Report No. EPA 460/3-88-011
Analytical Support for
Emission Factors Development
and Air Quality Assessment
Work Assignment No. 0-01:
Analysis of California
I/M Review Committee Data
Task 6 Report
Supplemental Analysis of
Random Roadside Survey Data
prepared for:
U.S. Environmental Protection Agency
September 30, 1988
prepared by:
Sierra Research, Inc.
1521 I Street
Sacramento, California 95814
(916) 444-6666
-------
EPA 460/3-88-011
ANALYTICAL SUPPORT FOR
EMISSION FACTORS DEVELOPMENT
AND AIR QUALITY ASSESSMENT
EPA Contract No. 68-03-3474
Work Assignment No. 0-01:
Analysis of California
I/M Review Committee Data
Task 6 Report:
Supplemental Analysis of
Random Roadside Survey Data
prepared for:
U.S. Environmental Protection Agency
September 30, 1988
prepared by:
Thomas C. Austin
Thomas R. Carlson
Kathryn A. Gianolini
Sierra Research, Inc.
1521 I Street
Sacramento, CA 95814
(916) 444-6666
-------
ANALYTICAL SUPPORT FOR
EMISSION FACTORS DEVELOPMENT
AND AIR QUALITY ASSESSMENT
Task 6 Report:
Supplemental Analysis of
Random Roadside Survey Data
Table of Contents
page
1. Summary 1
2. Introduction and Methodology 4
3. Results 19
-------
1. SUMMARY
During 1985 and 1986, the California Highway Patrol (CHP) detained a
random sample of vehicles at the roadside so that they could be
inspected by employees of the California Air Resources Board (CARB)
and the California Bureau of Automotive Repair (BAR). The results of
the "Random Roadside Surveys" appeared to show that the California
vehicle inspection and maintenance (Smog Check) program was having a
relatively minor impact on the overall rate of defects observed during
"underhood" inspections.
In response to a task assigned by EPA, an analysis was conducted to
determine whether the apparently small effect from I/M on underhood
failure rates is caused by certain high failure rate components
masking significant effects on lower-rate but more emission-critical
components. The overall results of the analysis are illustrated in
Figure 1. The results shown in the figure compare the observed
tampering rates for vehicles that had already received an I/M test
("I/M Sample") to vehicles located in the same geographic area that
had not yet been required to be inspected ("Non-I/M Sample"). As the
figure shows, there does not appear to be a substantial and consistent
reduction in tampering rates for individual components on the I/M
Sample that is masked by one or more of the components. A similar
analysis for non-tampering related defects shows the same trend.
-1-
-------
Figure 1
Underhood "Tampering" Rates
I/M vs. Non-l/M Vehicles
1985 Roadside Survey
I/M Sample
Non-l/M Sample
AIR CAT SPARK EVAP EGR LEAD O2S PCV TAG ANY
Device Type
Note: 1984 and later models removed from •, ,.
sample to eliminate age bias.
AIR = air injection
CAT = catalytic converter
SPARK = spark advance controls
EVAP = evaporative emission controls
EGR = exhaust gas recirculation
LEAD = fillpipe lead restrictor
02S = exhaust oxygen sensor
PCV = positive crankcase ventilation
TAG = thermostatic air cleaner
-2-
-------
Not shown in the figure are the results of a "Plumbtesmo" test for
tailpipe lead deposits. (This is not one of the standard underhood
inspections performed under the California I/M program.) During the
Roadside Survey, Plumbtesmo results indicated a similar pattern to the
lead restrictor check. Of those vehicles designed for the exclusive
use of unleaded fuel, 7.0% of I/M vehicles showed tailpipe lead
deposits and 4.6% of the non-I/M vehicles showed lead deposits. The
reason for this anomalous result was not clear.
More detailed analysis of the available data shows that it is the pre-
1975 model vehicles which appear to benefit the most from I/M (as far
as underhood failure rates and tampering correction are concerned).
However, the analysis indicates that all vehicle age groups
experienced less of a reduction in underhood failure rates than would
have been expected with properly performed inspections and repairs.
Although the I/M program was having a beneficial effect on underhood
failure rates, there appears to be room for significant improvements.
-3-
-------
2. INTRODUCTION AND METHODOLOGY
Under a contract with the U.S. Environmental Protection Agency (EPA)
for "Analytical Support for Emission Factors Development and Air
Quality Assessment," Sierra Research, Inc. (Sierra) performs a variety
of Work Assignments for the Emission Control Technology Division
(ECTD) of EPA's Motor Vehicle Emissions Laboratory in Ann Arbor,
Michigan. Work Assignment 0-01 directed Sierra to perform an analysis
of California I/M data for the ECTD Technical Support Staff (TSS).
Task number 6 of that Work Assignment required supplemental analysis
of data obtained from a random sample of vehicles stopped by the
California Highway Patrol.
With references to an earlier report that Sierra prepared for GARB
("Technical Appendix"), the general direction provided by TSS was as
follows:
Section 3 of the Technical Appendix presents roadside
"underhood" inspection results by component and by I/M or non-
I/M, but not by the combination. The contractor shall provide
this breakdown of results for the 1985 and 1986 roadside
surveys, and also show for comparison the Smog Check failure
rates by component. The analysis should be performed on a
statewide basis, unless the sample size distribution among
regions is much different for the roadside sample than for the
Smog Check sample, in which case the contractor shall apply
weighting factors to one data set or the other. Adjustment
for mileage differences is not required, but substantial
mileage differences between corresponding I/M and non-I/M
strata should be addressed. The purpose of the analysis is to
determine whether the apparent lack of effect from I/M on an
overall underhood basis is the result of high failure rate
-4-
-------
components masking significant effects on lower-rate but more
emission-critical components.
Background
Under a contract with the California Air Resources Board, Sierra
obtained and analyzed data on the condition of vehicles in customer
service from the results of inspections of thousands of vehicles
detained at CHP roadblocks. Teams of technicians from ARE and BAR did
the inspections, which came to be referred to as the "Random Roadside
Surveys". In order to determine possible regional differences in the
condition of the vehicle fleet, road blocks were set up in a variety
of areas from San Diego north to Sacramento.
During the Random Roadside Surveys, each vehicle was subjected to the
same visual and functional inspections specified under the California
I/M (Smog Check) program. In addition, chemical tests (Plumbtesmo) of
vehicle tailpipes were used to detect the use of leaded gasoline.
Emissions from each vehicle were measured using the same idle and 2500
rpm test procedures used in the Smog Check program.
The results of the Random Roadside Surveys indicated that visual and
functional inspection failure rates are very high. During the 1986
survey, about 40% of all vehicles inspected contained visual or
functional defects. A 14.6% visual/functional defect rate for 1980
and later models was the lowest of the three model year groups
analyzed. For 1975-1979 models and for pre-1975 models, visual and
functional defects were much higher at 62.1% and 70.3%, respectively.
-5-
-------
More detailed information from the 1985 Random Roadside Survey is
shown in Table 1. Table 1 is a reproduction of Table 3-6 from
Sierra's earlier report for GARB ("Evaluation of the Smog Check
Program, Technical Appendix," Sierra Research, Inc., April, 1987).
Individual component failure rates are presented for air injection
systems (AIR), catalytic converters (CAT), spark advance control
systems (SPARK), evaporative emission control systems (EVAP), fillpipe
lead restrictors (LEAD), exhaust gas oxygen sensors (02S), positive
crankcase ventilation systems (PCV), and thermostatically controlled
air cleaners (TAG). These are the same emission control related
components that are required to be inspected visually under the Smog
Check program. The sample sizes for the various categories of
vehicles shown in the table were as follows:
Model Year Range I/M Sample Non-I/M Sample
pre-1975 236 216
1975-79 235 261
1980 and Later 174 5 246
Total Sample 645 723
The underhood failure rates shown in Table 1 are expressed in several
different ways. For each category, the failure rate is shown for
"broken" and "tampered" devices, as well as for the "total" defect
rate. A "broken" component is one that was determined to be non-
functional for reasons other than "tampering". In other words, there
was no evidence that the device had been intentionally removed or
disconnected. An example of a "broken" device would be an air pump
-6-
-------
Table 1
Underhood Inspection Results - 1985 Roadside Survey
All
Model
Years
With
Device
All
Model
Years
All
Vehicles
Pre-75
Model
Years
With
Device
Pre-75
Model
Years
All
Vehicles
75-79
Model
Years
With
Device
75-79
Model
Years
All
Vehicles
1980+
Model
Years
With
Device
1980+
Model
Years
All
Vehicles
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
AIR
3.32
6.04
9.36
63.2%
2.10
3.31
5.91
7.05
23.08
30.13
29.4%
2.08
6.79
8.87
5.99
6.22
12.21
77.0%
4.61
4.79
9.40
0.18
1.08
1.26
77.2%
0.14
0.84
0.98
ra.j.j.ur
CAT
0.08
2.03
2.11
65.5%
0.06
1.33
1.39
--
--
--
--
0.00
4.43
4.43
84.0%
0.00
3.72
3.72
0.00
0.28
0.28
98.9%
0.00
0.28
0.28
e rerce
SPARK
1.80
6.63
8.43
89.2%
1.60
5.91
7.51
3.45
14.78
18.23
98.3%
3.40
14.53
17.93
1.62
4.86
6.48
98.4%
1.60
4.79
6.39
0.37
0.56
0.93
75.1%
0.28
0.42
0.70
ntage D
EVAP
2.60
6.37
8.97
89.3%
2.32
5.69
8.01
6.55
19.37
25.92
66.2%
4.34
12.83
17.17
3.44
5.43
8.87
97.9%
3.37
5.32
8.69
0.00
0.70
0.70
99.6%
0.00
0.70
0.70
y J-ype
EGR
14.58
8.07
22.65
71.2%
10.39
5.75
16.14
40.14
23.81
63.95
27.7%
11.13
6.60
17.73
21.38
12.83
34.21
87.1%
18.62
11.17
29.79
3.69
0.92
4.61
90.9%
3.35
0.84
4.19
or croK
LEAD
0.17
4.33
4.50
66.4%
0.11
2.87
2.98
—
--
--
--
0.41
9.13
9.54
!>
85.5%
0.35
7.80
8.15
0.00
0.98
0.98
99.4%
0.00
0.98
0.98
en or ii
02S
0.19
0.56
0.75
29.6%
0.06
0.17
0.23
—
--
0.00
0.00
0.00
4.26%
0.00
0.00
0.00
0.20
0.40
0.60
70.7%
0.14
0.28
0.42
unperea
PCV
2.55
5.04
7.59
99.7%
2.54
5.03
7.57
4.92
13.07
17.99
99.6%
4.91
13.02
17.93
2.49
3.37
5.86
99.8%
2.48
3.37
5.85
0.84
0.42
1.26
99.6%
0.84
0.42
1.26
uevice
TAG
8.76
9.07
17.83
88.3%
8.76
8.01
15.74
13.97
24.83
38.80
85.1%
11.89
21.13
33.02
12.23
5.63
18.86
91.3%
11.17
5.14
16.31
2.21
0.63
2.84
88.4%
1.96
0.56
2.52
ANY
22.15
22.65
37.24
30.75
42.08
60.38
34.22
27.13
49.11
6.28
4^75
10.75
-7-
-------
that was "frozen" or an EGR valve that was stuck closed but still
connected. An example of "tampering" would be an air pump that had
been removed or disconnected, or an EGR system with a vacuum line
removed or plugged.
The defect rates shown in Table 1 are also presented separately for
vehicles "with device" and for "all vehicles". The "with device"
failure rates are computed based on the number of vehicles that were
supposed to have a particular device installed. That is, the failure
rate for air pumps was calculated by dividing the number of vehicles
with broken or tampered air pumps by the total number of vehicles that
were factory equipped with air pumps. The failure rates for "all
vehicles" were computed by dividing the number of vehicles with broken
or tampered air pumps by the total number of vehicles, regardless of
whether they were supposed to be equipped with air pumps.
The relationship between the "with device" failure rates and the "all
vehicles" failure rates is determined by the percentage of inspected
vehicles that were supposed to be equipped with a particular emissions
control device. In Table 1, the rows entitled "usage" show the
percentage of all vehicles in the sample that were supposed to be
equipped with a particular device. For example 63.2% of all vehicles
in the sample were factory-equipped with air injection systems and
65.5% of all vehicles were factory-equipped with catalytic converters.
As Table 1 shows, usage rates are approaching 100% for evaporative
emission control systems and PCV systems. The lowest overall usage
rate for the devices listed is for exhaust oxygen sensors since the
-8-
-------
first 02 sensors were not introduced until 1977 (and not in
substantial numbers until 1980) .
Several of the defect rates shown in Table 1 are in boldface type.
These are the defects that are believed to have the greatest effect on
emissions. For pre-1975 model vehicles, boldface type is used for air
injection, evaporative controls, EGR, and PCV. Defects in these
components can have a large adverse effect on emissions, and the
defect rates reported are very high, ranging from 17.99% for PCV to
63.95% for EGR. Tampering with thermostatically controlled air
cleaner systems is also high; however, the effect on emissions of such
tampering is small (unless it has resulted in driveability problems
and further tampering).
For 1975-1979 models, defect rates are about half the rate for pre-
1975 models, but still significant. In the 1985 survey, one year
after the beginning of the Smog Check program, catalyst tampering was
4.43%. EGR defects were 34.21%, most of which is not related to
tampering. Almost 10% of the catalyst equipped vehicles have fuel
inlet restrictors large enough to allow the insertion of a leaded fuel
nozzle.
For 1980 and later models, defect rates are much lower. Catalyst
tampering is almost non-existent, but EGR defects total 4.61%.
Evaporative emission control system defects are over 90% lower than
for 1975-1979 models.
-9-
-------
Table 2 is a reproduction of Table 3-7 from the earlier Sierra report.
Table 2 shows an analysis of the 1985 Roadside data in which cars that
had already been through the Smog Check program are compared to those
which had not yet had an inspection. For 1979 and earlier models,
vehicles which had already been through the Smog Check program had
consistently lower failure rates. This trend is not apparent for
newer cars.
Table 2
1985 Roadside Survey Results
I/M vs. No I/M Vehicles
Tailpipe
Pre-1975 Models
No
1975-1979
1980-1983
All Model
Models
No
Models
No
Years
No
I/M
I/M
I/M
I/M
I/M
I/M
I/M
I/M
36
39
26
34
19
15
27
30
.6%
.8%
.3%
.1%
.5%
.1%
.7%
.1%
Underhood
56.
60.
53.
54.
15.
16.
42.
44.
4%
6%
6%
3%
7%
4%
5%
8%
Overall
66.
72.
60.
64.
31.
26.
53.
55.
8%
7%
3%
3%
4%
6%
4%
5%
Methodology
As shown in Table 2, there does not appear to be a major reduction in
underhood failure rates associated with a vehicle having been through
-10-
-------
the I/M program. However, without any more detailed information, it
is not clear whether the results are consistent for all types of
underhood defects. To investigate the possibility that some types of
defects are being corrected with greater efficiency, Sierra performed
a new analysis of the Random Roadside Survey data.
The objective of the analysis required that the sample of vehicles
captured in the Random Roadside Surveys be divided into those which
had already been through the Smog Check program before being captured
in the roadside sample (the "I/M" group), and those which had not yet
been subject to inspection and repair requirements (the "Non-I/M"
group). However, it should be noted that many of the "Non-I/M"
vehicles were tested under earlier (pre-1984) versions of the I/M
program when they went through change of ownership.
To segregate the vehicles in this manner, Sierra developed a
methodology under which vehicles from the Random Roadside programs
were tagged as "I/M" or "Non-I/M" based on the "Renewal Month", "Year
Due" and "Model Year" fields entered during the roadside inspection.
Renewal Month was presumably ascertained from the registration sticker
on the license plate. Year Due ("E" for even or "0" for odd) was
coded from the last digit of the each vehicle's vehicle identification
number (VIN). (In California's biennial program, vehicles are
"called" for inspection every other year based on the last digit of
their VIN: odd in odd years, etc.)
-11-
-------
Tables 3 and 4 detail the algorithms that were employed to produce the
I/M and non-I/M samples for the 1985 and 1986 surveys. The algorithms
were based on the fact that both the 1985 and the 1986 surveys were
conducted between March and May.
Table 3
Algorithm for Segregating 1985
I/M vs. Non-I/M Vehicles
1985 Roadside Survey
Renewal
Model Year VIN Month
65-82 E 1-4
5-12
0 1-2
3
3
4
4
5
Inspection
Month
-
-
_
3
4,5
3,4
5
-
I/M?
No
Yes
Yes
?
Yes
Yes
.'?
83
6-12
No
84 - 85
1-4
5-12
1-2
3
3
4
4
5
6-12
-
-
.
3
4,5
3,4
5
-
-
—
No
No
Yes
?
Yes
?
Yes
?
No
No
Comments
'84 renewal prior
to I/M start
1st cycle 5/84-
12/84
1st cycle 1/85-2/85
Uncertainty w/in
month
1st cycle 3/85
Uncertainty,IM=3
included due to
renewal notice lead
Car had I/M 4/85
Uncertainty, renewal
notice lead
Not till later in
'85
Renewal before I/M
No I/M req'd for 1st
renewal
1st cycle 1/85-2/85
Uncertainty
1st cycle 3/85
Uncertainty, renewal
notice lead time
Car had I/M 4/85
Uncertainty, lead
time
Not till later in '85
Not yet
-12-
-------
Table 4
Algorithm for Segregating 1986
I/M vs. Non-I/M Vehicles
1986 Roadside Survey
Renewal
Model Year VIN Month
66-82 E 1-2
3
3
4
4
5
6-12
0
83 E 1-2
3
3
4
4
5
6-9
Inspection
Month
3
4,5
3,4
5
-
-
-
_
3
4,5
3,4
5
-
-
I/M?
Yes
7
Yes
7
Yes
Yes
Yes
Yes
Yes
?
Yes
?
Yes
7
No
84
0
E
10
11-12
1-2
3
3
4
4
5
6-12
1-9
10
11-12
3
4,5
3,4
5
85 - 86
Yes
Yes
Yes
Yes
Yes
No
No
Yes
No
Comments
1st cycle 1/86-2/86
Uncertainty w/in
month
1st cycle 3/86
Uncertainty, renewal
notice lead time
1st cycle 4/86
1st cycle 5/84
2nd cycle 6/86-12/86,
1st cycle 6/84-12/84
Had I/M in 1985
1st cycle 1/86-2/86
Uncertainty
1st cycle 3/86
Uncertainty
1st cycle 4/86
Uncertainty
1st I/M will be 6/86-
9/86
Uncertainty,w/year
of purchase (82or83)
assumed
1st cycle 11/84-12/84
(purchase in 82)
Had I/M in 1985
1st cycle 1/86-2/86
Uncertainty
1st cycle 3/86
Uncertainty
1st cycle 4/86
Uncertainty
Not until 6/86-12/86
Assume veh purchased
'84, no I/M till '87
Some veh's purchased
'83, some '84
Assume purchased
'83, 1st cycle 11/85-
12/85
Not yet
-13-
-------
The algorithm shown in Table 3 represents and improvement over the
algorithm used in Sierra's earlier analysis of the Roadside Survey
data for the California I/M Review Committee. The earlier algorithm
did not accurately treat 1983 model vehicles with "even" Vehicle
Identification Numbers. None of the "even" VIN 1983s should have been
in the I/M sample because no I/M test was required before their first
registration renewal. In addition, the old algorithm lumped the "odd"
VIN vehicles with March-May renewal months together and treated them
all as "uncertain" . The new algorithm uses the inspection month
information to estimate whether the vehicle is likely to have
completed an I/M cycle.
Table 5 summarizes the results of the segregation effort. It is
apparent that the "Non-I/M" sample for 1986 is very small as most of
the vehicles old enough to be subject to I/M would have already been
through the program. The sample sizes for the various categories of
vehicles in the 1986 survey were as follows:
,->
Model Year Range I/M Sample Non-I/M Sample
pre-1975 469 0
1975-79 677 0
1980 and Later 562 103
Total Sample 1708 103
Obviously, it was only possible to do "I/M" vs. "Non-I/M" comparisons
for 1980 and later models using the 1986 Roadside Survey.
14-
-------
EPA's direction to Sierra was that "The analysis should be performed
on a statewide basis, unless the sample size distribution among
regions is much different for the roadside sample than for the Smog
Check sample, in which case the contractor shall apply weighting
factors to one data set or the other." To determine whether weighting
,of the results would be required, the geographical distribution of
Random Roadside data was compared to the geographical distribution of
Test Analyzer System (TAS) data recorded at Smog Check stations.
Table 5
Description of Random Roadside Survey Sample
1985 1986
Roadside Survey Roadside Survey
Identified as "I/M" 645 1708
Identified as "Non-I/M" 723 103
Other 442 602
Total 1810 » 2413
Where:
"I/M" = Vehicles expected to have had an I/M test before
the roadside inspection.
"Non-I/M" - Vehicles expected to have not had an I/M test
before the roadside inspection (See Other).
"Other" - Vehicles for which the occurrence of an I/M test
before the roadside inspection is uncertain.
Also, newer vehicles known to have not had an I/M
test which were removed from the analysis to avert
a data bias. For the 1985 Survey, this category
would include '84 & later vehicles, for 1986
Roadside, '85 & later vehicles.
-15-
-------
Table 6 shows the distribution of the sample of vehicles obtained
through the Random Roadside Survey. Note that the first three areas
(Los Angeles, San Francisco, and San Diego) listed for the 1985 survey
have a similar fraction of the Random Roadside Survey Vehicles even
though they have substantially different populations.
Table 6
Distribution of Roadside Vehicles by District
District I/M
1985 Survey:
Los Angeles
San Francisco
San Diego
Sacramento
Ventura
Fresno
Total 645
1986 Survey:
Los Angeles
San Francisco
San Diego
Sacramento
Ventura
Fresno
Total 1,708
Non-I/M
723
1,368
103
1,811
Fraction
166
183
146
66
84
0
200
197
162
85
79
0
366
380
308
151
163
0
0.268
0.278
0.225
0.110
0.119
0.000
1.000
380
492
264
132
137
303
20
36
22
11
5
5
400
528
286
143
142
308
0.221
0.292
0.158
0.079
0.078
0.170
1.000
To compare the Random Roadside Survey sample to a vehicle population
weighted sample, Sierra performed an analysis of TAS data. As shown
in Table 7, the fraction of the TAS total records in each district is
consistent with what might be expected from population differences.
For example, about twice as many Smog Check tests are performed in Los
-16-
-------
Angeles as in San Francisco. Based on this analysis, weighting of the
sample was done using the factors shown in Table 7-
Based on the direction received from EPA, "Adjustment for mileage
differences is not required, but substantial mileage differences
between corresponding I/M and non-I/M strata should be addressed."
Table 7
Development of District Weighting Factors
Dist (ARB~)
1
2
3
4
5
6
(LA)
(SF)
(SD)
(SACTO)
(VENT)
(FRESNO)
April 87
337
280
65
88
5
29
,817
,253
,401
,411
,045
,288
IAS
Mav 87
325
143
61
28
23
29
,185
,371
,301
,026
,282
,936
bampie Size By District -
June 87 Julv 87 AMP-
233,196
103,972
57,407
18,138
17,111
16,161
448,
194,
85,
33,
20,
18,
947
018
806
334
173
069
300,
174,
65,
42,
20,
12,
87
666
053
790
160
159
510
Sept 87
328,906
165,146
26,953
26,373
17,849
Dist
TOTALS (w/Fresno)
N Fraction
1 1,974,717 0.5048
2 1,060,813 0.2712
3(x6/5) 402,846 0.1030
4 237,022 0.0606
5 112,513 0.0288
6 123,813 0.0317
Total
3,911,724
TOTALS (w/o Fresno)
Dist N Fraction
1 1,974,717 0.5213
2 1,060,813 0.2801
3(x6/5) 402,846 0.1064
4 237,022 0.0626
5 112,513 0.0297
Total 3,787,911
Table 8 shows mileage distributions of the four Roadside samples (I/M
and non-I/M for both years). No significant differences in mileage
distribution are observed between the I/M & non-I/M samples for the
1985 Survey vehicles. Therefore, no weighting factors were required.
-17-
-------
For the 1986 Survey, the substantial mileage difference between I/M &
non-I/M vehicles occurs because almost all vehicles (excluding new
models) had been I/M tested by the time of the 1986 Survey (3/86
through 5/86). However, the 1986 survey has so few non-I/M vehicles
that it provides no meaningful information about differences between
I/M and non-I/M vehicles regardless of whether weighting factors are
applied.
Table 8
Distribution (in %) of
Roadside Data by
Mileage Interval
Sample Size
Mileage Interval
10,
20,
30,
40,
50,
60,
70,
80,
90,
100,
0-
000-
000-
000-
000-
000-
000-
000-
000-
000-
000-
150,
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
90,000
100,000
150,000
000
1985 Roadside Survey
I/M Non-I/M
645
0.0%
0.2%
3.9%
3.4%
7.0%
9.5%
7.4%
8.2%
8.5%
8.7%
29.6%
723
0.1%
2.6%
4.8%
7.3%
6.4%
7.9%
8.9%
7.1%
7.6%
8.7%
28.4%
1986 Roadside Survey
I/M Non-I/M
13.6% 10.2%
1708
0.0%
0.6%
1.8%
4.0%
4.7%
7.9%
8.6%
8.5%
8.7%
8.8%
34.3%
12.0%
103
2.9%
15.5%
23.3%
23.3%
17.5%
8.7%
1.9%
2.9%
1.0%
1.0%
1.9%
0.0%
-18-
-------
3. RESULTS
Detailed results from the Random Roadside Survey underhood inspections
for both the "I/M" and "Non-I/M" cases are presented in Tables 9-12.
The format is the same as the component-specific results presented in
the earlier analysis, however, the samples have been segregated into
"I/M" and "non-I/M" subgroups.
Figure 1, repeated from the Summary, and Figures 2-4 show the
comparison for all model years and three different model year groups
based on the 1985 Random Roadside Survey. Readers should note that
Figure 1
Underhood "Tampering" Rates
I/M vs. Non-I/M Vehicles
19S5 Roadside Survey
I/M Sample
Non-I/M Sample
AIR CAT SPARK EVAP EGR LEAD O2S PCV TAG ANY
Device Type
Note: 1984 and later models removed from
sample to eliminate age bias.
-19-
-------
Table 9
Underhood Inspection Results - 1985 Roadside Survey
I/M Vehicles
All
Model
Years
With
Device
All
Model
Years
All
Vehicles
Pre-75
Model
Years
With
Device
Pre-75
Model
Years
All
Vehicles
75-79
Model
Years
With
Device
75-79
Model
Years
All
Vehicles
1980+
Model
Years
With
Device
1980+
Model
Years
All
Vehicles
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
AIR
2.61
6.79
9.40
59.4%
1.55
4.03
5.58
6.45
16.13
22.58
26.3%
1.69
4.24
5.93
3.30
7.69
10.99
77.4%
2.55
5.96
8.51
0.00
1.44
1.44
79.9%
0.00
1.15
1.15
rairur
CAT
0.00
2.12
2.12
58.4%
0.00
1.24
1.24
--
--
--
0.00
3.40
3.40
87.7%
0.00
2.98
2.98
0.00
0.59
0.59
97.7%
0.00
0.57
0.57
e rerce
SPARK
2.67
6.84
9.51
92.9%
2.48
6.36
8.84
4.33
13.42
17.75
97.9%
4.24
13.14
17.38
2.19
4.39
6.58
97.0%
2.13
4.26
6.39
0.71
0.00
0.71
80.5%
0.57
0.00
0.57
ntage b
EVAP
3.07
6.15
9.22
85.7%
2.64
5.27
7.91
6.04
14.77
20.81
63.1%
3.81
9.32
13.13
3.48
4.78
8.26
97.9%
3.40
4.68
8.08
0.00
0.57
0.57
100%
0.00
0.57
0.57
y Type
EGR
17.62
9.15
26.77
67.8%
11.94
6.20
18.14
36.11
22.22
58.33
30.5%
11.02
6.78
17.80
20.67
10.58
31.25
88.5%
18.30
9.36
27.66
5.10
1.27
6.37
90.2%
4.60
1.15
5.75
of Brok
LEAD
0.26
6.49
6.75
59.7%
0.16
3.88
4.04
--
--
--
0.48
9.09
9';57
88.9%
0.43
8.09
8.52
0.00
2.89
2.89
99.4%
0.00
2.87
2.87
en or T<
02S
0.87
0.00
0.87
17.8%
0.16
0.00
0.16
--
--
0.00
0.00
0.00
4.70%
0.00
0.00
0.00
0.98
0.00
0.98
58.6%
0.57
0.00
0.57
unpered
PCV
3.26
4.50
7.76
99.8%
3.26
4.50
7.76
3.83
10.21
14.04
99.6%
3.81
10.17
13.98
3.83
2.13
5.96
100%
3.83
2.13
5.96
1.72
0.00
1.72
100%
1.72
0.00
1.72
Device
TAG
9.49
9.84
19.33
88.2%
8.37
8.68
17.05
13.64
22.73
36.37
83.9%
11.44
19.07
30.51
11.11
4.63
15.74
91.9%
10.21
4.26
14.47
1.94
0.65
2.59
89.1%
1.72
0.57
2.29
ANY
24.34
24.96
42.33
29.24
38.56
58.05
31.06
24.68
46.38
8.62
6.90
15.52
-20-
-------
Table 10
Underhood Inspection Results - 1985 Roadside Survey
Non-I/M Vehicles
All
Model
Years
With
Device
All
Model
Years
All
Broken
Tamper
Total
Usage
Broken
Tamper
Total
AIR
5.23
7.19
12.42
63.5%
3.32
4.56
7.88
raij.ur
CAT
0.00
2.44
2.44
62.4%
0.00
1.52
1.52
e rerce
SPARK
1.05
7.06
8.11
92.1%
0.97
6.50
7.47
ntage b
EVAP
2.63
8.50
11.13
89.5%
2.35
7.61
9.96
7 Type
EGR
15.97
9.58
25.55
69.3%
11.07
6.64
17.71
of Brok
LEAD
0.22
4.35
4.57
63.6%
0.14
2.77
2.91
en or Tampered
02S
0.00
1.82
1.82
22.8%
0.00
0.41
0.41
PCV
2.23
6.95
9.18
99.4%
2.21
6.92
9.13
Device
TAG
9.49
9.95
19.44
88.9%
8.44
8.85
17.29
ANY
24.07
25.86
40.80
Vehicles
Pre-75
Model
Years
With
Device
Pre-75
Model
Years
All
Broken
Tamper
Total
Usage
Broken
Tamper
Total
5.97
26.87
32.84
31.0%
1.85
8.33
10.18
--
--
--
--
1.88
14.55
16.43
98.6%
1.85
14.35
16.20
6.08
24.32
30.40
68.5%
4.17
16.67
20.84
46.30
22.22
68.52
25.0%
11.57
5.56
17.13
—
50.00
50.00
--
0.46
0.46
4.65
15.81
20.46
99.5%
4.63
15.74
20.37
14.05
24.86
38.91
85.6%
12.04
21.30
33.34
30.09
42.13
60.19
Vehicles
75-79
Model
Years
With
Device
75-79
Model
Years
All
Vehicles
1980+
Model
Years
With
Device
1980+
Model
Years
All
Vehicles
Broken
Tamper
Total
Usage
Broken
Tamper
Total
Broken
Tamper
Total
Usage
Broken
Tamper
Total
9.50
5.50
15.00
76.6%
7.28
4.21
11.49
0.52
2.08
2.60
78.0%
0.41
1.63
2.04
0.00
4.81
4.81
79.7%
0.00
3.83
3.83
0.00
0.41
0.41
98.8%
0.00
0.41
0.41
1.16
5.41
6.57
99.2%
1.15
5.36
6.51
0.00
1.03
1.03
78.9%
0.00
0.81
0.81
3.15
6.30
9.45
97.3%
3.07
6.13
9.20
0.00
1.22
1.22
99.6%
0.00
1.22
1.22
20.54
14.73
35.27
85.8%
17.62
12.64
30.26
4.04
1.35
5.39
90.7%
3.66
1.22
4.88
0.47
8.92
9*39
81.6%
0.38
7.28
7.66
0.00
0.41
0.41
99.2%
0.00
0.41
0.41
0.00
0.00
0.00
4.20%
0.00
0.00
0.00
0.00
1.32
1.32
61.8%
0.00
0.81
0.81
1.92
5.00
6.92
99.6%
1.92
4.98
6.90
0.41
1.23
1.64
99.2%
0.41
1.22
1.63
12.66
6.75
19.41
90.8%
11.49
6.13
17.62
2.26
0.90
3.16
89.8%
2.03
0.81
2.84
36.40
30.27
51.72
5. '69
6.91
12.20
-21-
-------
Table 11
Underhood Inspection Results - 1986 Roadside Survey
I/M Vehicles
All
Model
Years
With
Device
All
Model
Years
All
Broken
Tamper
Total
Usage
Broken
Tamper
Total
AIR
3.25
8.21
11.46
64.9%
2.11
5.33
7.44
ranure percentage by Type of Broken or Tampered
CAT
0.00
2.59
2.59
67.9%
0.00
1.76
1.76
SPARK
1.19
8.26
9.45
88.6%
1.05
7.32
8.37
EVAP
3.21
8.19
11.40
93.0%
2.99
7.61
10.60
EGR
22.30
12.94
35.24
76.9%
17.15
9.95
27.10
LEAD
1.37
4.78
6.15
68.6%
0.94
3.28
4.22
02S
0.00
0.84
0.84
20.8%
0.00
0.18
0.18
PCV
2.76
7.57
10.33
99.7%
2.75
7.55
10.30
Device
TAG
10.45
9.02
19.47
90.2%
9.43
8.14
17.57
ANY
30.21
30.21
50.18
Vehicles
Pre-75
Model
Years
With
Device
Pre-75
Model
Years
All
Broken
Tamper
Total
Usage
Broken
Tamper
Total
3.68
25.74
29.42
29.0%
1.07
7.46
8.53
--
--
--
1.31
17.65
18.96
97.9%
1.28
17.27
18.55
7.41
19.37
26.78
74.8%
5.54
14.50
20.04
41.42
29.59
71.01
36.0%
14.93
10.66
25.59
--
--
--
--
--
--
4.50
14.99
19.49
99.6%
4.48
14.93
19.41
13.29
19.81
33.10
88.3%
11.73
17.48
29.21
31.56
46.91
65.67
Vehicles
75-79
Model
Years
With
Device
75-79
Model
Years
All
Broken
Tamper
Total
Usage
Broken
Tamper
Total
4.27
8.93
13.20
76.1%
3.25
6.79
10.04
0.00
3.64
3.64
89.4%
0.00
3.25
3.25
1.06
5.88
6.94
97.9%
1.03
5.76
6.79
2.81
8.00
10.81
99.7%
2.81
7.98
10.79
28.92
15.67
44.59
91.4%
26.44
14.33
40.77
1.48
8.22
9t70
89.8%
1.33
7.39
8.72
0.00
4.55
4.55
3.20%
0.00
0.15
0.15
2.96
7.11
10.07
99.7%
2.95
7.09
10.04
12.50
8.06
20.56
89.8%
11.23
7.24
18.47
40.18
34.12
59.23
Vehicles
1980+
Model
Years
With
Device
1980+
Model
Years
All
Vehicles
Broken
Tamper
Total
Usage
Broken
Tamper
Total
1.97
2.19
4.16
81.3%
1.60
1.78
3.38
0.00
1.45
1.45
98.4%
0.00
1.42
1.42
1.28
1.28
2.56
69.8%
0.89
0.89
1.78
1.07
1.42
2.49
100%
1.07
1.42
2.49
8.37
4.37
12.74
93.6%
7.83
4.09
11.92
1.25
1.07
2.32
99.5%
1.25
1.07
2.32
0.00
0.60
0.60
59.3%
0.00
0.36
0.36
1.07
1.96
3.03
99.8%
1.07
1.96
3.03
5.78
1.54
7.32
92.3%
5.34
1.42
6.76
17.08
11.57
26.33
-22-
-------
Table 12
Underhood Inspection Results - 1986 Roadside Survey
Non-I/M Vehicles
All
Model Broken
Years Tamper
With Total
Device
Usage
All
Model Broken
Years Tamper
All Total
Vehicles
AIR
0.00
1.45
1.45
67.0%
0.00
0.97
0.97
raj. .LUX
CAT
0.00
0.00
0.00
100%
0.00
0.00
0.00
e rercei
SPARK
0.00
0.00
0.00
44.7%
0.00
0.00
0.00
itage b
EVAP
0.00
0.97
0.97
100%
0.00
0.97
0.97
y Type c
EGR
0.00
1.16
1.16
83.5%
0.00
0.97
0.97
)t Broken or Tampered
LEAD 02S PCV
0.00 0.00 0.98
0.00 0.00 0.00
0.00 0.00 0.98
100% 89.3% 99.0%
0.00 0.00 0.97
0.00 0.00 0.00
0.00 0.00 0.97
Device
TAG
0.00
0.00
0.00
82.5%
0.00
0.00
0.00
ANY
0.97
1.94
2.91
Pre-75
Model Broken
Years Tamper
With Total
Device
Usage
Pre-75
Model Broken
Years Tamper
All Total
Vehicles
75-79
Model Broken
Years Tamper
With Total
Device
Usage
75-79
Model Broken
Years Tamper
All Total
Vehicles
1980+
Model
Years
With
Device
1980+
Model
Years
All
Vehicles
Broken
Tamper
Total
Usage
Broken
Tamper
Total
0.00
1.45
1.45
67.0%
0.00
0.97
0.97
0.00
0.00
0.00
100%
0.00
0.00
0.00
0.00
0.00
0.00
44.7%
0.00
0.00
0.00
0.00
0.97
0.97
100%
0.00
0.97
0.97
0.00
1.16
1.16
83.5%
0.00
0.97
0.97
0.00
0.00
0.00
100%
0.00
0.00
0.00
0.00
0.00
0.00
89.3%
0.00
0.00
0.00
0.98
0.00
0.98
99.0%
0.97
0.00
0.97
0.00
0.00
0.00
82.5%
0.00
0.00
0.00
0.97
1.94
2.91
-23-
-------
Figure 2
Underhood "Tampering" Rates
Pre-'75 I/M vs. Non-l/M Vehicles
1985 Roadside Survey
I/M Sample
Non-l/M Sample
SPARK
EVAP EGR PCV
Device Type
TAG
ANY
Figure 3
Underhood "Tampering" Rates
1975-79 I/M vs. Non-l/M Vehicles
1985 Roadside Survey
I/M Sample
Non-l/M Sample
AIR CAT SPARK EVAP EGR LEAD PCV TAC ANY
Device Type
-24-
-------
Figure 4
Underhood "Tampering" Rates
1980-83 I/M vs. Non-l/M Vehicles
1985 Roadside Survey
8
f •
o
O)
.£
&
CO
6.96.9
I/M Sample
Non-l/M Sample
AIR CAT SPARK EVAP EGR LEAD O2S PCV TAG ANY
Device Type
these figures show only those underhood defects classified as
"tampering". The detailed results shown in Tables 9-12 contain the
defect rates for non-tampering and "total" defects as well.
As illustrated in Figure 2, the tampering rates for pre-1975 models
are consistently lower for vehicles that have already been through the
I/M program. However, Figures 3 and 4 indicate that the results are
mixed for 1975 and later models.
As illustrated earlier in Tables 6 and 7, there was a substantial
discrepancy between the fraction of the Random Roadside sample that
was drawn from each region and the fraction of the total number of
-25-
-------
Table 13
Underhood Inspection Results Weighted by District
All
Model
Years
With
Device
All
Model
Years
All
Vehicles
Broken
Tamper
Total
Broken
Tamper
Total
AIR
2.96
6.34
9.30
1.68
3.60
5.28
r eij-iure
CAT
1985
0.00
1.48
1.48
0.00
0.83
0.83
; rercer
SPARK
itage D}
EVAP
' Type
EGR
of Broken or Tampered
LEAD 02S PCV
Device
TAG
ANY
Roadside Survey- I/M Vehicles
3.48
8.80
12.28
3.21
8.17
11.38
1985 Roadside
All
Model
Years
With
Device
All
Model
Years
All
Vehicles
All
Model
Years
With
Device
All
Model
Years
All
Vehicles
Broken
Tamper
Total
Broken
Tamper
Total
Broken
Tamper
Total
Broken
Tamper
Total
3.99
7.76
11.75
2.54
4.92
7.46
3.44
8.39
11.83
2.24
5.47
7.70
0.00
2.09
2.09
0.00
1.32
1.32
1986
0.00
2.59
2.59
0.00
1.75
1.75
0.60
7.93
8.53
0.58
7.20
7.78
3.72
6.71
10.42
3.17
5.73
8.90
Survey
2.68
9.29
11.98
2.42
8.41
10.83
16.50
8.35
24.85
11.05
5.64
16.70
- Non
15.46
10.28
25.74
10.78
7.27
18.05
0.57
6.38
6.95
0.31
3.70
4.01
0.72
0.00
0.72
0.15
0.00
0.15
2.77
5.25
8.02
2.76
5.24
8.00
8.44
10.77
19.21
7.42
9.38
16.79
23.82
26.05
42.47
I/M Vehicles
0.41
4.02
4.43
0.26
2.54
2.80
0.00
2.48
2.48
0.00
0.66
0.66
2.01
6.14
8.15
2.00
6.11
8.11
9.92
10.67
20.59
8.76
9.43
18.19
22.94
27.10
40.93
Roadside Survey - I/M Vehicles
1.20
9.65
10.85
1.05
8.37
9.41
1986 Roadside
All
Model
Years
With
Device
All
Model
Years
All
Broken
Tamper
Total
Broken
Tamper
Total
0.00
0.45
0.45
0.00
0.35
0.35
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
3.91
9.37
13.28
3.63
8.70
12.33
Survey
0.00
0.55
0.55
0.00
0.55
0.55
22.43
13.73
36.16
17.15
10.50
27.65
Non
0.00
0.40
0.40
0.00
0.35
0.35
1.91
4:69
6.60
1.30
3.20
4.50
0.00
0.74
0.74
0.00
0.14
0.14
3.07
8.61
11.68
3.06
8.57
11.63
10.05
9.43
19.47
9.15
8.62
17.77
30.09
31.95
50.64
I/M Vehicles
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.52
0.00
2.52
2.52
0.00
2.52
0.00
0.00
0.00
0.00
0.00
0.00
2.52
0.90
3.43
Vehicles
-26-
-------
Smog Check tests performed in each region. In order to eliminate any
possible biases in the data due to differences in underhood failure
rates between I/M districts in California, the Random Roadside sample
was weighted by the number of Smog Check tests performed in each
district. The weighted component failure rates at the roadside for
I/M and non-I/M vehicles (all model years) are shown in Table 13.
Figures 5 and 6 show how the tampering rates are affected by the
use of "unweighted" vs. "weighted" samples. As these figures show,
there is no significant difference between the tampering rates for
vehicles that have been through the I/M program and those that have
not.
Figure 5
Weighted and Unweighted "Tampering" Rates
for Vehicles Already Subject to I/M
1985 Roadside Survey
Device Type
2.1 m
6.8
6.2
9.2
6.5
0.0
25.0 .
r"*""™—'-'"-". i i , i ^_
30 25 20 15 10 5 0
Tampering Rate (%) Unweighted
Note: 1984 and later models removed from
sample to eliminate age bias.
6.3
26.1
0 5 10 15 20 25 30
Tampering Rate (%) Weighted
-27-
-------
25.9
Figure 6
Weighted and Unweighted "Tampering" Rates
for Vehicles Not Yet Subject to I/M
1985 Roadside Survey
Device Type
7.2
7.1
8.5
9.6
2.4
4.4
7.0
10.0
27.1
30 25 20 15 10 5 0
Tampering Rate (%) Unweighted
Note: 1984 and later models removed from
sample to eliminate age bias.
0 5 10 15 20 25 30
Tampering Rate (%) Weighted
-28-
------- |