EPA-AA-IMS-81-11
                                Technical  Report
               A  Study of  the Effectiveness of Mechanic Training
           For Vehicle Emissions Inspection and Maintenance Programs
                                R.  Bruce Michael
                                  April,  1981
                                     Notice

Technical  Reports  do  not  necessarily   represent   final  EPA   decisions   or
positions.  They  are  intended  to  present technical  analysis  of issues  using
data  which are  currently  available.   The  purpose  in the  release  of  such
reports is to  facilitate  the exchange of  technical  information and  to  inform
the public of  technical developments which may  form  the basis for a  final  EPA
decision, position or regulatory action.
                        Inspection and Maintenance Staff
                      Emission Control Technology Division
                       Office of Air, Noise and Radiation
                      U.S. Environmental Protection Agency
                              Ann Arbor, Michigan

-------
                                       2

                               TABLE OF CONTENTS

                                             /
                                                                Page

1.0  SUMMARY AND CONCLUSIONS                  .                   1


2.0  BACKGROUND                                                  2


3.0  TRAINING COURSE USED                                        3


4.0  FACILITY TRAINING                                           3
5.0  STUDY DESIGN                                                5
5.1  Vehicle Selection and Recruitment                           5
5.2  Vehicle Testing                                             6
5.3  Repairs by Participating Facilities                         7
6.0  RESULTS                                                     7
6.1  Emissions and Fuel Economy                                  7
6.2  Cost and Amount of Repairs by Commercial                    15
       Facilities
6.3  Quality of Repairs                                         16
6.4  Repair Effectiveness by Facility Type                      24
7.0  DISCUSSION AND CONSLUSIONS:  POSSIBLE REASONS FOR THE
     OBSERVATIONS IN PORTLAND AND GENERAL IMPLICATIONS FOR
     MECHANIC TRAINING BENEFITS
7.1  Possible expectations of training,  actual                  25
       results and possible explanations
7.2  General implications for mechanic training                 26
       benefits and limitations of this  study
8.0  MECHANIC TRAINING EMISSION REDUCTION BENEFITS
     CONTAINED IN MOBILE2
APPENDIX - Vehicle Test Fleet                                   33

-------
1.0 SUMMARY          .

This  report  describes a  study to  determine  if  a  short,  practical  training
course for emissions  repairs has a  supplementary  emission reduction and  fuel
economy  benefit   to  an  Inspection   and  Maintenance  programi   The  study  was
performed in the early part of 1980  in Portland,  Oregon.

Two matched  samples  of  21  vehicles each  were sent  to  ten repair  facilities
before and after  the facilities  received  training.   The experiment was  blind
in that the ten repair facilities did  not  know that the study was  in  progress
or  that  the  42 vehicles  were in  any way  connected with  EPA.   The  training
course used  was  a 16  hour  course  in repairing  vehicles  which  fail  an  idle
emissions test.

Vehicles were  tested  before and  after  repairs on  the  Federal Test Procedure
and various short tests for  emissions  and  fuel economy.  Contractor laboratory
personnel then attempted  to restore the  vehicles  to their specifications  and
vehicles were retested.

Results indicate that:

    1.  Training  in  Portland  did  not  produce  greater   emission  reduction
    benefits  from  repairs  than  the  emission   reduction  benefits   without
    training.  Similar emission reductions  of approximately 41% HC and  53% CO
    on the Federal Test  Procedure occurred from repairs both  before and  after
    training.

    2. A  0.8%  improvement in  fuel  economy from repairs  (average  of city  and
    highway  driving)  was  seen   after  training.   This  improvement  was  not
    statistically significant, however.

    3.  The  average  cost  of repairs  was  nearly  identical before  and  after
    training, about $19.50.

    4. Several measures  of the quality of  repairs indicate a  slight  improve-
    ment  with  training.   The  quality  of  carburetor  adjustments,  which  is
    highly important,  did not  improve  with training, but  spark  timing  adjust-
    ments and driveability improved slightly.

    5.  Emission  reductions  from  repairs   performed  by  different  types  of
    facilities  (gas  stations,  independent  garages and chain  stores)  did  not
    differ  significantly  in  emission  reductions   either  before  or  after
    training.

    6. Contractor repairs to the vehicles  show  that the potential  exists  for
    significant further emissions reductions  beyond  the  reductions  achieved by
    commercial  facility  repairs.    This  potential  may  not be realizeable  in
    practice.  Significant potential for further  fuel economy  improvement also
    exists.

-------
There  are  two major  reasons why  this study  may  not  have  shown  significant
emission reduction benefits  from  training.   First,  the Portland idle  emission
standards are  more  stringent than  almost any  other  I/M  program  will  adopt,
thus  there  is  not  as  much  "room"  for   improvement  that   training   could
accomplish in Portland, compared to  other I/M programs in which the mechanics
do not have  to adjust the vehicles' emissions.so low.   For  example,  the  idle
CO standard  for  most  late model  vehicles in Portland  is  1.0% (including  the
allowed tolerance) and the average after  repair  idle  CO level  is 0.1%;  in  New
Jersey the  standard  for those vehicles is  3.0% and  the  average after  repair
idle  CO  level  is 0.8%.   Second,  the I/M  program  had  been  operational  in
Portland for over four  years prior to the study.   It  is expected  that  during
that time mechanics would have gained  some on-the-job  training  and  "grapevine"
information which would lessen the  effect  of  a training course.   Also,  the
mechanics would have  formed habits  for repairing the  emission  failure  vehicles
which  a  short  training  course  could  not greatly  change;  for  example,  many
facilities  in  Portland always  perform,   for  a  set  fee,  a  "DEQ  adjustment",
meaning standard adjustments to  pass the  state idle  emissions test.

Additional discussion and conclusions are  contained  in Section  7.0.

2.0 BACKGROUND

Numerous studies  in the past have indicated  that mechanic  training is  a  very
important part  of an Inspection  and Maintenance (I/M)  program.   In  1973  the
National Academy  of Sciences[1]*  stated  its  concern that  the  service  industry
may  not  be able  to  adequately  service cars  from an  emission control  stand-
point.  Since  that time many other studies  have agreed that mechanic  training
is essential to an I/M  program in  order to  achieve  the full benefits, both in
terms  of emissions  reduction and  fuel economy  improvement.  This  need  can be
seen by observing the effort the  auto manufacturers put into  training dealer-
ship mechanics.  A service  training  manager  from one of the major  manufactur-
ers  summarized  the general  need:   "Engineers continually make  refinements  and
improvements  so technicians  need continual  refreshers to  keep  up with  the
changes in existing technology;  it's a never-ending  process."(2]

Twenty-nine  states  will  have operating  I/M  programs  in 1987.   States  will
receive "credits", representing expected  emission reductions due to I/M, which
they will use  in  1982 to  demonstrate compliance with the National  Ambient Air
Quality Standards.  EPA estimated  in 1978 that additional emission reductions
could  be achieved  from a mechanic  training  program;  many states  are  planning
to institute  training partly in order to claim more credits.    Since  training
is a relatively low cost strategy, it  is  very attractive.  The  1978 EPA  credit
estimates  were based  (in  MOBILEl)  on   a mathematical model  using  emission
results from a large  number of in-use  vehicles.   The  credits  associated  with
mechanic training were based on the  assumption  that  trained  mechanics  would be
able  to  repair failed  I/M  vehicles  to  the  Federal  standards  for which  they
were  designed.   This  assumption  had not  been  tested  in  a  field  experiment.
Therefore, in  1979 EPA  made  plans  for testing the effectiveness of a  training
course on  emissions  in an  area which currently had  I/M,  but  whose mechanics
had  little or  no  formal training  in the  area of emissions.  Results from that
study were to be used in a revision of MOBILEl, called MOBILE2.


* Numbers in brackets designate references at the end of the  report.

-------
The Portland, Oregon area was chosen as  the  site  to test the effectiveness  of
training.  This was a logical choice for EPA since the area had  an  established
I/M program  with  very  little previous  formal training, and  EPA already had  a
contractor-operated emissions testing  laboratory  present.   The  contractor  had
experience in  recruiting  vehicles  and  testing  them using a  variety of  tests
including  the  Federal  Test  Procedure  (FTP), the  same procedure EPA uses  to
certify new cars.

3.0 TRAINING COURSE USED

In  order for  emissions  training  to  be  widespread and  accepted  by a  large
number of  auto technicians,  it would  have  to be  relatively  short  in  length.
EPA had  a  short,  practical course  designed  by the National Center  for  Vehicle
Emissions  Control and  Safety  of  Colorado  State  University  (CSU)  over  the
1978-79  period.  Design for  this course  began more  than  one  year prior to  its
use in Portland  and  it had already been  field  tested, although CSU  felt  that
minor refinements may still have been needed.

The course was  designed  to  be  a maximum of  16 hours  in length, which  could,
for example, be  taught over several evening  sessions.   In this way  mechanics
would  not have   to  take  leave from  work,  a  design  which  was   considered
essential to the course becoming widespread.

The  course mainly  teaches  the  proper  diagnosis  of  I/M  failures;  in  other
words, hov the mechanic should  proceed in order to repair a  vehicle  which  has
failed the I/M idle  test.   The titles  of each  of the  six units taught in  the
course follow.

  Unit                               Topics

    1.                  I/M Programs and  Vehicle Emissions;  Short Tests.
    2.                  Equipment Used  (Infrared Analyzer).
    3.                  Types  of  Emission  Failures;  General  Troubleshooting
                        Information.
    4.                  Correction Procedures for Excessive  HC Emissions.
    5.                  Correction  Procedures for Excessive  CO Emissions.
    6.                  Proper Carburetor Adjustment Procedures.

4.0 FACILITY TRAINING

Five  categories  of repair facilities  which  perform work for emission  repairs
were  designated.   The  original  design  called for  the  recruitment of three of
each  type,  for a  total  of  15  facilities to be  used  in  the  study.   Fifteen
facilities were  recruited, but  five later had to be dropped  due to  low  or no
participation in  the training,  including  the  only new  car dealership which  had
been  signed  up.   (Dealership  mechanics   were  generally  not  agreeable  to  the
training, because  they already  attended  many  training  sessions  provided by  the
manufacturers.)   The facility  categories and the number  of facilities  of each
that were actually used throughout  the  study are listed below.

-------
                        Facilities  Used                    No.
               1. Gasoline-service  stations   s               2

               2. Independent repair garages                2

               3. Chain stores (e.g.,  Wards, Goodyear)       2

               4. New car dealerships                        0

               5. Facilities advertising I/M emissions       4
                  repairs (one was  a gas station,  the
                  other three independent garages).        	
                                                           10

Approximately 100 facilities  were  originally contacted.   Facilities were  told
that  EPA  was sponsoring  a  pilot  offering   in  the  Portland  area  of  a  short
course  in  emissions  repair  that  had  been  developed   by   Colorado   State
University  (CSU).   The  course  was   offered  at  no   charge   to   mechanics.
Incentives to the mechanics were  free  training,  a  certificate  of  completion,
and a monetary   incentive  of $50 for  completing  the course.   They were  also
told  that  we wanted  to  include only  those  facilities  in which all  tune-up
mechanics could take the course.   (The reason for this was  to assure that  test
vehicles  sent   to  the   facilities   would  be   repaired   by   participating
mechanics.)   Although  this  was not   strictly  adhered  to  for  one  facility,
results from that facility were not used  in the final analysis.   All mechanics
in the 10 facilities listed above received the training.

Many  facilities  contacted were either  not interested  or  could not get all  the
tune-up mechanics  there  to take the  training.   The EPA Project Officer  also
asked  questions  concerning previous training in  order  to  identify  facilities
which  had  an  unusually  high  amount  of  training  and  would  therefore  be
unrepresentative  of the  norm.  Nearly  all facilities  reported  that  their
mechanics had at least some previous training,  but none were judged as  being
outside  the   norm.   The  mechanics  from the  facilities  used  in  the  final
analysis  had varied  experience and   prior training.    The median  amount  of
tune-up experience  was five years   and  the  median amount of  training was  one
prior  course.   The  range of experience was  from  one  to  twenty-nine years  and
the range of  the amount  of courses taken was from  none  to seven  (nearly  half
of the mechanics had previously taken one prior  course).

The mechanics were trained  in  February  and  March  of 1980.   Five  classes  at
three  community  colleges  were held, averaging  7 mechanics per class.  Four of
the classes were held  in  five evening sessions  over two-week periods, and the
fifth  class was given  on two consecutive Saturdays  in all-day  meetings.   All
but three of  the mechanics who started attended all sessions.   A total of 23
mechanics  were   trained   from  the   10  facilities  used   in  the  study.    Their
average test  scores on the same written  test at  the  beginning  and  end of  the
training were 71  and 90  respectively.   CSU  felt  that  a minimum  final score of
80 should be  attained  by anyone who had  successfully  learned material in  the
course.  All of  the mechanics achieved scores of at least 80 at  the  end of the
training.

-------
Prior to the training,  a two-day workshop was given by CSU for the instructors
who would be  teaching  the classes.   Orientation was given  to familiarize the
teachers with  the  course  and the CSU  instructional materials and to  give an
idea of how much  time  to spend on  each  section.   EPA and CSU  felt  that  this
was necessary, because  the instructors  had not seen the course material and at
that time there was also no mention  of  time to be spent on each section of the
course -  since the course  covers  a lot  of  material  over a  relatively short
period  of  time, mention  of the  time  to  be spent on  each  section  was more
important than with most courses.  Because teaching the course  would  be a new
experience for the teachers, we felt that the workshop was necessary and would
not  give  them  more  knowledge  or  expertise  than  teachers in  an actual  I/M
mechanic training program would have after  teaching the  course  once  or  twice.
Some  experience in  teaching  the  course  would be  the  normal  situation to
evaluate anyway, since it can be expected  that  in  I/M  areas which offer a lot
of training, each teacher would hold several  classes.

5.0 STUDY DESIGN

5.1 VEHICLE SELECTION AND RECRUITMENT

Two groups of  1974-77 model year light duty passenger vehicles were used.  The
Scope of Work called  for the two groups to match each other by vehicle  type
and  emission  failure  type.   Vehicle type  includes  model year,  make,  engine
size,  fuel  system  (1, 2,  or  4  venturi  carburetor  or  fuel  injection)  and
transmission  type.  Emission failure type means  the failure of  HC,  CO or  both
HC and CO, on the  state idle  emissions test.   The  first  group of vehicles was
sent to  the  facilities for repairs  before training had  occurred,  the  second
group after training.   Each facility was sent the same types  of vehicles after
training as had been sent before training.

The pre-training group was selected in order to be representative of a failing
fleet  of  vehicles.  For  the 1975-77  model years  this  was  a  simple matter.
Another EPA study  in Portland  had  already obtained a  representative sample of
failing vehicles  which were  used  for  this  training  study.   Another  approach
was  needed  for  the   1974  model  year  cars.   First,  a  group  of   vehicles
representing  high  sales  was  chosen.  From  this group were  recruited  enough
willing owners  of  vehicles such that  the assumed  failure rate  of  50 percent
for  that model year would yield the required  number  of  failing vehicles, and
they would be representative  of a failing  fleet.   The  post-training  group of
vehicles was recruited in order to match  the first  group.

The  pre-training  group  consisted  of  43  vehicles.   Due  to  several  factors
including four  facilities  dropping  out  of the training program and miscellan-
eous  vehicle  problems, only  21 were  used in  the final  match-up.    Matching
these  vehicles  exactly,  in  order   to  produce  the  post-training  group,  was
somewhat  difficult.    In  total,  55  vehicles  were recruited for  the  second
group.  This  number was  large mainly for two  reasons.  First,  EPA planned on
there being 31  vehicle pairs and was recruiting  vehicles  accordingly.   Second,
matching  exactly  for  emission  failure  type  caused  more  vehicles  to be
recruited than could be used in the  study.

-------
                                       8

One of the 21 pairs used was not an exact match for emission failure type, but
was considered  acceptable.   In  that one case-the  pre-training vehicle failed
the state idle test for CO only, and the postftraining vehicle failed for both
HC and CO.   The HC level was not  extremely high and  no  misfire was present,
however,  so that maintenance to reduce the  CO  level was all that was required
in  order to  also  reduce  the  HC   level,   the  same  type  of  maintenance the
pre-training vehicle  needed.   Nine of  the  21 pairs did  not match  for model
year, but matched in everything  else.

5.2 Vehicle Testing

Owners brought  their vehicles to the EPA contractor test  laboratory, completed
a  questionnaire,  and  had no further involvement.   Contractor  personnel  drove
the vehicles to the state  test  lane for inspection.   If  a vehicle failed the
state inspection  test  (SIT)  for emissions,  it entered the  program.  Otherwise
it was returned to the owner.

A series of tests was performed on each study vehicle under a minimum of  three
conditions:  as  received, after maintenance by the commercial  repair facility
and  after restorative  maintenance (RM)  by the  contractor.   Each  series  of
tests consisted of the following:

    0 Diagnostic Inspection
    0 Federal Test Procedure
    0 50 MPH Cruise Test
    0 Highway Fuel Economy Test
    0 Four-Mode Idle
    0 Loaded Two-Mode
    0 State Idle Test at Laboratory

Separate  from  the purpose  of this  study,  tire pressures   and  spark timing  at
curb idle conditions were recorded  during one or more of the series  of tests.

The purpose of  the  RM  sequence  was  to  return the vehicles  to  their specifica-
tions concerning  tune-up parameters and emission  controls, within reasonable
limits.  The sequence  was  designed to  follow diagnosis and repair  information
that was  given in  the  training course,  so that the trained mechanics  in the
study should have been able to perform  them.   No extremely expensive  repairs
were  made,   but  repairs  were  more extensive than  mechanics  might   usually
perform.  For example, no major engine work was performed  even if diagnosed  as
probably being  needed,  however,  carburetor  repairs  and replacements were made
if diagnosed as being needed.

In some cases, more than one RM sequence was performed.  For the final  matched
list  this occurred  four times  in the  pre-training  sample and once  in the
post-training.  All of the  additional  RM sequences in the  pre-training sample
were performed  due  to  a concern that  carburetor  idle  mixture  adjustments may
not have  been  performed according  to the specifications of the  manufacturers.
There had  been a  misunderstanding  on  the  part  of the  contractor concerning
this adjustment, which was  easily  cleared  up.   The one instance of additional
repairs to a post-training vehicle  was  done because of  driveability  problems.

-------
The problem  was  not serious,  but  was noticeable,  and  was solved  by further
adjustment and repairs.  These additional repairs were acceptable to the study
design since repair procedures do take driveability  into consideration.

5.3 Repairs by Participating Facilities

Contractor personnel, rather than the owners, drove the vehicles to the repair
facilities.  This was  done  so that the  best  documentation of vehicle repairs
could be  achieved,  consistent instructions  to  mechanics   could  be  given,  and
laboratory scheduling would  be smoothest.   The  drivers acted as if they were
the owners,  simply  requesting  that  the cars be repaired in order  to  pass the
state emissions  test.   The state  inspection sheet  which gives  the failure
condition(s) was always  in  the  car.   No mention of  a price limit was made.
However, if asked,  the driver would instruct the mechanic  to perform the least
expensive  adjustments  or repairs  in order  to  simply make  the car  pass  the
emissions test.  Due to  the low  average repair costs seen for  I/M repairs in
other studies, it is assumed that  this low-cost instruction is  commonly given
by owners.

6.0 RESULTS

Average  emission  changes due to  repairs  before  and after  training  are  an
obvious highly important result,  but  is not  the  only important result.  Also
of importance in determining the  degree of success of the  training  is the type
and quality  of  repairs,  cost of  repairs,  fuel economy changes, the emission
reductions  versus  the  potential  reductions  of  the  vehicles,  individual
facility  performances, and  more.  The  motivation of  the  type   and amount of
repairs  must also  be  considered  when  determining  the  success  of  training.
What may appear as  lack of success by the training program may actually be due
to a  failure  to  change economic  motivations, rather  than  a  failure to impart
new skills.   For example,  if low  cost  is  always  the top  priority, repairs
which are diagnosed as needed to restore the vehicle's emission  control system
to full function,  but not necessary in order to pass  the state emissions test,
will not be done.

6.1 Emissions and Fuel Economy

Average  emissions   and  fuel  economies  are  presented for six  categories  in
Tables  1  and 2.   The emissions  are  from   the  Federal  Test  Procedure (FTP)
driving cycle and from the  idle  emissions  test (performed at the laboratory).
The fuel economies are from  the FTP and  the Highway Fuel  Economy Test  (HFET).
Table 1  is  for  the before-training  group  of  vehicles,  Table  2   is  for  the
after-training  group.   The  FTP  HC,  CO  and fuel  economy  levels  are  shown
graphically in Figures 1-3.   The percent and absolute changes in FTP emissions
and fuel economies are presented in Tables  3 and  4.   All  changes are from the
as-received condition.

-------
As-Received

After Repairs
by Facilities

After Repairs
by Laboratory
                                       10

                                    Table 1

                                Before Training
                       Emissions and Fuel Economy Levels
                                 (21 Vehicles)

                                              Idle Emissions**
                  Federal Test Procedure       (Using Garage-
                  Emissions (grams  per  mile)   Type Analyzer)
                  HC          CO       NOx    HC -(ppm)  CO (%)
2.94

1.77


1.73
47.4

22.7


16.9
2.57   252

2.39   107


2.53   106
3.02

0.38


0.20
                                               Fuel Economy***
                                               (miles per gallon)
                                                 FTP     HFET
15.15

15.16
20.95

20.59
15.82   21.69
Federal/State*
Standards
(Average for all
vehicles)
1.78
18.6
2.84   258
1.2
15
21
                                      Table 2

                                   After  Training
                         Emissions and Fuel Economy Levels
                                   (21 Vehicles)

                                              Idle  Emissions**
                  Federal Test Procedure       (Using Garage-
                  Emissions (grams per mile)   Type  Analyzer)
HC CO NOx HC (ppm) CO (%)
As-Received 3.33
After Repairs 1.92
by Facilities
After Repairs . 1.64
by Laboratory
45.1 2.93 307 4.2
20.2 2.93 64 0.15
13.9 2.65 60 0.07
                                               Fuel  Economy***
                                               (miles  per  gallon)
                                                 FTP    HFET
                                                                    14.91

                                                                    15.23
                                                          20.58

                                                          20.32
                                                                    15.41    21.03
* Federal emission standards are based on the 1975 FTP  - the 1974 Federal  emission
standards were converted in order to average  them with  the 1975-77 standards.   The
fuel economies  are  based on the new  car certification values;  fuel  economy  stan-
dards  for  a few  vehicles  were estimated.  "State  standards" refers  to  the  state
idle test cutpoints.
** From the second idle portion of  the  SIT taken  at the  contractor's laboratory.
*** The conclusion stated in Section  1.0 that training produced a 0.8% improvement
in fuel economy is based on changes in combined  FTP  (city) and HFET  (highway)  fuel
economy.

-------
                                 11
 u.o
 3.5
-3.0
a.
 2.5
DC

UJ
en
ac

cl.O




 0.5




 0.0
                      Figure 1


                    HTDROCflRBONS

       MEflN LEVELS  BEFORE  flND  flFTER TRfllNING
                    (21 Vehicles)
            2.9^
                            3.33
            fl/R  fl/M R/M



         BEFORE  TRAINING
                             fl/R  fl/M  R/M



                            flFTER TRfllNING
                                                 fl/R - flS RECEIVED


                                                 fl/M - flFTER MflINT,



                                                 R/M - RESTORflTIVE

-------
                               12
 60
 50
                      Figure 2

                 CflRBON  MONOXIDE

      HERN VALUES BEFORE RNO  RFTER TRfllNING
                     (21  Vehicles)
= 30

oc
IU
ti-

er, 20
ac
ec
cc
O

 10
           47.
22.7
    16.9
                           US. 1
           fl/R fl/M R/M

          BEFORE  TRAINING
             fl/R  fl/M R/M

             flFTER TRfllNING
fl/R - flS RECEIVED


fl/M - flFTER  MfllNT.


R/M - RESTORflTIVE

-------
                                 13
 21.0
                        Figure  3

                  CITY  FUEL ECONOMY

        HERN VflLUES  BEFORE  flND  flFTER TRfUNING
                       (21 Vehicles)
  18.0  .




  15.0  ;




£12.0
t-
u.
o 9.0
e_
   6.0
   3.0
   0.0
15.1 15.2
        15.8
                 14.
             fl/R fl/M  R/H      fl/R fl/M R/H


            BEFORE TRfllNING   flFTER  TRRINING
                                     fl/R - RS RECEIVED


                                     fl/M - flFTER MfllNT.


                                     R/M - RESTORflTIVE

-------
BEFORE TRAINING

    After Repairs
    by Facilities

    After Repairs
    by Laboratory

AFTER TRAINING

    After Repairs
    by Facilities

    After Repairs
    by Laboratory
BEFORE TRAINING

    After Repairs
    by Facilities

    After Repairs
    by Laboratory

AFTER TRAINING

    After Repairs
    by Facilities

    After Repairs
    by Laboratory
                                       14

                                    Table 3

                        Percent Changes In Emissions  and
                           Fuel Economy From Repairs*.
                                  (21 Vehicles)
                     Federal Test Procedure
                       HC
          CO
 NOx
 Fuel Economy**
 FTP       HFET
-40%    -52%
-41%
-2%
                 +0.1%
+4.4%
          -1.7%
+3.5%
-42%    -55%
-51%    -69%
-10%
                 +2.1%
+3.4
          -1.3%
+2.2%
                                     Table  4

                        Absolute  Changes  In Emissions  and
                           Fuel Economy From Repairs*
                                  (21 Vehicles)
                     Federal Test Procedure
                        (grams per mile)
                       HC
          CO
 NOx
-1.17   -24.7   -0.18
-1.21   -30.5   -0.04
Fuel Economy**
(miles per gallon)
 FTP       HFET
                 +0.01     -0.36


                 +0.67     +0.74
-1.41   -24.9   0.00
-1.69   -31.2   -0.28
                 +0.32     -0.26


                 +0.50     +0.45
* All changes are relative to the as-received levels.
** The conclusion stated in Section 1.0 that training produced a 0.8% improve-
ment in fuel economy is based on changes in combined FTP (city) and HFET
(highway) fuel economy.

-------
                                       15

Several statistical  tests  were  used to analyze  the  data in  Tables  1-4.   The
standard  two-sample  t-test  and  the  paired  t-test  are  the  two  tests most
appropriate (the latter pairs  the matching vehicles of  the  two groups  before
comparison  of  the  groups).    These  tests  all  indicate  that  there   is   no
significant difference (at even the 90% confidence level) between  the emission
and fuel economy levels of the  before and  after-training  groups at any  of  the
repair  stages.   In other  words,  the  two groups were  so similar  in  both  the
as-received condition  and  the  after maintenance condition,   whether  from  the
trained or the untrained facilities, that  the differences which appear  in  the
tables  could  easily  be due to  random variations.  The latter  statement holds
true whether  the comparison is made using the  after-maintenance levels,  or  the
changes due to maintenance (the ideas of Tables 2 and 4).

The as-received  emission  levels  compare  favorably  to samples  of vehicles  in
other test programs, thus  giving  credibility to the representativeness  of  the
vehicles  used in  the  training  study.   Comparisons  of   1975-1977  model year
vehicles  from this training  study  and  the  EPA Portland Study  (which  was  an
evaluation of the  Portland  I/M  program)  are given  in Table 5.   Both  the
aatched,  paired  sample of  34  vehicles  (17  vehicles  each  in  the before  and
after-training samples were 1975-77 model years), and all 75  vehicles from  the
mechanic training  evaluation which  are  1975-1977 model years are  shown.   (The
Portland Element III program  used only 1975-77  model  year  vehicles,  which is
the reason for selecting only these years  from the  training  study for  compar-
ison.)  The slightly lower HC and CO emissions of  the Portland Element  II  and
III vehicles  is expected,  since these  vehicles were  tested  when  they were
younger and had less mileage  than the vehicles in the Mechanic Training  study.

-------
                                       16

                                    Table 5
                                              /
                     As  Received Status of Mechanic Training
                 Evaluation Vehicles and  Portland Study Vehicles
                           1975-1977 Model Years Only

                             Federal Test Procedure
Study

Portland
Element II & III

Mechanic
Training
Evaluation (Matched)

Mechanic                75    44,931        3.13      44.0      2.67
Training
Evaluation (All)
N
207
34
Odometer
31,100
41,722
HC
2.69
3.17
CO
39.7
47.3
NOx
2.72
2.53

-------
                                       17

6.2 Cost and Amount of Repairs by Commercial Facilities

The average maintenance  costs  were  similar for the  two  fleets and  are  listed
in Table  6.   The types  and  amounts  of  repairs  for the  two  fleets were  also
very similar.   The repairs performed are listed in Table  7.
                                     Table  6

                      Average  Cost of Repairs  by Facilities

                        Before Training     $19.14
                        After Training      $19.81
                                     Table  7
                    Type and Amount of Repairs by Facilities
Item

Spark Plugs
Carburetor
Carburetor
Idle Speed
Timing
Air Filter
Oil
EGR System
Choke
Vacuum Hoses
Type of
Repair
Amount
Before
Training
Amount
After
Training
Amount By
Laboratory*
Replace         2
Adjust          20
Overhaul        1
Adjust          20
Adjust          2
Replace         0
Change          0
Repair/Replace  0
Repair/Adjust   0
Reroute/
Replace/
Repair          ?
1
20
1
20
3
1
1
0
0
5
17
4
18
5
6
0
8
2
* Average for the before and after training groups.

-------
                                       18

6*3 Quality of Repairs
                                              >
The contractor performed  various  inspections ,.of the  vehicles  after each  test
sequence.  These  inspections  gave  an evaluation of  the  maintenance primarily
in  the  areas  of  the carburetor  and  ignition  timing,  which  are  the  most
frequent maintenance items performed.

Three measurements were taken which reflect  the  quality  of maintenance  to the
carburetion system:  (1)  the  position of  the carburetor  idle mixture  screw(s)
relative  to  the  laboratory setting,  (2)  the maximum amount  of  engine  speed
increase  at  idle  due to  the  addition of  propane gas, called  "propane  gain",
and  (3)  the idle  emission  levels,   particularly  idle  CO.   For  commercial
repairs, each of these measurements can be compared to the  laboratory  measure-
ments  for determining maintenance quality.  (However, this method  assumes  that
the  laboratory adjusted  the  carburetors  perfectly  which may  not always  have
been the  case.)   The  carburetor idle  mixture screw positions were measured  in
terms  of the number of  1/4  turns  necessary  to  reach  the  seated position.
Propane gain correlates with the leanness of the air-fuel mixture;  the greater
the engine  speed increase with  propane added,  the  leaner  the  setting.   The
idle CO is important, but less so than  the two measurements  previously stated;
this is  because  nearly all  of the  vehicles have catalytic converters  which
substantially decrease HC and CO resulting in readings at the  tailpipe that  do
not reflect well  the emissions from the engine itself.  Results of these three
measurements, shown  in Tables 8-10 do not show any  trend  in the quality  of
carburetor adjustments after training.

Concerning Table 10,  training had  no  noticeable  effect on the distribution  of
idle CO  levels.   The  mean  levels  and  standard  deviations  are  very  similar,
when leaving out  the one  noted vehicle in Table 10.   Also,  17 of 21  vehicles
in  each  group  were  adjusted to  between  0.0  and  0.3%  CO,   confirming  the
similarity in distribution.

-------
                                       19

                                    Table 8

                      Evaluation of Carburetor Idle Mixture
                     Screw  Settings by Commercial Facilities
A. Number of 1/4 Turns Different from Contractor Setting*
(Positive numbers mean the commercial settings  were  richer  than  the  contractor
settings.)
          BEFORE TRAINING
          AFTER TRAINING
      Mean

      .45
      .24
        Standard
        Deviation

          1.7
          1.5
B. Absolute Value of Number of 1/4 Turns Different from Contractor Setting**
          BEFORE TRAINING
          AFTER TRAINING
      Mean

      0.9
      1.3
        Standard
        Deviation

          1.5
          1.4
C.  Frequency  of  Commercial  Settings  In  Directions  Relative  to  Contractor
Settings.
          BEFORE TRAINING
          AFTER TRAINING
Richer

  11
  11
Same

 15
 12
Leaner

   7
  10
* Example:  If  one  screw were  2 1/2 turns out (from the  seated  position) after
commerical  repair  and only  2  turns out after contractor repair,  the  number of
1/4 turns different would be +2.

**  Example: If the  number  of  1/4 turns different for  one screw  was  +2  and
another was -2 the simple  mean  would  be zero;  the absolute  mean would  be 2,
though, because only  the numbers are considered in computing  absolute values,
not the signs.  Using absolute values results in a higher mean for this reason.

-------
                                       20

                                    Table 9
                                             v
                 Evaluation of Carburetor  Idle Mixture  Settings
                        Using Propane Gain Measurements
              (Idle Speed  Increase  Due  to  Addition of Propane Gas)

                                                            Standard
                                                            Deviation
                                 Mean*         Mean         of  the
                                Increase     Difference    Differences
                                 (rpm)

   BEFORE TRAINING
     Commercial Facilities         144
     Contractor                    130            -14             81

   AFTER TRAINING
     Commercial Facilities         104
     Contractor                    112             8             82
* Note that although the mean increases from the commercial facilities and  the
contractor are very similar, the standard deviation is large, showing that  for
individual  vehicles  the  propane  gains  from  the  commercial  facilities were
often  quite different  from  the  contractor  setting.   The  differences were
positive one-half  of the time  and  negative the  other half, both  before  and
after training.
                                    Table 10
                              Idle CO Settings  (%}

                                                       Standard
                                      Mean             Deviation

     BEFORE TRAINING
       Commercial Facilities       0.38 (.18)*         0.92  (.29)*
       Contractor                     0.20                0.39

     AFTER TRAINING
       Commercial Facilities          0.15                0.24
       Contractor                     0.07                0.17
* Excluding one vehicle in the group which  had  a very high CO reading  results
in the  amounts shown  in  parentheses.   The  facility  apparently  adjusted  this
vehicle to pass  the  state I/M test, got the  certificate  of compliance at  the
state I/M  lane,  and  then readjusted  the vehicle  before  returning  it  to  the
"owner".  The  assumed  reason  was that the  vehicle  had poor driveability  when
adjusted to pass the  DEQ.  With  more  thorough repairs the  contractor achieved
acceptable driveability and low emissions, however.

-------
                                       21

The contractor  measured three types  of ignition  timing which  are useful  in
evaluating  the  quality of  repairs  to  this  system by  the  commercial  facili-
ties.   The three types  are  basic, mechanical, and  total  advance.   Basic timing
is  the  initial  setting,  usually made  at  idle.   Mechanical  (or  centrifugal)
advance  is  the  increase  in  advance  when  the  engine  is  operated at  higher
speed.  Total advance measures the first two  types  plus  engine  vacuum advance,
and is  also  measured  at  a moderately  high engine  speed.  Since the  basic
timing is  the only one  which affects  idle  emissions,  that  will  be  the  main
type of  timing  noted  here;  it also affects  fuel  economy significantly  and, is
an  important  measure  of  the  quality  of repairs  for that  reason.   Table  11
shows  the number  of  vehicles  with  basic  timing  retarded  before and  after
repair.


                                    Table 11

               Vehicles With Basic Timing Retarded  By  More Than  2°

                                     BEFORE TRAINING     AFTER TRAINING

           As-Received                     2                    1
           After Repair by
           Commercial Facilities           5                    1
In  the before  training group,  the  two  vehicles  in Table  11 with  retarded
timing  as-received were  unchanged  by  repair.   Therefore,  a  total  of  three
vehicles  were  retarded by  repair.   In  the  after  training   group,  the  one
vehicle with  retarded  timing as-received  was  corrected  by repair and one other
vehicle which had timing properly  set  as-received was  retarded by  the repair
facility.

Considering  both  mechanical  and  total  advance together,  two vehicles less  in
each  group were  retarded  (by  more  than  4°)   after  repair,  thus  some correct
repairs were  mada in each group.   It was not apparent  what repairs were made
to  the before-training  cars to  correct  the  mechanical and total retard, but  in
the after-training group it was  evident that  two cars had their vacuum advance
mechanisms fixed.

A fourth  measure  of  the quality of repairs can be made by looking at the type
of  repairs performed  for the  different types  of  failures.   A  previous concern
with  I/M  repairs  was that mechanics  did  not  know what  the  possible causes  of
HC  and   CO   failures  were and  would  perform   some  needless  repairs.   One
particular  problem  cited was   that  of  performing  repairs to  the  ignition
system,  such as  spark plugs  and wires,  distributor cap,  rotor,  etc.,  on  a
vehicle   which  failed  for CO   only.   The  repairs  just mentioned  would   be
unnecessary  because  they do not have an effect on CO.  Earlier EPA studies  of
the Portland I/M  program showed  that there was  a significant  amount  of the'se
unnecessary  repairs performed.

In  this  study almost no unnecessary  repairs  were noted.  In  the pre^training
group,  one  vehicle  which  failed  for  CO-only  received  a  timing  adjustment.
This  was  unnecessary,  but  is  a  reasonable  and often recommended adjustment  to
make  when  doing  carburetor  adjustments,  and   therefore   is   of  almost  no

-------
                                       22

concern.  No other adjustments or  repairs  to any part  of  the ignition system
were performed on any of the  other  vehicles  which failed for CO-only, showing
good  diagnosis  even  without  training.    Similar results  were  seen in  the
post-training sample.   Only one  unnecessary, repair  item  was  performed,  the
replacement  of  spark  plugs  to  a  vehicle  which  failed   for   CO-only.    The
majority of the  vehicles failed for both HC and CO and  therefore no  judgements
could be  made as  to the types  of  repairs performed  except  to  say  that  they
appeared to be correct.

The above refers to only the paired vehicles used in  the analysis.   Inspecting
the repairs  of  all other vehicles  in  the  study, 'there were  two instances  in
the   pre-training   sample   of  expensive  misdiagnoses,   but   none  in   the
post-training sample.  One pre-training vehicle was diagnosed by a mechanic as
needing a  new  timing chain and gears  with an estimated repair  cost of  $220.
This vehicle was sent  to another facility and  received a  carburetor overhaul
which  apparently  was  what  it  needed,  although this  cost  $111.   A second
vehicle was adjusted to pass  the idle  test,  but  the  mechanic thought it  might
have  a  bad catalytic converter.   The  laboratory found that  the real problem
was leaking  intake manifold  gaskets;  in  all  fairness, however,  this  was  a
difficult  repair  case  with  the  laboratory  reportedly spending  6  hours  in
repairs and  noting  that one  cylinder  had  low  power  which could  have been  a
valve problem.

One expensive repair that  the faciities often did not  diagnose as  necessary,
but which  the laboratory found  necessary,  was  carburetor overhauling.  Of all
98 cars in the  study,  25 received carburetor overhauling or  replacements.   Of
these 25, only 8 were diagnosed as  being needed  by  the  facilities (in 3  cases
outside of the matched,  paired  groups, EPA did not let them  do the work  even
though  the laboratory  later  agreed that  the  carburetor  needed  to be  over-
hauled).   Apparently, repair  facilities do  not  generally  diagnose  carburetor
overhauling unless they cannot get the vehicle to pass  the idle  test.  This is
expected with the assumption that owners usually  try  to obtain the lowest cost
repairs which will allow their vehicles to  pass  the test.

As expected, vehicles which need  carburetor  repairs,  but do  not receive  them,
still have high FTP emissions  after  repairs even  though  they  pass the  idle
test.  Table 12  shows the FTP and idle emission  levels  at  the different repair
stages of  the 17  vehicles  which needed carburetor work (as  determined by the
contractor), but did not receive work by the repair facilities.  Only 8 of the
17 were from the final matched samples (4 from each group in  A and B).  As can
be seen, FTP HC and CO remain quite high after repairs  by  the facilities.   One
vehicle  (not from  the  matched  pairs)  also had very  high  emissions   after
laboratory  repairs.   The  carburetor  was  overhauled,  but  the   repair  had  no
effect  on FTP HC  or CO  emissions.   The  contractor  should  have  had further
repairs  performed,   because  CO  emissions   remained   high  at the  high   speed
portion of the  state  idle  test,  which is an indication that  there is still  a
carburetor problem.  This is a diagnostic  check  that  was used in  other cases,
but  was  not  used  on  this  vehicle  after  the  carburetor   overhaul,   for
unexplained reasons.  Without this one vehicle the mean FTP  HC and  CO for the
"combined" group  (after laboratory  repairs) would have  been  similar to  the
mean  for all the other  vehicles after  repairs  by the laboratory (see Tables 1
and 2).

-------
                                      23

                                   Table  12
Emissions and Fuel Economy of Vehicles Needing
Carburetor Repairs But Not Receiving Them By The Commercial Facilities*
Idle Emissions**
Federal Test Procedure (Using Garage- Fuel Economy
Emissions (grams per mile) Type Analyzer) (miles per gallon)
BEFORE TRAINING (n=7)
As-Received
After Repairs
by Facilities
After Repairs
by Laboratory
AFTER TRAINING (n=10)
As -Received
After Repairs
by Facilities
After Repairs
by Laboratory
COMBINED (n=17)
As-Received
After Repairs
by Facilities
After Repairs
by Laboratory
HC

4.08
(3.88)
2.97
(2.44)
2.26
(1.75)

3.78
2.80
1.72

3.91
(3.83)
2.87
(2.71)
1.94
(1.73)
CO

72.1
(65.0)
57.5
(45.4)
32.7
(19.0)

63.5
46.4
17.5

67.0
(64.0)
51.0
(46.1)
23.8
(18.1)
NOx

2.37
(2.55)
2.11
(2.29)
2.46
(2.63)

2.37
2.29
3.22

2.37
(2.44)
2.22
(2.29)
2.91
(3.00)
HC (ppm) CO (%)

334
(366)
169
(162)
120
(118)

365
73
75

352
(365)
112
(106)
94
(92)

2.8
(3.1)
0.70
(0.72)
0.33
(0.24)

2.7
0.20
0.09

2.7
(2.8)
0.41
(0.40)
0.19
(0.15)
FTP

15.58
(16.46)
15.66
(16.89)
15.98
(16.98)

14.96
13.90
15.53

15.21
(15.49)
14.57
(14.88)
15.71
(16.04)
HFET

20.69
(21.52)
20.66
(22.05)
21.55
(22.62)

20.30
18.63
21.64

20.46
(20.74)
19.42
(19.78)
21.60
(21.99)
*  Emissions and fuel economies without  the one vehicle  that still had a carburetor
problem are shown in parentheses.

** From the second idle portion of  the SIT  taken at the contractor's laboratory.

-------
                                       24

A final measurement  which would reflect  the  quality of repairs  was a  drive-
ability evaluation at  each repair stage.   The  main problems  looked for  were
engine  surging,   stumbling,  backfiring,  stretchiness,  misfiring  and   run-on
("dieseling").  The number of vehicles having at least one problem  was  greater
before training than after training, as  shown  in Table 13.  The  reader  should
be aware of the fact that this evaluation was quite  subjective and  that  only a
trend can be  deduced  from the results,  i.e.,  that driveability  remained  good
after repairs  by  trained  mechanics,  whereas it was  worse  in some  cases  after
repairs by untrained mechanics.   The after repair frequencies of problems  are
not statistically different.
                                    Table 13

                  Number of Vehicles With Driveability Problems

                                         After Repairs     After Repairs
                         As-Received     by Facilities     by Laboratory

BEFORE TRAINING (n=21)        2                63

AFTER TRAINING (n=21)         2                1                 2

-------
                                       25

6.4 Repair Effectiveness by Facility Type

As was  mentioned  in Section  III,,  four types of  repair  facilities were  used:
gasoline-service  stations,  independent  repair  garages,   chain  stores,  and
facilities  advertising  I/M  repairs.   In  order  to  increase  the  number  of
vehicles  in  some  of  the categories for  statistical  comparison purposes,  the
facilities  in the  "advertising" category  will  be  put  into  the  appropriate
other three  sections.   This will particularly help the  "independent"  category
which formerly had only  two facilities and  three vehicles sent to  it,  because
three of  the four  facilities which advertised  were  independents  (the  fourth
advertiser  was a gas   station).   The  FTP  HC  and  CO  emissions,  and  fuel
economies are shown for  each  of  the categories at the different  repair  stages
in Table  14.   The two-sample t-test and  the  paired t-test  show that  there  is
no significant difference  between any  of  the  different  facility  types'  repairs
before  and  after training.   The   tests  were  performed for  both  the  after
maintenance levels and the changes  due to maintenance.
                                    Table 14
                Effectiveness of Different Repair Facility Types
                          BEFORE TRAINING
                         As     After     %
                      Received  Repair  Change
                               AFTER TRAINING
                             As      After     %
                           Received  Repair  Change
Gas Stations (n=7)
FTP HC
    CO
    mpg*

Independents (n=8)
FTP HC
    CO
    mpg*

Chain Stores (n=6)
FTP HC
    CO
    mpg*
 2.68
47.0
15.30
 3.17
44.1
14.65
 2.93
52.4
15.68
 1.65
24.5
15.35
 2.06
20.1
14.65
 1.51
24.1
15.67
-38%
-48%
0.3%
-35%
-54%
  0
-48%
-54%
  0
 2.85
35.7
15.72
 3.79
47.2
14.32
 3.26
53.3
14.83
 1.25
 9.38
15.88
 2.16
22.1
14.76
 2.38
30.2
15.14
-56%
-74%
+ 1.0%
-43%
-53%
+3.1%
-27%
-43%
+2.1%
* City  fuel  economy.  Highway  fuel  economy is not  shown  by facility type  in this
table  but  is shown  in Tables  1-4.   The  conclusion in Section  1,0  that  training
produced a 0.8%  improvement  in  fuel  economy is based on combined  city/highway fuel
economy.

-------
                                       26

7.0 DISCUSSION AND CONCLUSIONS:   POSSIBLE  REASONS FOR THE OBSERVATIONS
    IN PORTLAND AND GENERAL IMPLICATIONS FOR MECHANIC TRAINING BENEFITS

The  results  of  this  study  showed   small  and   statistically   insignificant
incremental benefits due to  training  in  the  important areas  of  emissions and
fuel economy.  This conflicts with the higher expectations  of  EPA,  as well  as
the  expectations  of  many  people  in  state  agencies,  industry,  and  the
commercial  repair  industry.   This   section  will attempt  to   resolve   this
apparent conflict.

Part A will  discuss  the  optimum results  of training,  the  actual results, and
the possible explanations for the differences between the expectations and the
results.   These  three   topics  will  cover  several  important   areas   in  the
training, e.g., diagnosis of the cause of  the  I/M failure.

Part  B   will  discuss the  implications  that  this study  has for   full-scale
mechanic training programs conducted  as  part  of an operating  I/M program and
the limitations of this  study in determining the benefits of training.

7.1 Possible expectations of training,  actual  results and possible
    explanations

    7.1.1 Maximum expectations of what the training should accomplish

    The  ideal  situation  following an  I/M mechanic training  program would  be
    that all diagnoses and repairs which affect emissions and fuel economy are
    completely  accurate,  no vehicles  are  left unrepaired  to such  an extent
    that they  are unable  to pass the I/M test,  and vehicle  driveability  is
    good.  The  only repairs which are performed  are  ones  which are  in  fact
    necessary  to  correct  an  emissions  or  fuel   economy  problem; wasteful
    replacement  of  parts   that   are  still  good  would  be  eliminated.   All
    repairable emissions problems would be  corrected,  not just those which are
    directly  causing the  I/M  test  failure.   The  entire  vehicle emission
    control  system  would   ideally  be  working  at  its  full  function after
    repair.   The  emission  levels  of  individual  repaired  vehicles  would
    therefore be  close to  the  Federal Test Procedure  certification  standards,
    and fuel economy would be optimal.

    Most I/M areas  will  have  cost  waivers.    In  areas which  do  allow  cost
    waivers,  the  number  of  waivers would  ideally  be   less   with   trained
    mechanics due to proper diagnoses and repairs resulting  in lower repair
    costs  (assuming  that  the  repairs  necessary  to  pass  the  I/M  test  are
    performed  first).   Also,   fewer   I/M-avoidance  measures  (registering   a
    vehicle in  a non-I/M area, fraud, cheating) would  occur,  contributing  to
    lower  fleet  emission  levels.  The  acceptable  driveability of vehicles
    after repair  would result  in  fewer vehicles  receiving  readjustments after
    repair, thus further contributing  to  low fleet  emission  levels.

-------
                                      27

    7.1.2 Actual results

    The   repair  facilities  appeared  to  perform   nearly   identical  proper
    diagnoses  both  before  and  after  training.   In  the pre-training  sample
    there were  only two cases  of mis-diagnoses and  none  in  the post-training
    sample;  however since problems which are  difficult  to diagnosis are rare,
    the  mechanics  were  not  tested  rigorously on  their  diagnosis  abilities.
    Repairs  were generally  correct,  but not  entirely accurate,  resulting  in
    similar  emissions and  fuel  economies  before  and  after  training.   Judging
    from  the   emissions  and   fuel  economies   achieved  by  the  contractor
    personnel,  significant improvement  could  be  made  in the  accuracy  and/or
    quantity of repairs by commercial  facilities.   The  commercial  facilities
    did  not  correct all  of  the  emissions problems that were in fact  repairable.

    No vehicles were  left  unrepaired  to  such an extent  that they would not
    pass the   I/M  retest,  either before  or  after  training.   However,  one
    instance of  a  facility  readjustment  occurred  in  the  before-training
    sample.   The readjustment was apparently  due  to poor driveability.   It is
    not  possible to  determine  if this would  have recurred  after  training  if
    the  same circumstances  had  appeared.

    7.1.3 Possible  explanations

    Mechanics  in the  Portland  area  had been  familiar with   I/M  repairs  for a
    few  years.   Through experience  on-the-job  they  seem to have  learned  to
    diagnose the common problems of  failing vehicles  well..   Unusual problems
    continue to give  them  difficulty,  however, which result  in some  incorrect
    diagnoses.

    Concerning  repair procedures, mechanics  learned  the correct  procedures
    (particularly carburetor adjustment procedures) well  in  the training, but
    there is no evidence in  the FTP  or other data  that  they took the time to
    do  them precisely at  the  repair facilities.  This  is  partly  due  to the
    mechanics  being  satisfied  with adjustments which enabled  the  vehicles to
    pass the I/M test with  a margin of  safety  and fairly good driveability.

    Although the training  course did try  to  teach mechanics  to repair  other
    emissions   problems  not  necessary to pass the I/M   test,   not  enough
    incentive   existed for  them  to  do  so.    The  overriding  incentives  from
    owners   seem  to be  low  cost  and acceptable  driveability.   Also,  due  to
    competition,  repair  shops want to have low average repair charges to  their
    customers  and hope to  have  a reputation for inexpensive repairs.

7.2 General  implications for mechanic training benefits and limitations of
    this study

    Little   Likelihood of  Complete  Repair -  In  order  to  get  the  vehicles'
    emission  control system  as close  as  possible  to a   fully  functional
    condition,  the  contractor's  laboratory personnel often spent more than one
    hour in diagnoses  and  repairs.    Because of  the substantial  extra  cost
    necessary  for these repairs, it is doubtful  that  a training  program would

-------
                                   28

ever  result  in commercial mechanics being  as  thorough.   A legal  require-
ment plus a  large enforcement effort might  be  successful  in mandating  very
thorough repairs of  all  repairable problems,  but  this  would  likely be  an
unacceptable option to all I/M areas.

Learning Curve - There may exist a  learning  curve  such that mechanics  will
have learned as much after a period of time  of on-the-job training as  they
would  have  with  a  formal  training course.   Consider  the   hypothetical
curves in Figure 4.  The  top curve  may represent the  knowledge  gained  when
formal  training is  given  to  mechanics  in  the  first  year  of  an  I/M
program.   That curve continues to  rise  due to continued  knowledge gained
through experience,  i.e.,  on-the-job type training.   The bottom curve may
be  the  learning  associated  with  on-the-job  training  (and  no  formal
training).   In this  hypothetical case,  the  difference  between  the curves
is  greatest  during  the first  year and  then decreases  steadily until the
third  or  fourth year  at  which  time  they  are  approximately  equal.   The
value  of   formal  training is  the  area   between  the  two curves   and  may
translate,  to  some  extent,  into  emissions  benefits  during  the first few
years.  Since  the  formal  training course  in  Portland  occurred after I/M
had been in  effect for about four  years,  the value of early  training would
not have been  detected.

The  implication is  that  in any  I/M program,  in the  early  years formal
training will  (1) reduce  emissions  and repair  costs,  (2)  reduce the number
of  cars which  are  not  successfully repaired  at  first  attempt  and  must
therefore  get  re-repaired (the ping-pong problem) or obtain waivers, and
(3)  prevent cases   of  poor  driveability  after  repairs  with  subsequent
readjustments.  The  results of  (2)  and (3)  would  lead to further emissions
benefits during the early years.

The  effect of  training  in the later  years of an I/M program  appears  to
depend on  the  stringency of the cutpoints used in the inspections.

Effect of  Stringent  Cutpoints - In mature I/M  programs with  cutpoints like
Portland's,  there   appears  to  be  no  emissions,  cost  or  driveability
benefits from  training.   The mechanics may  perform slightly more  accurate
repairs,  but  if  so,  the difference  in terms  of emission  levels  after
repair appears to be minimal.

Effect of  Lenient  Cutpoints  -  In I/M programs  with more  lenient cutpoints
than in Portland the implications are less clear.

One  possibility is  that  the  learning  curve  mentioned  above   will occur,
with  associated temporary benefits  from  formal training.   Other than that,
mechanics  may  not  perform any  better  than  they  would  have  after a few
years of on-the-job  training, and  at  that point in time,  trained mechanics
will  not do  any better  than  untrained  mechanics.   Mechanics may not adjust
cars  to a  point  far below  the  cutpoints,   but will  stop at  a margin  of
safety point with which they are comfortable.

-------
                                   29

Another  possibility is  that mechanics  can  be  convinced  to  adjust  the
vehicles to far below the cutpoints,  to  about the same levels as  Portland
mechanics  achieve  (with the  same  FTP  emissions).   No  extra  time   or
equipment  is  needed,  making this  idea plausible.   Evidence from  Portland
shows it  is  easily possible and can be  done with  low  repair costs.   If
this  possibility  occurs, there  will be  a permanent  benefit  from  formal
training.

EPA  believes  that  the  second  possibility  is likely  to occur.   However,
only a training experiment in an I/M area with cutpoints much higher than
in Portland would verify this.

-------
                            30
               EMISSION REPfUR  KNOWLEDGE VERSUS  TIME

                   (Conceptual Diagram)	
UJ
o
o
UJ
o
z
ae
                                            UNTRflJNED
                                            TRfllNED
                            TIME  (YEflRS)

-------
                                       31

8.0 MECHANIC TRAINING EMISSION REDUCTION BENEFITS  CONTAINED  IN MOBILE2

Based  on this  study of  mechanic training,  I/M emission  reduction  benefits
associated with  mechanic  training were derived and  are  available in  MOBILE2..
As would be expected  from  the  above  analyses,  the incremental benefits due  to
mechanic training  are substantially  smaller than those  appearing in  MOBILE1.
Generally,  the   assumption  used   in  the derivation  of MOBILE2  was  that  the
after-maintenance  mean  FTP  levels  are approximately  equal to  those  seen  in
this mechanic training study if Portland short test cutpoints are  assumed.   If
the  cutpoints  differ  from  Portland's,  an adjustment  is  made.   For  a  full
explanation  of  the  derivation,   see  the  documentation  for   derivation  of
pre-1981 I/M credits[3].

Mechanic training  only  effects I/M benefits  for  pre-1981  model  year cars  in
MOBILE2.  In brief,  it  is felt that  1981  and  later model year  cars will  tend
to  experience  component  failure  instead   of  the  maladjustments  typical  of
pre-1981 model   year  cars.   Thus,  a  car  of  advanced  technology experiences
quantun  instead  of continuous emission  reductions.   For more  detail see  the
documentation for derivation of 1981  and later  I/M credits[4].

Table 15 gives examples of  the incremental  fifth-year benefits due to  mechanic
training  as  seen  in MOBILE2.   These examples  show  typical  results  of  0-2
percent  HC and  0-6  percent  CO  reductions   associated  with  emissions  from
pre-1981 model year  cars.   After averaging these effects with  the portion  of
the  fleet experiencing no effect  from mechanic training, incremental  benefits
due  to mechanic  training  are typically 0-1  percent  for  HC and 0-4 percent  for
CO.
                                    Table 15

                                    MOBILE2
                  Incremental Benefits Due to Mechanic Training
                                Start  I/M 1/1/83
                                Evaluate 1/1/88
                 Pre-1981              Pre-1981             Total
                 Stringency            Fleet*              Fleet**
                                       HC    CO             HC   CO

                     20                05              03
                     30                2614
                     40                26              14
*  Incremental  benefits in  percent  reduction to pre-1981  model year  emission
factors.

** Incremental  benefits  in percent reduction to  total  fleet emission  factors
assuming a 50% Identification Rate for 1981  and  later model year cars.

-------
                                       32

References
1.  National  Academy  of  Sciences,  Report by  the  Committee  on Motor  Vehicle
    Emissions, February 12, 1973, pp 74-75.

2.  Automotive News, December 19, 1978, pp 18-19.

3.  "Derivation  of I/M  Benefits  for Pre-1981  Light  Duty  Vehicles  for  Low
    Altitude,  Non-California  Areas",  EPA-AA-IMS/81-4  (In  preparation at  the
    time of this report).

4.  "Derivation  of I/M  Benefits  for  Post-1980  Light  Duty  Vehicles  for  Low
    Altitude, Non-California Areas", EPA-AA-IMS/80-8, January 1981.

-------
        33
     APPENDIX




Vehicle Test Fleet

-------
                                                      Vehicle List
                                Pre-Training  Vehicles  Used  For  The  Final  Matched  Analysis
            Vehicle #
Make
CID
CYL
CARS     TRANS.     AIR
ODOMETER
I/M FAIL
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
60506
40602
42001
61057
42002
61002
61006
62002
62102
60151
61000
41501
71015
71105
52057
51016
52018
52019
72111
72005
71120
Chev
Olds
Ford
Ford
Ford
Chev
Chrys
Ford
Chrys
Chrys
Chev
Chrys
Pont
Ford
Chrys
Chev
Chev
Chrys
Chrys
Chrys
Ford
250
350
302
250
302
305
318
140
318
318
305
225
301
140
225
350
140
225
318
225
250
6.
8.
8.
6.
8.
8.
8.
4.
8.
8.
8.
6.
8.
4.
6.
8.
4.
6.
8.
6.
6.
1-V
4-V
2-V
1-V
2-V
2-V
2-V
2-V
2-V
2-V
2-V
1-V
2-V
2-V
1-V
2-V
2-V
1-V
2-V
1-V
2-V
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
MANUAL
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
MANUAL
AUTO
AUTO
AUTO
AUTO
NO
NO
NO
YES
NO
NO
NO
YES
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
YES
YES
NO
34615
42503
74020
25541
75391
47364
30329
75065
51893
25304
93687
63158
41184
38866
42182
51366
49444
75571
33306
24192
20676
HC,
CO
CO
HC,
CO
CO
HC,
HC,
HC,
CO
HC,
CO
HC,
CO
HC,
HC,
... -- JIG,
"HC,
CO
HC,
HC,
CO


CO


CO
CO
CO

CO

CO

CO
CO
CO
CO

CO
CO




>
X)
•a
0>
3
a
X













-------
                                                      Vehicle List
                               Post-Training Vehicles Used For The Final Matched Analysis
            Vehicle #
Make
CID
CYL
CARB     TRANS.
                                                      ATR?
                                                                                             ODOMETER
                                                                    I/M FAIL
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
52001
40605
42003
60568
42005
60416
62201
70398
60302
60507
76012
41503
70262
61101
60584
60308
60385
60388
51055
70287
70508
Chev
Olds
Ford
Ford
Ford
Chev
Chrys
Ford
Chrys
Chrys
Chev
Chrys
Pont
Ford
Chrys
Chev
Chev
Chrys
Chrys
Chrys
Ford
250
350
302
250
302
305
318
140
318
318
305
225
301
140
225
350
140
225
318
225
250
6.
8.
8.
6.
8.
8.
8.
4.
8.
8.
8.
6.
8.
4.
6.
8.
4.
6.
8.
6.
6.
1-V
4-V
2-V
1-V
2-V
2-V
2-V
2-V
2-V
2-V
2-V
2-V
2-V
2-V
1-V
2-V
2-V
2-V
2-V
2-V
1-V
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
MANUAL
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
AUTO
NO
NO
NO
YES
NO
NO
NO
YES
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
YES
YES
58393
56942
32763
39789
95156
46938
26254
58393
37182
44303
36401
35147
42703
23668
32107
29508
61747
30424
27652
45990
16517
HC,
CO
CO
HC,
CO
HC,
HC,
HC,
HC,
CO
HC,
CO
HC,
CO
HC,
HC,
HC,
HC,
CO
HC,
HC,
CO


CO

CO
CO
CO
CO

CO

CO

CO
CO
CO
CO

CO
CO





•§•
*o
m
«
a.
H»
X













-------