EPA-AA-IMS-82-1
Public Opinion Polls for Inspection and Maintenance Programs:
Some Technical Considerations
January 15, 1982
Rocco DePietro, Ph.D.
Inspection and Maintenance Staff
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Ann Arbor, Michigan
-------
Table of Contents
Page
I. Introduction 3
II. Purpose of I/M Public Opinion Polls 3
III. Planning I/M Public Opinion Surveys 4
A. Timeliness of Surveys 4
B. Types of Survey Methodologies 5
1. Face-to-Face 5
2. Telephone Survey 6
3. Mail Survey 6
C. Design of Survey Questionnaires 6
1. Reviewing Question Wording 8
2. Some Measurement Issues 10
3. Pretesting the Questionnaire 10
4. Sample I/M Questions 11
D. Sample Selection and Size 11
E. Accuracy of Survey Data 12
IV. Analysis, Interpretation and Presentation of Survey Data 13
A. Data Analysis 13
B. Interpreting the Data 13
C. Presentation of Data 14
V. Costs for I/M Public Opinion Polls 15
VI. Determining Who Should Conduct a Poll 17
A. Staff Experience and Expertise 17
B. Methodological Considerations 17
C. Costs 18
VII. Summary 19
References 20
Appendices
Appendix I Determination of Sample Size 21
Appendix II Accuracy of Survey Data 22
Tables
Table 1 - Uses of Polling Data 23
Table 2 - Ratings of Survey Methodologies 24-26
Table 3 - Deciding on Question Format 27
Table 4 - Sample I/M Survey Questions 28-30
Table 5 - Sampling Tolerances 31
Table 6 - Sampling Errors 32
Table 7 - Inspection Attitudes by Background Factors 33
-------
I. Introduction
Since the passage of the 1977 Amendments to the Clean Air Act, several state
and local governments have sponsored or have considered sponsoring opinion
polls to determine public attitudes and knowledge about air quality issues and
programs. In some cases, they have conducted these polls themselves; in other
cases, they have sought the assistance of professional polling firms. The
quality of these surveys varys widely and affects their potential usefulness
for air planners and policymakers alike. The purpose of this report is to
provide state and local governments with information that could be used in the
planning and implementation of public opinion polls, especially those with
respect to inspection and maintenance (I/M) programs.*
II. Purposes of I/M Public Opinion Polls
I/M public opinion polls can be undertaken to determine the public's: (1)
perceptions of air pollution including its seriousness, sources, linkage to
cars and trucks, and proposed solutions; (2) awareness and knowledge of the
concepts of emission testing and repair to reduce auto pollution and help
clean the air; (3) attitudes about I/M including acceptance of specific
program elements and feelings about government regulation in this area; and
(4) behaviors such as previous participation in voluntary I/M testing programs
and care of the car's emission equipment and engine. It is useful to consider
taking I/M public opinion polls during the three phases of an I/M program: the
developmental phase, the startup phase, and the operational phase. For
information regarding surveys that could be taken during these phases see
Table 1.
Data from I/M public opinion polls can be used for program planning, feedback,
and evaluation. Polling data for program planning include attitudinal,
knowledge and behavioral data about motorists as well as input from the public
in terms of interests, preferences, and suggestions for improving program
effectiveness. Polls provide a way for the public to participate in actual
program decision-making. Examples of how polling data can be used for program
planning include:
- data on I/M awareness and knowledge levels could be used to assess the
informational needs of specific target audiences; and
- data on perceptions about the causes of air pollution, emission testing
and maintenance of motor vehicles can be used to identify any
misconceptions about these topics.
*Polls refer to a survey research methodology that employs a standardized
measurement instrument, or questionnaire, that is administered by phone,
face-to-face or mail. The terms polls and surveys are used interchangeably
throughout this report.
-------
Polling data can also be used for program evaluation. Program managers want
to know how effective their public awareness programs are in informing and
educating the public about I/M. They also want to know whether the program
changes or modifications which they make help to improve program
effectiveness. Examples of how polling data can be used for program
evaluation include:
- data showing the number persons who are reached by the public awareness
program, e.g., the percentage of persons who actually read, saw or
heard I/M messages in the mass media;
- data showing new information that people acquire from exposure to a
public awareness program; and
- data collected at two points in time (e.g., two separate surveys before
and after a public awareness program) showing changes in awareness and
knowledge levels.
Finally, polls can also be used for program feedback. This process enables
the public and other groups to respond to proposed or actual program changes.
Such information can be used to make program modifications or adjustments.
Examples of how polling data can be used for program feedback include:
- data showing public approval or disapproval of a proposed I/M program
change, such as performing tire pressure checks in addition to emission
testing;
- data showing public reactions to an actual program change such as the
use of computerized emission analyzers with a printout of testing
results; and
- data showing any problems encountered by motorists or mechanics during
the early implementation of the program.
III. Planning I/M Public Opinion Surveys
A. Timeliness of Surveys
The decision when to take a poll is almo.st as critical as the decision whether
to take one. To be useful, polling data must be timely. For example, polling
data for planning the overall public awareness program should be taken at
least three months prior to program startup. If polling data is to be used
for evaluating the impact of a public awareness campaign, it should be
gathered the month immediately before and after the campaign. (See Table 1
for polls that could be taken during different phases of an I/M program.)
Some state or local planners may wish to "update" an old poll or wonder
whether data from a poll taken 4-5 years ago is still valid today. If
populations remained stable demographically and had little exposure to
external influences, then polling data would probably remain valid
indefinitely. For the most part, however, the demographic composition of
-------
populations is continuously changing and large populations are constantly
being affected by powerful outside influences. Change in population attitudes
and behaviors is a complicated phenomenon. There is no one-to-one
correspondence between exposure to external influences and attitudinal or
behavioral change. External events are simply not that deterministic. Often
attitudes and behaviors change slowly over time. For this reason, it is not
too surprising to find some polling data relatively stable over a five-year
period. On the other hand, it is equally possible to find some dramatic
shifts in attitudes for the same time period, especially if several factors
have been at work in the environment to produce these changes.
Several factors can influence the stability of public opinion data over time.
They are: changes in the general state of economy which may affect public
feelings about the relative cost and benefits of I/M; changes in the political
climate and mass media coverage and treatment of air quality and I/M related
issues which will affect awareness of these issues and feelings about them.
Depending on external events, a one-year old poll may be invalid while one
five-years old might still be an accurate indicator of how people feel about
certain issues. There is no formula for determining when a public opinion
poll is out of date. Specific knowledge of the local situation and good
judgment must be used to make case by case decisions.
B. Types of Survey Methodologies
The three types of survey methodologies are commonly used for public opinion
surveys. They are: face-to-face surveys, where an interviewer who is
physically present conducts the interview; mail surveys, where the
questionnaire is sent in the mail and completed and returned by a respondent;
and telephone surveys, where interviews are conducted over the phone by an
interviewer. Most major I/M public opinion surveys to date have employed
telephone interviews. Dillman (1978) has rated the three survey methodologies
for selected performance characteristics (see Table 2). Telephone interviews
are rated higher than the others in avoiding interference by others in the
interviewing process and in speed of implementation. They are rated lower
than the others in not allowing for complex types of questions.
1. Face-to-Face Surveys
Where the goal is to obtain a sample that is representative of a large
population, such as a city or state, face-to-face surveys can be quite
expensive due to the high cost of identifying and contacting respondents.
Where the goal is to obtain a representative sample from a relatively small
and easily contacted population, such as I/M program managers or motorists
that pass through an inspection lane, face-to-face surveys can be undertaken
quite inexpensively. More complicated and in-depth questions can be asked in
face-to-face surveys since the interivewer is present to give complete
instructions. In general, these surveys are better than the others in holding
respondents' interest and attention for a long period of time, i.e., 30-45
minutes.
-------
2. Telephone Surveys
The major reason telephone surveys are used more often than face-to-face
ones is their relative cost advantage. Compared to other methodologies,
telephone interviewing is an "easy" method to administer. A representative
sample of the population can be identified through the random generation of
telephone numbers, i.e., random digit dialing. Interviewers also work from
central headquarters where their work can be closely supervised. This helps
to eliminate the effects that individual interviewers could have on the
interviewing process. The questionnaire does not have to look "professional,
since respondents never see it. All these factors contribute to the low cost
of telephone surveys. Telephone surveys are recommended for major I/M surveys
where the goal is to generalize survey findings to a large population. Their
only drawbacks are: their length has to be kept short (15-20 minutes) to keep
respondents from hanging up the phone; their questions, formats and response
categories have to be kept simple to enable persons to comprehend them.
Studies also show that respondents tend to give less complete answers to
open-ended questions on telephone surveys than on other survey methodologies
(see discussion of question formats in section III C).
3. Mail Surveys
Mail surveys have considerable potential for use by I/M program planners.
Though not widely used, their convenience and low cost make them an attractive
option to other methods in some instances. Also,new techniques in surveying
respondents by mail (Dillman, 1978) have been developed that obtain response
rates comparable to telephone and face-to-face methods. In general, though,
mail surveys seem to work best with well-educated respondents who have the
reading and writing skills to complete them without difficulty. They are also
easier to administer to select groups than to the general population. For
example, they seem to work quite well with professional groups that have a
governing body and an up-to-date mailing list. This makes contacting and
recontacting members of these groups fairly simple. Group members are also
more likely to complete a mail questionnaire if they know that their
organization is supporting it. However, some mail surveys can also be sent to
members of the general population such as licensed motorists. A random sample
of this group could be reached through applications for license renewals. To
increase the response rate to mail surveys, it is important to make them
attractive, simply worded, short, and logical. They should also have
explicitly worded instructions, easy to follow formats, and convenient return
procedures. They should take no longer than 15-20 minutes to complete.
C. Design of Survey Questionnaires
Information in this section is designed to assist persons faced with the tasks
of either developing a questionnaire or working with a firm or group that has
been hired to construct one. Comments pertain to all three types of survey
questionnaires: face-to-face, telephone, and mail. Obviously, each survey
type will also have some special requirements. To learn more about these
requirements, we recommend that you consult Dillman (1978) for telephone and
-------
mail surveys and Babbie (1973) for face-to-face ones. Below are a few
important steps to follow in constructing a survey questionnaire.
Step 1 - Make a list of all the topics for which you would like to obtain
information. Make sure the list is as complete as possible. It can be
pared later on.
Step 2 - Decide whether any of these topics can be conveniently
combined. This will help to shorten the list.
Step 3 - Rank the topics in terms of their relative importance,
select topics to be included in the questionnaire.
and
Step 4 - Under each topic, outline a list of all questions you want
answered. At this point, be concerned only with question content and not
the format or specific wording.
Step 5 - Arrange topics in a logical order for inclusion in the
questionnaire.
Step 6 - Select items to be included in the questionnaire based on the
priority of information and constraints on questionnaire length, i.e.,
time and funding constraints. A good rule to apply when deciding on
questions to include is: "Collect only data that can be used."
Step 7 - Decide on the question format. Rewrite questions in these
formats. You have options between a range of open and closed-ended type
questions (see Table 3 for a discussion of five such formats).
Step 8 - Review question wording (see section 3A below).
Step 9 - Order the questions in the questionnaire keeping in mind the
following principles. Begin with something easy and important. Include
more important questions ahead of less important ones so that respondents
do not lose interest in the interview process or get bored with it.
Within topic or content area, group questions by formats for easy
processing by respondents. Build a sense of continuity or flow to the
questionnaire by taking advantage of linkages that respondents are likely
to perceive among groups of questions. Position questions that are
likely to be objectionable after ones that are likely to be
unobjectionable so that respondents do not terminate the interview.
Step 10 - Prepare appropriate introduction, transitions and close for the
questionnaire.
An important consideration in the construction of an I/M survey is how a topic
such as I/M is introduced to the respondent. If I/M is not a widely known or
understood concept, then it is necessary to provide enough detail about it so
people are clear about what they are responding to. It is also important to
place I/M in the context of other air pollution control strategies that people
-------
8
are more familiar with. This will provide an indication how people feel about
I/M relative to other strategies.
An I/M survey undertaken by the Highway Safety Research Institute of the
University of Michigan accomplished just that. The questionnaire starts out
by asking respondents if they ever have been bothered by air pollution. It
then asks whether they feel air pollution is a problem in Michigan. It then
mentions that the state is thinking about reducing air pollution by setting up
yearly inspections of Michigan motor vehicles to test their exhaust emissions
and to require owners to fix those engines which are causing too much air
pollution. In trying to decide if such a vehicle inspection program is
desirable, it cautions that one has to think about the value of reducing air
pollution against the total program costs in time, travel, inspection fees,
and repairs. Then it asks for a general opinion: "Do you favor or oppose a
required emissions inspection program in your county." This question is
followed up with one that specifically addresses some proposed parameters of
the program. Several other I/M public opinion surveys which EPA reviewed used
a similar approach.
1. Reviewing Question Format and Wording
At some point prior to conducting the actual I/M public opinion poll, the
agency will be asked to comment on a draft questionnaire. Questions will have
to be written in appropriate formats. Five such formats with examples are
found in Table 3. Basically, questions can be either open or closed-ended.
In open-ended questions, the respondent supplies the entire response. There
are no pre-established response categories such as "approve" or "disapprove".
In closed-ended questions, the researcher supplies the response categories
which are sometimes ordered, i.e., categories which indicate degree of
approval or disapproval.
During the editing of the draft questionnaire, question wording will also have
to be checked. Simple rules for writing questions abound such as those
recommended by Payne (1951) and Babbie (1973). Some common rules are: use
simple words; keep it short, be specific; do not talk down to respondents;
avoid bias; avoid objectionable questions; and avoid hypothetical questions.
Dillman (1978) suggests that a number of questions should be asked about each
item considered for inclusion in the questionnaire. The questions are:
Q: WILL THE WORDS BE UNIFORMLY UNDERSTOOD?
For example, the phrase "maintenance of motor vehicles" could be replaced by
"tune-up of cars."
(Solution: generally substitute simple for complex or difficult words.)
Q: DO THE QUESTIONS CONTAIN ABBREVIATIONS OR UNCONVENTIONAL PHRASES (OR
JARGON)?
(Solution: Generally avoid abbreviations, foreign phrases or slang unless they
refer to something that can be communicated more efficiently by the use of
such terms, e.g., tampering instead of removal or rendering inoperative of
emission control devices.)
-------
9
Q: ARE THE QUESTIONS TOO VAGUE?
(Solution: Replace vague words with more precise ones, e.g., replace regularly
with "number of times per week".)
Q: IS THE QUESTION TOO PRECISE?
(Solution: Revise the question to include categories that will aid recall of
specific information, e.g., instead of asking person for number of times in a
year their car has received a tune-up, include the categories such as none,
1-2, 3-4, five or more.)
Q: IS THE QUESTION BIASED?
Here's an example of the problem of establishing a behavioral expectation.
"Most people have tune-ups done on their cars at least once a year. How about
you? How often do you have your car tuned?" An improved question is: "How
often do you have your car tuned?" An example of the problem of unequal
comparison is a preference question where one category is made to appear more
attractive. "Where would you prefer to take your car for inspection? To your
local garage mechanic or to a centralized inspection station that has built-in
quality control procedures?" An example of bias from unbalanced categories
would be a questionnaire item with three categories representing favorable
feelings towards a topic and only two categories representing unfavorable
feelings towards it.
(Solution: Make sure the question does not; create a behavioral expectation,
make one response category more attractive than others or more likely to be
selected because of unbalanced categories, or contain affective terminology
that might predispose respondents to answer in one way or another.)
Q: IS THE QUESTION OBJECTIONABLE?
Sources of objection include requests for personal or potentially embarrassing
information. For example, it is objectionable to ask outright whether or not
a person has tampered with the emission equipment on his or her car. In such
a case, an indirect approach is preferable. "Do you happen to know whether
the emission control equipment on your car has ever been worked on for any
reason? (If yes, ask: "What was done to it as far as you know?") Similarly,
the question could be prefaced with contextual material to overcome
objections. For example, it could be prefaced with some background
information. "In some cars, it is possible to remove or adjust the emission
control equipment so it doesn't work. Are you aware that this can happen?
(If "Yes") Do you approve of this happening?
(Solution: Obtain personal or sensitive information by using indirect
questions, broad categories, background material, and series of questions to
overcome objections.)
0: IS THE QUESTION TOO .DEMANDING?
For example, avoid asking persons to estimate the percent of air pollution
caused by several major contributors. It's hard work, and some respondents
might not be up to the task.
(Solution: Redesign questions that are difficult to answer, e.g., questions
with more than 10 things to rank, questions requiring mathematical
calculations, or recall from the distant past.)
-------
10
Q: IS IT A DOUBLE QUESTION?
Some questions actually are two questions rolled into one. For example, it's
ambiguous to ask in a single question whether people approve or disapprove of
I/M and safety programs. They could favor one and not favor the other. Two
separate questions are required.
(Solution: Rewrite question so that a respondent replies to a single point.
Q: DOES THE QUESTION HAVE A DOUBLE NEGATIVE?
(Solution: Reword the question in the affirmative to eliminate confusing
language.)
Q: ARE THE ANSWER CHOICES MUTUALLY EXCLUSIVE?
(Solution: Make sure respondents cannot select more than one category and be
correct; if necessary, break the question into parts.)
Q: DOES THE QUESTION ASSUME TOO MUCH KNOWLEDGE?
Some I/M and related areas are highly technical. To ask public opinions about
air pollution regulations, pollution control devices, and airborne pollutants
presupposes knowledge of these topics on the part of respondents. In some
cases, this knowledge might not be present. Prior to asking opinions about
highly technical topics, it is important first to establish what, if anything,
people know about them, even if it is only self-reports of awareness or
knowledge.
(Solution: Explain a topic or ask respondents if they are familiar with a
topic before asking their attitudes about it.)
Q: IS THE QUESTION TECHNICALLY ACCURATE?
(Solution: During the editing of the questionnaire, identify and eliminate any
questions with technical inaccuracies.)
Q: IS AN APPROPRIATE TIME PERIOD PROVIDED?
(Solution: Specifying a precise time period is usually easier for respondents
than mental averaging, e.g., how many times in the past month or so as opposed
to an average number of times per year.
Q: CAM THE RESPONSES BE COMPARED WITH EXISTING INFORMATION?
(Solution: Use the same questions and response categories as in other surveys.)
Q: ARE THE QUESTIONS TOO CRYPTIC?
(Solution: Use complete sentences rather than leaving words out for
simplicity.)
2. Some Measurement Issues
A type of attitudinal scale commonly used in public opinion surveys is a
5-point Likert scale, named after Rensis Likert, the researcher who developed
it. Typically respondents are asked the extent to which they approve (or
favor) a particular proposal. On a 5-point scale, they could answer:
"Strongly approve, somewhat approve, neutral (or no opinion), somewhat
disapprove, or strongly disapprove." No more than a 5-point scale is
recommended for telephone interviews because of the difficulty that some
-------
11
respondents have in recalling accurately the response categories. The
advantage of the 5-point scale over the 3-point one (e.g., approve,
disapprove, and neutral) is that it obtains more information, i.e., the
strength of approval and disapproval. For later analysis, a 5-point scale can
be collapsed into three categories: approve, disapprove and neutral.
Very important is the treatment of the "neutral" or "unsure" category in
surveys. It should be included in the questionnaire to permit persons who do
not have knowledge or opinions about I/M to indicate this feeling. Not
including a "neutral" or "unsure" category makes it appear that these persons
have opinions when in fact they do not. This creates "pseudo data" which
should be avoided. A significant number of respondents answering "neutral" or
"unsure" for a particular question could suggest that the public awareness
effort has not reached a significant number of motorists.
3. Pretesting the Questionnaire
The questionnaire should be pretested to make sure that it is clear to
respondents and that it can be completed within a set time frame. Adding five
to 10 minutes to a survey could increase its cost by 10-20% and decrease
participation. The pretest could also pinpoint other problems with the
questionnaire such as awkward or difficult wording, illogical sequencing, and
boring formats. An important point to discuss with the research firm or group
preparing the survey is the purpose and amount of pretesting that is
proposed. Pretesting is sometimes done in small focused groups (8-12 persons)
where the objective is to have persons react to topics and specific
questions. This technique involves paying focused group members $15-$25 for
their participation in about a two-hour session. A more elaborate pretest
involves administration of a draft questionnaire to about 20-50 persons under
actual field conditions. This form of pretest, even with machine tabulated
results, should not add more than one week to the survey process. Depending
on which type of survey methodology is used, the cost of pretesting is about
$5-$25 per interview. Questionnaires should always be pretested. Some
pretesting, regardless of how little, is always better than no pretesting.
4. Sample I/M Questions
Sample I/M questions that have been used in several public opinion surveys to
date are included in Table 4. The questions attempt to determine whether I/M
as an air pollution control strategy is favored or opposed, whether I/M is
more or less acceptable than other air pollution control strategies, and
whether I/M is acceptable in view of its costs. Other items commonly covered
in a questionnaire are attitudes about the recommended parameters of the I/M
program such as the inclusion of buses and large trucks, state versus private
inspection stations, the maximum distance people want to travel to an
inspection station, the longest permissible time for someone to spend at the
inspection station waiting for inspection, the cost of inspections, the length
of time to give the owner of a failed vehicle to get the emissions system
working properly, the treatment of people who fail the emission test,
exemptions for older cars and for persons who cannot afford to pay, combining
-------
12
I/M with a safety or noise program, the I/M informational needs and
preferences of motorists, and adding features to an I/M program such tire
pressure checks which can improve fuel economy.
D. Sample Selection and Size
Where the goal is to generalize findings from a sample of respondents to a
large population, such as people of a city or state, it is necessary to use
scientific procedures to select the sample. These procedures help ensure that
each person included in the sample has a known probability of being selected.
(Different respondents may also purposely be chosen with unequal chances of
selection, but.,if scientific procedures are followed each respondent will have
a specific known probability of selection. For example, a researcher may wish
to obtain a large subsample of young males to study their emissions attitudes
and tampering behavior. In the analysis of the entire sample, this subsample
would be weighted to reflect its actual number in the total population.)
Following these procedures makes it possible to generalize sample findings to
the parent population from which the sample is drawn.
Sample selection involves a decision about who to survey, such as adults
18-65, car owners, or persons who have a valid driver's license. Deciding on
the sampling unit is a decision that should be made by the agency, not the
polling firm. Surveys to determine the attitudes or knowledge of the general
population should include all persons 18 or older currently residing in the
geographical area of interest, rather than only licensed drivers or car owners.
Larger samples increase survey precision up to a point, but they also cost
more. Obviously a balance must be struck between precision and cost. Larger
size samples also permit in-depth analysis of population subgroups that might
be of interest to program planners. These types of analyses would be useful
if program planners suspect considerable variation in knowledge, attitudes and
behaviors among these subgroups. The goal of these analyses might be to learn
more about the information needs and misconceptions of these subgroups so that
improved strategies for informing and educating them about I/M could be
developed. A sample size of about 800 is usually large enough to permit a
number 'of subgroups to be compared. As the size of subgroups increase,
difference between them (in percentages) does not have to be as great to reach
statistical significance (see Table 5).
An optimal sample size can be computed from formulas based on sampling theory
statistics. With knowledge of the standard error of a sampling distribution
of means (a measure of sampling error), and the desired confidence level and
interval, the sample size can be easily computed (see Appendix I).
E. Accuracy of Survey Data
Sample size affects the accuracy of survey data. In general, the smaller the
sample size, the larger the sampling error or degree of error that can be
expected from a given sample design. A statement of sampling error must
contain two essential components: the confidence level, which by convention is
usually set at 95%; and the confidence interval, which by convention is sought
-------
13
at about ^5%. Commonly, researchers will say that in 95 cases out of 100 the
results based on the entire sample differ by no more than 5% points from what
would have been obtained by interviewing all members of the population.
(Sampling errors do not take into account a margain of additional error
resulting from the various practical difficulties of taking a survey of public
opinion, e.g., a low response rate.)
The relationship between sample size and accuracy of survey data is complex.
It is valuable to increase sample size only up to a point to reduce sampling
error. Beyond that point, sampling error only decreases slightly. Doubling
the sample size does not reduce the sampling error by half, but only by the
inverse of the square root of two. For example, going from a sample size of
400 to 800 only reduces sampling error by 29% (see Table 6). It is necessary
to quadruple the sample size to halve the sampling error. Sampling error
imposes restrictions on interpreting the data, especially with regards to the
precision of results. (See Appendix II for a more complete explanation of how
sample size affects the accuracy of survey data.)
The size of the population from which samples are drawn, whether it be a state
of 12 million or a city of 600,000, only affects sampling error very
slightly. The reason is because sampling error is linked to sampling theory
which treats theoretical distributions of samples from an entire population
regardless of its size.
Finally, response rate can affect the accuracy of polling data. Response rate
is the percentage of all those who responded to the questionnaire from the
pool of those who were eligible. If only half of those who were eligible
responded, the response rate would be 50%. (See Appendix II for a more
complete discussion of how response rates affect the accuracy of polling data.)
In general, response rates are higher for face-to-face and telephone surveys
(70-85%) than for mail ones (50-80%). However, a mail survey that has good
follow-up procedures can achieve a response rate as high as 80%. Response
rates of 70% or above are considered acceptable. A rate lower than that can
introduce considerable error into estimates of sample characteristics.
IV. Analysis, Interpretation and Presentation of Survey Data
A. Data Analysis
To facilitate their interpretation, polling data is usually analyzed and
presented in two simple forms. One form of data presentation is marginals
which show the absolute numbers (Ns) and percentages for each response
category to every question on the questionnaire. For example, for a question
on an emissions inspection attitude the data presented might show that 41.0%
of a sample of 800 strongly favored I/M; 16.9% somewhat favored it; 12.5% felt
neutral about it; 20.9% somewhat opposed it; and 8.7% strongly opposed it.
Another way to present data is in tables with crosstabs, which show the
numbers (Ns) and percentages for each response category on key questions for
different background factors such as age, sex, geographical location, income,
and education. Sometimes opinions about emission testing would be examined
-------
14
"controlling" for other opinions, such as those about clean air and pollution
control. The tabular data would show the cases where there were significant
differences in I/M attitudes by background factors and general attitudes (see
Table 7).
More elaborate analyses of I/M data are possible and sometimes desirable, but
they are sometimes difficult for planners to interpret and can add
considerable costs to the survey. An example of a potentially useful anlaysis
is a demographic profile of persons who engage in tampering with their car's
emissions equipment or fuel switching. The objective of such an analysis
would be to attempt to determine characteristics of these persons
(personality and attitudinal) that would make it easier to design educational
messages and media specifically for them. Other analyses of motorists'
information needs or preferences could reveal groupings or clusters of needs
or preferences that could be addressed separately by leaflets or booklets.
For example, the survey could identify a number of consumer confidence issues
that could be addressed in a publication.
B. Interpreting the Data
Data do not speak for themselves. They have to be interpreted. In
interpreting the data from I/M public opinion polls, here are a few things to
consider.
1. The most frequent statistic used in the analysis of
polling data is the percentage. A typical analysis shows
the percentages of respondents in the overall sample or
subsample who favor or oppose some issue or who agree or
disagree with some proposal. In interpreting these data,
it is important to note the overall percentages for each
response category. For example, does the majority (more
than 50%) of respondents favor or oppose I/M? Does a
plurality favor or oppose it (e.g., less than 50% favor
but more favor than oppose)? What percentage is neutral
or unsure about it? (If more than 10% fall into the
"unsure" category, it is a fact worth noting.) Is the
difference in percentages between those favoring and
opposing I/M large or small? Does the difference reach
statistical significance? (See Table 5 below.) Do I/M
attitudes differ significantly depending on one's sex,
geographical location, age, education, political
thinking, perceptions of air quality, and car ownership?
2. In interpreting the data, it is useful to look for
sizable differences in percentages and patterns among
responses. Some differences will turn out larger than
expected. That's worth noting. Others will turn out
much smaller than expected. That's worth noting, too.
For example, one I/M poll showed only very small
differences in I/M attitudes when controlling for
political thinking ('e.g., for persons who classified
-------
15
themselves as liberal, conservative or
middle-of-the-road). It is also important to look for
patterns in the data such as the consistency of responses
to similar or related questions. The expectation is that
respondents with positive attitudes towards one issue or
proposal, should have positive attitudes towards other
similar or related issues or proposals. Response
consistency across several similar items adds support to
one's interpretation of the data.
3. As an aid to interpreting data, it is often useful to
compare findings with similar polls conducted elsewhere.
Ideally, these comparisons can be made with earlier
surveys that asked identical questions to a similar
population. Such comparisons can provide two useful
kinds of information. First, they can indicate whether
your findings are within the range of what has been found
elsewhere. If your findings are considerably different
from those obtained elsewhere, you should be alerted to
possible methodological problems in data collection,
e.g., biased questions or interviewer effects. Second,
they can indicate trends in the data. This is a
technique commonly used by national polling firms that
are able to show trends in public opinion on important
issues. The University of Louisville has the most
complete trend data on public opinions about I/M. Since
1978, it has conducted five air quality surveys in
Jefferson County. These five surveys consistently show
strong support for I/M (in the range of 61%-68%). The
reporting of this trend data adds considerable weight to
the conclusion of popular support for I/M in Jefferson
County.
C. Presentation of Data
Careful consideration should be given to the presentation of data to various
audiences. Each audience has its own information interests and needs. For
example, a writeup of results for legislators should include a convenient
summary of the main findings with their implications for public policy. A
writeup of results for the press and broadcast media should include a
convenient summary of main findings with emphasis on what is newsworthy and
the relationship between survey findings and important issues of the day. In
general, the main findings should be stated clearly, succinctly, and
highlighted where appropriate. Their implications should be discussed fully.
Audiences should not be overwhelmed by tables of statistics that are difficult
to understand. Charts and graphs that are easy to interpret should be used
whenever possible.
Often little care is taken in releasing data to various audiences. For state
and local I/M planners, the implications of the results for program
implementation should be stressed in a written report that could serve as an
-------
16
internal working document for them. Planners should have a more sophisticated
grasp of data so it is not a problem presenting them with a fully detailed
report as long as data implications are clearly spelled out. Political
leaders are interested more in public support for I/M and related issues.
Their report should be brief and attractively presented. Public opinions
about key program issues should be summarized. Special legislative briefings
could be used as a vehicle to present survey results. Groups including lung
association chapters, chambers of commerce, and service industry organizations
will all have different interest in such surveys. For example, members of the
service industry would be very interested in public sentiment regarding car
maintenance, tampering with emissions control devices, and any proposed
modifications of emissions testing procedures. The leaders of these groups
should receive a brief, attractively presented report that summarizes public
opinions about key I/M program issues. In cases where the group plans to
assist with the I/M educational effort, a workshop or seminar is a highly
appropriate format for presenting survey findings.
V. Costs for I/M Public Opinion Polls
Costs for an I/M public opinion poll are usually figured per completed
interview. These costs can range anywhere from $10-$25 for each completed
telephone interview, $25-$75 for face-to-face interviews, and $5-$15 for mail
interviews. The range in price per completed interview will vary depending on
two factors: (1) the total number of interviews completed, with cost
decreasing as this number rises; and (2) options included as part of the
survey. Below is a list of key items that are factored into the costs of such
public opinion polls. Some of these items are standard features of any
survey. Others, such as extent of pretesting, additional quality control
measures, and special analyses, are options that increase survey costs.
A. Questionnaire Construction
The two aspects of questionnaire construction that directly affect its cost
are length and question structure. Long-worded questions or ones with many
branches or response categories add to questionnaire length. Questions which
invite respondents to supply their own answers, i.e., open-ended questions,
also add length. Respondents vary in their quickness and completeness of
response to such questions. Interviewers also take extra time recording
responses to such questions since they must write down the entire answers.
Coding schemes for open-ended questions are usually elaborate and take
considerable time to develop. This adds to survey costs. Questionnaires can
be streamlined so that only the most important questions are asked. In
deciding which items to include or discard, ask the questions: "Is the data
from this item really important? How will it be used once it is collected?"
Applying the criteria of importance and ultimate use will often show that data
from some items have only marginal implications for the program.
-------
17
B. Choice of Interview Methodology
Mail questionnaires are less expensive to administer than telephone
interviews; and telephone interviews are less expensive to administer than
face-to-face interviews.
C. Number of Followups to Obtain a Completed Interview
For example, some telephone surveys include up to seven callbacks, at varying
times of the day and evening and on varying days of the week, before being
considered a non-contact. Non-contacts are combined with refusals in
computing response rate.
D. Pretesting the Questionnaire
Focused groups are used to test language, format, and design aspects of the
questionnaire. Surveys are sometimes administered to about 25-50 respondents
to check their length, test for any problems with the interview process, and
determine the response distribution to specific survey items.
E. Coding and Keypunching
This includes transferring the data from the questionnaire to codesheets and
punching data cards or entering data directly into a computer.
F. Data Management
This includes preparing the data for data analysis, i.e., making sure that the
data set does not have "wild codes" or inappropriate response categories.
Additional charges for computer time are made for the handling of complex data
bases such as those with large samples, many cards of data, and two waves of
data where samples have to be merged for some analyses.
G. Quality Control and Recordkeeping
This includes supervising the work of interviewers, coders, key punch
operators, computer programmers, and data analysts.
H. Data Analysis
Additional charges may be made for detailed cross-tabulations and multivariate
analyses of the data.
I. Reporting Results
Factors affecting costs are: number of reports; whether reports are typed,
photocopied, or printed; distribution of reports; and holding meetings to
discuss survey results.
-------
18
J. Travel
This includes costs for meetings between the survey firm or group and
contractor to discuss questionnaire development and survey results.
In negotiating with the research firm or group, it is important to identify
the costs of the items above to determine whether they are included in the
cost per interview. Knowing the cost of these options will help the agency to
obtain the best product at the most reasonable price.
VI. Determining Who Should Conduct a Poll
Public opinion polls can be done either by outside consultants, such as
professional polling firms and survey research groups at universities, or by
state and local agencies themselves. Outside consultants should be used for
major surveys where the objective is to generalize findings from a sample of
respondents to a large population such as people of a city or state. The use
of outside consultants adds considerable credibility to polling results. The
issue of polling data credibility is important when there is a good chance
findings will be closely scrutinized for their accuracy or public policy
implications. (Of the major public opinion surveys completed to date, all but
one was conducted by a polling firm or university group.)
Inhouse surveys can be undertaken by the personnel of state and local
agencies for program planning purposes, especially when there are fewer
restrictions on the quality of data to be collected. These are typically
small, specialized surveys of easily reachable population subgroups, such as
motorists at inspection lanes or garage mechanics. More often than not their
goal is to detect rough patterns in attitudes or behaviors in these subgroups
than to generalize findings to a large population. Issues of the statistical
accuracy of the data and the "representativeness" of samples are often less
important than issues of the timeliness and cost of data collection.
The decision to hire a polling firm requires careful consideration. The three
most important things to consider are: the experience and expertise of the
people who will conduct the poll; the methods to be used in conducting the
poll; and the cost for conducting it.
A. Staff Experience and Expertise
Experience counts. Experience with telephone methodology, state-wide surveys,
environmental issues are all valuable assests of a polling firm. Well-trained
staff also help to eliminate "interviewer bias" that can occur in some
surveys. The presence of skilled data analysts is also a big plus. The
recommendations of former clients can be helpful in assessing the experience
of the polling firm.
-------
19
B. Methodological Considerations
Methods include the choice of an appropriate survey methodology and the use of
quality control measures. If polling is to be conducted by phone (the
technique that has been used for most I/M public opinion polls to date), make
sure that the polling firm is using random digit dialing a proven method
that ensures that almost every person in the population has a known
probability of being selected for the sample or is sampled from some kind of
complete listing such as driver license records. The use of telephone
directories is inadequate in view of numerous unlisted numbers. (There is
usually a very small proportion of the population which completely lacks
telephone service, but this proportion is usually so small that it does not
substantially affect the representativeness of the overall results.) If a
mail survey is planned, make sure enough follow-up mailings will be made to
achieve a satisfactory response rate.
In cases where there is likely to be more than one eligible respondent per
household, it's important to have a sampling procedure to decide who gets
interviewed. Better polling firms have such procedures. Without them there
is a danger of biasing the sample by over-representing those persons, e.g.,
housewives and the elderly, who tend to be home more often than other
household members.
The use of quality control measures such as interviewer supervision, callback
procedures to contact persons identified through sampling techniques, and
checking the coding of questionnaires and keypunching of cards (data records),
help to distinguish between the good and mediocre polling firms. A good
question to ask representatives of a polling firm is: "What quality control
measures do you employ to help avoid biasing survey results?"
C. Costs
In many cases, the budget j^s the bottom line. While it might make sense in
some instances to select the more expensive polling firm because it offers
some special options, more often than not the agency with the lowest bid will
get the contract. All things being equal, there is usually no reason not to
select the lowest bidder. When things are unequal, such as when a firm offers
better quality controls than another or has more survey experience, one has to
estimate the value of these differences in deciding which firm to choose. In
general, the lower bidder should not be selected when there is reason to
suspect any insurmountable problems with their staff expertise or training,
their application of survey methodology, and their quality control measures.
Some administrative considerations also figure into the selection of a polling
firm. Here are some questions that should be answered prior to actual
selection. Is the firm conveniently located so that meetings with their
representatives can be held as needed, i.e., at least once to review the draft
questionnaire and another time to discuss project results? (If the firm is
located in another part of the country, then provisions for travel will have
to be built into the budget.) Will the firm seek any input in designing the
questionnaire? What form will this input take? What analyses of data will be
-------
20
undertaken? What will the final report look like? How soon after data
collection will the final report be issued? How good is the firm at meeting
reporting deadlines?
Here are a few tips on working effectively with a polling firm. Be highly
directive initially so that the firm has a clear understanding of project
goals and expected project outcomes. In developing a scope of work for the
project, identify the key topic areas that should be covered in the survey.
It is good to be as specific as possible, short of actually drafting
questions. Leave that work to the polling firm. It's more important to
specify all the areas in which questions should be asked so that no
information is inadvertently left out of the questionnaire. In preparation
for actual questionnare construction, it is useful to send a memo to the firm
stating detailed suggestions for content to be included in the questionnaire.
At that time, it is also useful to attach samples of questionnaires that have
been used for I/M surveys in other states. (See attached samples of
questionnaires developed by the Highway Safety Research Institute of the
University of Michigan, the Survey Research Laboratory of the University of
Illinois, and AHF Marketing Research of New York.)
VII. Summary
High quality, scientific polls require careful planning and attention to
numerous details. There is considerable flexibility in the development and
implementation of a poll. The decisions which are made about sample selection
and size, the choice of a polling methodology, questionnaire design and
wording, and the use of quality control measures will ultimately affect the
cost and accuracy of polling results.
-------
21
REFERENCES
Babbie, E.R. Survey Research Methods, Belmont, California: Wadsworth
Publishing Company, 1973.
Dillman, D. A. Mail and Telephone Surveys;The Total Design Method, New
York: Wiley, 1978.
Payne, S. The Art of Asking Questions, Princeton, New Jersey: Princeton
University Press, 1951.
Wolfe, A.C., Michigan Public Opinion Toward Motor Vehicle Inspection,
Ann Arbor, Michigan: Highway Safety Research Institute
(UM-HSRI-79-60), 1979.
-------
22
APPENDIX I
DETERMINATION OF SAMPLE SIZE
Sample size can be calculated to insure that a desired confidence interval
level is obtained, plus or minus a range of percentage points. By convention
in public opinion research, the confidence level is set at 95% with a
confidence interval of ^5%. This means that the chances are about 95 in 100
that, using the same survey procedures, the results in the study would not
vary by more than five percentage points from the result that would be
obtained from a complete coverage of the sample universe. The sample size is
determined by the formula for the standard error (SE) of a sampling
distribution of means and the formula for the confidence interval for the
estimate of the population mean. The former is given by the formula:
SE = square root of PQ where
N-l
P = the proportion of positive responses (e.g., .5 is 50%)
Q = (1-P) or the proportion of negative responses
N = sample size
To determine the sample size from the above formula, first determine the
standard error. From sampling theory, it is known that a 95% confidence
interval can be translated into a t score of 1.96 which represents a
standardized measure of the number of deviational units a sample mean is from
the population mean. Setting a confidence interval of 95% insures that in 95
out of 100 cases the sample mean will be no more than 1.96 units from the
actual population mean. With this information, the SE can be easily calcu-
lated for the sample. SE=the confidence interval (.05) divided by t (1.96).
SE=.0255. This information can then be used to determine the sample size.
SE = square root of PQ
N-l
.0255= .5 x .5 where p =.5 (the worst possibility)*
N-l
.0006508= .25
N-l
N = 385
* The reader should bear in mind that there is a relationship between the
sample size needed for a certain confidence interval and the distribution of
answers for specific question items. For example, given a question with a YES
or NO answer, the number of respondents needed decreases with the movement
away from an even distribution of yes/no answers. Given a random survey with
a 50%/50% split on YES/NO answers, the minimum number of respondents needed
would be in the area of about 390. However, should the distribution of
answers obtained from respondents change to an 85%/15% split, then the needed
sample size would drop to just less than 200 respondents. By convention in
public opinion research, the "worst situation" (where there is a 50%/50%
split) is usually used in the determination of sample size. This assures at
least a 5% confidence interval for all combinations of response percentages.
(Source: Adcock & Associates, Inc., Albuquerque, New Mexico, 1981)
-------
23
APPENDIX II
ACCURACY OF SURVEY DATA
With a sample size of 800, a survey finding that 55% of respondents favored
I/M as a strategy for air pollution control would have a sampling error of
about 3.5, assuming the responding sample was a simple random sample of the
population of interest (see Table 6). This means that the true percentage in
the population has a 95% probability of being within the range of 51.5% to
55.5%, assuming no non-response bias. Consider for a moment the implications
of a sample size of only 200, where 55% of the public were found to favor
I/M. This sample would have a sampling error of 6.9, meaning that the true
percentage in the population has a 95% probability of being within the range
of 48.1% to 61.9%. In this case, one is less certain that the majority of the
population actually favor I/M.
Response rate is calculated as the percentage resulting from completed
interviews divided by the number of sample members known to be eligible.
Inability to get an answer from a ringing number does not take that number out
of the sample. In a sample of 400 persons with valid driver's licenses, if 40
households contacted by phone had no drivers, and 40 telephone numbers were
non-working or commercial, the original sample of potential respondents would
be adjusted to 320. Eighty numbers would be ineligible. Then, if 200
interviews were obtained, the response rate would be 62.5%.
If the actual feelings of non-respondents are extreme in either direction,
then each 10% decrease in the response rate increases by 10 percentage points
the range by which the distribution could be affected. An example will help
to clarify this point. Let's say that we determined that the response rate to
a recent mail survey was 50%. Suppose further that for a key I/M survey
question 75% favored I/M and 25% opposed it. Suppose that all non-respondents
(representing 50% of all those who were mailed a questionnaire but did not
return it) were opposed to I/M. This would mean that the actual percent in
the population who favor I/M is 37.5% (.75x.5) and the percent who oppose it
is 62.5% (.50 + .25x.5).
Often there is no reason to suspect that non-respondents and respondents to a
survey have substantially different attitudes. This is especially true for
surveys treating non-controversial topics. I/M is a fairly controversial
topic, as evidenced by the considerable amount of postive and negative
treatment it has received in mass media nationwide. As a result, we could
expect that at least a small percentage of non-respondents will be persons who
refuse to be interviewed because they disapprove of the topic. For this
reason, it would be important to obtain a high response rate to an I/M survey
to protect against this source of potential bias. A way to reduce the number
of potential refusals is to present I/M in the context of a general environ-
mental survey which, presumably, would appeal to a broader base of .the general
public. This approach would add cost to the survey, but would also increase
the response rate and data accuracy.
-------
24
Table 1
Uses Of I/M Polling Data During I/M Program Phases
Phases
of an I/M
Program
Program
Planning
Uses of Polling Data
Program
Evaluation
Program
Feedback
Developmental
Phase
(one-year
prior to
startup)
Startup
Phase
(period
immediately
before and
Assess informational
needs of target
audiences; also any
misconceptions about
I/M program
Continue to assess
informational needs
of target audiences;
identify public's
concerns about I/M
Determine
baseline of
public attitudes
and knowledge
Obtain reactions
from public to
proposed program
features or
elements
Determine whether Obtain reactions
public awareness
program is
reaching intended
audiences
after startup) program
to actual
program features,
(e.g., quality
control and con
venience aspects)
Operational
Phase
(the life
of operating
program)
Program evaluation
data is fed back
into the planning
process to make
any needed changes
in I/M public
awareness strategies
Determine changes Obtain reactions
in awareness and
knowledge of I/M
as result of the
public awareness
program
to proposed
program changes
or modifications
-------
25
Table 2
Rating of Face-to-Face Interviews, Telephone Interviews, and
Mail Questionnaires for Selected Performance Characteristics
(Dillman, 1978, pp. 74-75)
METHOD
Performance
Characteristics
I. Obtaining a Representative
Sample
A. Known opportunity for
all members of popula-
tion to be included
in the sample.
1. Completely listed
populations.
2. Populations which
are not completely
listed (e.g., house-
hold occupants).
B. Control over selection
or respondents within
sampling units.
C. Likelihood that selected
respondents will be
located.
D. Insensitivity to sub-
stitution of respondents
and households.
Face-to-Face
Interviews
Mail Telephone
Questionnaires Interviews
High
High
High
Medium
High
Medium
High
Medium
Medium
Medium
High
Low
High
High
Low
E. Response rates.
1. Hetrogeneous samples
(e.g., general public).
High
2. Homogeneous specialized High
samples (e.g., agency
directors, ministers,
students).
F. Likelihood that unknown
bias from refusals will
be avoided.
High
Medium
High
High
High
Low
High
-------
26
Table 2 (cont.)
Face-to-Face
Interviews
Performance
Characteristics
II. Questionnaire Constuction
and Question Design
A. Allowable length of High
questionnaire
B. Type of question.
1. Allowable complexity. High
2. Success with open- High
ended questions.
3. Success with screen High
questions.
4. Success with controll- High
ing sequence.
5. Success with tedious High
or boring questions.
C. Success in avoiding item High
non-response.
D. Insensitivity to question- High
naire construction pro-
cedures.
III. Obtaining Accurate Answers
A. Likelihood that social Low
desirability bias can be
avoided.
B. Likelihood that inter- Low
viewer distortion and
subversion can be
avoided.
C. Likelihood that contamin- Medium
ation by others can be
avoided.
Mail
Questionnaires
Medium
Medium
Low
Medium
Low
Low
Medium
Low
High
High
Medium
Telephone
Interviews
Medium
Low
High
High
High
Medium
High
Medium
Medium
Medium
High
-------
Performance
Characteristics
27
Table 2 (cont.)
Face-to-Face
Interviews
IV. Administrative Requirements
A. Likelihood that personnel Low
requirements can be met.
B. Potential speed of Low
implementation
C. Keeping costs low.
1. Overall potential Low
for low per inter-
view costs.
2. Insensitivity of Low
costs to increasing
' geographical dis-
persion.
Mail
Questionnaires
High
Low
High
High
Telephone
Interviews
High
High
Medium
Medium
-------
28
Table 3
Deciding on Question Format
QUESTION FORMATS
OPEN-ENDED (Respondent supplies
entire answer)
OPEN-ENDED WITH FIXED
PARAMETERS (Respondent supplies
answer within a range)
CLOSED-ENDED WITH ORDERED
CHOICES (Researcher supplies
response categories which
are ordered in some way,
e.g., category 1 reflects a
difference in amount or degree
than category 2)
CLOSED-ENDED WITH UNORDERED
RESPONSE CHOICES (Researcher
supplies response categories
which are unordered, e.g., cat-
egories do not reflect differ-
ences in amount or degree)
PARTIALLY CLOSED-ENDED
(Researcher supplies response
categories but also permits
respondent to provide addi-
tional information)
EXAMPLES
What are some of your reasons for favoring
an emissions inspection program? (asked
only to persons who responded that they
favored one)
What do you think would be a reasonable
longest distance for someone in your area
to have to drive for an emissions
inspection?
MILES 98 NO OPINION
To what extent do you agree or disagree
with the following statement? Heavy duty
trucks and buses should be included in an
emissions inspection program.
1 Strongly disagree
2 Somewhat disagree
3 Neutral
4 Somewhat agree
5 Strongly agree
Who would you prefer to conduct an emissions
test on your car? An inspector in a state-
operated station or a private .inspector in
a state-licensed garage or service station?
1 STATE INSPECTORS
2 PRIVATE INSPECTORS
3 NO DIFFERENCE
4 NO OPINION
If a vehicle owner could not afford to pay
the costs of repairing the emissions
system, do you think the vehicle should be
exempted from the emissions requirements,
or should the state set up a special fund
for the repair of such vehicles, or should
the vehicle be taken off the road, or what?
1 EXEMPTED
2 SPECIAL STATE FUND
3 TAKEN OFF THE ROAD
4 OTHER:
-------
29
Table 4
Sample I/M Survey Questions
The following questions were included as part of I/M public opinion
surveys in several states.
Arizona has had an emissions inspection program for cars and trucks for
about three years now. Do you feel the program should be kept or
repealed?. (Arizona State University Survey Research Center)
In the near future, everyone may be required to have their cars inspected
for emissions each year. The way you feel now would you tend to oppose
or favor annual emissions inspections? (California Air Resources Board)
Several ways are being considered all over the country for dealing with
air pollution. For each of the following strategies under consideration,
I would like you to tell me whether you are very much in favor, somewhat
in favor, somewhat opposed or very much opposed to it. (A program of
vehicle emissions inspection and maintenance was one of 22 items
mentioned.) (Colorado Market Research Services Inc., University of
Colorado Applied Sociological Research Team)
Should annual and safety inspections include tests to control air
pollution?" (Metropolitan Bade County and the Dade-Monroe Lung
Association, Florida)
Specific proposals have been made to cutdown the amount of pollution
caused by cars and trucks. Using the other side of the green card
[face-to-face interview] please read with me as I review one possible
proposal.
a. All car and truck owners in the country would be required to get
their vehicles inspected to see if they were causing too much
pollution.
b. This inspection would cost $5-10.
c. Those who fail the inspection would have to get their vehicle
fixed and return for another inspection.
d. Those who passed the inspection would receive a sticker for their
windshield. Anyone driving without a sticker would be subject to
a fine.
Do you favor or oppose this program?
(Urban Studies Center of the University of Louisville)
Some states have a combined program of inspection for both safety and
emissions. Do you favor such a combined program for Michigan? (Highway
Safety Research Institute, Ann Arbor, Michigan)
A yearly inspection of motor vehicles for pollution emission levels will
be mandatory in the State of Nevada. Are you in favor of this law?
(University of Nevada, Las Vegas)
-------
30
Table 4 (cont.)
In the past, for economic reasons, there have been recommendations to
eliminate our mandatory vehicle inspection program. Do you agree or
disagree? (New Jersey Motor Vehicle Inspection Study Commission)
Starting in January 1981, cars will have to be inspected to find out
whether they are safe and if they are polluting the air too much. The
cost of the inspection will go up from $6.00 to $12.00. Do you think
that this program is:
1. Good idea
2. Bad idea
3. Don't know
(Social Data Analysts, Inc., New York)
Do you think that exhaust emissions tests on automobiles are important?
(Research Corporation of New England, Rhode Island)
Would you favor legislation requiring periodic inspection of emission
systems on all cars? (Nashville AAA Mid-South Auto Club)
As mentioned in the letter, the state is thinking about reducing air
pollution by setting up yearly inspections of Michigan motor vehicles to
test their exhaust emissions and to require owners to fix those engines
which are causing too much air pollution. In trying to decide if such a
vehicle inspection program is desirable, one has to think about the value
of reducing air pollution against the total program costs in time,
travel, inspection fees, and repairs. What is your general opinion? Do
you favor or oppose a required emissions inspection program in your
county?
1 FAVOR 3. NO OPINION - (GO TO NEXT QUESTION)
2 OPPOSE
Do you feel strongly about that, or not very strongly?
1 STRONGLY
2 NOT VERY STRONGLY
(Highway Safety Research Institute, The University of Michigan)
A recent New York State program aimed at controlling air pollution is the
Auto Emissions Inspection and Maintenance Program. As of January 1,
1981, all gasoline fueled vehicles in the New York metropolitan region
(i.e. , New York City, Rockland, Westchester, Nassau, and Suffolk
counties) which weigh 8,500 Ibs. or less must be inspected annually for
emissions. In 1981, the Auto Emissions Inspection and Maintenance
Program only requires the inspection of vehicles. Beginning in January,
1982 vehicles that fail the emissions test must be repaired in order to
be registered. Based on what I've just read to you or what you may have
-------
31
Table 4 (cent.)
previously head about the Auto Emissions Inspection and Maintenance
Program, would you say that you (READ LIST)?
Strongly favor it
Somewhat favor it
Neither favor nor oppose it
Somewhat oppose it
Strongly oppose it
(Source: AHF Marketing Group)
To reduce air pollution from cars, the State of Illinois is considering a
law that would require all cars, vans and pickup trucks in the Chicago
metropolitan area to be tested for emissions. That means that you would
have to take all cars into a designated place to be checked for pollution
levels. This inspection would check only pollution levels and not safety
problems. Each car would have to pass the test in order to get license
plates. Would you favor or oppose such a program?
1. Favor
2. Oppose
3. DependsDon't Know
(University of Illinois)
In the spring of 1983, the twice a year safety inspection for cars in
Massachusetts will be replaced by a new inspection program. Under this
new program, cars will have to be inspected just once a year, but the
inspection will include not only mechanical safety, but also tailpipe
emissions. In order to pass the tailpipe emissions inspection, a car's
emissions would be measured against a standard set for that particular
car's make and year. A 1980 car, for example, would have to meet a
higher standard than a 1971 car. If the car fails either the safety or
emissions inspection, it must be repaired before an inspection sticker
will be issued. The cost for this once a year inspection will be $10 and
will be done in gas stations as it is now. Were you aware of this new
inspection program, or hadn't you heard about it before?
1. AWARE
2. HADN'T HEARD
3. DON'T KNOW
Based on the description I have just read, is your opinion of this new
car inspection mainly favorable or mainly unfavorable?
1. MAINLY FAVORABLE
2. MAINLY UNFAVORABLE
3. DON'T KNOW
(Becker Research Corporation, Boston, Mass.)
-------
32
Table 5
Sampling Tolerances for Comparisons Between Two Samples
Tolerances are also involved in the comparison of results from different
parts of any one sample and in the comparison of results between two
different, but comparable, samples. A difference, in other words, must
be of at least a certain size to be considered statistically
significant. The table below is a guide to the sampling tolerances
applicable to such comparisons.
Differences required for significance
at or near these percentage levels
Size of samples
compared 10% or 90% 30% or 70% 50%
1000 and 1000 3 44
1000 and 750 345
1000 and 500 3 55
1000 and 250 467
1000 and 100 6 9 10
750 and 750 355
750 and 500 356
750 and 250 477
750 and 100 6 10 10
500 and 500 466
500 and 250 578
500 and 100 6 10 11
250 and 250 589
250 and 100 7 11 12
100 and 100 8 13 14
Source: Becker Research Corporation, Boston, Mass.
-------
33
Table 6
Sampling Errors at a 95% Level of Confidence
for Various Percentages and Sample Sizes,
Assuming a Simple Random Sample
(Percentages)
Sample or Subsample Size
Approximate
Percentages 1600 1000 800 400 250 200 150 100 50^
50%
40% or 60%
30% or 70%
25% or 75%
20% or 80%
15% or 85%
10% or 90%
5% or 95%
Note: For a given sample size, the chances are 95 in 100 that the value being
estimated lies within a range equal to the reported percentage, plus or minus
the number of percentage points shown above.
Source: Internal document, Institute for Social Research, The University of
Michigan, Ann Arbor, Michigan.
2.50
2.45
2.29
2.17
2.00
1.78
1.50
1.09
3.16
3.10
2.90
2.74
2.53
2.26
1.90
1.38
3.54
3.48
3.24
3.06
2.83
2.52
2.12
1.54
5.00
4.90
4.58
4.34
4.00
3.58
3.00
2.18
6.32
6.20
5.80
5.48
5.06
4.52
3.80
2.76
7.08
6.92
6.48
6.12
5.66
5.04
4.24
3.08
8.16
8.00
7.48
7.08
6.54
5.84
4.90
3.56
10.00
9.80
9.16
8.67
8.00
7.14
6.00
4.36
14.14
13.86
12.96
12.24
11.32
10.10
9.48
6.16
-------
34
Table 7
Emissions Inspection Attitude by Background Factors
(Percentages)
AGE
15-24
25-34
35-54
55+
SEX
Male
Female
GEOGRAPHICAL AREA
Wayne
Oakland, Macomb
Other Detroit SMSA
Genesee, Ingham
Kent
Other Central SMSA
Counties
Other Surburban
SMSA Counties
Other Counties w
Cities | 15,000
Rural Counties
INCOME
Over $30,000
$20,000 - $30,000
$10,000 - $20,000
Under $10,000
TOTAL
N
171
218
259
159
417
390
179
175
86
92
81
50
42
104
174
238
237
105
809
Source: Wolfe, A.C. , Michigan
Strong
Favor
38.0
38.5
28.6
17.6
29.3
33.1
39.7
37.7
23.3
34.8
29.6
16.0
19.0
21.2
36.2
29.4
32.5
28.6
31.0
Public
Weak
Favor
22.8
14.7
12.7
20.1
16.3
17.4
16.2
14.3
23.3
21.7
13.6
14.0
23.8
14.4
17.2
16.8
16.9
15.2
16.9
Opinion
Weakly
Oppose
10.5
11.5
11.6
15.1
12.0
12.1
13.4
10.3
12.8
9.8
11.1
18.0
14.3
10.6
12.1
12.2
13.1
9.5
12.0
Toward
Strong
Oppose
21.6
26.1
39.8
32.7
36.0
25.4
21.2
29.1
34.9
23.9
38.3
44.0
35.7
39.4
31.6
32.8
28.3
29.5
30.9
Depends
0
1.8
0.8
3.1
1.2
1.5
1.1
1.7
1.2
4.3
0
0
0
1.0
0.6
2.9
0.4
1.9
1.4
No
Opinion
7.0
.7.3
6.6
11.3
5.3
10.5
8.4
6.9
4.7
5.4
7.4
8.0
7.1
13.5
2.3
5.9
8.9
15.2
7.8
Motor Vehicle Inspection,
Ann Arbor, Michigan: Highway Safety Research Institute, UM-HSRL-79-60,
1979.
-------
ATTACHMENTS
SAMPLE QUESTIONNAIRES FOR I/R SURVEYS IN THREE STATES
-------
SURVEY OF MICHIGAN DRIVERS
CONCERNING MOTOR VEHICLE INSPECTIONS
A BRIEF QUESTIONNAIRE
Sponsored by: Michigan Department of Transportation
Survey Director: Dr. Arthur C. Wolfe
111 H.S.R.I.
University of Michigan
Ann Arbor, Mich. 48109
Tel. (313) 764-0248
-------
-1-
DIRECTIONg; For moat of the following questions please circle the number in
front of your answer choice. On a few questions you will need to fill in a
blank line or write out an answer in your own words.
1. Have you personally ever been bothered by air pollution in the county
where you live?
r
11 YES
5 NOKgo to Q.2)
la. In the past year would you say you have been bothered by air pollution:
1 A LOT
2 SOME
3 NOT MUCH
4 NOT AT ALL ?
[11]
2. How widespread do you think the air pollution problem in Michigan is?
you say that air pollution.Is a serious problem: (circle one answer)
Would
1 NOWHERE IN MICHIGAN
2 ONLY IN A FEW BIG CITIES
3 THROUGHOUT SOUTHERN MICHIGAN
4 ALL OVER THE STATE ?
8 DON'T KNOW, NO OPINION
[12]
3. You probably know that using regular gas in cars designed for no-lead gas
increases the air pollution from such cars. Do you favor or oppose a law
prohibiting motorists from using regular gas in no-lead cars?
1 FAVOR
2 OPPOSE
8 NO OPINION
[13]
As mentioned in the letter, the state is thinking about reducing air pollution
by setting up yearly Inspections of Michigan motor vehicles to test their
exhaust emissions and to require owners to fix those engines which are causing
too much air pollution. In trying to decide if such a vehicle inspection
program is desirable, one has to think about the value of reducing air
pollution against the total program costs in time, travel, inspection fees,
and repairs. What is your general opinion? Do you favor or oppose a required
emissions inspection program in your county?
F
1 FAVOR
2 OPPOSE
80 NO OPINION^(go to Q.5)
[14-15]
4a. Do you feel strongly about that, or not very strongly?
1 STRONGLY [15]
2 NOT VERY STRONGLY
5. Another idea which has been mentioned for reducing air pollution in some places
is to prohibit parking in certain areas with heavy air pollution in order to
encourage people to use public transportation. If you had to choose between a
no parking program and an emissions testing program, which would you prefer?
1 NO PARKING
2 EMISSIONS INSPECTION
8 NO OPINION
[16]
-------
-2-
6. In thinking about a possible exhaust emissions inspection program for Michigan
there are a number of decisions which would have to be made about how the
program might work. One choice has to do with licensing a large number of
private garages and service stations to do the inspections, or with setting
up a limited number of new state stations to do the inspections. Do you
think it would be better to set up special state inspection stations, or
to license private garages to do the inspections?
P
1 STATE STATIONS
2 PRIVATE GARAGES AND SERVICE STATIONS
800 NO OPINIONKgo to Q.7)
6a. Why do you feel that way?_
7. If an emissions inspection program is set up, do you favor or oppose
including buses and large trucks in the program?
[17-19]
[18.19J
1 FAVOR
2 OPPOSE
8 NO OPINION
[20]
8. If an emissions inspection program is set up, do you think it should be
required statewide, or only in the areas with the worst pollution problems?
F
1 STATEWIDE
WORST POLLUTION AREAS 'ONLY
80 NO OPINION-Kgo to Q.9) [21-22]
8a. Do you feel strongly about that, or not very strongly?
1 STRONGLY
2 NOT VERY STRONGLY
9. What do you think would be a reasonable longest distance for someone in
your area to have to drive for. an emissions inspection?
[22]
MILES
98 MO OPINION
10. What do you think would be a reasonable longest time for someone to spend
at the inspection station waiting and getting inspected?
MINUTES
998 NO OPINION
11. What do you think would be a reasonable inspection fee, or do you think
the inspection should be paid for by other state funds?
$ DOLLARS
00 FREE, OTHER STATE FUNDS
98 NO OPINION
[23-24]
[25-27]
[28-29]
12. If a vehicle did not meet the emissions standards f»r vehicles of its age
and type, what do you think would be a reasonable length of time to give
the owner to get the emissions system working properly?
DAYS
98 NO OPINION
[30-31]
-------
13. It has been suggested that some older cars should be exempted from having to
repair their emissions systems if the repair cost was a large percentage of
the vehicle's value. Do you favor or oppose this idea?
fl FAVOR 800 NO OPINION-Kgo to Q.14)
2 OPPOSB-Kgo to Q.14) [32-34]
13a. What percentage of the vehicle's value should the repair cost be
in order to be exempted?
(at least) _____ PER CENT 98 NO OPINION [33-341
14. If a vehicle owner could not afford to pay the costs of repairing the
emissions system, do you think the vehicle should be exempted from the
emissions requirements, or should the state set up a special fund for
the repair of such vehicles, or should the vehicle be taken off the road,
or what?
li
f:
10 EXEMPTEDKgo to Q.15) 80 NO OPINION-Kgo to Q.1S)
SPECIAL STATE FUND [35-36]
30 TAKEN OFF THE ROAD-*(go to Q.15)
40 OTHER; -*(RO to Q.15)
14a. Should this be a loan fund which the owner has to pay back, or should
it be a welfare grant?
1 LOAN 8 NO OPINION [36]
2 WELFARE GRANT
15. Who do you think could be trusted more to do a proper emissions test,
inspectors in state-operated stations or inspectors in state-licensed
garages and service stations, or don't you think there would be any difference?
1 STATE INSPECTORS 8 NO OPINION [37]
2 PRIVATE INSPECTORS
3 NO DIFFERENCE
16.Now we have a question about safety inspections. As you probably know, many
other states require regular safety inspections of their registered vehicles
for such components as lights, brakes, steering, and tires. Do you favor
or oppose such a regular safety inspection program for Michigan?
F
>1 FAVOR 80 NO OPINION-Kgo to Q.17)
OPPOSE [38-39]
16a* Do you feel strongly about that, or not very strongly?
1 STRONGLY [39]
2 NOT VERY STRONGLY
17. Some states have combined programs of inspection for both safety and
emissions. Do you favor or oppose such a combined program for Michigan?
1 FAVOR 8 NO OPINION [40]
2 OPPOSE
-------
-4-
18. A combined emissions and safety inspection would cost somewhat more than
an emissions inspection alone. What do you think would be a reasonable
combined inspection fee, or do you think the combined inspection should be
paid for by other state funds?
S DOLLARS 98 NO OPINION [41-421
00 FREE, OTHER STATE FUNDS
19. How much do you think required vehicle safety inspections would reduce
motor vehicle accidents in Michigan?
1 A LOT 8 NO OPINION [431
2 SOME
3 NOT MUCH
20. It has also been suggested that an inspection program should include testing
of vehicle noise and should make any noisy vehicle be fixed. Do you favor
or oppose required state testing of motor vehicle noise?
FAVOR 80 NO OPINION-*go to Q.21) [44-45]
OPPOSE
20a» Do you feel strongly about that, or not very strongly?
1 STRONGLY 1*5]
2 NOT VERY STRONGLY
21« Are you ever disturbed by the noise made by individual noisy vehicles?
til YES
77 NO4
-------
-5-
24. Now there are a few background questions about your household and its
vehicles? How many licensed drivers are there in your household?
NUMBER
[50]
25. How many personal motor vehicles of what types are owned by your household?
^B
VANS, PICKUPS, JEEPS, UTILITY VEHICLES . .
MOTORCYCLES
OTHER:
.s the model year of the oldest vehicle?
NUMBER
MMV
NUMBER
NUMBER
NUMBER
YEAR
T51I
[52]
f531
[54]
[55-56]
27. Would you say that this vehicle is in safe operating condition now, or
that there are some safety components which need repair?
100 SAFE NOW_Kg0 to Q.28)
r-2 REPAIRS NEEDED
27a. What components are these?
800 DON'T KNOWfr(go to Q.28) [57-59]
i [58,59]
28. What do you think are the main types of vehicle defects which sometimes
contribute to accidents?
29. Have you ever had occasion to use regular gasoline in a vehicle designed
to use no-lead gasoline?
1 YES
5 NO
8 DON'T KNOW, NOT SURE
[60,61,62]
[63]
30. How closely would you say you follow your owner's manual for regular service
and repair of your vehicle(s)?
1 VERY CLOSELY
-2 SOMEWHAT CLOSELY
L300000 NOT VERY CLOSELYf(R° to (J.31)
800000 DON'T KNOW-Kgo Co Q.31) [6*-69]
30aT About how often do you get your oldest vehicle tuned up?
MILES or .MONTHS 99998 DON'T KNOW
[65-69]
-------
-6-
31. Do you, or does someone else in your household, do routine maintenance and
repair of your vehicle(s), or do you usually have it done by a garage or
service station?
I. 1 DO IT MYSELF 80000 DON'T KNOW-*go to Q.32)
\20000 DONE BY SOMEONE ELSE IN HOUSEHOLD OR. BY FRIEND-*go to Q.32)
\30000 DONE BY PRIVATE GAKAGE OR.SERVICE STATIONKgo to Q.32) [70-74]
31a. Do you consider the emissions control devices when you do a tune-up?
1 YES 8 DON'T KNOW [71]
3 SOMEWHAT
5 NO
31b. If your vehicle exhaust system needed a major repair to meet the emissions
standards, how likely is it that you would be able to make the repair yourself?
1 VERY LIKELY 8 DON'T KNOW [72]
2 SOMEWHAT LIKELY
3 NOT VERY LIKELY
31c. Do you check your vehicle's brakes during the course of normal maintenance?
pi YES [73-74]
3 SOMETIMES
50 NO(go to Q.32)
31d. How do you check the brakes?
[74]
32. Do you believe that your vehicles' emissions control systems are working
properly now?
10 YES-Kgo to Q.33) 80 DON'T KNOWKgo to Q.33) [75-76]
p5 NO
32a. Why is that? .
[76J
33. Finally there are 3 background questions for statistical purposes.
In what year were you born?
YEAR [77-78]
34. What is the highest level of education you have completed?
1 NO FORMAL EDUCATION 6 SOME COLLEGE [79]
2 GRADE SCHOOL ONLY 7 COMPLETED A FOUR-YEAR COLLEGE
3 SOME HIGH SCHOOL 8 SOME GRADUATE WORK
4 COMPLETED HIGH SCHOOL 9 COMPLETED A GRADUATE DEGREE
5 VOCATIONAL SCHOOL
35. In which of the four categories below does your total yearly family income fall,
before taxes?
5 LESS THAN $10,000 ,RQ,
4 $10,000-$20,000 l J
2 $20,000-$30,000
1 OVER $30,000
-------
Is there anything else you would like to say about the problem of air
pollution or about motor vehicle inspections, or do you have any suggestions
concerning appropriate questions to include in a future questionnaire on
this topic? If so, please use the space below for these additional comments.
THANK YOU VERY MUCH FOR EXPRESSING YOUR VIEWS ON THESE ISSUES OF GREAT
IMPORTANCE TO YOUR STATE GOVERNMENT AND TO ALL OF THE CITIZENS OF MICHIGAN.
-------
ilF-MARKETING RESEARCH
801 Socord Avenue
New York, New York 10017
RESPONDENT INFORMATION BOX;
AUTO EMISSIONS TRACKING STUDY
RATIONAL CONSENSUS SURVEYS
026771F WAVE I
JULY, 1981
L3TART CARD I
NAME: (PRINT)
ADDRESS:
Cm: STATEl
TEL.* (AREA CODE): ( } TIME STARTED:
INTERVIEWER:
ZIP CODE:
TIME ENDED:
DATE:
CIRCLE COONTY:
7"
9
CIRCLE SEX:
Male
Female
1
2.
I ASK TO SPEAK TO AN ADULT 18 YEARS OF AGE OH OLDER. [
Cello. I'm
of National Consensus Surveys. We «re interviewing New Yorlc
residents and I'd like co aak you a few questions.
A. First, please cell n* the age and sex of all persons in your household 18 years
of age or older.
Please start with the females, from oldest to youngest, then the ages of the
ales, from oldest to youngest.
(ASK TO SPEAK TO THE PERSON NEXT TO TBS LOWEST "X" AND ASK Q.la.)
NAME ACE
SEX
IF DESIGNATED RESPONDENT REFUSES, TERMINATE. DO NOT INTERVIEW OTHERS IN
HOUSEHOLD.
IF DESIGNATED PERSON NOT AVAILABLE, MAKE CALLBACK APPOINTMENT AND RECORD
BELOW:
NAME
PATE
TIME
1ST ATTEMPT:
2ND ATTEMPT:
_(AM/PM)
CAM/PM)
Augur, 1911
S M T W T F S
1
2345679
10 11 12 13 14 IS
IS IT 18 19 20 21 22
23 24 25 26 27 28 29
3031
IF NECESSARY. RE-INTRODUCE YOUUSELF AND RE-READ INTRODUCTION.
-------
AUTO EMISSIONS TRACKING STOW
026771* WAVE I
1. I'm going Co read you a short list of possible health problems facing residents
of the New York metropolitan region. For each problem we'd like to know whether
you view the problem as extremely serious, very serious, somewhat serious,
not so serious, or. not at all serious.
Let's start with ... (READ HEALTH PROBLEM NEXT TO "X"). Would you say Chat ...
(FIRST HEALTH PROBLEM) is an extremely serious, very serious, somewhat serious,
not so serious, or not at all serious problem to residents of the-New York
metropolitan region? (RECORD ANSWER BELOW. CONTINUE ONTO. ALL PROBLEMS HAVE
BBEH ASKED ABOUT.)
Extremely Very Home what Mot So Not At All
(START HERE) Serious Serious Serious Serious Serious
Toxic wastes
Water pollution
Air pollution
Chemicals sad additives
in foods
( ) Noise pollutic
4-
4
4
4
4
3
3
3
3
3
2
2
2
2
2
2a.
There are several major contributors to air pollution in the New York metropolitan
region. What percent of air pollution would you say is caused by ... (READ LIST) t
SCORCH
Pollution caused by industry
Pollution caused by motor vehicles
Pollution caused by public
utilities
Pollution caused tUrongh residential
use, such as home heating "
jr
z
Pollution caused by other sources. (SPECIFY) t
TOTAL - 100Z
17 TOTAL DOES HOT ADD TO 100Z, SAY TO RESPONDENT: "This adds Co (more/less)
than 100Z. Can we please go over what percent of air pollution is caused by
(READ SOURCES NAMED IN Q.2a.)
2b. Now, chinking only about motor vehicle pollution, what percent is caused by ...
(READ LIST)?
TYPE OP VEHICLE
Commercial vehicles, such as
trucks and taxis
Passenger cars
Buses
PERCENT
Z
Z
Z
0?her (SPECIFY^:
TOTAL * 100Z
IF TOTAL DOES NOT ADD TO 100Z, SAY TO RESPONDENT: "This adds Co (more/less)
than 100Z. Can we please go over what percent of air pollution is caused by
(READ SOURCES NAMED IN Q.2b.)
-------
ADTO EMISSIONS TRACKIHC STOW
#26771? WAVE I
3a.
Daring Che pest year, several regulatory programs aimed at controlling air
pollution in the Hew York metropolitan region have gone into effect. Please
tell me the name* of any city, state, or federal regulation* to control air
pollution in the New York metropolitan region which have gone into effect in
the past year. (PROBE:) What others have you heard of? (DO NOT RKAD LIST.)
New York State Auto Emissions Inspection/
Maintenance Program ''. 1
All other* 2
IF RESPONDENT MENTIONS "NEW YORK AUTO EMISSIONS INSPECTION AND MAINTENANCE
PROGRAM" SKIP TO Q.Aa. OTHERWISE. READ Q.3b. . .
3b.
3c.
A recent New York State program timed at controlling air pollution is the
Auto Emissions Inspection and Maintenance Program. At of January 1, 1981,
all gasoline- fueled vehicles in the New York metropolitan region (i.e.,
New York City, Rocklsnd, Westcheater, Nassau, and Suffolk counties) which.
weigh 8,500 IDS. or less most be inspected annually for emissions.
In 1981 the Auto Emissions Inspection and Maintenance Program only requires
the inspection of vehicles. Beginning in January, 198Z vehicles that fail
the emissions test must be repaired in order to be registered.
Based upon what I've just read to yon or what yon may have previously heard
about the Auto Emissions Inspection and Maintenance Program, would yon say
that you ... (READ LIST)?
Strongly favor it
Somewhat favor it
Neither favor nor-oppose it
Somewhat oppose it
Strongly oppose it
Why do you feel that way?
saying that?
(PSOBBi) What other reason do yon have for
4a.
(SKIP TO q.57)
(ASK ONLY IF ADTO EMISSIONS INSPECTION AND MAINTENANCE PROGRAM IS MENTIONED
IN Q.3a:)
You mentioned that during the past year New York State has begun Che Auto
Emissions Inspecton and Maintenance Program. Based upon your personal
experience or what yon may have heard about it, would you say Chat you ...
(READ LIST)?
Strongly favor it
Somewhat favor it
Neither favor nor oppose it
Somewhat oppose it
Strongly oppose it
4b. Why do yon feel that way? (PROBE:) What other reasons do you have for
saying chat?
5.
(ASK EVERYONE:)
Have you seen or heard any advertising about Che New York State Auto Emissions
Inspection and Maintenance Program?
(ASK Q.6) « -Yes 1
(SKIP TO q.7)
-------
AUTO EMISSIONS TRACKING STUDY 026771F WATS I
6. What do you recall seeing or hearing about the Auto Emissions Inspection and
Maintenance Program? (PROBBi) What else did the ad say or show?
7. I'm going to read a short liet of statements which might describe particular
aspects of the Auto Emissions Inspection and Maintenance Program. Please
tell me if you strongly agree, somewhat agree, neither agree: nor. disagree,
somewhat disagree, or strongly disagree) with each statement.
Let's begin with ... (READ STATEMENT NEXT TO "X".) would yon say yon strongly
agree, snmswhat agree,, neither agree nor disagree, somertut disagree, or strongly
diaagree? (RECORD BELOW AND CONTINUE UNTIL. ALL STATEMENTS HAVE BEEN ASKED ABOUT.)
Neither
Agree
Strongly Somewhat Nor Somewhat Strongly
(START HERB) Agree Agree Disagree Diaagree Disagree
( ) Automobile mechanics have the
knowledge and training
necessary to repair emissions
control devices -' ' 54321
( ) The auto emissions inspection
and maintenance program will
significantly improve air
quality in the New York metro-
politan region " 54 3 2 1
( ) There ie a sufficient number of
auto emissions inspection
stations 3 4 32
( ) Automobile mechanics are generally
honest about needed repairs- 5432.
( ) Certain cars should be exempt
that is, not have to
undergo auto emissions
inspection - -
5 4
( ) Automobile mechanics will gen-
erally charge a fair price for
needed emissons control repairs
( ) All motor vehicles not just
passenger ears should be
inspected for auto emissions
( ) Auto emissions inspection
stations are conveniently
located '- "
( ) Automobile mechanics who work on
emissions control devices
should be certified by
New York State
( ) The auto emissions inspection
and maintenance program is
« unnecessary government
regulation
-------
AUTO EMISSIONS TRACKING STUDT #2677IF WAVE I
(FOR ALL RESPONDENTS WHO ANSWERED "STRONGLY" OR "SOMEWHAT" AGREE BOXED AREAS
IN Q.7 ASK Q.&. OTHERWISE, SKIP TO Q.9a.)
8. Ton said Chat certain cara should be exempt Chat ia, not have to undergo
auto emissions inspection. Under what circumstances or what types of cara
should be exenpt from being inspected for auto emissions? (PROBE:) Are there
any other circumstances or other types of cars which should be exempt?
9a. Some people have proposed that if a car fails the auto emission* inspection,
a ceiling be put OB the coat of repairs that have to be mad*. That ia, no
matter what the coat, of needed, repair* are, a ear owner would only be
required to spend a set amount for. repairs. A ear would be registered whether
or not all the needed repair* on emiasions rcntrol device* were mad*.
Do- you think that there should be such a ceiling on the cost of repair* thi.it
have to be made?
Yes 1
No 2
9b. (IF "YES":) What is the ""Tf' amount of money that a ear owner be required
to pay for repairs to emissions control devices?
(RECORD AMOUNT TO NEAREST DOLLAR): $
10. Some people believe that removing emissions control devices such as
catalytic converters improves gas mileage and car performance. Do you
think that removing emiasions control devices actually does improve gaa
mileage and ear performance?
Yea 1
No 2
11. As you may know, sine* 1937 New York State haa required that each, year all
passenger vehicles undergo a mandatory safety inspection. This inspection
involves the cheeking of brakes, tires, lights, steering, suspension, etc.
to see if they are in proper working order. We would like to know whether
yon feel that this safety inspection program haa contributed to highway
and driving safety. Would you say that the program has contributed ...
(READ LIST)?
Very much 4
Somewhat 3
A little, or : 2
Not at all to highway and driving
safety? 1
12a. Do you think that this annual mandatory safety inspection program should be
continued?
" Yea 1
(CONTINUE)
12b. Why do you say that? (PROBE:) What other reasons do you have for feeling I
this way? I
-------
< I
AUTO
13*.
EMISSIONS TRACKING STUDY
Are you a licenced driver?
426771F WAVE I
Tec
Ho -
13b. Do yea personally own a ear?
(SKIP TO Q.1S) « Yea
(ASK Q.14) * Bo -
14* Do you pay for any of tte expenses for maintaining or repairing a cart
Yea
Bo -
I ASK EVERYONE* I
13* Including any eara chat yon may own, in. total how many earii da people in
thia hooaehold own?
(HUTS IK NUMBKRj)
1
2
I ASK Q'a. 16a-20 OHLY IF ONE OR MORE CAH3 IH 0.15. OTHERWISE. SEP TO 0.21. I
16a. Are any of the ear or eara that people in your honaehold own 1976 models
or newer?
Yea
No
16b. Are any of these eara older than 1978?
Yea
Ho -
17.
18.
19.
Please think about manufacturers' warranties covering emissions control
devices. To the best of your knowledge, do these warranties generally
extend beyond the first year of owning the car?
Yea
Bo
1
2
1
2
Have any of the ears that members of your household own been inspected under
the New York State Anto Emissions Inspection and Maintenance Program?
Yea
Bo -
Who usually repairs or maintains the cars owned by members of your household?
(READ LIST. CIRCLE AS MANY AS APPLY.)
Other (SPECIFY):
0
20. What would you estimate Che average bill for the repair of auto emissions
control devices to be?
(RECORD AMOUNT TO NEAREST DOLLAR:) $
-------
-AUTO EMISSIONS TRACKING STUDY
#2677IF WAVE I
1 ASK EVERYONE;|
Finally, some questions for classification purposes only.
21. Which of the following groups includes your age? (RKAD LIST.)
18 to 24
25 to 34
35 to 44
45 to 54
55 to 64
65 or over
22. Whet i» your marital status? (BEAD LIST.)
Married
Single
Separated
Divorced -
Widowed
23. Including yourself, what is the total number of people in your household?
One
Two
Three
Pour -
Five -
Six or more
24.- What was the last grade of formal school you completed? Wae it ...
1
2
3
4
5
6
1
2
3
4
5
1
2
3
4
5
6
(BEAD LIST)?
Grade school or less
Some high school
Completed high school -
Vocational or technical school
beyond high school
college
Completed college or more
25a. Do you work outside your home at a paid job?
(ASK Q.2Sb)
(SKIP TO Q.26)
25b. In an average work week, do you work 35 or more hours per week, or less
than 35 hours per week?
35 hours or more
Less than 35 hours
26. Finally, which of Che following income groups includes your total annual
family income before taxes? (READ LIST.)
Under $5,000
$5,000 to under $8,000
$8,000 Co under $10,000 -
$10,000 co under $15,000
$15,000 co under $20,000
$20,000 co under $25,000
$25,000 co under $30,000
$30,000 or more
(DO NOT READ)
Refused --- '-
Don't know -'
1
2
3
4
5
6
3
4
1
2
3
4
5
6
7
8
x
y
THANK RESPONDENT FOR HIS/HER COOPERATION.
BE SURE ALL INFORMATION IN BOX ON TOP OF FIRST
PACE IS COMPLETE.
-------
1-3
Study 385
Su.rvey Research Laboratory
University of Illinois
Attitudes on Air Pollution
385
2/50
f 1 AM
Time Interview Began : .
[ ] PM 7"10
Hello, is this (phone #) ? My name is ___» an<^ Ilm calling from the
Survey Research Laboratory at the University of Illinois in Urbana. We are
doing a study in the Chicago area on peoples' attitudes towards air pollution.
1. How serious do you feel air pollution is in the Chicago area? Would you say
it's . . .
Very serious, 1 11
Somewhat serious, . . . 2
Not too serious, or .....3
Not at all serious? 4
Don't know 8
2. Which do you think causes more air pollution in the Chicago areafactories
or motor vehicles such as buses, trucks, and autos?
Factories 1 12
Motor vehicles 2
Equal . . . . 3
Don't know 8
3. What in your opinion could be done to cut down on pollution from
automobiles? (Circle all that apply)
More public transportation/reduce use of cars 1 i a
Manufacturers should improve care 2 i<*
Better emissions inspections of cars ...... ?. is
Other (Specify)
4 IB
Don't knoti 8
-------
A.- Would ye«- saythat poUut-ion control devices-on cars help^cs^lmprovei air ^=i-
quality ...
Very much ...1 17
Somewhat, . . . 2
Very little, or 3
Not at all? 4
Don't know . 8
5. How much do you think the government is doing to control pollution from
cars? Would you say it's doing too much, what it should be doing, or doing
too little?
Too much 1 1B
What it should be doing ........2
Too little 3
Don't know 8
6. Do you think it's all right for people to remove pollution control devices
from their cars or should they be fined if they do so?
All right 1 19
Fined 2
Neither 3
Don't know . 8
7. Do you think it's all right for people to use leaded gas in a car that
should take unleaded or should they be fined if they do so?
All right 1 20
Fined 2
Neither ~. ... 3
Don't right 8
-------
8a. To reduce air pollution from cars, the State of Illinois is considering a
law that would require all cars, vans and pick-up trucks in the Chicago
metropolitan area to be tested for emissions. That, means that you would have
to take all cars into a designated place to be checked for pollution levels.
This inspection would check only pollution levels and not safety problems.
Each car would have to pass the test in order to get license plates. Would
you favor or oppose such a program?
Favor ...1 21
Oppose (Skip to d) . . . 2
Dependsdon't know
(Skip to e) . 8
b. How much do you think it would cost to have a car inspected for
emissions?
$ 22-23
(If $7 or more skip to Q.9)
Don't know 98
c. If the inspection cost about $9 would you favor it or oppose it?
Favor 1 21*
Oppose ....2
Don't know ......... 8
(Skip to Q.9)
d. Why would you oppose such a program?
25
26
(Skip to f)
e. What would it depend on?
; 27
- ; 28
f. How much do you think it would cost to have a car inspected for
emissions?
$ 29-30
(If $14 or less ekip to Q.9)
Don't know 98
-------
g. If the Inspection cost about $9 would you favor it or oppose it?
Favor 1 31
Oppose . . . 2
Don't knou 8
9. If an emission inspection program were set up ...
a. Should it include buses?
Yes 1 32
No 2
Don't know 8
b. Should it include large trucks?
Yes 1 33
No 2
Don't know 8
c. Should it be required statewide or only in areas where pollution is very
bad?
Statewide 1 3u
Bad areas .2
Don't know 8
d. Should the program be funded by an inspection fee or paid for by other
state funds?
Inspection fee 1 35
Other state funds 2
Both 3
Don't know . 8
e. Should the testing be done by state inspectors or through private
garages and service stations?
State 1 36
Private 2
Both 3
Don't know 8
10. Would you favor or oppose an emissions inspection program if it also included an
inspection for safety problems?
Favor 1 37
Oppose 2
DependsDon't know 8
-------
11. How many cars, vans and pickup trucks belong to you and the other members"
of your household?
vehicles
38
. .
12. (For each) What is the
make, model and year?
a. Make
b. Model
c. Tear
Don't know .
13. How many cylinders does
it have?
Six .'..
Eight
Don't know ....
14. On the average, about
how often is the
engine tuned-up?
15. When it needs engine
repairs, who usually
works on it?f Circle all
that apply)
Self (Skip to.Q.17). . .
Service station mechanic
Independent garage
Not app (Skip to Q.I 7) .
(A) Car 1
39-"»0
. 98
. . 1
, . 2 41
. . 3
. . 4
, . 8
42-«f 3
miles
( times/
year) ^
New car 97
DK . . 98
... 1 i« 5
o
... 3 ^7
. . . 7 51
None (Skip to Q.20) . . . . 0
(B) Car 2
5657
98
1
2 58
3
4
8
59-60
miles
( times/
year) ei
New car 97
DK . . 98
1 62
2 63
3 64
4 65
5 66
6 67
7 68
(C) Car 3
73-7i»
98
1
2 75
3
4
8
76-77
miles
( times
year) 7e
New car 97
UK ... 98
79/BK
l-3/Dup
1 *
2 5
3 6
4 7
5 8
6 9
7 10
(P) Car 4
15-16
98
1
2 17
3
4
8
18-19
miles
( times/
year) 20
New car 97
DK . . . 98
1 21
2 09
3 23
4 24
5 25
6 9 fi
7 27
-------
16. Are you usually
satisfied with the
repair service you
receive?
Yes
17. Has it ever been taken
in for a specific
emission inspection?
No ........
18a.Do you think it
would pass or fail an
emissions inspection if
it were done today?
Pass (Skip to Q.19)
Don't know
b. Why do you think it
would fail?
Old-Becauee of age
Other
(A) Car 1
.... 1 5
. . . . 2
.... 3
.... 1 53
.... 2
.... 8
1 sk
... 2
... 8
... 1 55
... 2
3
(B) Car 2=
1 69
2
3
1 70
2
8
1 71
2
8
1 72
2
3
CO CarJi_.
i i
2
3
1 12
2
8
1 13
2
8
I !»
2
3
. CD) Car A
1 2
2
3
1 2<
2
8
1 30
2
8
2 31
2
3
19.
On the average how much do you think it would cost to get a car fixed if it
failed to meet the inspection?
$
3231*
20,
Don't know 998
If someone's car failed inspection and the owner could not afford to make
repairs, what do you think should be done? Do you think ...
The vehicle should be taken off road
after a reasonable time, 1 3S
The vehicle should be exempt, or 2
Should state funds be available to help pay? . . 3
Other (Specify) 4
Don't know 8
-------
Now I'd like to ask you some information about your household, so that we can com-
pare your answers with the answers of others like you.
2 la. Do youxlive in the City of Chicago or outside of Chicago?
. " . Chicago (Skip to Q.22) .... 1 35
Outside ........... 2
b. In what county do you live?
Cook (Skip to Q.22) 1 37
\
DuPage (Skip to Q.22) .... 2
Kane (Skip to Q.22) 3
Lake (Skip to Q.22) 4
McHenry (Skip to Q.22) .... 5
Will (Skip to Q.22) 6
Don't knov 8
c. In or near what town do you live?
22. Including yourself, how many people currently live in this household?
38
23. Do you have a current driver's license?
Yes 1 39
No 2
24a. What is your present job title?
b. What kind of work do you do; that is, what are your duties on this job?
c. In what business or industry is this; that is, what product is made or
what service is given?
-------
23. «What is the highest grade or year in school you have completed?
None . . 00 i»2~i»3
\ Elementary ..... 01 02 03 04___05_ 06-OT 08 ,
High School 09 10 11 12
College 13 14 15 16
Some graduate school ....... 17
Graduate or professional degree ........ 18
26. In what year were you born? 19
(Code 00 for 1900 or earlier) ^"^s
27. Was your total family income before taxes last year (1979) . . .
(Repeat until "Yes"; then circle)
Less than $5,000? Yes
Less than $10,000? Yes
Less than $15,000? Yes
Less than $20,000? Yes .
Less than $25,000? Yes .
No .
Refused
Don't know 8
THANK YOU FOR YOUR COOPERATION.
. . 1
. . 2
. . 3
. . A
. . 5
. . 6
. . 7
28. SEX
Circle eex of respondent.
Male .
Female
1
2
Time Interview Ended
L J AM
[ J m
Int. ID#
Coder ID#
Check coder Wit-
Keypunch ID#
1*6
U8-63/BK
68-70
71-73
7i»-76
77-79
60/2
------- |