the least squares
criterion. The resulting estimated regression line is generally
written as: y b0 b1 ( x) . The estimates, b0 and b1 are referred to as
the least squares estimates of 0 and 1.
There are also several assumptions that must be checked in order
for inferences to be valid. First, the response variable must have a
normal distribution with a mean that varies linearly with the
predictor variable and a standard deviation () that does not
depend on the predictor variable. We also assume that the error
157
terms are normally distributed with mean 0 and a standard
deviation .
158
Formula Card:
WarmUp: Check Your Understanding
Which scatterplot(s) indicates a linear regression analysis would be
appropriate?
159
Plot 1
Plot 2
Plot 3
160
ILP: Using a Scatterplot to Display the
Relationship
Background: The U.S. Census Bureau collects many different
kinds of data, including demographics, housing data, and
economic indicators. Data are collected at the individual, county,
state, and national levels. The data set poverty.sav contains
poverty rates, teen birth rates, and violent crime rates for the 50
states and the District of Columbia for the year 2000.
The poverty rate variable (PovPct) gives the percentage of the
population living in households with income below the poverty
level for each state and the District of Columbia (Location). The
teen birth rate variable (TeenBrth) contains the birth rate for
females 15 to 19 years old, recorded as number of births per 1,000
females in that age group. The violent crime rate (ViolCrime), the
birth rate for females 15 to 17 years old (Brth15to17), and the
birth rate for females 18 to 19 years old (Brth18to19) are also
recorded. (Source: Mind on Statistics, 3rd edition, by J. M. Utts, R. F.
Heckard. 2007)
Task: Is there a relationship between the poverty rate (PovPct)
and the teen birth rate (TeenBrth) by state? Produce a scatterplot
to help examine the plausibility of a linear relationship between
the poverty rate and the teen birth rate. Note: You are interested
in predicting teen birth rate from poverty rate.
Before conducting any test, here are a set of questions to ask
yourself:
How many populations are there?
One
Two
More
than two
How many variables are there? One
What is the response variable?
What type of variable is the response?
Two
Categorical
Quantitative
What is the explanatory variable?
What type of variable is the explanatory variable?
Categorical
Quantitative
161
What type of parameter would be useful for summarizing
this response, considering the explanatory variable?
(See Supplement 3)
Proportion Mean
Other
The appropriate inference procedure for this scenario is simple
linear regression.
162
1. Open the data set and produce a scatterplot: Graphs> Legacy
Dialogs> Scatter/Dot. Note the explanatory variable should
be on the xaxis (independent), and the response variable
should be on the yaxis (dependent). Sketch the general
pattern of the plot below. Make sure to label your axes.
2. Based on the scatterplot, does there appear to be a linear
relationship between teen birth rate and poverty rate? Explain.
3. Are there any unusual observations or outliers present?
4. Which states have the highest poverty rates? Highest teen
birth rates? Do states with low poverty rates tend to have low
teen birth rates also? What does this tell you about the
direction of the association between these two variables?
(Note: You can find this information quickly by choosing
Location for Label Cases by and selecting Display chart
with case labels under Options.)
5. Does there appear to be a strong relationship between poverty
rates and teen birth rates?
163
ILP: Describing a Linear Relationship with a
Regression Line
If we are satisfied that a linear model seems to describe the
relationship between the poverty rate and the teen birth rate, we
are ready to estimate that model and use it to predict TeenBrth on
the basis of PovPct.
Task: Fit a linear model to the data. Refer to the first part of the
ILP for response/explanatory variables. If you have questions
about the regression output after the ILP activities, refer to
Supplement 8 at the beginning of this workbook for more details.
1. Obtain the linear regression output using Analyze>
Regression> Linear. (Enter the appropriate dependent and
independent variables.) Report the estimated regression line
(predicting equation): _________________________________
2. How does the estimated regression line differ from the equation
for the population regression line?
3. Interpret the estimated slope b1. Clearly explain what the slope
says about the change in the teen birth rate.
4. Report the coefficient of correlation, r, between TeenBrth and
PovPct:
r = _______________
5. Report the coefficient of determination, r2, and interpret it: r2 =
____________
Interpretation:
6. Use the regression line to predict the teen birth rate for New
Mexico (with a poverty rate of 25.3%) and for Michigan (with a
poverty rate of 12.2%). How do they compare to the observed
TeenBrth values for the states of New Mexico and Michigan?
164
165
ILP:
Is There a Significant (NonZero)
Linear Relationship Between Teen Birth
Rate and Poverty Rate?
In this ILP activity, you will assess if the explanatory variable,
poverty rate, is a useful linear predictor for the teen birth rate. In
other words, you will test to see if there is a significant, nonzero
linear relationship between poverty rate (PovPct) and teen birth
rate (TeenBrth). Remember that another way to make inferences
about the significance of the linear relationship is through a
confidence interval for the population slope. Further, recall the
basic form of a confidence interval: point estimate (a few)
standard errors. Most standard computer regression output
provides the slope estimate and its standard error, and the few
will correspond to a t* value for the corresponding confidence level
with degrees of freedom for regression of n 2. Since a
confidence interval provides a range of reasonable values for the
parameter, it can be used to perform twosided hypothesis tests
by seeing whether the hypothesized value falls in the interval or
not.
Task: Perform a test to assess if the explanatory variable PovPct
is significant in the linear model.
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Hypotheses:
H0: ___________ = ___________ and Ha: ___________
___________ , where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
2. Checking the Assumptions and Computing the Test
Statistic
Assumptions: Covered in the next ILP activity.
TestStatistic:
a. Using the SPSS output already produced, which two test
statistics could you use to test these hypotheses?
166
b. Give the value for each test statistic listed in part (a).
c. Which of the test statistics in part a would not be
appropriate for conducting a onesided version of the
alternative hypothesis?
3. Calculate the pValue:
What is the SPSS reported pvalue for both test statistics?
________________
Are these the pvalues you want? ______________________
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
Note: Conclusions should include a reference to the population
parameter of interest. Your conclusion should not be too strong;
you can say you have sufficient evidence or something equivalent,
but NOT that you have proven anything.
6.
Confidence Intervals (CI):
a.
What is the formula for a confidence interval for the
population slope?
b.
Generate the appropriate confidence intervals using
Analyze> Regression> Linear. Under Statistics, choose
Confidence intervals, which has a default level of 95%.
Give the 95% confidence interval for the population slope.
c. Provide an interpretation of the 95% confidence interval for
the population slope.
167
d. Provide an interpretation of the 95% confidence level.
e. Based on the confidence interval, would you reject the null
hypothesis at a 5% significance level? Circle one: Yes No
Explain.
Did your conclusion here match the one you made in part 4?
ILP: Is the Linear Model Appropriate?
Are the Assumptions Met for
Inference?
In this ILP activity, you will examine the residuals from the
regression line as well as create some plots to assess the fit of the
linear model. This will also serve to evaluate the validity of the
testing and confidence intervals performed in previous ILP
activities. Regression assumptions may be stated in terms of the
response variable or in terms of the error terms. In general, the
statistical model for simple linear regression assumes for each
value of x, the observed values of the response are normally
distributed with some mean (that may depend on x in a linear
way) and a standard deviation that does not depend on x. That
is, for each x, Y is N(E(Y), ), where E(Y) = 0 + 1x. Thinking
about the error terms, we can say the true error terms (those
that we do not observe) are the difference of the response and the
true mean (for a given x). These errors are to have a normal
distribution with mean 0 and standard deviation of (that does not
depend on x).
Task: Obtain the residuals for the regression of TeenBrth on
PovPct, and generate the appropriate plots to check the
assumptions of the linear model. For additional information on
checking the regression assumptions, refer to Supplement 8.
1. What assumptions have to hold for the inferences in the
previous ILP activities to be valid? (Be sure you state them in
the context of the problem.)
168
2. Write the expression for residuals in terms of observed and
predicted values.
3. Obtain the residuals in SPSS using Analyze> Regression>
Linear, but instead of running it right away, click on the Save
button. In the box that will open, select Unstandardized
under Residuals, and click Continue. This saves the residuals
as a new variable. Now, construct a scatterplot of the residuals
against the explanatory variable, PovPct. Sketch the general
pattern of the plot.
4. What assumption of the error terms in the linear model do you
think the plot in part 3 is useful for assessing?
What conclusion can you draw from this plot?
5.
Construct a QQ plot of the residuals.
pattern of the plot.
Sketch the general
6. What assumption of the error terms in the linear model do you
think this plot is useful for assessing?
169
Based on the plot what is your conclusion about this
assumption?
7. If there were an ordering present in the data, it would be
appropriate to construct a time (or sequence) plot of the
residuals. Although we have no order here to worry about,
what assumption of the error terms in the linear model do you
think such a time plot is useful for?
CoolDown: Think About It
Why would you not want to use this model to predict the teen birth
rate for a state that has a poverty rate of 2%?
Example Exam Question on Regression
A doctor wanted to study the relationship between a males age
and HDL (socalled good) cholesterol. She randomly selects 18
male adult patients and records their x = age and y = HDL. A
scatterplot of the data indicated an approximately linear
relationship overall, so she performs a linear regression analysis
using SPSS.
Model Summary
Model
1
R
.598
R Square
.358
Adjusted
R Square
.318
Std. Error of
the Estimate
6.807
ANOVA
Model
1
Regression
Residual
Total
Sum of
Squares
413.01
741.43
1154.44
df
1
16
17
170
Mean Square
413.01
46.34
F
8.91
Sig.
.009
Coefficients
Model
1
(Constant)
Age
Unstandardized
Coefficients
B
Std. Error
61.05
6.08
.38
.13
Standardized
Coefficients
Beta
t
10.05
2.99
.598
Sig.
.000
.009
x x 2883.131 .
We also have: x 46.22 and Sxx =
2
a. What is the correlation between age and HDL cholesterol
levels? ___________
b. What is the least square regression line for predicting HDL from
age?
final answer: _______________________________________________
c. Based on this model, predict the HDL cholesterol for a 30year
old male.
final answer: __________________________
d. Compute the residual for the observation (x = 30, y = 53).
final answer: __________________________
e. Use a 1% significance level to assess if there is a significant
linear relationship between age and HDL cholesterol for male
adults. State the hypotheses, the value of the test statistic, the
corresponding pvalue, and your decision.
Hypotheses:
Ha:________________________
H0:_____________________
Test Statistic Value: __________________
____________________
Decision: (circle)
pvalue:
Fail to reject H0
H0
171
Reject
Exam Question Example on Regression continued
f.
The standard error for the estimated slope is given as 0.13.
Interpret this standard error in terms of repetitions of this study.
g. Calculate a 95% confidence interval for the mean HDL
cholesterol level for all 30yearold male adults.
final answer: ____________________________________
f.
Consider the residual plot shown at the right.
Does this plot support the conclusion that
the linear regression model is appropriate?
Yes
No
172
20
Explain:
10
Residual
10
20
20
25
30
35
40
45
50
55
60
65
70
75
Age
173
Lab 13: ChiSquare Tests
Objective: In this lab, you will learn how to perform three Chisquare tests (the test of goodness of fit, the test of independence,
and the test of homogeneity) that are used to analyze categorical
responses.
Overview: There are three ChiSquare tests presented in this
lab: the tests of goodness of fit, independence, and
homogeneity.
For all three tests, the data are generally
presented in the form of a contingency table (a rectangular
array of numbers in cells). All three tests are based on the ChiSquare statistic:
(Oi Ei ) 2
Ei
, where Oi is the observed count and
Ei is the expected count under the corresponding null hypothesis.
Goodness of Fit Test: This test answers the question, Do the
data fit well compared to a specified distribution? It considers
one categorical variable, and assesses whether the proportion of
sampled observations falling into each category matches well
enough to the null distribution for the given problem. The null
hypothesis for the goodness of fit test specifies this null
distribution which describes the population proportion of
observations in each category.
Test of Homogeneity: This test answers the question, Do two or
more populations have the same distribution for one categorical
variable? It considers one categorical variable, and assesses
whether this variable is distributed the same in two (or more)
different populations.
The null hypothesis for the test of
homogeneity is that the distribution of the categorical variable is
the same for the two (or more) populations.
Test of Independence: This test answers the question, Are two
factors (or variables) independent for a population under study?
It considers two categorical variables, and assesses whether there
is a relationship between these two variables for a single
population. The null hypothesis for the test of independence is
that the two categorical variables are independent (that is, they
are not related) for the population of interest.
There are a few properties of the Chisquare distribution that you
might find useful. The expected value of a Chisquare distribution
174
is its degrees of freedom (mean = df ), and its variance is 2
times its degrees of freedom. Thus, its standard deviation is the
2
square root of 2 times the degrees of freedom ( 2 * df
so
2 * df ). This frame of reference can help us assess if our
observed statistic is unusual under the null hypothesis or
somewhat consistent with the null hypothesis.
For details on how to enter crosstabulated data and perform these
tests in SPSS, please refer to the How To Do It in SPSS section
for ChiSquare Tests found online in CTOOLS. Output will be
provided for all activities.
Formula Card:
WarmUp: ChiSquare Name That Scenario
Think about the three different chisquare tests (Independence,
Homogeneity, and Goodness of Fit). Fill in the blank with the most
appropriate Chisquare test to address each presented research
question.
1. A researcher wants to determine if scoring high or low on an
artistic ability test depends on being right or lefthanded.
___________________________________________
2. A national organization wants to compare the distribution of
level of highest education completed (high school, college,
masters, doctoral) for Republicans versus Democrats.
175
___________________________________________
3. A preservation society has the percentages of five main types
of fish in the river from 10 years ago. After noticing an
imbalance recently, they add some fish from hatcheries to the
river. How can they determine if they restored the ecosystem
from a new sample of fish?
___________________________________________
ILP: Is There a Different Pattern in the
Distribution of Accidental Deaths in a
Certain Region Compared to the
Pattern in the Entire United States?
Background: According to the records of the National Safety
Council, accidental deaths in the United States during 2002 had
the following distribution according to the principal types of
accidents.
Motor
Vehicle
45%
Falls
15%
Drownin
g
4%
Fire
Poison
Other
3%
16%
17%
Suppose that an accidental death data set from a
particular geographical region yielded the following frequency
distribution for the principal types of accidents:
Motor
Vehicle
442
Falls
161
Drownin
g
42
Fire
Poison
Other
33
162
150
Do these data show a significantly different pattern in the
distribution of accidental deaths in the particular region compared
to the pattern in the entire United States? Use a 5% significance
level. (Source: National Safety Council Website, 2005)
Task: Perform a Chisquare goodness of fit test to asses whether
the data fit well with the model specified in the null hypothesis.
1. State the hypotheses:
a. Explain in one sentence why this is a goodness of fit test.
176
b. State the null hypothesis H0: _________________________ ,
where __________ represents:
Remember: Your hypotheses and parameter definition should
always be a statement about the population(s) under study.
177
2. Checking the Assumptions and Computing the Test
Statistic
a. Find the expected counts for the different categories of
accidental deaths, and enter them in the table that has the
null hypothesis percentages and the observed counts.
Null %
Observ
ed
Expect
ed
MotorVeh
icle
45%
442
Fall
s
15
%
161
Drowni
ng
4%
Fir
e
3%
Poiso
n
16%
Othe
r
17%
42
33
162
150
Tota
l
100
%
990
b. Do all cells have expected counts greater than 5? Yes
No
c. Complete the calculation of the test statistic based on your
table above.
442 445.5 2 161 148.5 2 42 39.6 2 33 29.7 2
X2
445.5
148.5
39.6
29.7
162 158.4 2 .......................... 2
158.4
3.663
3. Calculate the pvalue: Based on the SPSS output, answer
the following:
Test Statistics
Catego
ry
Chi3.663
Square
df
5
Asymp.
.599
Sig.
a. The pvalue is _____________ .
b. The expected value of the test statistic assuming the H0 is
true is _______ .
c. The large pvalue is consistent with the fact that our
observed test statistic value is
greater than
less
than
the expected test statistic value (under the null
hypothesis).
178
4. Decision:
What is your decision at a 5% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion: What is your conclusion in the context of the
problem?
ILP: Comparison of the Distribution of
Academic Degrees: Males Versus
Females
Background: How do women and men compare in the pursuit of
academic degrees? The table below present counts (in thousands)
from the Statistical Abstract of degrees earned in 1996
categorized by the level of the degree and the sex of the recipient.
Female
Male
Bachelor
Master
642
522
227
179
Profession
al
32
45
Task: Perform a Chisquare test of homogeneity.
significance level.
Doctorate
18
27
Use a 1%
Hypothesis Test: You can now implement the Five Steps for
hypothesis testing.
1. State the Null Hypotheses:
H0: __________________ _________________________________________
2. Checking the Assumptions and Computing the Test
Statistic
179
a. Show how the expected count 531.8 (first cell for males)
was computed.
b.
Based on the table, do the assumptions appear to be met
to perform the test? (Are all expected counts greater than
5?)
Yes
No
c. The expected value of the test statistic
assuming the null hypothesis is true is _________.
ChiSquare Tests
Pearson ChiSquare
Likelihood Ratio
LinearbyLinear
Association
N of Valid Cases
Value
9.514a
9.485
3
3
Asymp. Sig.
(2sided)
.023
.023
.024
df
5.099
1692
a. 0 cells (.0%) have expected count less than 5. The
minimum expected count is 20.56.
d. What is the value of the test statistic? ____ = ____________
3. Calculate the pValue:
Based on the SPSS output, the pvalue is _____________ .
4. Decision:
What is your decision at a 1% significance level? Reject H0
Fail to Reject H0
Remember:
significant
significant
Reject H0
Fail to Reject H0
Results statistically
Results not statistically
5. Conclusion:
What is your conclusion at a 1% significance level
180
in the context of the problem?
6. Would your decision and conclusion change if the significance
level was:
5% instead of 1%?
3% instead of 1%?
2.3% instead of 1%?
2% instead of 1%?
Based on your answers above, you can see that
the pvalue represents the _________________ significance
level
at which the results would be statistically
significant.
181
CoolDown: Frame of Reference
Comparing Package Designs As part of a survey of 70
consumers, each was presented with 3 possible designs for
packaging a product (Design 1, Design 2, Design 3) and asked
"Which package design do you prefer?" The data will be used to
assess whether or not there is a preference for any of the three
designs.
a. The data gathered from this survey question would be what
type of data?
Circle your answer(s). categorical
quantitative
continuous
b. Name the test that could be conducted to assess if the model
of no preference (that is, that the three designs are equally
preferred) is reasonable.
Name of the test: ___
_______________________
c. State the null hypothesis that the three designs are equally
preferred in the population of all such consumers.
H0: _____
___________
d. If there is no preference for any one of the three designs in the
population, how many of the consumers surveyed would you
have expected to prefer Design 1? Show your work.
Final Answer: _____ _________
CoolDown: ChiSquare Name That Scenario
Match each research question with the appropriate ChiSquare test
that should be used to answer the question.
a. Is student status (instate versus outofstate) associated with
ones eventual graduation outcome (graduating versus not
graduating)?
Answer: ChiSquare test of _____ _
________
b. To test a theory that people have no preference among four
different outdoor activities, you ask 100 people to select among
jogging, bicycling, hiking, or swimming.
Answer: ChiSquare test of _____ _
182
________
c. A biostatistician would like to determine if the ratio of the blood
type in the storage for transfusions should be different in
Hawaii from the main land. She collected a sample of blood
types of 10,000 people in Hawaii and that of 100,000 people in
the mainland. She wishes to see if the breakdown of blood
types (A, B, AB and 0) is the same for both populations.
Answer: ChiSquare test of _____ _
Example
Tests
Exam
Question
on
________
ChiSquare
A study was performed to examine the attendance pattern and
exam performance of students in an intro statistics course. A
sample of 96 students enrolled in an intro stat course was selected
(these can be considered a random sample of such students).
These 96 students were classified by their attendance status
Regularly Attend? * Exam Performance Crosstabulation
(regularly
attend
lecture or not) and
Count
their
performance
Exam Performance
Low
Middle
High
Total
on
the
midterm
Regularly Yes
12
20
28
60
exam, classified as
Attend?
No
16
12
8
36
low (below 50%),
Total
28
32
36
96
middle
(50%
to
80%), and high (above 80%). The data and SPSS output are
provided.
ChiSquare Tests
Pearson ChiSquare
Likelihood Ratio
LinearbyLinear Association
N of Valid Cases
Value
8.195a
8.298
8.067
96
df
2
2
1
Asymp. Sig.
.017
.016
.005
a. 0 cells (.0%) have expected count less than 5. The minimum
expected count is 10.50.
a. Give the name of the chisquare test for assessing if there is a
relationship
between
attendance
status
and
exam
performance.
_________________________________________________________
b. Based on the above data, what proportion of regular attendees
performed high on the exam?
Final answer: ________________________
c. Assuming there is no relationship between attendance status
and exam performance, how many regular attendees would
you expect to perform high on the exam? Show all work.
183
Final answer: ________________________
d. Assuming there is no relationship between attendance status
and exam performance, what is the distribution of the test
statistic? (Be specific.)
Final answer:
___________________________________
e. Use a level of 0.05 to assess if there is a significant relationship
between attendance status and exam performance.
Test Statistic Value: __________________
pvalue:
____________________
Thus, there (circle your answer):
does
does not
appear
to be an association between attendance status and exam
performance.
Exam 1
Questions
184
Statistics 250 Fall 2006 Exam 1
1.
A study is designed to determine whether acupuncture really works
(Source: Parade Magazine: July 9, 2006). A total of 300 migraine patients
in Germany were randomly divided into three groups. In one group,
acupuncture needles were placed at prescribed sites; the second group also
was needled, but randomly (a procedure known as sham acupuncture). The
third group received no treatment. All the needled patients underwent 12
treatment sessions, each lasting 30 minutes. The researchers found that
both the real and the sham acupuncture recipients reported 50% fewer
headache days, while only 15% of those untreated felt better. It was
concluded that the placement of the needles made no difference.
a. This study was: (circle one) an observational study
an experiment.
[1]
b. Was a control used in this study? (circle one)
yes
no
cant tell
[1]
c. For all patients in the study, the number of headache days was recorded.
response
[1] This variable is the: (circle one)
explanatory
variable.
d. Fill in the blank. Suppose the younger patients generally respond better to
acupuncture and that the patients in the untreated group were mostly
older (over 50), while those in the real and sham acupuncture groups were
generally younger.
[1]
The variable age would be an example of a(n)
variable.
2. The
following
graph
for
the
color
of
jelly
by the JB Company.
Which of the following
description(s) are appropriate
for this distribution?
(Circle all that apply.)
[2]
Symmetric
shows
beans
Uniform
Rightskewed
Leftskewed
None of these is appropriate
185
the
distribution
in
bags
produced
3. According to Starbucks International, 32.2% of its entire staff that work
at the local coffee shops is Hispanic. This value of 32.2% is a: (circle all
that apply)
[2]
sample statistic
sample parameter
population statistic
population parameter
4. A law firm in the DallasFort Worth area prepares contracts and other
legal documents and uses a courier service to deliver the documents to its
many clients. Recently a partner reported that a few complaints have come
in from some of their best clients about delayed contract deliveries. The
current courier service used is Metro Delivery. There are two other
courier services that have opened up in past 2 years. Should they keep
using Metro or consider a new one?
To help address this question, a study was conducted over a month period
in which all three couriers were used. When a delivery was required, one of
the couriers was randomly selected. One of the responses of interest was
the total delivery time, that is, the time (in minutes) from when the order
is phoned in to when the documents are delivered to the destination.
Below are the boxplots and some additional numerical summaries for
comparing the total delivery times for the three couriers.
One Variable Summary
Mean
Variance
Standard Deviation
Count (n)
Total Delivery Time (minutes)
DFW Express
Carborne Carrier
Metro Delivery
41.7
46.9
52.0
269.1
274.5
227.5
16.4
16.6
15.0
56
65
61
For any answers below, when appropriate, INCLUDE YOUR UNITS.
186
187
a.
What type of variable is total delivery time? Circle one:
[1]
Categorical
Quantitative
b. Compute the approximate IQR of the delivery times for DFW Express.
[2]
IQR = ___________________________________________
c.
[2]
Give the value of the upper boundary (or fence) that would be used to show
why the longest delivery time for DFW Express is an outlier. Show all work.
Upper boundary or fence = ______________________
d. Fill in the blank. The Q1 of the delivery times by Metro is equal to the
[1]
____________ of the delivery times by DFW Express.
e. Fill in the blank. For Carborne Carrier, 25% of the deliveries took longer
[2]
than ________________.
f.
How many deliveries by DFW Express were made in under 41 minutes?
[2]
_______________
g.
One of the Metro deliveries resulted in a standard zscore of 1.4. What is
the corresponding actual delivery time?
[2]
Final answer: ________________
h. The following is an incorrect interpretation of the standard deviation of
the Metros delivery times. Enhance the sentence to make it correct by
inserting a few words
[2]
The average distance between the Metro delivery times
is roughly 15 minutes.
i.
DFW Express appears to be the
winner as it had the lowest mean
delivery time. Another variable
measured was mileage, the distance
in miles from the law firm to the
client. At the right are boxplots
that compare the mileage for the
deliveries by these three firms in
the study. Which of the following
statements
is
an
appropriate
conclusion? Clearly circle one. [2]
DFW Express is still the winner as it had the lowest median mileage.
DFW Express is still the winner as mileage was not a response variable.
188
The response variable of delivery time should be adjusted for mileage
before any winner is selected.
189
5. A survey was sponsored by the Ann Arbor News. A random sample of 200
Ann Arbor residents resulted in 110 stating they support extending the Art
Fair by one day.
a.
[1]
What is the value of the population proportion of Ann Arbor residents who
support extending the Art Fair by one day?
Final answer: ____________________
b. What is the margin of error for this survey, that is, a measure of accuracy
of the sample proportion of Ann Arbor residents who support extending
the Art Fair by one day?
[2]
Final answer: ____________________
c. Do these numbers allow the Ann Arbor News to conclude that a majority of
its citizens advocate extending the Art Fair? Circle your answer and
provide a brief rationale.
[3] Circle one:
Yes
No
Because:
d. Suppose the survey will be repeated and the director of the Art Fair would
like to ensure a 95% margin of error of at most 3%. What is the minimum
number of such people that should be surveyed?
[2]
Final answer: ____________________
6. Determine whether each of the following statements is true or false.
Clearly circle your answer.
a. If X has a Binomial (50, 0.7) distribution, then the criteria to use the
normal approximation are met.
[2]
True
False
b. Two independent researchers are each trying to estimate a population
proportion p. Researcher I takes a random sample resulting in a sample
proportion of 0.60. Researcher II takes a random sample that is twice as
big and also results in a sample proportion of 0.60. Then, the standard
errors for the two sample proportions will be the same.
[2]
True
False
190
c.
A 95% confidence interval for p, the population proportion of college
students who drive a foreign car, was found to be (0.524, 0.676). Thus,
there is a 95% chance that the population proportion of college students
who drive a foreign car will be between (0.524, 0.676).
True
[2]
False
7. Suppose that the amount of time spent waiting for your bus to our campus
each day is a uniform random variable between 0 to 20 minutes.
a.
[3]
Sketch a picture of the model for waiting time for the bus. Provide labels
for each axis and some values along each axis.
Lets define the following events:
A is the event that you wait at least 10 minutes, that is your waiting time is
in the interval [10,20].
B is the event that you wait at most 15 minutes, that is your waiting time is
in the interval [0,15].
C is the event that you wait at most 10 minutes, that is your waiting time is
in the interval [0,10].
Answer the questions based on the information given above. Show all work.
b. What is P(A)?
[1]
Final answer: ________________
c. What is P(A and B)?
[1]
Final answer: ________________
d. What is P(BA)?
[2]
Final answer: ________________
191
e. What is P(A and C)?
[2]
Final answer: ________________
f.
Are the events A and C mutually exclusive?
[2] Circle one:
Yes
No
Explain briefly.
8. The demand for a certain weekly magazine at a newsstand is a discrete
random variable, X, with an expected value of 3 magazines sold per week.
Furthermore, the probability distribution function of X is symmetric about
the value of 3.
The magazine is sold for $6.00 per copy to the customers and it costs
$4.00 per copy for the owner of the newsstand. At the beginning of each
week, the owner of this newsstand buys 6 magazines to sell during the
week.
Value of X
Probability
0.05
0.10
0.20
a. Fill in the missing values in the table above.
[2]
b. In dollars, what is the expected amount of money the owner of the
newsstand will take in from the sales of this magazine per week?
[2]
Final answer: ________________
c.
[2]
Explain briefly why the owner is not wise to buy 6 magazines at the
beginning of each week.
Note: longwinded explanations will be penalized!
192
9. Consider the population of all UM students. Suppose 30% of all UM
students have their own laptop computer. A random sample of n = 100 UM
students will be selected and the sample proportion that have their own
laptop computer will be computed.
a.
[4]
If you consider all possible samples of 100 UM students, what is the
approximate distribution that describes the possible values for the sample
p
proportion ? Be complete, give all aspects of the distribution.
b. Suppose that 35 of the 100 UM students surveyed stated they have their
own laptop, for a sample proportion of 0.35. If the rate of students that
have their own laptop is indeed 30%, how likely would it be to a sample
proportion as large as 0.35 or larger? Show all work.
[3]
Final answer: ________________
10. The BioPharm company has developed a stress test for collegeaged
students. Scores on the stress test for such students are approximately
normally distributed with a mean of 55 points and a standard deviation of 2
points. Scores of 58 points or higher indicate a high level of stress and are
of concern to doctors. A random sample of collegeaged students will be
selected and each will be given this stress test. Answer the following
questions based on the information given. Show all work.
a.
Complete the sentence.
About 25% of collegeaged students will have a stress test score of at
most ________ points.
[3]
Final answer: ________________
193
b. What is probability the score for the first randomly selected student will
be at least 58 points?
[3]
Final answer: ________________
c.
[2]
Given that the first randomly selected student has a score of at least 58
points, what is the probability that the next randomly selected student
scores below 58 points?
Final answer: ________________
d. Suppose the stress test is given to six randomly selected collegeaged
students. Let Y be the number of students with a high level of stress (a
score of at least 58 points).
i.
What kind of a random variable is Y? Clearly circle your answer.
[2]
[3]
Normal random variable with a mean of 55 and a standard deviation
of 2.
Normal random variable with a mean of 55 and standard deviation of
12 (6 students times 2).
Binomial random variable with n = 6 and p = probability a score is at
least 58 points.
Binomial random variable with n = 6 and p = .
ii. What is the probability that exactly 2 of the 6 students will have a
high level of stress (a score of at least 58 points)?
Final answer: ________________
194
Statistics 250 Fall 2007 Exam 1
1.
This past September the results of a study were reported by the
Associated Press. The headline read Study: Dieting for dollars works.
The study involved about 200 overweight employees at several campuses in
North Carolina, which were randomly divided up into three groups. One
group received no incentives while the other two groups received $7 or $14
for each percentage point of weight lost. (For example, someone in the
middle group weighing 200 pounds who lost 10 pounds, or 5 percent, would
get $35.) In the end, employees who received the most incentives lost
the most weight, an average of nearly 5 pounds after three months.
a. This study is: (circle one) an observational study
an experiment
[2]
b. Weight loss (in pounds) is: (circle all that are appropriate)
[2]
an explanatory variable a response variable a confounding variable
a categorical variable
variable
c.
a quantitative variable
a normal
The study reported that those offered no incentives lost an average of
2 pounds; whereas those in the $7 group lost about 3 pounds on average.
[2] The value of 2 pounds is: a parameter
d. Use the boxplots to answer the
remaining questions.
[6]
i. For the no incentive group, 75% lost
at least _________ pound(s).
ii. The most weight lost by
an employee in the $14 incentive
group was ________ pound(s).
iii. True or False: The boxplot
of the weight losses for
the $7 incentive group
confirms the distribution
is bellshaped.
True
False
195
a statistic
a sample
a population
2. A study of the behavior of a large number of drug offenders after
treatment for drug abuse suggests that the likelihood of conviction within
a 2year period after treatment may depend on the offenders education.
Use the following table of results to answer parts (a) and (b):
Status within 2 years of treatment
Education
Level
a.
[2]
Convicted
Not Convicted
Total
At least 10 years
50
150
200
Less than 10 years
Total
135
185
165
315
300
500
What is the probability that a randomly selected drug offender will be
convicted within 2 years of treatment?
Final Answer: ______________________
b. The researcher would like to assess if conviction status is independent of
education level.
[5] i. To check for independence, the probability found in part (a) should be
compared to which of the following probabilities?
Clearly circle your answer.
o
P (education level of at least 10 years)
P (education level of at least 10 years given convicted within 2 years)
P (convicted within 2 years given education level of at least 10 years)
ii. Find the probability you circled and circle the appropriate conclusion.
The probability =______________________________________
Thus it appears that conviction status
is
is not
independent of education level.
3. Based on the graph
adapted from the USA
Today graph "Seeking a
cure", is it appropriate to
say that the distribution
of the number diagnosed
with diabetes is right
skewed?
Circle one:
Yes
No
[2] Briefly explain:
196
4. The scores for a recent pretest were recorded for the two sections of a
class (all scores were integer values). A summary of the two score
distributions is provided below.
a. What is the mean score for the students in class A?
[2]
Final answer = __________________
b. The standard deviation for the scores in Class A is 3.2. The standard
deviation for the scores in Class B would be (circle one)
[2]
larger
smaller
the same
c.
[2]
Which of the following is an appropriate interpretation of that standard
deviation for Class A?
i.
All of the individual scores are within 3.2 points, on average.
ii.
The average difference between the individual scores is roughly 3.2 points.
iii. We expect 95% of the scores to be within 3.2 points of the mean.
iv. The individual scores vary from the mean by about 3.2 points, on average.
d. At the right we have one more plot but the title is missing the letter for
the Class. For which class is this a plot of?
[2]
Clearly circle your answer:
Class A
Class B
197
5. For this problem,
consider the chart at
the right which can be
found on the Gallup
Organizations website
(www.galluppoll.com).
a.
What type of chart is
this?
[1] ___________________
b. Suppose
that
this
chart
summarizes the results of a
survey of a random sample of
1,000 adult college football fans.
How many people in the sample reported that they prefer the current bowl
game system?
[2]
Final Answer: ______________________
c. Create a 90% confidence interval to estimate the proportion of all adult
college football fans who prefer the current bowl game system.
[3]
Final Answer: ______________________
d. Does the interval from part (c) provide evidence that a minority of all adult
college football fans prefer the current bowl game system?
[2] Circle one:
yes
no
Explain in one simple sentence.
e.
Does the interval from part (c) provide evidence that a minority of all
adults prefer the current bowl game system?
[2] Circle one:
yes
no
Explain in one simple sentence.
f.
[2]
Consider the following statements below. Clearly circle all which correctly
explain the meaning of the 90% confidence level.
i
Confidence intervals computed by using the same procedure will
include the true population value for 90% of all possible random
samples taken from the population.
198
ii
The probability that the population parameter falls between the
bounds of an already computed confidence interval is roughly
90%.
iii If we consider all possible randomly selected samples of the
same size from a population, the 90% is the percentage of those
samples for which the corresponding confidence interval will
contain the population parameter.
f.
[2]
The 90% confidence interval for the population proportion of all adult
college football fans who would prefer a playoff tournament to the current
system was found to be (0.67,0.72).
Which of the following is a correct interpretation of this 90% confidence
interval? Clearly circle your one answer.
i
There is a 90% probability that the proportion of all
adult college football fans who would prefer a playoff tournament
is between 0.67and 0.72.
ii
If this study were to be repeated with a sample of the
same size, there is a 90% probability that the sample proportion
would be between 0.67and 0.72.
iii
We estimate, with 90% confidence, that the sample
proportion of all adult college football fans who would prefer a
playoff tournament is between 0.67and 0.72.
iv
We estimate, with 90% confidence, that the population
proportion of all adult college football fans who would prefer a
playoff tournament is between 0.67and 0.72.
h. Suppose another survey is to be conducted for estimating the population
proportion of all adult college football fans prefer the current bowl game
system. The sample size should be large enough so that, with 90%
conservative confidence, we are within 2% of the population proportion.
How large a sample size do we need?
[3]
199
Final answer:_________________
200
6. A study at the United States Postal Service suggests that the time taken
to serve an individual customer at a post office is normally distributed with
a mean of 4 minutes and standard deviation of 45 seconds.
a.
Complete the following sentence: (show all work)
[4] 20% of the customers will have a service time shorter than _____ minutes.
b. What is the probability that a customer will take longer than 5 minutes?
[3]
Final Answer: ______________________
c.
[3]
Suppose we track 4 randomly selected customers. What is the probability
that exactly two of the four customers will take more than 5 minutes?
Show all work.
Final Answer: ______________________
201
7. Suppose that among all of the many cars parked at the stadium on a game
day, 20% are outofstate. Let X represent the number of cars in a random
sample of size 6 that are outofstate. The probability distribution for X is
given below.
X=x
P(X = x)
a.
0
0.26214
1
0.3932
2
2
0.2457
6
0.08192
0.01536
5
0.00153
6
6
0.00006
4
What is the distribution for the random variable X = the number of outofstate cars in a random sample of 6 cars? (Be specific.)
[3]
b. What is the probability of finding at least 1 but no more than 4 outofstate cars in your random sample of 6 cars?
[2]
c.
[2]
Final Answer: ______________________
What is the expected number of outofstate cars in a random sample of
size 6?
Final Answer: ______________________
8. On August 7, 2007, Barry Bonds became the alltime home run leader in
major league baseball. Bonds record is controversial due to his use of
steroids, which are not allowed in professional baseball. Mark Ecko
purchased the record making ball and set up an internet survey to let the
public decide what he should do with it. Below are excerpts from the
September 27th New York Times article giving the results of that poll.
[Quote removed: Access article online at: http://www.nytimes.com/2007/09/27/sports/baseball/27ball.html]
[2] Based on the above, circle all of the statements below that are true.
i.
With such a large sample size, there would be a very large margin of
error, so this poll gives very accurate information about the population
of all baseball fans opinion on Bonds achievements.
ii. Since the sample was a volunteer sample, it may not give information
that is representative of the population of all baseball fans.
202
iii. The responses are not independent, therefore this data should not be
used to create a confidence interval.
203
9. The distribution of heights of adult men is approximately normal with a
mean of 69 inches and a standard deviation of 2 inches. Bob's height has a
zscore of 1 when compared to all adult men.
[2] Consider the following statements. Circle all true statements based on the
above information.
i.
Bob is taller than 69 inches tall.
ii. Bob's height is 1 standard deviation above the mean.
iii. Bob is 70 inches tall.
iv. About 16% of adult men are taller than Bob.
10. The overbooking problem: Airlines find that a small percentage of ticket
holders fail to show up to board a flight. Assume that the percentage is
10%. As a result, the airlines often sell more tickets than the capacity of
the plane. Suppose for planes with 120 seats, the airlines sells 130 tickets.
Let X be the number of passengers out of 130 that actually show up.
a.
[2]
If 130 tickets are sold, how many passengers are expected to actually show
up?
Final Answer: ______________________
b. If 130 tickets are sold, what is the standard deviation for the number of
passengers that actually show up?
[2]
Final Answer: ______________________
c.
[4]
If 130 tickets are sold, what is the probability that every passenger that
actually shows up will get a seat? Show all work.
204
Final Answer: ______________________
205
Statistics 250 Fall 2008 Exam 1
1. A doctor collects a large set of heart rate measurements that
approximately follow a normal distribution. He provides you
with the following summaries from SPSS:
Statistics
Heart Rate
N
Valid
Missing
Mean
Minimum
Maximum
1000
0
110
65.0
155
a. Which of the following is most likely to be the median of the
distribution? Clearly circle your one answer.
[1]
5
15
35
90
110
135
b. Which of the following is most likely to be the standard
deviation of the distribution?
[2] Clearly circle your one answer.
5
15
35
90
110
135
c. What would a qqplot of this data look like (roughly)?
[1]
Expected
Normal
Value
206
Observed value
207
2. Todays typical undergraduate student is often characterized as
preferring teamwork, experiential activities, and the use of
technology. An ECAR (Educause Center for Applied Research)
study was published on technology use among undergraduate
students. The study used survey and interviewer data to create
a portrait of todays students experiences with and skill using
information technology.
a. Listed below are some of the response variables that were
measured in this study. For each of these determine whether it
is categorical, quantitative discrete, or quantitative continuous.
[1 pt each] Clearly circle your answer.
Technology ownership: Do you own a computer?
Categorical
continuous
quantitative
Time (per week in minutes) spent using a computer for
writing documents (word processing).
Categorical
continuous
quantitative discrete
quantitative discrete
quantitative
Which social networking site(s) are you a member?
(facebook, myspace, friendster, etc.)
Categorical
continuous
quantitative discrete
quantitative
b. Another question on the survey asked for the number of friends
on their social networking account. Below and at the right is a
summary of some responses. Use these results to answer the
following questions.
208
4
0
3
5
0
2
5
0
1
5
0
5
0N
u
m
b
e
ro
fF
rie
n
d
s
Statistics
Number of Friends
Mean
Std. Deviation
i.
226
87
Complete the sentence:
[2]
The shortest list had about ____ friends on the social
networking account, while 75% of the lists had at least _____
friends.
ii. Report the average number of friends on a social
networking account. Include the appropriate symbol in your
answer.
[2]
=
iii. The standard deviation reported above is 87. Consider the
following interpretations of this value and clearly circle the
correct interpretation(s).
[2]
The number of friends on a social networking
account varies by about 87, on average.
The average distance between the numbers of
friends on social networking accounts
is roughly 87 friends.
On average, the number of friends on a social
networking account varied from the mean by
about 87 friends.
209
3. Obama versus McCain: A mock presidential election is
planned on the UM college campus. An enterprising Statistics
350 student plans to use the results of the mock election to
test the hypothesis that Obama would have a majority of votes
among all UM college students. The significance level of the
test is set at 5%. A random sample of 3000 college students at
the UM was selected to take part in this mock election which
resulted in 1578 voting for Obama.
a. State the hypotheses to be tested and completely define the
parameter.
[4]
Ho:________________________Ha:___________________________
where the parameter _____ is defined (in the context of this
problem) as
b. Provide the value of the test statistic (including its symbol) and
the corresponding pvalue. Include a sketch to show how the pvalue is found.
[5]
Sketch (include labels):
Symbol
Test Statistic = ______ = ______________________
pvalue= ________________________
c. Give the decision and corresponding conclusion by circling the
appropriate statements:
[2]
The statistical decision at a 5% significance level is:
Reject
Ho Fail to reject Ho
210
Therefore, there
is
evidence to say that,
is not
sufficient
a majority of all students (represented by this sample) would
vote for Obama.
d. If this decision in part (c) is incorrect, it would be a (circle one)
[1]
Type 1
Type 2
error.
4. The website http://global.mms.com/cai/mms/faq.html reports
that in a package of milk chocolate M&M's, 20% of the
candies are red. Let X = the number of red candies in a
randomly selected bag of milk chocolate M&M's.
a. Suppose you randomly select one small bag that contains 15
M&M candies.
i. What is the distribution of X? Include all relevant features to
completely specify the distribution.
[3]
ii. How many red candies would you expect to find in this bag?
[1]
Final Answer: ____________________
iii. What is the probability of finding at least one red candy?
[3]
Final Answer: ____________________
b. Suppose now that you randomly select a larger bag (one bag is
never enough...) with 100 M&M candies.
211
i. What is the approximate distribution of X? Include all
relevant features to completely specify the distribution.
[3]
ii. What is the approximate probability of finding no more than
15 red candies?
[3]
Final Answer: ____________________
iii. Without doing any calculations, what is the approximate
probability of finding more than 30 red candies? Circle your
one answer.
[1]
Less than 0.5
Equal to 0.5
More
than 0.5
5. Appreciation for cTools: In a recent survey, a random
sample of 3824 UM students resulted in 87% agreeing that
cTools is a valuable tool.
a. Calculate the 95% (general, not conservative) confidence
interval for estimating the population proportion of all UM
students who agree that cTools is a valuable tool.
[3]
Final Answer: ___________________
b. In a similar survey of instructors, 81% of the 1639 instructors
agreed that cTools is a valuable tool. The 6% difference in
rates (87% for students less 81% for instructors) would lead to
a 95% confidence interval for the difference in population rates
(students less instructors) of (3.8%, 8.2%).
Based on this interval, does there appear to be a higher
population proportion of students who appreciate cTools as
compared to that for all instructors?
[2]
Circle one:
Yes
No
Explain briefly:
212
6. For a survey about American diets a random sample of 1000
people were contacted. Of the 1000 people, 140 people
completed the questionnaire. The results of this study, if
applied to all Americans, are questionable because of (circle
one)
[1]
a large margin of error
nonresponse bias
selection bias
response bias
7. Blackberry subscribers
was the title of this graph
based off a USA Today
Snapshot
regarding
Blackberry email devices.
Based on the graph, is it
appropriate to say that
the distribution of the
number
of
Blackberry
subscribers
is
right
skewed?
[2] Circle one:
Briefly explain:
Yes
No
213
8. Determine whether each statement is True or False.
Clearly CIRCLE your answer. [1 point each]
a.
The significance level, , is determined
prior to the study
by considering the consequence of incorrectly rejecting H 0.
True
False
b. If the test of H0: p = 0.75 versus Ha: p > 0.75 resulted in a pvalue of 0.083,
then the probability that H0 is true is 0.083.
True
False
c. If the time to wait for pharmacy help has a uniform distribution
from 0 minutes to 20 minutes, then 25% of the customers
are expected to wait more than 15 minutes.
True
False
d. One condition for using a Z statistic for testing hypotheses
about a
population proportion p is that the population has a normal
distribution.
True
False
9. A researcher selects 100 subjects at random from a population,
observes 50 successes, and calculates three confidence
intervals. The confidence levels are 90%, 95%, and 99%, and
the
intervals
are
(0.402,
0.598),
(0.371, 0.629), and (0.418, 0.582). Match each interval with its
confidence level. Write your values clearly.
[2]
Interval (0.371, 0.629) ________ % confidence level
Interval (0.402, 0.598) ________ % confidence level
Interval (0.418, 0.582) ________ % confidence level
10.
Read the following description and indicate whether it is an
observational study or an experiment, and which variable is the
response and which is the explanatory variable.
An employer is interested in determining if there is a
relationship between the age of his employees and the number
of sick days that the employee takes. In order to study this
relationship, he collects this information from the personnel file
for a random sample of employees.
[2]
This is an (circle one) observational study
experiment.
214
Number of sick days is
explanatory
variable.
the
Age of employee is the (circle one)
explanatory
variable.
215
(circle
one)
response
response
11.
Does popcorn enhance a movie or not? In August 2008,
the BBC reported that the UK's largest arthouse cinema chain is
planning to take popcorn off the concession stand menu. A
spokesman for the movie chain claims that many people have
asked them to ban popcorn, citing that it is messy, smelly, and
loud. In an effort to save her favorite snack from being banned,
Jessica decided to take a survey of people at one of the chains
theaters on a busy Friday night. The following table partially
represents the data that she collected.
Bought
popcorn?
Enjoyed the movie?
Yes
No
180
72
40
Yes
No
a. Suppose that the total number of people that
Jessica surveyed is 400. What is the probability that someone
randomly selected from this sample of moviegoers enjoyed the
movie?
[2]
Final Answer: ___________________
b. What is the probability that someone randomly selected from
the sample enjoyed the movie, given that he or she bought
popcorn?
[2]
Final Answer: ___________________
c. How many of the people that Jessica surveyed did not buy
popcorn and did not enjoy the movie?
[2]
Final Answer: ___________________
d. Assuming this is sample is representative of the movie going
population, do these results support Jessica's goal of convincing
the theater owners not to ban popcorn? Do the above results
suggest that popcorn enhances the moviegoers experience?
Give your answer and completing one of the following
statements and include numerical support for your conclusion.
[2]
Yes, these results do suggest that popcorn enhances
the moviegoers experience because
OR
216
No, these results do not suggest that popcorn enhances
the moviegoers experience because
217
12.PainRelief: You are on a project team to consider whether
your companys new painrelief drug (Drug X) should go to
market. A study was conducted to compare the new painrelief
drug against a placebo. The measured response is whether the
pain from a dental procedure was relieved in 10 minutes (yes
or no). Patients were assigned to one of the two treatment
groups using a 2:1 randomization plan. The results are to be
used to test the hypotheses H o: p1 = p2 (no difference in the
population pain relief rates) versus Ha: p1 > p2 (the new drug
has a higher pain relief rate over placebo). Use a 5%
significance level.
Group 1 =
Group 2 =
Study Results
Drug X
Placebo
Number of patients
100
50
assigned
Number of patients that
had pain relieved in 10
44 (44%)
16 (32%)
minutes (%)
a. This is an (circle one) observational
study
experiment.
[1]
b. In computing the test statistic, we assume H o is true, that is, p1
= p2 = p (the common population proportion). Using the
results, provide the estimate of that common proportion.
[2]
Final Answer: ___________________
c. Provide the value of the test statistic (including its symbol) and
the corresponding pvalue. Include a sketch to show how the pvalue is found.
[4] Sketch (include labels):
Symbol
Test Statistic = ______ = __________________
pvalue = ________________________
d. Give the decision by circling the appropriate statements:
[1] The statistical decision at a 5% significance level is: Reject Ho
Fail to reject Ho
218
e. You are asked to report the conclusion of this study to the Vice
President of your company. With the significance level of 5%,
what should your conclusion be? As your boss appreciates
brevity, provide 12 well structured sentences.
[2]
f.
What could be done in a follow up study to produce a test with
higher power? Circle all correct answers.
[2] Repeat the study many times
Use more patients
219
Use a 10% significance level
Stats 250 Fall 2009 Exam 1
1. Use Twitter? Despite the hype surrounding Twitter, most
higher education professionals are not using the microblogging
service at all. In a survey of 1,958 higher education
professionals conducted by Faculty Focus in July and August
2009, 69.3 percent of respondents said they do not use
Twitter in any capacity.
a. The population of interest for this survey is:
[2]
Circle one.
All higher education professionals in the United States.
The 1,958 higher education professionals who answered the
question.
All higher education institutions in the United States.
The 69.3% who answered "Yes" to the question
b. Which of the following pertain to the value of 0.693 or 69.3%?
[2] Circle all that apply.
Population proportion
parameter
p
^p
statistic
sample proportion
2. Coffee and Heart Attacks: A recent medical study found
that people who drink more than 4 cups of coffee a day have
more heart attacks than people who drink less coffee or no
coffee. This led some doctors to suspect that coffee may be a
contributing factor in causing heart attacks. However, more
careful analysis of the data showed that heavy coffee drinkers
tend to smoke more than other people. This is an example of:
[2]
Circle your answer:
the placebo effect a response bias
doubleblind technique
confounding
the
3. Application levels: This graph was part of the St. Olaf Annual
Report, in the section providing information about applications
to the college.
Based on this graph, is it appropriate to say the distribution of
application levels is skewed to the left?
[2] Circle one:
Briefly explain:
Yes
No
220
4. Eating Habits: A study
was conducted to learn
about eating habits for
American families.
A
random sample of 200
families was selected
and an adult head of
household was asked to
complete a survey.
a. Listed below are some
questions
that
were
asked as part of the
survey. For each, determine whether it is
quantitative discrete, or quantitative continuous.
Clearly circle your answer.
categorical,
[1 pt each]
How long would you say dinner usually lasts when your
family eats dinner together (in minutes)?
categorical
quantitative discrete
quantitative continuous
In the last week, how many evenings did your family eat
dinner together?
categorical
quantitative discrete
quantitative continuous
Did your family eat dinner together last Sunday evening?
categorical
quantitative discrete
quantitative continuous
b. Another question on the survey asked
for the time it typically takes to
prepare a home dinner for the family
(in minutes). Below and at the right is
a summary of the responses. Use
these results to answer the questions.
Prep_Time (minutes)
Mean
Std. Deviation
35.6
8.9
i. Complete the sentence:
[4] The longest reported time to prepare
a home dinner for the family was
about ___________ minutes, while 25%
221
of those responding reportedly take at least__________ minutes
to prepare a home dinner for the family.
ii. The standard deviation for preparation time is reported as
8.9. Consider the following interpretations of this
value and clearly circle the correct interpretation(s).
[2]
The average distance between the times to prepare a home
dinner is roughly 8.9 minutes.
The times to prepare a home dinner vary by about 8.9
minutes, on average.
On average, the times to prepare a home dinner varied from
the mean by about 8.9 minutes.
222
5. Mom Survey: A research center conducted a telephone poll
based on a random sample of 1,000 women who are 18 and
older and have at least one child. The women were asked two
questions: (1) whether she was a stayathome mom or a
working mom; and (2) how she would rate herself as a parent
on the scale of 1, 2, 3, , 10; where 1 represents poorest
and 10 represents best.
The table below summarizes the poll results: the number of
women who rated themselves at the top of the scale (9, 10)
and below the top of the scale (18). Use these results to
answer the following questions.
Rate herself at the top
Stayathome Mom
212
Rate herself below the top
Total
Working Mom
198
402
600
a. Provide the values for the 3 missing cells to complete the
above table.
[2]
b. Which variable plays the role of the explanatory variable in this
study?
[1] Circle one:
Rating as Parent
Working
Status
(work or stay at home)
c. What is the probability that a randomly selected woman rates
herself at the top as a parent?
[2]
Final Answer: ______________
d. Knowing the mom is a working mom, what is the probability
she rates herself at the top as a parent?
[2]
Final Answer: ______________
e. Does there seem to be an association between whether a mom
staysathome or works and how she rates herself as a parent?
Explain your answer with appropriate numerical support.
[3]
Yes, there does appear to be an
association,
Circle one:
No, there does not appear to be an
association,
223
because
224
6. Voter Support: Consider the population of all registered
Michigan voters. Suppose 45% of all registered Michigan voters
support a particular candidate for political office. A random
sample of n = 400 registered Michigan voters will be selected
and the sample proportion that support the candidate will be
computed.
a.
If you consider all possible random samples of 400
registered Michigan voters, what is the approximate
distribution that describes the possible values for the sample
proportion? Be complete; give all aspects of the distribution.
[3]
b.
The approximate distribution in part (a) holds only if certain
data conditions are met. We already have that a random
sample was selected. State the remaining condition and verify
that it is met for this situation.
[2]
c.
What is the probability that a random sample of 400
Michigan voters will result in at least half supporting the
political candidate? Show your work.
[3]
Final Answer: _____________________
7. Lost Wages: Suppose the probability that a college student
leaves Las Vegas after 3 days as a winner is 0.20. Consider
seven such students (strangers to one another) who go to Las
Vegas for a 3day break. What is the probability that at least
two of the seven leave Las Vegas after 3 days as a winner?
[3]
225
Final Answer: _____________________
226
8. Tea or Coffee: A national coffee chain wants to estimate the
proportion of its customers that prefer tea over coffee. The
chain collected a random sample of 250 customers.
All
customers in the sample were asked if they preferred tea over
coffee. The 99% confidence interval for the population
proportion of all customers that prefer tea over coffee was
found to be (0.32, 0.48).
a.
What is the margin error for the confidence interval of (0.32,
0.48)?
[2]
b.
Final Answer: ______________
How many customers in the sample reported that they
preferred tea over coffee?
[2]
Final Answer: ______________
c. Which of the following are valid interpretations of the 99%
confident level? (Circle all that apply.) [2]
If the sampling procedure were repeated many times,
then approximately 99% of the resulting intervals
would contain the population proportion of all
customers that prefer tea over coffee.
There is about a 99% chance that the population
proportion of all customers that prefer tea over coffee
lies in the interval (0.32, 0.48).
For repeated samples of 250 customers from the same
population, the proportion of all customers that prefer
tea over coffee will fall in the interval (0.32, 0.48) 99%
of the time.
d. What sample size would be required to estimate the population
proportion with 99% confidence and a margin of error of 4%?
Show your work.
[2]
Final Answer: ______________
9. Got the vote? A politician claims that she will receive more
than 70% of the vote in an upcoming election. To test this
claim, a random sample of 20 voters will be surveyed. The
significance level is set at 10%.
227
a. In the sample of 20 voters, 17 stated they would vote for her.
The pvalue is 0.11. Which of the following is a reasonable
interpretation of these results?
[2]
This proves that the politicians claim is incorrect.
The politicians claim could be correct, but the sample size was too
small to detect it.
The null hypothesis should be rejected.
b. What distribution was used to compute the pvalue in part (a)?
[3] Give all details.
Distribution:_____________________________________
10.
Office Visits: For the past 10 years, Dr. Werner has
recorded the number of students who come to his Friday 12
noon to 1 PM office hours. Based on his observations he has
developed the following model.
X = number of
0
1
2
students
Probability
0.4
0.5
0.1
a. How many students would you expect during this
scheduled office hour?
Include the appropriate symbol in your answer and the units.
[3]
= ________________________________
b. What is the probability of at least one student coming to office
hours?
[2]
Final Answer: ______________
c. Given that at least one student will come to office hours on a
given Friday, what is the probability that exactly two students
will come?
[2]
Final Answer: ______________
11.
Disturbed Sleep: A psychologist claims that more than
25% of all American adults experienced disturbed sleep in the
past 6 months because they are worried about money. Let p =
the population proportion of all American adults that
228
experienced disturbed sleep in the past 6 months because they
are worried about money.
a. State the appropriate hypotheses to be tested in terms of p.
[3] H0:____________________________
Ha:____________________________
b. The significance level was set at 10%. A random sample of 500
American adults resulted in 140 stating they had experienced
disturbed sleep in the past 6 months because they are worried
about
money.
The
test
statistic
value
is
z = 1.55. Draw a picture that shows what the pvalue
corresponds to and report its value.
[3]
pvalue =
________________
c. True or False? One of the conditions required for the Z test for
the above hypotheses to be valid is that the model for the
original population be normal.
[2]
True
False
229
12.
Decision Making: A researcher prepares a Null Hypothesis
(H0) and an Alternative Hypothesis (Ha), runs an experiment,
and then makes a decision to reject or fail to reject the Null
Hypothesis.
In each case, use the given information to
determine which decision was made and circle it.
a. The pvalue was 0.08852 and the significance level was 0.05.
[1]
Decision: (circle one)
Reject H0 Fail to
Reject H0
b. A Type I error was made.
[1]
Decision: (circle one)
Reject H0
Reject H0
Fail
to
c. The result was statistically significant.
[1]
Decision: (circle one)
Reject H0
Reject H0
Fail
to
13.
Weights of Male High School Seniors: Suppose the
weights for high school male seniors are normally distributed
with an average of 166 pounds and a standard deviation of 7
pounds. Use this model to address the following questions.
a. Jim weighs 175 pounds. How many standard deviations is he
from the mean?
[2]
Final Answer: ____________________
b. What proportion of the male high school seniors weigh at most
170 pounds?
[2]
Final Answer: ____________________
c. Find the value to complete the following statement (show your
work).
40% of the male high school seniors weigh _________________
pounds or more.
[2]
230
Exam 2
Questions
231
Statistics 250 Fall 2006 Exam 2
1.
The Major League Baseball Organization would like to assess whether or
not players have the same batting averages during the nighttime games and
the daytime games. Eight players are selected at random and their
nighttime and daytime batting averages are collected for a given period of
time. Let d represent the population mean difference between the
nighttime minus the daytime batting average. Assume a 5% significance
level. Some SPSS output is provided.
Paired Samples Test
Paired Differences
Pair 1
a.
Nighttime  Daytime
Mean
.006375
Std. Deviation
.006545
Std. Error
Mean
.002314
95% Confidence
Interval of the
Difference
Lower
Upper
.011847
.000903
t
2.755
df
7
Sig. (2tailed)
.028
The QQ plot helps us check the
assumption that (circle one) [1]
The batting differences come
from a normal population.
The batting differences are
increasing on average.
The populations are normal.
Not helpful alone. We need two
more QQ plots to check the
normality of the daytime and
nighttime batting averages.
b. State the hypotheses to be tested.
[2]
H0: _________________ Ha: ___________________
c. Give the observed value of the test statistic.
[1]
Test Statistic: _________________
d. Sketch the distribution of the test statistic when we assume the null
hypothesis to be true, label this distribution and show the pvalue is in
terms of an area under this distribution.
[3]
Give the pvalue corresponding to the hypotheses in (b) and the observed
test statistic value in (c).
[2]
pvalue = ______________
f. At the 5% significance level, the decision is: (circle one)
e.
232
[1]
reject H0
fail to reject H0
g. Conclusion: There is (circle one) sufficient
insufficient
[1] evidence to conclude that the batting averages for players appear to be
different for daytime versus nighttime games, on average.
2. For a certain headache painreliever, about 25% of patients taking this
medicine experience nausea.
A new formulation of the headache painreliever has been approved which
should lower this nausea incidence rate. The research team planned a study
to test this theory using a 5% level of significance.
a. State the hypotheses to be tested and identify the parameter of interest.
[4]
H0: ______________________ Ha: _______________________
where the parameter ____________ represents
b. A random sample of 50 patients who experience headaches is selected and
will be given the new formulation to take at the onset of the next
headache. The number of patients that report nausea after taking the
medication will be recorded. State and verify the additional data condition
for testing the hypotheses in part (a).
[2]
c.
The study results: a total of 10 of the 50 patients reported nausea after
taking the medication.
[6] i.
Compute the observed value of the test statistic.
Test Statistic: __________________
ii. Provide the corresponding pvalue (show all work).
pvalue: _____________
iii. Circle the decision and give a one sentence realworld conclusion.
Decision:
reject H0
fail to reject H0
233
Conclusion:
d. Based on your decision in part (c), what mistake might have been made?
(circle one) [1]
Type I error
Type II error
234
3. Company records show that drivers get an average of 32,500 miles on a set
of AllWeather tires before an unacceptable level of wear on tires results.
Hoping to improve that figure, the company has added a new polymer to the
rubber that should help protect the tires from wear caused by extreme
weather. The results of a preliminary study from twenty randomly
selected drivers who tested the new tires gave a sample mean of 33,824
miles and a standard error of the sample mean of 323 miles. Let =
population mean tire mileage with the added polymer. Assume = 0.05 and
that all necessary conditions are met to carry out the procedure.
a. Clearly state the hypotheses to be tested.
[2]
H0: ______________________ Ha: _______________________
b.
[4]
Compute the value of the test statistic and give (bounds for) the
corresponding pvalue. Show all work.
Test Statistic Value: _____________ pvalue: ________________
c.
[2]
Five different representatives from the company make the following
claims. Circle all of the comments that are supported by the statistical
evidence and the knowledge you have learned:
Our results have proven that the new polymer increases the tire mileage.
Our results could have been due to chance.
Our results are inconsistent with the hypothesis that the population mean tire
mileage with the new polymer is 32,500 miles.
Oh my goodness, we got a pvalue of nearly zero which proves the null is false!
d. The company would like to conduct a larger experiment (with more drivers)
and keep the significance level at 5%.
The features of the new
experiment, before carrying it out, are: (Circle all that apply)
[2]
Higher statistical power.
Lower pvalue.
Lower Type II probability.
None of the above
235
4. A student researcher, Jackie, is interested in learning about how much
time students spend studying per week. She is curious to find out if study
time differs for female students versus male students. Data from a survey
of 92 undergraduate students is collected. One question asked students to
report the number of hours that they study in a typical week. The study
also collected some demographic data including sex. SPSS was used to
provide the following output.
Group Statistics
StudyHrs
Gender
Female
Male
Mean
15.000
12.925
52
40
Std. Deviation
9.852
9.135
Std. Error
Mean
1.366
1.444
Independent Samples Test
Levene's Test for
Equality of Variances
F
StudyHrs
Equal variances
assumed
Equal variances
not assumed
.216
Sig.
.643
ttest for Equality of Means
df
Sig.
(2tailed)
Mean
Difference
Std. Error
Difference
90% Confidence
Interval of the
Difference
Lower
Upper
1.033
90
.304
2.075
2.008
1.262
5.412
1.044
86.843
.300
2.075
1.988
1.231
5.381
a.
Jackie would like to compare the mean study time for female students to
the mean for male students using a 90% confidence interval.
Which confidence interval should she use? (circle one)
Unpooled
Pooled
[4] Give two reasons that support your choice (provide all details).
#1:
#2:
Report the confidence interval here: ___________________________
b. Based on the confidence interval, does there appear to be a difference
between the mean study time for these two college student populations?
[3] Circle one:
Yes
No
Explain using one simple sentence:
c.
Suppose Jackies researcher mentor requested a 95% confidence level be
used to compute the confidence interval based on the collected data
instead of a 90% confidence level. The width of the 95% confidence
interval would be:
[1] (circle one)
narrower
wider
the same
cant tell
236
5. A chain of coffee houses is trying to decide which of two bakers should
supply pastries for sale in the chains outlets. The managers of the chain
would like to test that the population mean sales of pastries from baker 1
will be higher than the population mean sales for the pastries from baker 2,
that is, baker 1 has higher mean sales appeal. They select 32 stores (in
comparable market areas) and randomly select half to display and sell the
pastries from baker 1 and the other half to display and sell the pastries
from baker 2. They obtain total dollar sales from a twoweek trial period
for each of the 32 stores. Assume = 0.01 and that all necessary
conditions are met to carry out the appropriate statistical procedure.
a.
[2]
b.
[2]
State the hypotheses to be tested.
H0:_________________________Ha:________________________
The 16 stores that sold pastries from baker 1 resulted in a sample mean
sales of $283 and a standard error of the sample mean of $7.50. What was
the sample standard deviation for the 16 baker 1 store sales? (Include units)
Sample standard deviation = _______________________
c.
In testing whether the mean sales of pastries from baker 1 will be
higher than the mean sales for the pastries from baker 2, the
corresponding
ttest statistic was 2.8 and the pvalue was 0.005. For each statement
below, clearly circle your answer. [1 point each]
i.
The results are statistically significant at the 1% level. True
False
ii. If this study were repeated many times and if the null hypothesis of no
difference between the population mean sales were true, we would see
a ttest statistic of 2.8 or larger in only 0.5% of the repetitions.
True False
iii. For this study the sample mean sales for baker 1 stores differed from
the sample mean sales for baker 2 stores by 2.8 standard errors (of
the difference in sample means).
True False
iv. The probability that is H0 true is 0.005.
True False
d. Circle the statistical decision and realworld conclusion that is appropriate.
[2] i.
Reject H0 (the population mean sales of pastries for baker 1 do not appear to
be higher than the population mean sales of pastries for baker 2)
ii. Reject H0 (population mean sales of pastries for baker 1 do appear to be higher
than the population mean sales of pastries for baker 2)
iii. Fail to reject H0 (there was insufficient evidence to demonstrate the
population mean sales of pastries for baker 1 is higher than the population mean
sales of pastries for baker 2)
237
iv. Fail to reject H0 (the population mean sales of pastries for baker 1 is equal to
the population mean sales of pastries for baker 2)
238
6. BioMatrix Inc. is conducting three different studies to assess the
effectiveness of a new medication that is to widen blood vessels and thus
reduce blood pressure. Two studies have been completed. Regretfully, the
3rd studys enrollment is so slow that the company has decided to end the
study and use the data collected at this point. This third study focuses on
learning about the population proportion of patients that will respond
positively to the medicine, denoted by p. The hypotheses to be tested are
given by H0: p = 0.40 versus Ha: p > 0.4. The study was terminated with only
11 patients but 10 of them responded positively. Use the appropriate
method and report the corresponding pvalue for this study.
[3]
pvalue: ______________________
7. A college dean plans a student survey to estimate the percentage of all
currently enrolled students who plan to take classes during the next spring
half term. She wants to be 90% confidence that the margin of error will
be at most 5%.
a.
What is the necessary sample size? Show all work.
[3]
Final
answer:
_________________
b. Suppose the survey is conducted and the resulting interval was found to be
(0.20, 0.30).
The following statement was written to summarize the
239
findings, but it is incorrect. Correct this statement by clearly crossing out
and replacing one word. (DO NOT rewrite the entire sentence.)
[2]
Based on the survey results, with 90% confidence, we would estimate the
sample proportion of currently enrolled students who plan to take classes
during the next spring half term to be between 0.20 and 0.30.
N
o
r
m
a
l
Q
P
l
o
t
f
L
i
e
_
T
i
m
e
4
,3
0
,2
0
,1
0
,0
0
1
,2
0
,02
,01
,00O
1
,e
0
2
,0
3
,04
,05
,0
b
s
rv
d
V
a
lu
e
E
x
p
e
c
td
N
o
rm
a
lV
u
e
8. A manufacturer of light bulbs selects at random 100 light bulbs from their
production line and measures the lifetime of each light bulb in hours (that
is, how long the light bulb stay on before it burns out). The manufacturer
would like to estimate the population mean lifetime of its light bulbs.
A histogram and a QQ plot of the lifetimes of this random sample are
shown below.
a. Do the data appear to come from a normal population?
[2] Explain briefly.
Yes
No
b. The Central Limit Theorem helps us in estimating the population mean
lifetime of the light bulbs because: (Circle all that apply)
[2]
It tells us that the test statistic will have a skewed to the right
distribution.
It tells us that the distribution of the sample mean will be same as the
population distribution.
240
It tells us that the distribution of the sample mean will be a t
distribution with n 1 degrees of freedom, t(n 1).
It tells us that the distribution of the sample mean will be
approximately normally distributed with its mean equal to the
population mean and its standard deviation equal to the population
standard deviation divided by 10.
9. Determine whether each of the following statements is true or false.
Clearly circle your answer. [1 point each]
a.
A researcher should consider the consequences of a Type I error in order
to choose .
True
False
b. If we were testing H0: p = 0.5 versus Ha; p > 0.5 and the experiment
resulted in a sample proportion of
p = 0.42, then the resulting pvalue
would be more than 0.5.
True
c.
False
A statistically significant finding does not necessarily have practical
significance.
True
False
d. Power of the test is the probability of correctly rejecting the null
hypothesis.
e.
True
False
The sampling distribution of a parameter is the distribution of all
possible values of the parameter if repeated samples are obtained.
True
False
10. For each of the following research questions determine whether the
primary method for analyzing the data should be a confidence interval or a
hypothesis test. If the primary method is a confidence interval, circle the
words Confidence interval, and state what the confidence interval is for
by giving the symbol on the blank line. If a hypothesis test should be done,
circle the words Hypothesis Test, and write out the corresponding null
and alternative hypothesis using the appropriate notation.
[6 points]
a.
Research question: What proportion of adults in the United States is in
favor of the death penalty for persons convicted of murder?
Primary Analysis Method: Confidence Interval for ______________
241
Hypothesis Test with H0:_______________
Ha:________________
b. Research question: Is the mean normal human body temperature less than
98.6 degrees Fahrenheit?
Primary Analysis Method: Confidence Interval for ______________
Hypothesis Test with H0:_______________
c.
Research question: On average, how much difference is there between the
adult heights of a father and his son?
Primary Analysis Method: Confidence Interval for ______________
Hypothesis Test with H0:_______________
Statistics 250 Fall 2007 Exam 2
1.
a.
[2]
Ha:________________
Ha:________________
A research paper in gerontology suggests that about half of all elderly
people prefer to live in an apartment rather than in a onefamily home. In a
recent sample survey of retired persons, 214 out of 400 stated that they
preferred an apartment.
What is the sample estimate of the population proportion of all retired
people who prefer an apartment?
Final answer: __________________
b. Compute the standard error of the estimate in part (a).
[2]
c.
Final answer: __________________
Using the above information, obtain a 90% confidence interval for the
population proportion of all retired people who prefer an apartment.
[2]
Final answer: ( ___________, ___________ )
d. We wish to assess if the population proportion of all retired people who
prefer an apartment differs from half. Use your confidence interval above
and give your decision.
242
[2] Decision: (circle one)
Reject H0
Briefly explain your reason.
Fail to reject H0
e.
Which of the following is an assumption for the inference above to be
valid? [2] Circle all relevant items.
i.
Original response is normally distributed.
ii.
The sample is selected randomly from the population of interest
iii.
The sample size is sufficiently large.
iv.
Histogram of sample values is symmetric and unimodal.
2. Bjork Larsen needed to determine whether to use a new racing wax for the
Swedish Nordic Ski Team. He thinks that the new wax will be worth the
price if it would lead to a faster race time, on average. In the past, the
average time for a standard length race was 1.40 hours. He selects a
random sample of 36 Swedish skiers and has each use the new wax system
for the upcoming standard length race.
a. Let = average race time for all Swedish skiers using the new wax system.
[2] State the appropriate hypotheses to be tested.
1
0
9
8
7
6
5
4
3
2
1
01
.01
.0R
1
.a
2
0
1
0
c
e
T
im
e
(h
o
u
r.3
s
)1.401.50
F
re
q
u
n
c
y
H0:_________________
Ha:________________
Some SPSS output based on the race
times for the 36 skiers is provided
above.
b. Complete the sentence.
The histogram does not look
approximately normal, rather, it
is somewhat skewed to the ____________. However, the ttest result can
still be considered valid due to what main result?
[2]
OneSample Statistics
Race Time (hours)
N
36
Mean
1.330
Std.
Deviation
.124
243
Std. Error
Mean
.021
OneSample Test
Test Value = 1.40
t
3.361
Race Time (hours)
c.
df
35
Sig.
(2tailed)
.002
Mean
Difference
.070
95% Confidence
Interval of the
Difference
Lower
Upper
.112
.028
What is the distribution of the test statistic if the new wax system does
not lead to faster race times on average?
[2]
Final answer: _________________________
d. The value of the test statistic is missing in the SPSS output. Provide its
value and the corresponding pvalue for testing the hypotheses in part (a).
[2] Test Statistic Value t = ___________
pvalue = _____________
e.
Based on the test results, which of the following is the appropriate realworld conclusion?
[1] Clearly circle your answer.
i.
The new wax system significantly decreased the average race time
supporting its use by the Swedish Nordic Ski Team.
ii.
There was insufficient evidence to demonstrate the new wax system
decreased the average race time.
3. Data was collected on the height of male college students and their
fathers. Suppose we are interested in testing the hypothesis that sons are
taller than their fathers, on average using a 5% significance level. The
researchers assistant had not taken Stat 350 and did not know which test
to perform, so he generated SPSS output for both the paired and the
independent samples ttests. You will need to determine which output is
appropriate to use as you conduct the test below.
Paired Samples Test
Paired Differences
Pair 1
Sons  Fathers
Mean
1.556
95% Confidence
Interval of the
Difference
Lower
Upper
.393
2.719
Std. Error
Mean
.584
Std. Deviation
5.090
t
2.665
df
75
Sig. (2tailed)
.009
Independent Samples Test
Levene's Test for
Equality of Variances
ttest for Equality of Means
1 = Sons
2 = Fathers
F
height
Equal variances
assumed
Equal variances
not assumed
4.138
Sig.
.044
Sig. (2tailed)
2.120
150
.036
1.556
.734
.106
3.006
2.120
116.202
.036
1.556
.734
.102
3.010
df
244
Std. Error
Difference
95% Confidence
Interval of the
Difference
Lower
Upper
Mean
Difference
[8]
State the hypothesis using the appropriate statistical notation.
H0:____________________
Ha:_____________________
Give the value of the test statistic and the corresponding pvalue.
Test Statistic = _____________ pvalue = _______________
Based on the pvalue, your decision would be (circle one):
reject H0
do not reject H0.
245
Question 3 continued
Based on the decision, the conclusion would be:
There is (circle one) sufficient insufficient evidence that sons are
taller than their fathers, on average.
Here are several plots that could be used
to check one of the assumptions required
for this test to be valid. Clearly state
that assumption and circle the plot(s)
that should be examined to check it.
The assumption is (be specific):
246
4. A medical doctor uses a diagnostic test to determine if her patient has
rheumatoid arthritis. The doctor will prescribe treatment only if she thinks
the patient has arthritis based on the test results. In a sense, the doctor
is using a null and an alternative hypothesis to decide whether or not to
administer treatment. The hypotheses might be stated as:
H0: The person does not have arthritis (no arthritis) versus
Ha: The person has arthritis.
a. The grid below presents the four possible combinations of the doctor's
decision and the "true" situation. There are two statements in each cell,
one about the doctor's decision and one about the patient's actual
condition. For each statement, clearly circle the word phrase in
parentheses that makes the statements match the doctor's decision and
the true state of the patient. There are eight statements all together, so
be sure to make a selection for each one.
[2]
b. Identify the cell that represents a Type 1 Error (circle your answer).
[1]
Cell A
Cell B
Cell C
Cell D
c.
Identify the cell that represents a Type 2 Error (circle your answer).
[1]
d.
[1]
Cell A
Cell B
Cell C
Cell D
Circle the realworld consequence of the doctor committing a
Type 2 Error.
Doctor
Doctor
Doctor
Doctor
administers an unnecessary treatment to the patient.
administers a necessary treatment to the patient.
does not administer a treatment that is necessary for the patient.
does not administer a treatment that is not necessary for the patient.
247
5. Regents of a large state university proposed a plan to increase student
fees in order to build new parking facilities. A news channel claims that
over 70% of the students are opposed to the plan. We wish to test this
claim, that is, test H 0: p = 0.70 versus Ha: p > 0.70. A random sample of 18
students is taken and 17 of them are opposed to the plan.
a.
[2]
What is the population parameter of interest here? Give the statistical
symbol and describe in 1 brief sentence what it represents in this situation.
Symbol: ________
Description:
b. True or False: In this example we can use normal approximation to binomial
distribution.
[2] Circle one:
True
False
Give specific support for your selected answer:
c. Provide the test statistic and pvalue for testing the hypotheses.
[4]
Show all work.
Test statistic: ____________
d.
[1]
pvalue: _______________
What is your decision at a 5% significance level? Circle one:
Reject H0
Fail to reject H0
e. Give a one sentence conclusion in the context of this problem.
[2] Therefore
f.
Suppose the alternative hypothesis is really true. Then the probability of
rejecting the null hypothesis would be called [1] Circle one:
Alpha
Type 1 error
Beta
Type 2 error
Level of significance
Power
g. What could the news channel do to produce a test with higher power?
[1] Circle all correct answers.
Sample more students
Repeat the study many times
Use a 1% significance level
248
6. A researcher will compare men and women on 100 different quantitative
variables. He will use a 5% significance level to carry out the 100
independent samples ttests (one for each variable). If, for each test, the
null hypothesis is actually true, about how many "statistically significant"
results will be produced?
[2]
Final answer: ______________
7. A large study was conducted about binge drinking behavior of college
students. Out of the 718 males surveyed, 163 revealed that they indulged
in binge drinking, while 169 of the 991 women surveyed responded
affirmatively. We wish to compare the two genders with respect to the
population proportion that indulge in binge drinking.
163
a.
0.227
The value 718
is a [2]
Circle all that apply:
Normal distribution
Parameter
Population proportion
Sample
Sample Proportion
b. We wish to test the null hypothesis that there is no difference between
men and women college students with respect to the population proportion
that indulge in binge drinking against the alternative that men are more
prone to binge drinking. Let 1=the population of all male college students
and 2=the population of all female college students.
[1]
c.
Finish the hypotheses to be tested. H0 p1 = p2 vs Ha: p1 ___ p2
Under the above null hypothesis, the population proportions are the same.
Compute the estimate of this common population proportion. Show all work.
[2]
Final answer: __________________
d. The test statistic is z = 2.91 and the pvalue is 0.002. Consider the
following statements and clearly circle all that are correct.
[2]
i. If this study were repeated many times, the null hypothesis would be
true only 0.2% of the time.
ii. If this study were repeated many times, we would see a z test statistic
of 2.91 or more extreme in about 0.2% of the repetitions.
iii. If this study were repeated many times and if there is no difference
between men and women college students with respect to the
population proportion that indulge in binge drinking, we would see a z
test statistic of 2.91 or more extreme in about 0.2% of the repetitions.
249
iv. The results are statistically significant at the 1% level.
8. A sample of 29 wines from a large distributor have been randomly selected
for a wine tasting event. Three responses that are commonly measured,
each on a quantitative scale, are aroma, flavor, and quality. The 29 wines
were also classified by the region where each was produced. You are asked
to provide a 95% confidence interval estimate for the difference in the
population mean flavor rating for the two regions. SPSS output is provided.
Assume all conditions for a twoindependent samples t interval are met.
Group Statistics
Flavor
Region
1
2
N
17
12
Mean
4.376
5.617
Std.
Deviation
.880
.915
Std. Error
Mean
.213
.264
Independent Samples Test
Levene's Test for
Equality of Variances
F
Flavor
a.
Equal variances
assumed
Equal variances
not assumed
.028
Sig.
.868
ttest for Equality of Means
df
Sig.
(2tailed)
Mean
Difference
Std. Error
Difference
95% Confidence
Interval of the
Difference
Lower
Upper
3.68
27
.0010
1.240
.337
1.932
.548
3.65
23.236
.0013
1.240
.340
1.942
.538
Based on the output, give two reasons why you should
report the pooled version of the two independent samples t interval for
these data. Be specific, that is, include values in each of your explanations.
[4]
Because
Because
b. Which of the following values must be the pooled estimate of the common
population standard deviation?
[1] Circle one:
c.
0.028
0.868
0.337
0.894
0.799
The pooled 95% confidence interval for the difference in the population
mean flavor rating for region 1 versus region 2 is given by (1.932, 0.548).
[2 points each]
i.
The probability that (1 2) lies between
1.932 and 0.548 equals 0.95.
ii.
True
If this procedure were repeated many times,
95% of the possible (1 2) values will fall
True
between 1.932 and 0.548.
250
False
False
9. For each scenario below, determine if the most appropriate statistical
analysis technique is conducting a hypothesis test (HT) or constructing a
confidence interval (CI). Then also give the symbol for the corresponding
parameter that is of interest. [2 points each]
a.
The supervisor of a tourist information desk at a local airport is interested
in estimating how long it takes an employee to serve a customer on average.
Using a stopwatch, he measures the amount of time (in minutes) it takes
for each of 10 randomly selected customers to be served.
Circle one:
HT
CI
Parameter of interest: ____________
b. In a recent survey, college students were asked the amount of time they
spend watching television on a typical day and the amount of time they
spend surfing on the Internet on a typical day. The researchers were
interested in determining whether the time spent surfing the Internet was
higher than the time spent watching television, on average.
Circle one:
c.
HT
CI
Parameter of interest: ____________
A study was done by randomly assigning 200 volunteers with sore throats
to either drink a cup of herbal tea or use a throat lozenge to ease their
pain. Each subject reported whether or not they experienced any relief.
The researchers were interested in assessing if the tea was more
effective as compared to the throat lozenge.
Circle one:
HT
CI
Parameter of interest: ____________
d. The state lottery office claims that the average household income of those
people playing the lottery is greater than $37,000. A sample of 25
households will be obtained and the incomes recorded to check
this claim.
Circle one:
e.
CI
Parameter of interest: ____________
A firm conducts a survey on a sample of employees. Management would like
to assess if a majority of employees use the onsite exercise studio during
the lunch hour.
Circle one:
f.
HT
HT
CI
Parameter of interest: ____________
Men and women shop at a retail clothing store. The manager would like to
estimate how much more (or less), on average, a woman spends on a typical
purchase occasion than a man.
Circle one:
HT
CI
Parameter of interest: ____________
251
Statistics 250 Fall 2008 Exam 2
1. The Wolf River is a river in Tennessee that flows past an
abandoned site once used by the pesticide industry. Various
toxins (e.g. hexachlorobenzene and aldrin) were released into
the river. Your supervisor has asked that you test whether the
mean aldrin concentration at the bottom of the river is greater
than the mean aldrin concentration at middepth, using =
0.10. To test H0: 1 = 2 versus Ha: 1 > 2, independent
random samples of river water were obtained at random sites
within 100 yards downstream from the site, resulting in the
following statistical summary:
a. You decide to perform the pooled version of the test, so you are
assuming
12 = 22. Give an estimate of that common population
variance. Show work.
[3]
Final
Answer
=
_____________________
b. If there is no difference between the population mean aldrin
concentrations, bottom versus middepth, what is the
distribution that allows us to calculate the pvalue?
[2]
Final Answer = _____________________________________
c. Give the test statistic value and corresponding pvalue for
testing the hypotheses about the mean aldrin concentrations.
[3]
Test Statistic Value: ________________ pvalue:
___________________
d. Circle the most appropriate conclusion at the 10% level:
252
[2]
There was insufficient evidence to show the mean
concentrations at river bottom was greater than at middepth.
The mean concentration at river bottom was significantly
greater than at middepth.
253
2. Feliz Da del Pavo! (Happy Turkey Day): Maya and her
family in San Antonio often eat enchiladas at Thanksgiving. The
website www.butterball.com reports that 90% of Americans eat
turkey at Thanksgiving. Due to the large Mexican American
population in her city, Maya guesses that the proportion of all
families in San Antonio that eat turkey at Thanksgiving will be
smaller than this national rate.
In order to test this hypothesis, Maya takes a random
sample of 10 families in her city, and asks them whether or not
they eat turkey at Thanksgiving. The results: 80% of the
families reported that they eat turkey. Assuming that Maya's
sample is representative of the population of interest, perform
a test of Maya's hypothesis at the 5% significance level.
a. State the hypotheses being tested.
[3] H0:________________________
Ha:_______________________________
b. Report the value of the test statistic and the corresponding pvalue. Show all work.
[5]
Observed test statistic: ____________
______________
pvalue:
c. Based on your results in (b), circle the appropriate answers
below:
[2]
Are the results statistically significant at a 5% significance
level?
Yes
No
Is there sufficient evidence to say that the proportion of all
families in
254
San Antonio that eat turkey at Thanksgiving is less than the
national rate?
Yes
No
3. Do Freshmen really gain weight? It is a commonly held
belief that freshman gain weight during their first year of
college. To test this claim, a Professor of Nutrition at a
University took a random sample of 16 freshmen enrolled in his
introductory health course. The weight for each of these
students (in pounds) was measured at the beginning and end
of the year.
There was no missing data and the same
calibrated scale was used for all 16 students before and after.
The difference in weight (After less Before) was computed and
some SPSS output using the resulting data is below.
Paired Samples Test
Paired Differences
Mean
Pair
1
a.
AfterWeight BeforeWeight
2.06250
Std. Deviation
Std. Error
Mean
2.26477
.56619
t
3.643
df
Sig. (2tailed)
15
.002
The data in this scenario are paired. In one simple
sentence, explain why.
[2]
b. State the appropriate null and alternative hypotheses to be
tested.
[2]
Ho:______________________ Ha:__________________________
c. Before reporting the test results,
the
data
were
examined.
Consider the plot shown at the
right and clearly circle the
appropriate statement below:
[2]
This plot helps us assess
if the population distribution
for the differences in weights
is normal
This plot shows that the differences in weights has an
increasing trend.
This plot is not needed due to the Central Limit
Theorem.
255
This plot is not helpful as we need to make two plots,
one for the weights at the beginning of the year and
another for weights at the end of the first year.
d. Report the test statistic value, the corresponding pvalue, and
the decision:
[3]
Observed test statistic: _______________ pvalue:
________________
Statistical decision at a 10% significance level: Reject Ho
Fail to reject Ho
e. The professor would like to use these results to infer about the
population of all freshmen college students. Based on the
above description of how the study was conducted, this would
not be appropriate due to:
[1] (circle one)
Nonresponse bias
Response bias
Selection Bias
4. Comparing Drug X versus Placebo:
An experiment is
conducted to compare a new Drug X to placebo. For the
measured response, higher values imply a better response.
The following hypotheses were to be tested at the 10%
significance level:
H0: The population mean response for Drug X is equal to that of
placebo (1 = 2)
Ha: The population mean response for Drug X is higher than that of
placebo (1 > 2)
a. To Pool or Not to Pool, that is the question Before testing the
above hypotheses we must decide whether or not it is
appropriate to pool the standard deviations from the
independent samples. Which of the following are acceptable
reasons that lead to deciding to pool? Circle all that apply.
[2]
Both QQ plots indicate normality.
The two sample standard deviations are similar.
Compare the two boxplots (one for each group) sidebyside.
The pvalue for Levenes test is 0.10.
b. True or False?
When appropriate, we prefer to pool since the resulting ttest is
more powerful than the unpooled ttest
[2] Circle one:
True
False
c. The test statistic for testing the above hypotheses is 2.34 and
the corresponding pvalue is 0.013. Which of the following are
256
appropriate statements regarding the value of 0.013? Circle all
that apply.
[2]
Since the pvalue < 0.10, we reject H0.
The probability that the population mean response for
Drug X is equal to that of placebo is 0.013.
If there is no difference between the population means,
the probability of seeing a test statistic
of 2.34 or larger is just 0.013.
The probability of rejecting H0 when in fact H0 is true is
0.013.
5. True or False?
a. If a sample size is large enough, then the shape of the
histogram of the same data will be approximately normal, even
if the population distribution is not normal.
[2]
Circle one:
True
False
b. If the null hypothesis is actually true, then it is not possible to
make a Type I error.
[2] Circle one:
True
False
257
6. How many vegetarians are there? A poll was conducted in
order to get an idea of the demand for vegetarian meals at
restaurants. The study was commissioned by the Vegetarian
Resource Group and was published in the Vegetarian Today. A
random sample of 1100 U.S. adults was asked whether or not
they sometimes order a vegetarian dish (one without meat,
fish, or fowl) when dining out.
Those
who
conducted
the
survey
collected
many
characteristics of the participants, including political party
affiliation. In this sample, 600 identified themselves as
Democrats and 500 identified themselves as Republicans. Of
those sampled, 252 Democrats and 200 Republicans said that
they sometimes order a dish without meat, fish, or fowl when
they eat out.
a. Find a 95% confidence interval for the difference between the
two population proportions of people that sometimes order a
vegetarian meal when going out to eat. Let group 1 refer to
Democrats and group 2 refer to Republicans. Show all work.
[4]
Final Answer: (______________ , ______________)
b. Based on the confidence interval, in part (a), is it reasonable to
assume that the proportion of Democrats who sometimes order
a vegetarian meal is different than the proportion of
Republicans who sometimes order a vegetarian meal?
[2]
Circle one:
Yes
No
Explain in one sentence:
c. In the article that was published in Vegetarian Today, the
author wrote, "The 95% confidence level tells us that if the
same survey were repeated 100 times, 95 times the same
responses would be returned." Is this a correct interpretation
of the 95% confidence level?
[2]
Circle one:
Yes
No
258
7. Radio ads: A company is willing to renew its advertising
contract with a local radio station only if the station can
demonstrate that more than 20% of all residents of the city
have heard the ad and recognize the company's product.
The radio station proposes conducting a phone survey of
400 randomly selected people in the city in order to test that
the proportion of all residents in the city for whom the ad was
effective is more than 20%.
a. The station plans to conduct the test using a 10% level of
significance, but the company wants the significance level
lowered to 5%. For which level of significance will the
probability of correctly concluding that the ad was effective for
more than 20% of all residents be higher?
[2] Circle one:
10%
5%
b. They finally agree to use = 0.05, but the company proposes
that the station call 600 people instead of the 400 initially
proposed. Will this make the probability of a Type II error higher
or lower?
[2] Circle one:
higher
lower
8. Confidence Intervals and testing Hypotheses:
A
researcher selected a fixed level of significance of 1% and
analyzed
a
data
set
to
test
the
hypotheses
H0: = 50 vs the alternative Ha: 50. The null hypothesis
was not rejected.
a. If the 99% confidence interval were constructed with the same
data, would the resulting interval contain the value of 50?
[2] Circle one:
Yes
No
Cant Tell
b. If the researcher had used a 5% level of significance, would the
decision still be to fail to reject the null hypothesis?
[2] Circle one:
Yes
No
Cant Tell
9. NEOPR Personality Inventory is a test designed to measure
various aspects of adult personality, one such aspect is
agreeableness. The scores for the measure of agreeableness
have a mean of 50 and a standard deviation of 10.
a. The test is administered to a class of 49 psychology majors and
their average score on the agreeableness scale was found to be
46. Suppose these psychology majors can be considered a
random sample of all psychology majors. If they are really no
difference in agreeableness from the population at large, what
is the probability of seeing a sample mean this low or even
lower? Show all work.
259
[4]
Final
Answer
=
________________
b. Give the name of the statistical result that allowed you to
answer part (a):
[2]
10.
Waiting for that Bus: A public bus company official
claims that the mean waiting time for bus number 14 during
peak hours is less than 10 minutes. Karen will take bus number
14 during peak hours on 18 different occasions and use the
results to test the officials claim.
a. In one simple sentence, explain what is wrong with the null
hypothesis
H0: x = 10.
[2]
b. Karens sample mean waiting time was 2.6 standard errors
below the null value of 10 minutes. What test statistic would
she use to perform this test? Provide the symbol and value of
the test statistic.
[3]
Test statistic: ________ = __________________________________
c. Suppose the pvalue was found to be 0.01 and the decision was
to reject the null hypothesis. What is the smallest possible
value for the significance level of the test that this researcher
could have used?
[2]
Final answer: __________________
11.
Maize and Blue Marbles: A statistician is going to play
the game What is in the Box?. He will win a monetary prize if
he can correctly identify the contents of a box (which he cannot
see inside). There are two possibilities for the contents of the
box shown below as competing hypotheses:
H0: The box contains eight maize marbles and two blue
marbles.
Ha: The box contains five maize marbles and five blue marbles.
The box of marbles will be thoroughly mixed. The statistician
will be allowed to select just 1 marble out of the box (without
looking at the full contents) and observe its color. He must
make his decision based on the color of that one selected
marble. He has picked the following reasonable decision rule to
use:
260
Reject H0 if the observed color of the selected marble is
BLUE.
a. For this situation, what is the probability of correctly rejecting
H 0?
[2]
Final answer: ___________________________
b. For this situation, what is the probability of incorrectly rejecting
H 0?
[2]
Final answer: ___________________________
c. What is the statistical name for the quantity described in part
(b)?
[2]
261
Stats 250 Fall 2009 Exam 2
1. High School Sports The Department of Education wants to
estimate the difference in rates at which high school boys and
girls participate in school sponsored sports. Let p1 represent
the population proportion of all high school girls that participate
in a school sport and let p2 represent the population proportion
of all high school boys that participate in a school sport.
Independent random samples were collected and the following
results were obtained.
Participate in
Sports?
Gender
Yes
No
Total
1=
57
43
100
Girls
2=
48
52
100
Boys
a. Provide an estimate of p1, including an appropriate symbol for
your answer.
[2]
Estimate of p1: _____ = _______________
b. A 95% confidence interval for the difference in population
rates, p1 p2, was calculated to be (0.05, 0.23). In order for
this confidence interval to be valid, certain conditions must
apply. The problem states that the samples are independent
random samples. Verify the remaining condition(s).
[2]
c. For each of the following statements, circle TRUE or FALSE.
points each]
i.
There is a 95% chance that p1 p2 is between 0.05 and
0.23.
True
False
ii.
iii.
[2
With repeated samples of the same sizes, wed expect p1
p2
to fall between 0.05 and 0.23 roughly 95% of the time.
True
False
With repeated samples of the same sizes, wed expect
95% of the resulting intervals to contain p1 p2.
True
False
262
iv.
With 95% confidence, we would estimate the population
proportion
of all high school girls that participate in a school sport to be
anywhere from 5% below to 23% above the corresponding
proportion for all high school boys.
True
False
263
d. Compared to the 95% interval of (0.05, 0.23), a 90% interval
would be
[2] circle one:
Wider
Narrower
Cant Decide
e. Suppose you want to test the hypothesis H 0: p1 = p2 versus Ha:
p1 p2 at the 5% level. Based on the confidence interval (.05, .
23), what would be your decision?
[2] Circle one:
Decide
Reject H0
Fail to Reject H0 Cant
2. Too much radiation? A government agency suspects that
the microwaves produced by Quasar emit more radiation than
the federal standard limit of 0.05 deciwatt (dW) on average. A
random sample of Quasar microwaves was selected and the
radiation emitted was measured. Those results were entered
into SPSS and some output is provided. Use a 5% significance
level.
a. Clearly state the hypotheses the government
agency wishes to test; include the definition of the parameter.
[4]
H0:
__________________________
Ha:
_____________________________
where the parameter ______ is defined to be
b. How many microwave ovens were randomly selected and used
in this study?
[1]
Final answer: ___________
c. Complete the sentence:
[2] The sample mean radiation was _________________ standard
errors above the standard limit of 0.05.
d. At the right, provide a
sketch and label of the
distribution for the test
statistic
if
the
null
264
hypothesis were true. Show the pvalue for testing your
hypotheses in part (a) in terms of an area under this
distribution.
[3]
265
e. Provide that pvalue and the decision at the 5% level.
[3] The pvalue for testing
___________________
the
hypotheses
in
part
(a)
is
At the 5% significance level, the decision is: (circle one)
reject H0
f.
fail to reject H0
Which one of the following is the correct conclusion? Circle one.
[2 points]
There is sufficient evidence to conclude that the mean
amount of radiation emission for all microwaves from this
manufacturer is greater than 0.05 dW.
There is sufficient evidence to conclude that the mean
amount of radiation emission for the sampled microwaves is
greater than 0.05 dW.
There is insufficient evidence to conclude that the mean
amount of radiation emission for all microwaves from this
manufacturer is greater than 0.05 dW.
There is insufficient evidence to conclude that the mean
amount of radiation emission for the sampled microwaves is
greater than 0.05 dW.
3. Bear Weight  Harry, a naturalist for the Alaska Fish and
Game department, studies grizzly bears with the goal of
maintaining
a
healthy
population.
He took a random sample of 51 grizzly bears and measured the
weights of those bears. The mean weight was 95.5 kg and the
standard error of the mean was 7 kg. Harry asks you, his
summer intern, to help him with some statistical analyses.
a. Harry would like to better understand what the value of 7 kg
means. Provide a clear interpretation of this value in terms of
an average distance.
[2]
b. Provide Harry with a 99% confidence interval estimate for the
population mean weight of such grizzly bears. Show all work.
266
[2]
Final answer: ( __________ ,
___________ )
c. You explain to Harry that there are certain conditions that are
required in order for the confidence interval to be valid. You
present him the following graph for checking one of those
conditions.
i.
[1]
Clearly state the condition that this graph is used to check.
ii.
Based on this plot, the required condition does not appear
to be met.
However you can assure Harry that the interval estimate is
still valid.
Why is that the case? Explain briefly:
[1]
267
4. Coronary bypass surgery is a procedure used to reduce the
risk of death from coronary artery disease. A random sample
of 60 patients who had emergency bypass surgery and an
independent random sample of 100 patients who had nonemergency surgery were obtained. Patients were classified
according to whether they survived the surgery. The results are
shown below:
Surgery Type
1 = Emergency
2
=
Nonemergency
Survived?
Yes
No
48
88
12
12
Tota
l
60
100
a. A medical researcher wants to determine if the proportion that
survive emergency surgeries is significantly less than the
proportion that survive nonemergency surgeries at the 5%
level.
Let p1 represent the population proportion of all
emergency surgeries that result in survival, and p2 represent
the population proportion of all nonemergency surgeries that
result in survival. State the hypothesis to be tested:
[2]
H0: __________________________ Ha: __________________________
b. Under the null hypothesis, the two population proportions are
the same. Compute the estimate of this common population
proportion. Show all work.
[2]
Final answer: _______________
c. Compute the test statistic for testing the hypotheses in part
(a). Show all work.
[2]
Final answer: _______________
d. Report the pvalue corresponding to the hypotheses in part (a)
and your test statistic in part (c). Show your work.
[2]
268
Final answer:
_______________
e. At the 5% level, do the data provide sufficient evidence to say
the population survival rate of emergency surgeries is less than
the population survival rate of nonemergency surgeries?
[2] Circle one:
Yes
No
f. If there is no difference between the two population survival
rates, and we were to repeat this study many, many times,
each time conducting the same hypothesis test at the same
significance level, then we would expect to reject H0 (circle
one)
[2] ... about 95% of the time. ... about 5% of the time.
Cannot be determined.
5. Does it make sense? Decide whether each statement is true
or false.
[2 points each].
a. If the model for employee salaries at a company is strongly
skewed right,
and a large number of employees are selected randomly, then
the
model for employee salaries will become approximately
normal.
True
False
b. The pvalue is the probability that the null hypothesis H0 is
correct.
True
False
6. Which Technique? For each scenario below, determine if the
most appropriate statistical analysis technique is conducting a
hypothesis test (HT) or constructing a confidence interval (CI).
Then also give the symbol for the corresponding parameter
that is of interest. [2 points each]
a. A fitness instructor is working on a project which involved
having a random sample of HS students participate in a fitness
course during their last semester in HS. For each student it
was determined whether or not their fitness level improved.
The instructor would like to assess if the improvement rate for
all male students is higher than that for all female students.
Circle one:
_________________
HT
CI
Parameter of interest:
b. An advisor at a college has been given the task to estimate the
percentage of all graduating seniors who were involved a
research project with a professor during the senior year. He
decides to send out a survey to a random sample of 200
269
graduating seniors and one of the questions will ask whether or
not they were involved a research project with a professor
during the senior year.
Circle one:
_________________
HT
CI
Parameter of interest:
c. The manager of an industrial plant is required to be sure the
plant discharges no more than the specified limit of 1000
gallons of wastewater per hour, on average, into a neighboring
lake. Members of an environmental action group plan to
observe the amount of wastewater for a random sample of 16
hours to check whether or not the company is exceeding the
1000 gallons/hour limit.
Circle one:
_________________
HT
CI
Parameter of interest:
7. Time to Graduate A small private college states that the
model for time to graduate with a bachelors degree (in years)
is somewhat skewed to the right, and that the mean time is 4.5
years with a standard deviation of 1.1 years. A random sample
of 100 recent graduates will be selected and the sample mean
time to graduation (in years) will be computed. Provide a
detailed sketch of the (approximate) distribution of the possible
values for the sample mean time; include labels and values.
[3]
4.0
4.8
4.1
4.9
4.2
5.0
4.3
4.4
4.5
4.6
4.7
Label for xaxis
8. Casino or Not? A lobbyist is in support of some new
legislation that would allow casinos in a conservative
Midwestern state. The lobbyist quotes the result reported in
USA Today that more than twothirds of the U.S. adults approve
of casino gambling. The governor of this conservative
270
Midwestern state decides to initiate a survey to assess whether
this rate actually applies to the adults in her state.
a. Her proposed study plan reports the hypotheses to be tested
are H0: p = 2/3 versus Ha: p > 2/3. The plan also includes the
following sentence.
The symbol p represents the population proportion of
U.S. adults
approving of casino
gambling.
[2] Is this definition of p appropriately stated?
Circle one:
Yes
No
If not, correct it by clearly marking up the above sentence.
271
b. The Governors Pilot Study Before spending the budgeted
$80,000 on a large scale survey, the research team conducted
a small pilot study to ensure the methodology was acceptable.
The sample size was just 10; however, only 7 people responded
(a 30% nonresponse rate). Of the 7 people who responded, 6
people approved of casino gambling. For this pilot study, what
would have been the appropriate pvalue to report? Show all
work.
[3]
Final answer: _______________
c. Confirming the methodology was acceptable, a larger survey
was conducted. A total of 348 people out of a random sample
of 500 stated they approve of casino gambling.
The
corresponding test statistic gave a pvalue of 0.0823. Using a
5% level of significance, clearly state the appropriate
conclusion.
[2]
9. Which JAR?: A Game of Chance Kevin pays $5 to play the
following game of chance. There is one jar which contains 10
coins, some combination of pennies (i.e., 1), nickels (5),
dimes (10), or quarters (25). Kevin cant see the coins in the
jar, but he is told the jar is either JAR A or JAR B with the
following two possible distributions of the 10 coins.
272
JAR A: 4 pennies, 3 nickels, 2 dimes, and 1 quarter
JAR B: 1 penny, 3 nickels, 3 dimes, and 3 quarters
The game of chance is as follows: One coin is selected at random
from the jar. On the basis of that coin, Kevin must decide whether
he thinks the jar is JAR A or B.
If hes correct, he will win $25, but if he is wrong, he loses the $5 it
cost to play the game. Kevin is a statistics student, so he views
this situation as choosing between two hypotheses:
H0: The jar is JAR A
Ha: The jar is JAR B
Knowing that the coins in JAR B are generally of higher value, he
decides on the following decision rule:
If the randomly selected coin is a quarter (i.e., 25), he will
reject H0
(and conclude JAR B).
If the randomly selected coin is not a quarter , he will fail
to reject H0 (and conclude JAR A)
a. For this decision rule, what is the significance level, also known
as ?
[2]
Final answer: _______________
b. For this decision rule, what is the power of the test?
[2]
Final answer: _______________
c. Suppose Kevin plays the game and the resulting coin is not a
quarter. Based on the decision rule, you would Fail to Reject H 0.
What type of error could he have made? (Circle One)
[2]
273
Type 1
possible
Type 2
No Error is
Final Exam
Questions
274
Statistics 250 Fall 2006 Final Exam
1.
Diabetic Treatment: Two treatments (A and B) are to be compared as to
their longterm effect for diabetic patients. One of the variables
measured is overall response to treatment (improved, no change, or
worsened). Of the total of 200 diabetic patients available for the study,
half are randomized to each of the two treatments. Two years later, the
following data is obtained.
Treatment A
Treatment B
Total
Improved
66
54
120
No change
24
26
50
Worsened
10
20
30
Total
100
100
200
a.
What test should be performed to assess if the distribution of the overall
response is the same for the two treatment populations?
[2]
b.
[2]
Name of the test: __________________________________________
If the distribution of the overall response is the same for the two
treatment populations, how many diabetic patients taking Treatment A
would have been expected to improve?
Final answer: ___________________
c.
[2]
The observed test statistic value is X2 = 4.61. Determine the contribution
to this value of 4.61 that came from the Treatment A patients that
improved (i.e. compute the first term corresponding to the first cell of the
table that led to the test statistic value).
Final answer: ___________________
d. Give the corresponding pvalue (show your work by including a sketch of
the pvalue).
[2]
Final answer: ___________________
e. What is the appropriate conclusion at a 5% significance level?
[2]
275
Thus, the distribution of overall response (circle one): does does not
appear to be the same for the two treatment populations.
2. Trash Bag Strength: Jay Krug works for a distribution company that
purchases trash bags from a vendor. The bags are required to have a
breaking strength of at least 47 pounds. The bag vendor claims that its
production process is relatively stable and produces bags with a mean
breaking strength of 53 pounds and a standard deviation of 4.2 pounds. A
week ago Jay consulted with you asking how he should go about checking
this claim. You suggested obtaining some data.
So Jay strikes an agreement with the vendor that permits him to
sample from the vendors production process. A sample of 49 bags was
randomly selected and the breaking strength of each was measured. The
mean breaking strength for the sample was found to be 2 pounds below the
process mean breaking strength cited by the vendor. Jay does remember
that sample means are statistics which vary from sample to sample, but he
is not sure if his result reflects more variation than due to chance. He
seeks your help in interpreting his result.
a.
What was the observed sample mean breaking strength for the 49
randomly selected bags?
[1]
Final answer: ___________________
b. Suppose the bag vendors claim is true and many random samples of 49 bags
were obtained and for each sample, the sample mean was computed. What
is the distribution (model) for the possible values for the sample mean?
[3] Be specific.
c.
Assuming the bag vendors claim is true, what is the probability of obtaining
a sample mean as far or even farther below the process mean than that
observed by Jay? Show all work.
[3]
Final answer: ___________________
d. What does your answer in part (c) suggest about the validity of the bag
vendors claim? (circle one)
[2]
The vendors claim is reasonable.
The vendors claim is not reasonable.
276
The sample size was too small.
The probability is high that the vendors claim is true.
e. Suppose the sample size had been larger. How would the distribution in
part (b) have changed?
[2] (circle all that apply)
the
the
the
the
mean would have increased
mean would have decreased
variability would have increased
variability would have decreased
3. Christmas Trees: A survey was conducted for the Department of Natural
Resources. For those respondents that indicated they do have a decorative
tree up in their home, a question was asked as to whether that tree was
natural or fake. One of the demographic variables measured was whether
they lived in a rural or urban area. This variable was used to create two
independent sets of responses (each assumed to be a random sample) from
the two populations of those who live in a rural area versus those who live
in an urban area. The data were used to provide the following information
about the number who have a natural tree.
The researcher was interested in assessing if the population proportion of
those with a natural tree differed between the two areas rural versus
urban. That is, they wish to use a 1% level of significance to test the
hypotheses: H0: p1 = p2 versus Ha: p1 p2.
a.
[2]
Since the two sample sizes above are different, it will be helpful to report
the two sample proportions. Give these two sample proportions.
Proportion of respondents in a rural area that have a natural tree: ______
Proportion of respondents in an urban area that have a natural tree: _____
b. Before the hypothesis test can be performed, you remember that there
are some conditions that need to be assessed. We can assume that these
two samples are indeed independent random samples from the larger
populations of interest. So you have another condition that should be
checked before computing the test statistic. Perform this check.
[2]
277
c.
[5]
You are ready to perform the hypothesis test. Provide the value of the
observed test statistic and the corresponding pvalue. Show all work.
Test statistic: ________________
pvalue = __________________
d. Are the results statistically significant at the 1% level?
[1] Circle one:
yes
no
4. The following graph based on a USA Today article shows the distribution
of favorite style of brownies.
a.
Which of the following descriptions is appropriate for this graph?
(You might think about rotating the plate 90 degrees counter clockwise.)
[1]
Symmetric
Uniform
Rightskewed Leftskewed
278
None is appropriate
b. A student in Stats 350 saw this distribution and wanted to assess if this
stated distribution applies to the population of UM undergraduates. She
takes a random sample of 100 such students and records their responses.
State the appropriate null hypothesis to be tested for assessing the USA
Today stated model applies to the population of UM undergraduates.
For expressing the hypothesis, let 1 = fudgy, 2 = cakey, 3 = both same,
4 = neither, and 5 = dont know.
[2]
H0: ________________________________________________
c.
[2]
If the null hypothesis were true, what is the expected value of the test
statistic?
Final answer: ___________________
d. Suppose the observed test statistic value turns out to be 4.0. Using your
answer in part (c), briefly explain why the decision will be fail to reject H 0.
[2]
5. Snack Foods: A maker of crackers and other snack food was markettesting two versions of a microwaved snack called Version K and Version
S. A total of 200 adult volunteers were available for a study. Assume
these represent a random sample of potential adult consumers. The 200
adults filled out questionnaire and the results were used to compute a
Target Customer Score (TCS) based on how desirable person would be as a
customer. Then the TCS scores were ordered from smallest to largest to
form pairs of similar customers. Within each customer pair, one was
randomly assigned one to taste snack version K and the other to taste
snack version S.
The primary response was the approval rating for the tasted snack (with a
higher rating indicating higher approval). The differences in approval
rating (Version K Version S) were computed for each customer pair and
are to be used to assess if there is adequate information to say that one
snack version is preferred over the other. A 5% significance level is to be
used. Some SPSS output is provided next.
279
Paired Samples Test
Mean
Pair
1
Version K Rating Version S Rating
1.66
Paired Differences
95% Confidence
Interval of the
Std.
Difference
Std.
Error
Deviation
Mean
Lower
Upper
7.07
.71
3.06
.26
t
2.35
df
Sig. (2tailed)
99
.021
a. State the hypotheses to be tested.
[2]
H0: ______________________ Ha: ______________________
b. The observed value of the test statistic is missing in the above output.
Provide its value (showing how you arrived at that value).
[2]
Test Statistic: ________________________
c.
Circle each of the following statements that are an assumption required for
performing the test.
[2]
i.
the standard deviation for the ratings for the version K snack is
equal to the standard deviation for the ratings for the version S snack.
ii.
the ratings for the version K snack are normally distributed.
iii.
the ratings for the version S snack are normally distributed.
iv.
the difference in ratings (diff = K S) is normally distributed.
v.
the ratings for the version K snack are independent of the ratings
for the version S snack.
d. Is there conclusive evidence that one snack version is preferred over the
other at the 5% level?
[2] Select just one and complete the statement.
Yes, and snack version ________ appears to be preferred.
No, the two snacks appears to be equally preferred because
_____________________________________.
280
6. An advisor looks at the schedules of his 200 students to see how many
math and science courses each has registered for. Answer the following
questions based on his summary table below.
0 Science
1 Science
2 Science
Totals
0 Math
44
36
20
100
1 Math
24
40
0
64
2 Math
16
12
8
36
Totals
84
88
28
200
Lets define the following events:
A is the event that a randomly selected student is taking an equal
number of math and science courses.
B is the event that a randomly selected student is taking more math
courses than science courses.
C is the event that a randomly selected student is taking one math and
two science courses.
a. What is P(A)?
[2]
Final answer: ________________
b. What is P(C)?
[2]
Final answer: ________________
c.
The events A, B, and C are: (circle one)
[1] Complementary Mutually Exclusive
Independent
None of the above
d. Suppose a random sample of 20 students will be selected. Let X be defined
as the number of students in the sample that are taking an equal number of
math and science courses. What kind of a random variable is X?
[2] Circle one:
i. Approximately Standard Normal
ii. Binomial with n = 20, p = 0.54
iii. Binomial with n = 20, p = 0.46
iv. Uniform random variable
e.
What is the expected number of students in a sample of 20 that take an
equal number of math and science courses?
[2]
Final answer: ________________
281
7. Waiting in Line: A supermarket experimented with three different
staffing policies to try to determine when to open new check out registers
or add more staff in service areas. The time of most concern is weekdays
between 5 and 6 pm when the store is crowded with impatient shoppers.
The manager of the store was told to vary the three different staffing
policies in a random way.
A hired shopper came in each weekday shortly after 5 pm and measured
the waiting time (in minutes) needed to purchase a standard set of items.
The data have been entered into SPSS and the following results obtained.
a. Complete the sentences:
[2]
The median wait time under policy 2 took was _____ minutes.
About 75% of the wait times under policy 1 were ____ minutes or less.
b. The manager would like to assess if there are any differences in the
population mean waiting time under the three different policies (1, 2, and
3) using a oneway ANOVA. There are
many assumptions to be checked. The
boxplots above and the output below
are used to check one of the
assumptions.
Clearly state that
assumption and determine if that
assumption is reasonably met.
[3] Assumption being checked:
Does the assumption seem reasonably met? Circle one:
282
Yes
No
c.
[2]
Since this data was collected over a series of days, the manager
remembered (from his days in Stats 350) that he should examine the data
using time plots (one for each policy). Each plot showed approximately a
random scatter with no apparent pattern in roughly a horizontal band. If a
pattern had been found in any of these plots (say a downward trend), which
assumption underlying ANOVA would be called into question? (Be specific.)
Assumption:
Some additional ANOVA output is provided below.
ANOVA
wait
Between Groups
Within Groups
Total
Sum of
Squares
152.445
257.922
410.367
df
2
33
35
Mean Square
76.223
7.816
F
9.752
Sig.
.0005
d. Which of the following statements is true about the value of 7.816 in the
ANOVA table?
[2] Circle one:
It is a parameter of ANOVA.
It is an estimate of the common population variance.
It is an estimate of the common population standard deviation.
It is an estimate of the common sample standard deviation.
e.
[2]
What hypotheses are being tested with the F statistic given in the above
ANOVA table?
H0: __________________________________________________
Ha: __________________________________________________
283
f.
Is there statistical evidence at the 5% level to suggest that the mean
waiting times are different for the 3 staffing policies?
[2] Circle one:
Yes
No
Explain in one simple sentence:
g.
Multiple comparisons were performed using the Tukey method.
[2] Use the results and clearly circle all pairs that are significantly different
(at a 5% level).
Multiple Comparisons
Dependent Variable: wait
Tukey HSD
(I) policy
1
2
3
(J) policy
2
3
1
3
1
2
Mean
Difference
(IJ)
4.91
1.13
4.91
3.78
1.13
3.78
Policy 1 vs. Policy 2
Std.
Error
1.13
1.13
1.13
1.25
1.13
1.25
Sig.
.0003
.5811
.0003
.0130
.5811
.0130
95% Confidence Interval
Lower Bound Upper Bound
2.14
7.67
1.64
3.89
7.67
2.14
6.85
.71
3.89
1.64
.71
6.85
Policy 1 vs. Policy 3
Policy 2 vs. Policy 3
8. The director of admissions at a university administered a newly designed
entrance test to 25 incoming freshman students selected at random. At
the end of the freshman year, the grade point average (GPA) was recorded
for these 25 students.
The focus of the study was to learn about the distribution of scores on
this entrance test, the distribution of end of freshman year GPA, and to
assess whether the grade point average (GPA) could be predicted from this
entrance test score.
Use the plots and the SPSS output to answer the various questions. In all
cases for testing in this problem, let = 0.05.
284
Descriptive Statistics
N
GPA
Entrance Test Score
25
25
Minimum
1.40
3.50
Maximum
4.00
6.30
Mean
2.59
5.03
Std. Deviation
.7769
.7385
Model Summary
Model
1
R
.645
R Square
.417
Adjusted
R Square
.391
Std. Error of
the Estimate
.606
ANOVA
Model
1
Regression
Residual
Total
Sum of
Squares
6.04
8.45
14.49
df
1
23
24
Mean Square
6.035
.367
F
16.425
Sig.
.0005
Coefficients
Model
1
(Constant)
Entrance Test Score
Unstandardized
Coefficients
B
Std. Error
.83
.851
.68
.168
285
Standardized
Coefficients
Beta
.645
t
.971
4.05
Sig.
.342
.0005
Question 8 continued:
a.
What is the overall range of the GPA values for this sample of
25 students?
[1]
Final answer: ___________________
b.
[1]
c.
[2]
d.
[1]
Complete the following sentence.
Based on the regression analysis above, approximately _________ % of
the GPA variation at the end of the freshman year can be explained by its
linear relationship with the entrance test score.
Give the equation of the estimated regression line for predicting GPA.
The unstandardized coefficient value of 0.68 given in the SPSS output
is an example of (circle one):
a sample statistic a population parameter a test statistic an intercept
e.
By how much would you expect a students GPA at the end of the
freshman year change for a two point increase in the entrance test score?
[2]
Final answer: ___________________
f.
What is the predicted GPA value for a student who got the lowest
score on the entrance test?
[2]
Final answer: ___________________
286
Question 8 continued:
g.
Compute the approximate residual for the student from part (f).
[2]
Final answer: ___________________
h.
The standard deviation for regression can be thought of as measuring
the average distance of the predicted GPA values from the population
regression line (that is, the average size of the prediction errors).
Complete the sentence:
[1]
i.
[4]
On average, actual GPA values are about ___________
from their predictions using regression.
The director of admissions would like to assess if there is a significant
positive linear relationship between GPA and the entrance test scores.
Give the appropriate hypotheses in the provided spaces, the value of the
test statistic, and the corresponding pvalue.
H0:____________________
Ha: ____________________
Test Statistic value = ___________ pvalue = ____________
j.
[2]
The plot at the right is used to check the assumption that: (circle one)
The entrance exam
scores are normally
distributed.
The true error terms
have constant
variance.
The residuals have
constant variance.
The GPA values are
randomly distributed.
287
Question 8 continued:
k.
A 95% confidence interval for the mean GPA at the end of the
freshman year for all students that have a certain entrance test score will
be the narrowest when the entrance test score is: (circle one)
[1]
l.
2.59
3.50
5.03
6.30
The study is criticized for not taking into account the socioeconomic
background of the entering freshman class. The socioeconomic background
would be an example of: (circle all that apply)
[1]
a response variable
a confounding variable
a dependent variable
a normal variable
9. Determine whether each of the following statements is true or false.
Clearly circle your answer.
[2 points each]
a.
Results that are statistically significant are always practically
unimportant.
True
False
b. Increasing the sample size increases power.
True
False
c.
Although subjects may be randomized to treatments, the researcher
still needs to consider whether the subjects are representative
enough for a particular population.
True
False
d. A 95% confidence interval for 1 2 was created based on independent
random samples from two populations. If the study were repeated many
times, we would expect 95% of the resulting intervals to contain 0.
True
False
e.
The Registrars Office of a college reported that 56% of all currently
enrolled students were out of state. It would be inappropriate to use
this information to test the hypothesis that the proportion of all
currently enrolled students are out of state.
288
True
False
289
Statistics 250 Fall 2007 Final Exam
1.
What contributes most to a persons general well being? Is it friends and
social life, job, or health and physical condition? This question was posed
to subjects that are independent samples from three adult populations in
the United States: single adults, married adults, widowed/divorced adults.
The results are given in the following table.
Primary Well Being Source
Friends and social life
Job
Health and physical condition
Total
a.
Singl
e
41
27
12
80
Marrie
d
49
50
21
120
Tota
l
132
110
58
300
Widowed or
divorced
42
33
25
100
The researchers goal was to assess if the distribution of the primary well
being source is the same for the three populations. Give the name of the
statistical test that should be performed.
[3] Name of the test: _______________________________________
b. If the distribution of primary well being source is the same for the three
populations, what is the expected number of single adults who would select
Job as the primary well being source?
[2]
Final Answer: _______________________
c.
SPSS was used to perform the
test and some output is
provided below.
Use it to report the following
information and your conclusion
at a 5% significance level.
ChiSquare Tests
Pearson ChiSquare
Likelihood Ratio
LinearbyLinear
Association
N of Valid Cases
Value
5.337a
5.210
2.794
df
4
4
Asymp. Sig.
.254
.266
.095
300
a. 0 cells (.0%) have expected count less than 5. The
minimum expected count is 15.47.
[3]
Test Statistic Value: _____________
pvalue: ________________
Thus, the distribution of primary well being source
(circle your answer):
does
does not
appear to be the same
for the three adult populations represented by these samples.
290
2. We have to decide whether to spend our advertising budget on placing our
commercials in a comedy show or during a televised football game. Assume
that the cost for these two options is the same.
A study was conducted to compare the product memory of viewers for
these two types of shows. The results: out of the 80 people in the study
who saw a commercial during a comedy show, 56 were able to remember the
product in the commercial two hours later; while out of the other 60
subjects in the study that saw a commercial during a football game, 42
were able to recall the product from the commercial two hours later.
We wish to use this data to assess if there is a significant difference
between the two memory rates for the two populations represented by
these samples using a 10% significance level.
a. State the appropriate hypotheses.
[2]
H0: _____________________
Ha: _____________________
b. Which of the following are required conditions for this test to be valid?
[2] Clearly circle all that apply.
Each sample is a random sample.
The two samples are independent.
Each population has a normal model for the response.
The two population variances are equal.
c.
[2]
Compute the value of the test statistic. Report its symbol and value and
show all work.
Test statistic ______ = ______________________
d. If there is no difference in the two memory rates for the two populations,
the test statistic has what distribution? (be specific)
[2]
Final answer: _______________________________
e.
What will be your statistical decision?
Circle one: Reject H0
Fail to reject H0
[2] Briefly explain:
291
3. Seasonal Influence? The director of a small psychotherapy clinic would
like to assess if there is any difference in the use of the clinic during
different seasons of the year. Last year there were 28 new clients in the
winter, 33 in the spring, 16 in the summary and 51 in the fall. You are
asked to use this data to assess whether or not season makes a difference.
Season
Obs Count
1 = Winter
28
2 = Spring
33
3 = Summer
16
4 = Fall
51
a. State the appropriate null hypothesis to be tested.
[4]
H0: __________________________________________________
b. If the null hypothesis is true, what is the expected value of the test
statistic?
[2]
Final answer: _________________________
c.
[4]
The observed test statistic value is 19.8. Sketch a picture to show the
pvalue of the test and then give the corresponding bounds for the pvalue.
Final answer: _______________________________
d. What is the decision at the 1% significance level?
[2]
e.
[2]
Circle one:
Reject H0
Fail to reject H0
What type of error could have been made? Circle one:
Type I Error
Type II Error
292
Cannot be determined
4. In Chicago, a recent study was conducted to assess the effectiveness of
new antibiotic for strep throat. The Rogers Park Clinic randomized 33
children (aged 614) to one of three treatments (the standard treatment,
the new antibiotic at a low dose, and the new antibiotic at a higher dose).
The children were observed as to the number of days to cure (lower
numbers being better). The table
at the right contains some results
of this study: the means and
standard
deviations
for
the
number of days to cure.
a.
This study is: (circle one)
an observational study
an
experiment
[2]
b. State the appropriate hypotheses for assessing if the three treatments
are equally effective in terms of the average days to cure.
[2]
H0: ________________________ Ha: _______________________
c.
True/ False Questions (circle T or F for each bullet). [4]
A large pvalue for Levenes test would support
the equality of the population variances.
True
False
The normality of the data is assured due to the central limit theorem
since our sample size is n = 33.
True False
d. Consider the following analysis of variance table.
[4] A number of values are missing. Complete these missing values.
e.
Based on the results of the test in part (b), the following three multiple
comparisons using Tukeys method are also provided from SPSS.
293
[2] Is there a significance difference between 1 and 2, the high and low dose
of the new treatment, at the 5% level? Circle one: Yes
No
Explain (be specific):
5. Do the major league baseball (MLB) teams that spend more money on
players salaries win more games? To investigate this question, a sports
fanatic regressed a teams winning rate (proportion of games that were won
by a team during the regular season) on the corresponding teams total of
the players salaries (in millions of dollars) for a random sample of 20 MLB
teams from the 2007 season. The SPSS output for this regression is
provided. Use this output to answer the following questions.
Model Summary
Model
1
R
.689
R Square
.474
Adjusted
R Square
.445
Std. Error of
the Estimate
.04366
ANOVA
Model
1
Regression
Residual
Total
Sum of
Squares
.0309
.0343
.0652
df
1
18
19
294
Mean Square
.0309
.0019
F
16.23
Sig.
.0008
Coefficients
Model
1
(Constant)
2007 Salary (in millions)
Unstandardized
Coefficients
B
Std. Error
.4335
.02204
.0010
.00025
Standardized
Coefficients
Beta
.689
t
19.670
4.029
Sig.
.0000
.0008
a. What is the value of the coefficient of correlation? ______________
[2]
b. The model summary provides a value of R square of 0.474 that is often
interpreted as a percent. Clearly state that interpretation.
[2]
c.
[3]
Write the equation of the least squares regression line.
Final answer: _______________________________________
d. Based on this equation, what is the predicted winning rate for the Boston
Red Sox, a team whose player salaries totaled $143,000,000?
[2]
Final answer: ____________________________________________
e.
The actual winning rate for Boston was 0.593. What is the value of the
residual for the Red Sox?
[2]
Final answer: ____________________________________________
f.
Complete the sentence: Actual winning rates are expected to vary around
the regression line
[2]
by about ___________________ on average.
g.
[2]
A predication interval for the winning rate of the Boston Red Sox would be
(circle one)
wider
narrower
than a confidence interval for the mean winning rate of all teams who spent
$143,000,000 on players salary.
h. Report the test statistic and the pvalue for testing if salary significantly
increases the winning rate. Then complete the conclusion by circling the
appropriate phrase.
295
[4]
Test statistic = _________________ pvalue = _______________
Thus, it (circle one) does does not
appear that the teams that spend
more money on players salaries win more games during the regular season
on average.
i.
Does this imply that spending more money on players salaries will cause
teams to win more games? Clearly circle the one correct answer.
[2]
l00..15Q
N
o
r
m
a
P
lo
tfU
n
s
ta
n
d
riz
e
d
R
e
s
id
u
a
l
Assumptions: In order for the test statistic and inference conducted in
part (h) to be valid, certain assumptions must hold. Complete the following
questions regarding these assumptions.
E
xp
ectd
N
o
rm
alV
u
e
j.
No, because the pvalue was insignificant.
No, it is only evidence of a correlation between salary and winning
percentage.
Yes, because the pvalue was significant.
Yes; clearly spending more on salary will result in more wins.
[4]
.0
0
.0
5
.10
.10
.5
0
.d
0
.50
.1
O
b
s
e
rv
V
a
lu
e
(i) This plot is used to check the
assumption of normality of:
And this assumption (circle one)
does does not appear to hold.
(ii) Another plot that is used to check assumptions regarding our linear
regression model is a plot of the residuals against the independent
variable x. Which of the following plots would support that our linear
model seems reasonable? Clearly circle your answer.
296
297
6. An educational psychologist was interested in whether using a childs own
name in a story affected the childs attention span while reading. Sixteen
children were randomly assigned to read a story under ordinary conditions
(using names like Dick and Jane that were not any of these childrens
names). Fourteen other children read versions of the same story, but with
each childs own name substituted for one of the children in the story. The
researcher kept a careful measure of how long it took each child to read
the story. The results were entered into SPSS and the following output is
provided:
Group Statistics
Reading time (minutes)
Story Type
Ordinary Story
OwnName Story
298
N
16
14
Mean
6.1
9.4
Std.
Deviation
2.17
2.87
Std. Error
Mean
.54
.77
a. Complete the sentence:
[2]
Of the children in the OwnName Story group, 25% read for at least
_________ minutes.
299
300
b. The researcher plans to create a confidence interval for the difference
between the mean reading time for the ordinary story population and that
for the ownname story population. He remembered that one assumption
had something to do with normality of the response. So he created the
QQ plot above of the reading times for the 30 children in the study. Is
this the correct graph to make for checking the assumption? If yes,
comment on the validity of the assumption. If no, explain why not.
[3]
YES, and
NO, because
c.
A 95% confidence interval for the difference between the mean reading
time for the ordinary story population and the mean reading time for the
ownname story population was found to be: (5.3, 1.5). Circle the one
answer below that correctly completes the interpretation of the 95%
confidence level:
[2]
If we repeated this experiment with the same design,
we would expect 95% of the resulting confidence
intervals to contain the
difference in the population means of 3.4 minutes.
difference in the sample means of 3.4 minutes.
difference in the population means.
reading times for 95% of all such children.
7. A researcher selected a fixed level of significance of 1% and analyzed data
that was paired to test the hypothesis H0: d = 0 versus the alternative
hypothesis Ha: d 0. The null hypothesis was rejected.
c.
If the 99% confidence interval were constructed with the same data,
would the resulting interval contain the value of 0?
[2] Circle one:
Yes
No
Cant Tell
d. If the researcher had used a 5% level of significance, would the decision
still be to reject the null hypothesis?
[2] Circle one:
Yes
No
301
Cant Tell
8. Family owned companies often lack welldocumented strategic business
plan. A large scale survey of privately held family firms is to be conducted
to estimate the proportion of all such firms that do not have a strategic
business plan.
a. Which is the most appropriate statistical analysis technique?
[2] Circle one:
Conducting a hypothesis test (HT)
Construct a confidence interval (CI)
b. Give the symbol for the corresponding parameter that is of interest:
[2]
9. About 10 years ago, 41% of companies had their own ethics code. A survey
will be conducted today using a random sample of 15 companies and
recording whether or not they have their own ethics code. The data is to
be used to assess if there is evidence that the rate of companies with
their own ethics code has increased from previous rate 10 years ago.
a.
[3]
A hypothesis test will be conducted at the 5% significance level. A test
statistic will be computed and a corresponding pvalue will be found. What
distribution will be used to find that corresponding pvalue? Be specific.
b. If the study were to be repeated with a larger sample size, what effect
would this have in terms of the power of the test? Circle one:
[1]
Increases
Decreases
302
No change
10. Which of the following relationships must be analyzed using a chisquare
test of independence?
[2] Circle one:
a.
The relationship between height (inches) and weight (pounds).
b. The relationship between satisfaction with school (satisfied or not) and
political party affiliation.
c.
The relationship between gender and amount willing to spend on a
computer system (in dollars).
d. The relationship between opinion on gun control and income earned last
year (in thousands of dollars).
11. Suppose a production line manager monitored the weights of one pound (16
ounce) cans of nuts at monthly intervals to test Ho: = 16 ounces versus
Ha: < 16 ounces by taking a random sample of cans from a days production.
The results for the sampled cans on a day this past November gave a mean
of 16.24 ounces and a standard deviation of 1.81 ounces.
a.
[2]
What test statistic will be computed to perform this test?
symbol and formula.
Give the
_______ = __________________________________
b. If the null hypothesis is true, what is the expected value of the test
statistic?
[2]
c. The only feasible value for the pvalue is: (circle one)
[2]
0.04
0.09
0.24
0.38
0.74
12. The Personnel Department for a company reports that 76% of all of its
2000 employees are registered for using the onsite Health Club. In one
brief sentence, explain why it would be inappropriate to construct a 95%
confidence interval using the value of 0.76 as its center.
[2]
303
Statistics 250 Fall 2008 Final Exam
1. Supermarkets often place similar types of cereal together
on a shelf. To see if there is a difference in the average
sugar content of cereals on each shelf, the sugar content
(in grams) for a random sample of cereal boxes from each
of three shelves was collected at local grocery store.
Some output from SPSS is provided below.
a. The ANOVA table is incomplete. Fill in the missing values.
[4]
b. State the appropriate null and alternative hypothesis for
this problem.
[2]
H0: _______________________ Ha:
___________________________
c. i. State the
specific
assumption
that the
output at the
right is testing.
[2]
ii.
In one simple sentence, explain
assumption does appear to be reasonable
level. Your sentence should include the
value from the output that is used to
assessment.
[2]
304
why this
at the 5%
numerical
make this
d. The pvalue for testing the hypotheses in part (a) is
reported as 0.001. Provide a complete, labeled sketch of
this pvalue, including the name of the distribution used to
find it.
[2]
Question 1 continued
Multiple Comparisons
SUGARS
Tukey HSD
(I)
(J)
SHEL SHEL
F
F
1
2
3
Mean
Difference
(IJ)
95% Confidence
Interval
Std.
Error
Sig.
Lower
Bound
Upper
Bound
4.8190
1.2857
.001
7.894
1.744
1.7278
1.1476
.294
4.473
1.017
4.8190
1.2857
.001
1.744
7.894
3.0913
1.1299
.021
.389
5.794
1.7278
1.1476
.294
1.017
4.473
3.0913
1.1299
.021
5.794
.389
e. The pvalue from part (d) is clearly significant at a 5%
level. Which pairs of means are significantly different from
each other at the 5% level?
[2]
Shelf 1 and Shelf 2
Shelf 1 and Shelf 3
Shelf 2 and Shelf 3
2. Drew works at a car wash and is paid according to the
number of cars that he washes. Below is a partial
probability model for the random variable X = the number
of cars that Drew washes on a Wednesday afternoon
between 4:00 and 5:00 PM.
Value
Probabil
0
0.1
1
0.2
2
0.2
305
4
0.1
ity
a. Find the missing probability in the table. What is the
probability that Drew washes 3 cars on a Wednesday
afternoon between 4 and 5 PM?
[2]
Final Answer: _______________
b. Drew is paid $7 for each car that he washes. What is his
expected earnings for a Wednesday afternoon between
4:00 and 5:00 PM? Show your work and include your
units.
[3]
Final Answer: _______________
Consider the following two graphs. Each was made on
data collected for performing a twoindependent samples
ttest to assess if there was a significant difference
between the two population means.
Data Set A: Data Set A:
Data Set B: Data Set B:
3.
Which data set would result in a larger pvalue for the ttest?
[2] Circle one:
Data Set A
Data Set B
4. Henry is totally unprepared for a three question multiplechoice quiz. He decides to randomly answer each question
and hope for the best. Since he is guessing, his response
on one question does not depend on his response to
another question. There is only one correct answer for
306
each question. Below is a table of the number of possible
answers to each question on the quiz.
Quiz question number
Q1 Q2 Q3
Number of possible answers to select
5
4
4
from
a. What is the probability that Henry answers Q1 incorrectly?
[1]
Final Answer:
_______________
b.
What is the probability that Henry answers Q2
correctly, given that he incorrectly answered Q1?
[2]
Final Answer:
_______________
c.
What is the probability that Henry answers all 3
questions incorrectly?
[2]
Final Answer: _______________
Henry needs to get at least one question correct to
pass the quiz. What is the probability that Henry passes?
[2]
Final Answer:
______________
d.
307
5. Does the cost to make a movie depend on its length? Mr.
X, a movie executive is interested in the answer to this
question. He hires a statistician to investigate. The
statistician, Dr. B, takes a random sample of major release
films of 2005, and then records the final budget (millions
of dollars) and the running time (minutes) for each film.
SPSS output from the analysis is provided.
a. How many movies were included in the statistician's
sample? ________
[1]
b. Write out the equation of the estimated least squares
regression line.
308
[2]
c.
Julio is making a movie and had a starting budget of
$50 million. He thinks his final film will be hour longer
than he had originally planned; he needs to adjust his
budget accordingly. Using the statistician's results, what
should he estimate to be the increase in his budget?
Include your units.
[2]
Final Answer:
____________________
d. The 10th movie in the sample had a running time of 1
hours and residual of 5.351. Find the actual final budget
for this movie. Show all work. Include your units.
[3]
Final Answer:
___________________
e. What is the value of the correlation between the running
time of a movie and its final budget?
[1]
Final Answer: ___________________
f. The statistician knows she
should examine the data graphically to
check various assumptions before
assessing significance of the
regression equation. One of the
plots she produced is given.
Consider the following three
statements. Circle all that are
correct.
[2]
This plot shows that it is
reasonable to assume that
the residuals are normally
distributed.
This plot shows that it is
reasonable to assume that
309
there is a linear relationship between the running time
of a movie and its final budget.
This plot can be used to check the normality of the
error terms.
g. What hypotheses would the statistician use to test for a
positive linear association between the running time of a
movie and its final budget?
[2]
H0: ____________________________ Ha:
____________________
h. Perform the test for the hypotheses in part (f). Give the
test statistic, the pvalue, and the conclusion at a 1%
level.
[3]
Test statistic: ______________________ pvalue:
__________________
Thus, there (circle one)
Does
Does not
appear to be sufficient evidence to support a
significant positive linear relationship between the
running time of a movie and its final budget.
i. Below are two intervals, one is the 99% prediction
interval for the final budget (in millions of dollars) for an
individual movie that is 2 hours long is the 99%
confidence interval for the mean final budget for all
movies with a running time of 2 hours. Which of the
following is the confidence interval?
[2]
Circle one:
(12, 147)
(45, 113)
j. Below are two intervals, one is the 95% prediction
interval for the final budget (in millions of dollars) for a
movie that is 2 hours long the other is the 90%
prediction interval for the final budget for a movie that
310
is 2 hours long. Which of the following is the 90%
prediction interval?
[2]
Circle one:
(38, 121)
311
(29, 129)
6. Older Jurors? A number of years ago the UCLA Law
Review assessed the belief that older people were
overrepresented on juries in a particular California county.
The distribution of age in this particular county is given
below.
2140
4160
61 years or
Age Group
years
years
over
Countywide
42%
39%
19%
%
A random sample of 66 jurors in this count was selected
and the data are presented below.
2140
4160
61 years or
Age Group
years
years
over
Observed #
jurors
5(
28
(
33 (
You are to use these results to assess whether it appears
that jurors were selected according to the countywide
distribution for age. Use a 1% significance level.
a. State the appropriate null hypothesis.
[3]
H0:
_____________________________________________________
b. Compute the corresponding expected counts and enter
them in the table in the parentheses. Show all work.
[3]
c. Compute the value of the statistic. Show all work.
[2]
Final
___________________
d. Report the pvalue of this test.
[2]
312
answer:
Final
_________________________
answer:
e. What is the decision at the 1% significance level?
[1]
Circle one: Reject H0
Fail to
reject H0
f. What type of error could have been made? Circle one:
[1]
Type I Error
Type II Error
Cannot be determined
7. A recent Reuters article reported that Americans plan to
spend less for holiday gifts this year than in the past.
Based on a survey of 5,000 households, the average
amount a household planned to spend on holiday gifts
was estimated to be $418 for 2008.
Suppose the
standard error for this estimate was 0.71.
a. Provide a range of reasonable estimates for the true
average amount a household plans to spend on holiday
gifts for 2008 using a 90% confidence level. Show all
work.
[3]
Final Answer: ( ____________, ____________)
b. Based on a survey of 5,000 households last year, the
2007 estimate of planned holiday spending was $471,
with a standard error of 1.1.
[2]i. Is the value of $471 a statistic or a parameter?
Circle one:
Statistic
Parameter
ii. Which of the following are appropriate method(s) to
determine if this years 2008 estimate is significantly
313
different from last years 2007 estimate? Circle all that
are appropriate.
[2]
Check if the value $471 is included in the confidence
interval calculated in part (a).
Calculate a test statistic according to the formula
x 0
t
s
n .
Calculate a confidence interval for the true difference
in average planned spending estimates and check if
that interval includes zero.
8. For each question clearly circle your answer. [2 points
each]
a. In regression analysis, if r2 = 0.62, then 62% of the
variation in y is explained by the linear relationship with x.
True
False
b. In a chisquare test of independence, if the table has 3
rows
and 4 columns, then the degrees of freedom is 3.
True
False
c. The mean of a set of data is always one of the data
values.
True
False
d. A standard deviation is never negative.
True
False
e. Which score is better? A score of 90 on a test with a
mean of 70
and standard deviation of 10; or a score of 650 on a test
with
314
a mean of 500 and a standard deviation of 100?
Score of 90
Score of 650
315
9. Eating Healthy? The Time magazine asked a random
sample of 950 American adults How much effort are you
making to eat a healthy nutritionally balanced diet? The
following data was obtained and gender was used to
divide up the results for men versus women.
Response
Men
Women
Very Serious
141
213
Somewhat
214
218
serious
Dont really try
100
64
Total
455
495
To assess if the distribution of the response is the
same for the population of men and the population of
women, the appropriate test would be:
[2] Circle one:
ztest
ttest
goodnessoffit
homogeneity independence
If the distribution of the response is the same for the
population of men and the population of women, what is
the expected value of the test statistic?
[2]
Final answer = _______________
If the distribution of the response is the same for the
population of men and the population of women, how
many women would you expect to respond with dont
really try?
[2]
Final answer = _______________
Another researcher decides to collapse the two
serious response categories to simplify the analysis. So
the data give 355 of the 455 men (78.0%) responding
with serious about their diet compared to 431 of the
495 women (87.0%) responding with serious. Use these
results to test the hypotheses H0: p1 = p2 versus Ha: p1
p2 where 1 = men and 2 = women.
Report the test
statistic value and corresponding pvalue.
[5]
316
Test statistic value = ______________________
pvalue = __________________
317
10. Match each research question with the best type of
analysis to answer it by writing the appropriate capital
letter from the list at the right on the blank line to the left.
[2 points each]
Is student status (instate
A. paired ttest
a.
versus outofstate)
_____ associated with ones
B. two
_
eventual graduation
independent
outcome (graduating versus
samples tnot graduating)?
test
b.
_____
_
Do children who are nice get
more presents on average
than those who are naughty?
c.
_____
_
Do fans of five different UM
sports differ in the average
level of loyalty to their team
(measured on a 50 point
scale)?
d.
_____
_
Does the reaction (measured
on a quantitative scale) of
the same person differ when
receiving clothes as a gift
versus receiving electronics
as a gift?
C. oneway
ANOVA
D. chisquare test
of
homogeneity
E. chisquare
test
of
independence
F. regression
analysis
11. In a study of the effects of marijuana during
pregnancy, measurements on babies of mothers who used
marijuana during pregnancy were compared to
measurements on babies of mothers who did not. A 95%
confidence interval for the difference in the population
mean head circumference (nonuse minus use) was found
to be 0.61 to 1.19 cms.
a. What can be said about the pvalue for testing the
hypothesis that the population mean head circumference
for babies of nonusers is different from the population
mean head circumference for babies of users? Circle your
answer.
318
[2]
pvalue > 0.05
pvalue 0.05
cant tell
b. The following interpretation of the 95% confidence level is
incorrect.
Correct it by clearly crossing out and replacing one word.
[2]
If we were to repeat this study many times, we would
expect 95% of the resulting confidence intervals to
contain
the
difference
in
sample
mean
head
circumferences (non  use minus use).
Stats 250 Fall 2009 Final Exam
1. Fashion Trends: Bella wants to study the fashion tendencies
at her new school in Folks, WA where the weather is usually
somewhat cool. During lunch one day she conducted an
observational
study to learn
about what some
of
the
hip
fashion
trends
were for females. Here is a summary of what she observed:
a. If a girl wearing sneakers is randomly selected, what is the
probability that she is wearing them with a skirt?
[2]
Final Answer = _____________________
b. Bella read in Vogue that wearing fur boots and wearing a skirt
were mutually exclusive. Assuming Vogue is correct, what is the
value for the missing cell value X?
[2]
Final Answer = _____________________
c. Based on the results above and the answer in part (b), the
events wearing fur boots and wearing a skirt are:
[2]
Circle one:
dependent
independent
2. Nausea or Not? An article in Ergonomics described a study
designed to assess if the incidence of nausea is the same
across the three seat
locations in a bus (front,
middle, or rear). The
following
categorical
data were available from
this study.
319
Fill in the blank:
[1] In this study the variable seat location plays the role of
the_________ variable.
Give the name of the test for assessing if the incidence of
nausea is the same across the three seat location populations.
[2] Name of the test: ________________________________________
If there is no difference in the incidence of nausea across the
three seat location populations, how many of the middle of the
bus passengers would be expected to experience nausea?
(show your work)
[2]
Final Answer: _________________________
d. The observed counts were entered into the statistical package
R and the test output is provided below. Circle the correct
words to complete the conclusion based on the results of this
study.
[2] Thus it appears that the incidence of nausea
is
is not
same for the 3 bus seat location populations.
3. Responding favorably to a Treatment A medical
researcher would like to test the hypothesis that a majority of
patients who receive a new treatment will respond favorably,
that is, test H0: p = 0.5 versus Ha: p > 0.5, where p is the
population proportion of all patients favorably responding to
the new treatment. If there is not enough evidence for a
majority, there will be no further funding of this treatment.
There are 50 patients available for the study.
The summary report stated that 56% of the treated patients
responded successfully to the new treatment. How many of the
patients responded favorably?
[1]
Final Answer = _____________________
b. The numerical value of 56% is a statistic which estimates a
parameter.
Circle one: True
False
[2]
c. Using the results stated in part (a), perform the hypothesis test
at a 5% level. Give the observed test statistic value, its
respective pvalue, and decision. Show all work.
[4]
Observed test statistic = __________________
pvalue = ______________________
320
At a 5% level the statistical decision would be:
(circle one) Reject H0
Fail to reject H0
d. Suppose this decision in (c) is incorrect. What is a consequence
of such a mistake?
[2] Circle as many as apply:
o
Patients may receive a treatment that doesnt work.
o
Patients may not receive a beneficial treatment.
o
The company that sponsored the development of this
treatment may not financially benefit.
e. Determine if each of the following statements is true or false.
[2 points each]
i. By choosing a 5% level, the researcher accepts a 5% risk
of incorrectly concluding a majority would respond
favorably.
True
False
ii. Power of a test is the probability that you correctly conclude
a majority would respond favorably.
True
False
4. Treating Migraines:
A drug company is interested in
determining if a new medication is effective in relieving pain
associated with migraine headaches. A random sample of 20
migraine sufferers is obtained. Each patient rates his/her pain
level (on a quantitative scale between 0 and 10) both before
and after taking the medication. A rating of 10 represents the
highest amount of pain possible and a rating of 0 represents no
pain. For each person, the difference is defined to be pain level
before taking the medication minus pain level after taking the
medication. The company wants to determine if pain levels
improve after taking the medication, on average.
a. Clearly define the parameter of interest and then state the
hypotheses to be tested.
[4] Let the parameter ________ =
______________________________________
The hypotheses to be tested are H0:______________ versus
Ha:____________
The differences were entered into SPSS and the results for one
sample ttest on the differences were obtained.
321
b. Give the symbol and value for the test statistic to test the
hypotheses in (a).
[2]
_______ = _____________________
c. Provide the pvalue for testing the hypotheses stated in (a),
circle your decision at the 5% level, and state your conclusion
in the context of the problem.
[3] pvalue = ______________
Decision? (circle one): Fail to Reject H0 Reject H0
Conclusion:
d. The plot below was also
produced. Clearly state
the assumption being
assessed through this
plot.
[2]
5. Does a Cars Color
Influence the Chance
of Being Stolen? According to the American Automobile
Association, 15% of all cars are white, 15% are blue, 35% are
red, 30% are black, and the remaining 5% are some other color.
The Associated Press reported the following color results for a
random sample of 830 stolen cars:
1=
2=
3=
4=
5=
TOT
Whi
Blu
R
Bla
Oth
A
te
e
ed
ck
er
L
# stolen
830
140
100
270
230
90
cars
The hypotheses to be tested using the above results are:
H0: p1= 0.15, p2= 0.15, p3= 0.35, p4= 0.30, p5= 0.05
322
Ha: at least one of the pi is not as listed in H0
where
p1 is the proportion of all stolen cars that are
white,
p2 is the proportion of all stolen cars that are blue,
and so on.
a. Give the name of the statistical test for testing the above
hypotheses.
[2]
Name of the test:
______________________________________________
If the null hypothesis is true, how many cars in the random
sample would be expected to fall in the other category? Show
your work.
[2]
Final Answer: _________________________
The test statistic value is computed to be 66.33. Determine the
contribution to this 66.33 that came from the other category.
Show all work.
[2]
Final Answer: _________________________
What distribution would be used to find the corresponding pvalue?
[2]
Final Answer: _________________________
The researcher conducting the test noted that the observed test
statistic value of 66.33 is highly significant as it is more than 2
standard deviations above the expected value for such a test
statistic. What number represents being 2 standard deviations
above the expected value in this case?
[2] Show your work.
Final Answer: _________________________
323
6. Tires:
A consumer advocacy group wants to assess the
difference between how long Michelin and Goodyear tires last,
on average.
A random sample of 30 Michelin Tires was
obtained, and an independent, random sample of 35 Goodyear
tires was obtained. The number of miles that each tire lasted
was recorded for each tire. The following table summarizes the
data that was collected (group 1 = Michelin, group 2 =
Goodyear).
Group 1
Group 2
Sample Mean
3120
3050
Sample Size
30
35
Sample Standard
200
195
Deviation
a. The pooled confidence interval assumes that the two
populations have equal variances. Citing appropriate evidence,
comment on the validity of this assumption.
[1]
b. An estimate of the common population standard deviation
using the above results is 197.3. Provide a 95% pooled
confidence interval estimate for the difference between the two
population mean miles. Show all work.
[4]
Final Answer: ( ___________ , ____________ )
c. Which of the following are correct interpretations of the 95%
confidence level?
[2] Circle all that apply.
With repeated random samples, 1  2 should fall in the interval
computed in part (b) 95% of the time.
With repeated random samples, 1  2 should fall in 95% of the
resulting intervals.
There is a 95% chance that 1  2 will fall in the interval
computed in part (b).
d. Based on the confidence interval, what would your decision be
about testing the following hypotheses:
H 0 : 1 = 2
versus
Ha: 1 2 at a 5%
significance level?
[2]
324
Circle one: Fail to Reject H0
Information to Decide
Explain briefly:
Reject H0
Not Enough
7. Car weight and Fuel
Efficiency: A researcher
wants to examine and
model the relationship
between the weight of a
car
and
its
fuel
efficiency.
To
investigate
this
relationship, the data on
the fuel efficiency (miles
per
gallon)
and
the
weight (in 1,000lb) for 38
randomly selected cars were collected
and
a
model was developed using linear regression.
Use the plot and SPSS output below to answer the following
questions.
325
a. Consider the following statements and clearly circle all that are
correct:
[2]
With each 1lb increase in car weight, we would expect to
see a decrease of 5.378 miles per gallon in fuel efficiency.
With each 2,000lb increase in car weight, we would expect
to see a decrease of 10.756 miles per gallon in fuel
efficiency.
Based on the analysis above, about 57.4% of the variation
in fuel efficiency can be explained by the linear relationship
with car weight.
Based on the analysis above, about 33.0% of the variation
in fuel efficiency can be explained by the linear relationship
with car weight.
b. Give the correlation between fuel efficiency and weight.
[2]
Final Answer: _____________________
c. Give the equation of the estimated least squares regression
line for predicting fuel efficiency from weight.
[2]
Final Answer: ______________________________________________
d. The unstandardized coefficient value of 41.764 given in the
SPSS output is an example of:
[2] Circle all that are correct:
a sample statistic
a population parameter
test statistic
a slope of the regression line for the population
an intercept of the regression line for the sample
e. The actual fuel efficiency of the car that weighs 4,054lb in the
data set was 15.5 miles per gallon. Compute the residual for
this observation. Show all work and include the units.
326
[2]
Final Answer: ____________________________
f.
Explain in one simple sentence why it would not be appropriate
to use the least squares regression line in part (b) to estimate
the fuel efficiency for a car that weighs 6,000lb.
[1]
327
g. The researcher would like to test if there is a significant
negative linear relationship between fuel efficiency and car
weights. Write out the appropriate hypotheses, the value of the
test statistic, and the corresponding pvalue. Then circle the
appropriate conclusion using a 5% significance level.
[4]
H0: ___________________________
Ha:
________________________
Test statistic value = ______________ pvalue =
____________________
Therefore, it appears that (circle one):
heavier cars tend to have lower fuel efficiency.
lighter cars tend to have lower fuel efficiency.
there is not significant linear relationship between car
weight and fuel efficiency.
h. The 95% confidence interval for the average fuel efficiency for
all cars that weigh 3,000lb would be
[2]
(circle one)
narrower
wider
than the 95% confidence interval for the average fuel efficiency
for all cars that weigh 4,000lb.
i.
At the right is a plot from the regression analysis above.
Consider the following sentences and circle the appropriate
one.
[1] The plot at the right is used to check the assumption that:
(circle one)
328
o The car weights
are randomly
distributed.
o The fuel efficiency
values are
randomly
distributed.
o The true error
terms have
constant variance.
o The residuals have
constant variance.
329
8. Which classroom environment is best for teaching
college precalculus? A total of 75 students were randomly
assigned to one of three classes. The first class (group 1)
emphasized group work. The second class (group 2) was a
traditional lecture. The third class (group 3) emphasized
individual instruction and practice. At the end of the year,
students knowledge was assessed with a standardized test. A
school district wants to determine if the instruction method has
an effect on student outcomes, on average. The data was
analyzed and some output is provided.
a. State the hypotheses to be tested.
[2]
H0: _______________________________________________
Ha: _______________________________________________
b. There are 4 missing values in the ANOVA table. Find the values
and enter them clearly in the table.
[2]
c. If there was no difference in the average scores for the three
instruction methods, what is the distribution of the test statistic
for testing the hypotheses in part (a)?
[2]
Final Answer: ____________________________
330
d. One of the assumptions in ANOVA is that all population
variances are equal. What is the estimate of the common
population standard deviation?
[2]
Final Answer: ____________________________
e. At the 10% significance level, the decision is:
[1]
f.
(circle one)
Reject H0
Fail to Reject H0
The school district superintendant would like to know more
about the individualized instruction and practice method used
by group 3. Provide the superintendant a 90% confidence
interval estimate for the population mean test score for
students taught using the individualized instruction and
practice method. Show your work.
[3]
Final Answer: ( ___________ , ____________ )
9. Absenteeism Study: The personnel director of a corporation
wishes to study absenteeism among clerical workers at the
central office. A random sample of 25 clerical workers was
obtained and the number of days absent over the previous year
time frame was recorded. The director comes to you with the
95% confidence interval for the population mean number of
days absent for all such clerical workers of (8, 11).
a. The confidence interval estimate is provided above. What is
the value of the sample mean? Include your units.
[2]
Final Answer:____________________________
b. Write a simple sentence to provide an interpretation of the
interval (8, 11).
[2]
331
10.
Name that Scenario: One important aspect in Statistics is
to understand which statistical methods or procedures are
appropriate to use to address the research problem or question
of interest.
The Janus Sisters make cupcakes and are considering
providing their specialty treats for individual sale at local coffee
shops. Before they embark on this endeavor, they have many
plans to formulate and decisions to make. They need your help
in deciding which statistical method to use to address their
questions. For each question, select the letter corresponding to
the statistical analysis technique most appropriate for
addressing that question.
[2 points each]
_____ 1.
_____ 2.
_____ 3.
Is the average number of
chocolate chips in our
chippy cupcakes is higher
than the average number
of chocolate chips in our
competitors chocolate
chip cupcake?
Do a majority of consumers
prefer the Janus cupcakes
over our leading
competitors cupcakes?
Is there an association
between the type of flour
used
(allpurpose flour, wholewheat flour, cake flour) and
A. Simple linear
regression
B. 1sample ttest for
a population mean
C. Paired ttest for a
population mean
difference
D. 2sample ttest for
the comparison of
two population
means
E. 1sample Ztest
for a population
proportion
F.
the crumbliness of the
cupcake (crumbly, not
crumbly)?
_____ 4.
How does the baking
temperature affect the
final moisture level of the
cupcake?
332
2sample Ztest
for the
comparison of two
population
proportions
H. Chisquared test
of independence
I.
Chisquared test
of homogeneity
_____ 5.
Is the average number of
calories for our death by
chocolate cupcakes more
than 280?
333
Viel mehr als nur Dokumente.
Entdecken, was Scribd alles zu bieten hat, inklusive Bücher und Hörbücher von großen Verlagen.
Jederzeit kündbar.