Sie sind auf Seite 1von 15

CE 282: ENGINEERING PRACTICE

Measurement Techniques
in Accident Investigations
Statistical Analysis
Alyce Rollins
4/16/2013

The purpose of this statistical analysis is to determine the relative accuracy and precision of
four different measurement techniques, including an Electronic Total Station (ETS), used in
vehicular accident investigations. This report will focus on the validity of the test simulation
and will determine whether the ETS method is an acceptable method of measurement in future
investigations.

Table of Contents
Executive Summary....................................................................................................................................... 3
Experimental Data ........................................................................................................................................ 4
Data Analysis ................................................................................................................................................. 4
Results ........................................................................................................................................................... 8
Appendix A .................................................................................................................................................. 10

Executive Summary
The objective of this statistical report is to analyze four different methods of taking
measurements at an accident investigation scene. The data used in this report
corresponds to a research project that involves using an Electronic Total Station (ETS)
to take measurements at a simulated accident scene and was compared to three other
well know methods of measurement. These methods include a cloth tape, a steel tape
and a measurement wheel. Each of the four methods was used to measure a specified
distance between two points and the resulting data was analyzed using an Excel
spreadsheet.
From the data it was determined that the ETS machine is a precise measurement, but
does not necessarily conform to a normal distribution of sample data. The measuring
wheel was also determined to be a precise instrument; however, it also did not conform
to normal distribution. Each of the tape methods was found to be more accurate and
while each of these methods did not conform exactly to a normal distribution, they were
found to be closer to the normal distribution than both the ETS and the wheel methods.
All methods were compared against a theoretical distribution using the S-curve
technique and figures and equations follow in the body of this report.
From the composite means of the data that were calculated, a standard mean could be
determined and this value is the closest possible value to the true value that can be
calculated from the data provided. This value was used to determine a standard error in
the sampling means and was also plotted in an S-curve against a theoretical mean. It
was determined from this graph that the overall data is skewed to the left and not
necessarily representative of all possible values in the population.
Finally, a corresponding velocity was calculated using a given equation and was used to
determine whether a slight variation in the mean value from each set of data could
cause a wide variation in the final calculated velocity. It is shown in Table 1 of this
report that the slight variation in the distance measured causes very little overall
variation in the final obtained velocity, and that for all practical purposes of this
experiment, any of the four methods would be acceptable. Given that the ETS is a very
precise method of measurement, and has been an acceptable method used in
surveying for the past 25 years, it would be very effective in providing measurements for
all future accident investigations.

Experimental Data
The data used in this analysis was collected as part of an experimental research
project. The objective of the research project focuses on using an ETS machine to take
measurements at a simulated accident scene, as opposed to using three other common
methods of measurement. The goal of this analysis is to determine whether using the
ETS system is prone to the same error as the other common methods, and to determine
if the client should use this method for future investigations.
Teams were assembled to take 35 measurements between two points at the Boise
State University campus which were specified by the client. The four measurement
methods used were an ETS system, a cloth tape, a steel tape and a measuring wheel.
All measurements were done to the best of the technicians ability and it is assumed
that any discrepancies in the data can be attributed to other factors besides human
error. All measurements were provided in units of feet and the raw data used in
calculations can be found in Appendix A at the end of this report.

Data Analysis
The data was compiled into an excel spreadsheet for the purpose of this analysis, and
all calculations were done utilizing the function capabilities of excel. The spreadsheet
calculations and data can be found in Appendix A at the end of this report.
Initially data was grouped into the respective measurement technique and then sorted
from smallest to largest. For each set of data the mean, variance, standard deviation,
range (minimum and maximum), and the median were calculated. Then the cumulative
frequency of all values was determined and an S-curve was generated using the
specified values plotted against their cumulative frequencies. In order to determine the
normality of the data for each technique, a theoretical curve was developed using the
standard values used to determine normal distribution. The theoretical curve was
plotted on the same graph, specific for each set of data, and shows how the actual data
points may or may not conform to a normal distribution. The graphs for each set of data
are shown in Figures 1-4 following this section.
From the graphs shown, it was determined that none of the sets of data conform exactly
to normal distribution. The ETS system data was skewed to the left of the graph, as
was the data obtained from the measurement wheel. Both the cloth tape and the steel
tape sets of data were shown to be closer to the theoretical curve and obtained values
that were both above and below the theoretical values. Furthermore, the percentage of
the data falling within plus or minus 1, 2, and 3 standard deviations was calculated and
also used to determine the normality of the data. From the graphs and the percentages,
shown in Appendix A, we can assume that the distribution of the data is not attributed to
randomness and that there is specific error associated with each method of
measurement.
4

From the composite data, a standard mean was then determined based upon the four
means calculated from each set of data and a standard deviation of the means was
calculated. The standard mean of the calculated means is used as the most likely value
for the actual measurement between the two specified points. Without knowing the
actual value, this is the best determination that can be made in the analysis and was
used to determine the standard error of the sample means. Using the standard error,
the normality of the distribution of sample means was found by plotting an S-curve
similar to those used in the sample data sets. The theoretical curve was plotted against
the actual values and this graph is shown in Figure 5 at the end of this section. The
percentage of the sample means was also calculated and is shown in Appendix A at the
end of this report. From the graph of the distribution of the sample means, we can see
that the data is skewed to the left

ETS Distribution Curve


100
90

Theoretical Curve
Actual Curve

Cumulative Percent

80
70
60
50
40
30
20
10
0
245.650

245.700

245.750

245.800

245.850

245.900

245.950

246.000

Distance (feet)
Figure 1 ETS Distribution Curve plotted against the theoretical values.

Cloth Tape Distribution Curve


100
90

Theoretical Curve
Actual Curve

Cumulative Percent

80
70
60
50
40
30
20
10
0
243.740

244.240

244.740

245.240

245.740

246.240

246.740

Distance (feet)
Figure 2 - Cloth Tape Distribution plotted against the theoretical values.

Steel Tape Distribution Curve


100
90

Theoretical Curve
Actual Curve

Cumulative Percent

80
70
60
50
40
30
20
10
0
244.490

244.990

245.490

245.990

246.490

Distance (feet)
Figure 3 - Steel tape Distribution curve plotted against the theoretical values.

Measurement Wheel Distribution Curve


100
Theoretical Curve

90

Actual Curve

Cumulative Percent

80
70
60
50
40
30
20
10
0
246.000

246.100

246.200

246.300

246.400

246.500

246.600

246.700

246.800

246.900

Distance (feet)
Figure 4 - Measurement Wheel distribution plotted against the theoretical values.

Standard Error Distribution Curve


100

Cumulative Percent

90
80

Theoretical Curve
Actual Curve

70
60
50
40
30
20
10
0
245.300

245.500

245.700

245.900

246.100

246.300

246.500

Distance (feet)
Figure 5 - Distribution of the sample means plotted against the theoretical values.

Results
By analyzing the graphs of each distribution from the four methods of measurement, it
can be determined that the cloth tape and steel tape methods of measurement more
closely follow a normal distribution than the ETS and the measuring wheel methods.
While the data obtained from the ETS and the measuring wheel were very precise, it
seems that the tape methods were more accurate based on the fact that they are more
evenly distributed. The graph of the distribution of sampling means also shows that the
data has been skewed to the left and can be attributed the distribution of the ETS and
the measuring wheel. If the distribution of both of these methods had followed closer to
the distribution of the tapes, then the distribution of the standard means would be very
close to a normal distribution and would be representative of all possible values of the
true specified distance. However, the fact that the data from the ETS and the wheel are
both skewed to the left is reflected in the standard distribution being skewed to the left
and is therefore not necessarily representative of all possible values.
Once the means had been determined that they were not normally distributed, the
sampling means were factored into an equation to determine the corresponding relative
velocity of a car, given a measured distance of a skid mark. That equation is given by
Equation 1:
Eq 1:

Where
= final velocity; assumed to be zero
= f * g (g=32.2 ft/sec2; f = friction factor of 0.7 to 0.9 for concrete
= measured distance, given in feet
Using Equation 1 and the mean values calculated for each measurement technique, a
corresponding velocity was determined for each value of f (0.7 and 0.9). These results
are shown in Table 1.
Table 1 - Corresponding Velocities

Device

Mean

Corresponding
Velocity (f=0.7) mph

Corresponding
Velocity (f=0.9) mph

ETS
Cloth Tape
Steel Tape
Wheel

245.83
245.905
246.047
246.404

71.78
71.79
71.81
71.86

81.39
81.40
81.42
81.48

Standard Mean

246.047

71.81

81.42

From the values shown in Table 1, it can be determined that the variation in mean
values given by each individual measurement method gives very little variation in the
velocity of the vehicle that can be calculated. In fact, the variation in the friction factor
for concrete gives a much greater variation in the relative velocity of the vehicle and
would alter final results more than the measured distance.
8

For all practical purposes of this experiment, the error of the measuring device will have
little overall effect on the final velocity of the vehicle to be calculated, and other outside
errors that can be attributed from other factors could have more significant effects on
the final results. All of the four measuring devices would provide acceptable data
ranges to satisfy the practical needs of this experiment, even though the data may not
be normally distributed and will have varying levels of accuracy and precision. The ETS
system has been shown that is will be the most precise method for obtaining
measurements at an accident investigation and would be valid tool to be used in future
investigations preformed by the client.

Appendix A
Excel Spreadsheets and Raw Data
ETS Machine:
ETS
(feet)
245.70
245.70
245.70
245.70
245.70
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.80
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.90
245.95

Actual Curve
Value
245.70
245.80
245.90
245.95

variance
(feet)
0.0169000
0.0169000
0.0169000
0.0169000
0.0169000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0009000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0049000
0.0144000

Frequency
5
15
14
1

n
Mean
Variance
Standard Deviation
Min
Max
Range
Median

35
245.83
0.005171429
0.071912645
245.70
245.95
0.25
245.80

Theoretical Curve
Location
X-Bar - 3
X-Bar - 2
X-Bar - 1.5
X-Bar - 1
X-Bar - 0.5
X-bar
X-Bar + 0.5
X-Bar + 1
X-Bar + 1.5
X-Bar + 2
X-Bar - 3

X-Value
245.614
245.686
245.722
245.758
245.794
245.830
245.866
245.902
245.938
245.974
246.046

Data Range
+1 to -1
+2 to -2
+3 to -3

Cumulative Frequency
0.1
2.27
6.68
15.86
30.85
50
69.15
84.14
93.32
97.73
99.9

Percentage
82.86
100.00
100.00

Cum. Frequency Percent Cum. Frequency


5
14.29
20
57.14
34
97.14
35
100.00

10

Cloth Tape:
CLOTH TAPE
(feet)
243.75
243.79
245.67
245.71
245.71
245.71
245.75
245.75
245.83
245.88
245.92
245.92
245.92
245.94
245.96
246.00
246.00
246.00
246.00
246.00
246.04
246.07
246.08
246.08
246.08
246.08
246.08
246.17
246.17
246.17
246.20
246.25
246.50
246.50
247.00
Actual Curve
Value Range
243.75 - 244.29
244.30 - 244.84
244.85 - 245.39
245.40 - 245.94
245.95 - 246.49
246.50 - 247.04

variance
(feet)
4.6446407
4.4738293
0.0552922
0.0380807
0.0380807
0.0380807
0.0240693
0.0240693
0.0056464
0.0006322
0.0002207
0.0002207
0.0002207
0.0012150
0.0030093
0.0089979
0.0089979
0.0089979
0.0089979
0.0089979
0.0181864
0.0271779
0.0305750
0.0305750
0.0305750
0.0305750
0.0305750
0.0701493
0.0701493
0.0701493
0.0869407
0.1189264
0.3538550
0.3538550
1.1987122
Average Range
244.02
244.57
245.12
245.67
246.22
246.77

n
Mean
Variance
Standard Deviation
Min
Max
Range
Median

35
245.905
0.340379265
0.583420316
243.75
247.00
3.25
246.00

Theoretical Curve
Location
X-Bar - 3
X-Bar - 2
X-Bar - 1.5
X-Bar - 1
X-Bar - 0.5
X-bar
X-Bar + 0.5
X-Bar + 1
X-Bar + 1.5
X-Bar + 2
X-Bar - 3

X-Value
244.155
244.738
245.030
245.322
245.613
245.905
246.197
246.489
246.780
247.072
247.655

Data Range
+1 to -1
+2 to -2
+3 to -3
Outside 3

Percentage
85.71
91.43
94.29
100.00

Frequency
2
0
0
12
18
3

Cumulative Frequency
0.1
2.27
6.68
15.86
30.85
50
69.15
84.14
93.32
97.73
99.9

Cum. Frequency Percent Cum. Frequency


2
5.71
2
5.71
2
5.71
14
40.00
32
91.43
35
100.00

11

Steel Tape:
STEEL TAPE
(feet)
244.50
245.08
245.17
245.67
245.71
245.75
246.00
246.00
246.00
246.00
246.00
246.00
246.00
246.00
246.00
246.00
246.04
246.04
246.08
246.08
246.08
246.08
246.08
246.17
246.17
246.25
246.25
246.25
246.33
246.42
246.50
246.67
246.67
246.79
246.83

Actual Curve
Value Range
244.49 - 244.96
244.97 - 245.44
245.45 - 245.92
245.93 - 246.39
246.40 - 246.87

variance
(feet)
2.3945352
0.9359180
0.7698809
0.1424523
0.1138580
0.0884638
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0022495
0.0000552
0.0000552
0.0010609
0.0010609
0.0010609
0.0010609
0.0010609
0.0150238
0.0150238
0.0410352
0.0410352
0.0410352
0.0798466
0.1388095
0.2048209
0.3875952
0.3875952
0.5514123
0.6124180

Average Range
244.725
245.205
245.685
246.160
246.635

n
Mean
Variance
Standard Deviation
Min
Max
Range
Median

Theoretical Curve
Location
X-Bar - 3
X-Bar - 2
X-Bar - 1.5
X-Bar - 1
X-Bar - 0.5
X-bar
X-Bar + 0.5
X-Bar + 1
X-Bar + 1.5
X-Bar + 2
X-Bar - 3

Data Range
+1 to -1
+2 to -2
+3 to -3

Frequency
1
2
3
23
6

35
246.047
0.199676245
0.44685148
244.50
246.83
2.33
246.04

X-Value
244.707
245.154
245.377
245.601
245.824
246.047
246.271
246.494
246.718
246.941
247.388

Cumulative Frequency
0.1
2.27
6.68
15.86
30.85
50
69.15
84.14
93.32
97.73
99.9

Percentage
77.14
94.29
100.00

Cum. Frequency Percent Cum. Frequency


1
2.86
3
8.57
6
17.14
29
82.86
35
100.00

12

Measurement Wheel
WHEEL
(feet)
246.08
246.08
246.17
246.17
246.25
246.25
246.25
246.25
246.25
246.33
246.33
246.33
246.33
246.33
246.33
246.33
246.42
246.42
246.42
246.50
246.50
246.50
246.50
246.50
246.50
246.50
246.50
246.50
246.50
246.50
246.58
246.58
246.58
246.75
246.83
Actual Curve
Value Range
246.06 - 246.21
246.22 - 246.37
246.38 - 246.53
246.54 - 246.69
246.70 - 246.85

variance
(feet)
0.1049760
0.1049760
0.0547560
0.0547560
0.0237160
0.0237160
0.0237160
0.0237160
0.0237160
0.0054760
0.0054760
0.0054760
0.0054760
0.0054760
0.0054760
0.0054760
0.0002560
0.0002560
0.0002560
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0092160
0.0309760
0.0309760
0.0309760
0.1197160
0.1814760
Average Range
246.135
246.295
246.455
246.615
246.775

n
Mean
Variance
Standard Deviation
Min
Max
Range
Median

35
246.404
0.027789714
0.166702472
246.08
246.83
0.75
246.42

Theoretical Curve
Location
X-Bar - 3
X-Bar - 2
X-Bar - 1.5
X-Bar - 1
X-Bar - 0.5
X-bar
X-Bar + 0.5
X-Bar + 1
X-Bar + 1.5
X-Bar + 2
X-Bar - 3

X-Value
245.904
246.071
246.154
246.237
246.321
246.404
246.487
246.571
246.654
246.737
246.904

Data Range
+1 to -1
+2 to -2
+3 to -3

Percentage
74.29
94.29
100.00

Frequency
4
12
14
3
2

Cum. Frequency
4
16
30
33
35

Cumulative Frequency
0.1
2.27
6.68
15.86
30.85
50
69.15
84.14
93.32
97.73
99.9

Percent Cum. Frequency


11.43
45.71
85.71
94.29
100.00

13

Composite Data
MEANS (feet)
ETS
CLOTH TAPE
STEEL TAPE
WHEEL

245.83
245.905
246.047
246.404

Mean
Variance
Standard Deviation
Standard Error

246.0465
0.04867525
0.220624681
0.110312341

Theoretical Curve
Location
X-Bar - 3
X-Bar - 2
X-Bar - 1.5
X-Bar - 1
X-Bar - 0.5
X-bar
X-Bar + 0.5
X-Bar + 1
X-Bar + 1.5
X-Bar + 2
X-Bar - 3

X-Value
245.385
245.606
245.716
245.826
245.937
246.047
246.157
246.268
246.378
246.488
246.709

Actual Curve
245.83
245.905
246.047
246.404

25
50
75
100

Data Range
+ or - 1 Std Error
+ or - 2 Std Error
+ or - 3 Std Error

Percentage
50.00
25.00
25.00

Device
ETS
Cloth Tape
Steel Tape
Wheel

Mean
245.83
245.905
246.047
246.404

Most Likely Mean

246.0465

Variance
0.0469
0.0200
0.0000
0.1278

Cumulative Frequency
0.1
2.27
6.68
15.86
30.85
50
69.15
84.14
93.32
97.73
99.9

Corresponding
Corresponding
Velocity (f=0.7) feet/s Velocity (f=0.9) feet/s
105.27
119.37
105.29
119.38
105.32
119.42
105.39
119.51
105.32

119.42

14

Device
ETS
Cloth Tape
Steel Tape
Wheel

Mean
245.83
245.905
246.047
246.404

Standard Mean

246.047

Corresponding
Velocity (f=0.7) mph

Corresponding
Velocity (f=0.9) mph

71.78
71.79
71.81
71.86

81.39
81.40
81.42
81.48

71.81

81.42

15

Das könnte Ihnen auch gefallen