RESE ARC H 'TR I AN G LE INSTITUTE
October 1988
Communicating Radon Risk Effectively:
Radon Testing in Maryland
Final Report
•v .
. • ' "• Prepared for
''.' Nancy Zahedt
Program Evaluation Division
: ' .; AnnFisner
Economics Analysis Division
Office of Policy. Planning, and Evaluation
• u S Environmental Protection Agency
401 M Street. S.W.
-" _ Washmgton.-DC 20460
Prepared by
'. William H. Desvousges
.'•• V. Kerry Smith
Hillery H. Rink. IH
Center for Economics Research
Research Triangle Institute
. " • P.O. Box 12194
Research Triangle Park. NC 27709
EPA Cooperative Agreement No CR-811075
RTI Project No. 3990-02
= :CESP.X. 1-2-94 P E 5E A R
H TR1 AN 3 L £ .P AR K . ,N ORTH 'C AR 0 LI N A 27 709-2^ 94
-------
-------
Communicating.. Radon Risk Effectively:
Radon Testing in Maryland
Final Report
Prepared for
Nancy Zahedi
Program Evaluation Division
Ann Fisher
Economics Analysis Division
Dftce of Poncy. Planning, 'and Evaluation
U.S. Environmental Protection Agency
, 401 M, Street,'.S.W.
Washington.,DC" 20460 '
' ' Prepared by •
William H. Desvousges,
V. Kerry Smith
Hiliery H. Rink. Ill
Center for Economics Research
Research Triangle Institute
P.O. Box 12194
Research Tnangie Park. NC 27709
Cooperative Agreement No. CR-811075
RTI Proiect No. 3990-02
October 198?
-------
A
-------
CONTENTS
CONTENTS • i
' • • ' '-..-. ' i -' • .'• ' ,
! • L
Section ; Page
Preface : ...., iji
Figures ..• : ; I...: .' ' vii
Tables ; •. ..,...- ; ix
Executive Summary :. i xi
1 Introduction .' .... ;. 1-1
2 Study Objectives , .' '. !.'.J '..' ...'....'.., 2-1
3 Study Design :.:..-. -. i :...'..-....,:...-.• 3-1
31 Introduction .- ; •.....'. :.' .•..-."....., 3-1
3.2 Focus Groups I -...'• 3-1
3.3 Three-Community Design -.-..: :... 3-2
34 Risk Communication Dfesign ..' ....;.... :,...., .-. 3-3
• ' .3,5 Sample Design • , 3-4
' 36 WJLA Campaign :..... .}. ,'. : :... • 3-8
. 4 Data Collection : '....:...[...,...: ; ,.;..• .- 4-1
4.1 .Overview •..-. • .........I ..., .: : 4-1
4.2 Response Rates ,.-....... .i.,: ...• 4-2
! ' • .
5 Radon Awareness and Attitudes : •. .' , :...- 5-1
5 1 Introduction ,.; ; -;...' •...; .' 5-1
5.2 Radon Awareness '.}.. :.- .r :.- 5-1
5.3 Exposure to EPA Risk Communication Programs i •...•.'. .........'.; 5-8
5.4 Attitudes Toward Radon '. :j - 5-10
6 Knowledge About Radon '.... '.>..'. ,. : • . 6-1
61 Introduction ... , '..-..' ...... ...., .-:• : :..;., 6-1
• 6 2 Changes in Knowledge About Radon ! '...:... 6-1 <
7 Radon Testing ; .". ~ '....':. ..: :.....:.... .•'... 7-1
. 7.1 Introduction ::, ;...,- ,.:... 7-1
7.2 Characteristics of Radon Testing Decision :...,;.....,... '. •:.-;.: -7-1
73 Changes in Radon Testing Levels ...-. •. .....'... ' 7-2
74 Reasons for Not Testing ....- ,...,;. ".:....- :...,•. .'..-. -7-6
8 ' , Measuring Effectiveness ?.... ' /8-1.
' • . '8.1 Introduction • 8-1
8.2 Overall Effectiveness ' • .• •..: •., 8-1
8 3 Isolating the WJLA Factor :..... :i : ,..,.-... 8-6
84 Implications .,; ,...'. ." '. ' :;..-- • r. 8-14
S • References . .... ...... . : ,. •. 9-1
RTI v
-------
CONTENTS
CONTENTS (continued)
Appendix Page
A Statistical Modeling A-1 .
B Interpreting Selection Effects ^ B-1
C Background on R2 ; . _ C-1
D Testing for Differences in Proportions -.., D-1
E Tests for Sensitization Effects E-1
F Questionnaires—Screener, Baseline, Independent Followup, and Dependent
Followup : ..-. p.-j
G Risk Communication Materials : Q.-\
RTI vi
-------
PREFACE
PREFACE
This report summarizes the findings from kn evaluation of risk communication
effectiveness performed under Cooperative Agreement CR-811075. This evaluation has
involved the joint efforts of many parties including Region 3 >of the Environmental
Protection Agency (EPA', several divisions from EPA's headquarters, and the State of
Maryland. As pan of this joint effort, EPA hals prepared a separate report, Region
3/OPPE'State of Maryland Radon Risk Communication Project: An Evaluation o/
. Radon Risk Communication Approaches. The EPA report provides a more indepth
description of the risk communication outreach materials and is aimed at risk
communicators. - j - '. '
Our report highlights the practical risk communication issues. However, it takes
another step by providing an indepth evaluation; of risk communication effectiveness.
In particular, we have provided more detailed statistical analyses of risk communication
indicators—awareness, knowledge, attitudinal changes, and behavior changes. We have
also drawn on a larger data base than the EPA report—by adding a panel of Man-land
homeowners who provided both baseline and' followup interviews. Because of the
additional-data and analyses, some of our conclusions differ from those in the EPA report.
Such differences are common when complex issues are involved. To make our report more
' accessible, we have summarized the findings within the main text, and the technical
statistical issues are discussed in the appendixes. ,
This report has'benefited substantially from the people who have participated in
the joint study. In particular, we appreciate the guidance provided by the project officers
from the Program Evaluation Division iPEDl, Nancy Zahedi and Carol Deck, and other
staff. Ann Fisher and Reed Johnson of the Economic Analysis Division, which sponsored
the panel portion of the study, also provided eonsitructive comments and guidance
throughout the study. Bob O'Brien and Mel Hollander of the Statistical Policy-Branch
contributed to the statistical design and the development of the survey questionnaires.
At Research Triangle Institute, several people have played important roles in the
. study. Steve Williams developed the sampling design and helped prepare our survey plan.
Steve also coordinated and supervised the sampling activities for the survey and
contributed to the analysis. Kathy Heller helped conduct the focus groups and contributed
several creative ideas that are reflected in the risk communication materials. Nancy
Monroe helped develop survey questionnaires .than were used in the evaluation. Our
telephone survey unit arid data entry groups performed very well under difficult deadlines.
Finally, our word processing staff, especially Jan Shirley, supervisor, word processor Cathy
Boykin. and typesetters Debbie Walker and Beth Tressler did their usual superb job of
finalizing this report. ;
t
RTi
-------
-------
FIGURES
FIGURES i
Number Page
3-1 ' Three-community design '. :.... : -. 3-2
3-2 Radon message theme: Test now and be sure ;' ' 3-5
5-1 i Radon awareness ..., '...'...[...... .:..... .,.' . 5-3
5-2 Panel sample radon awareness '..... ". ..' — '5-4
5-3 Awareness sources ,...." ...I •............'. 5-7
,5-4 EPA program ,,.-. •,..'., 5-9
5-5 Attitudes—importance of testing for radon J.... •• • -5-13'
5-6 Attitudes—mitigation cost....:,....... •... ! '....-....'. 5-14
5-7 Attitudes—mitigation effectiveness -1..'. '. 5-15
• 6-1 Key'knowledge questions ..;.. ..: 6-4
6-2 Panel sample1 radon knowledge , ':,....'. ......" ,, 6-6
7-1 Radon testing . ......,: : i : -. 7-4
- 7-2 . Informal communication channels ". ! •.-. ,.• 7-5
7-3 .. Reasons for not testing i...,. : 7-7
7-4 Knowledge questions answered correctly by nontester groups';.,.: 7-8
8-1 Overall performance ; ! '.........• \.'....:.. . 8--2
8-2 Awareness of radon in both surveys •....;...]>. '...-. 8-8
8-3 Correct answer probability—testing home for radon ......[ 8-10
8-4 Probability of correct answer—mitigation knowledge L. • 8-11
8-5 Testing probability ..'. ;.....,..., .'- : ;. , 8-12 _'
8-6 Discussing radon in followup ;........ ,. , 8-13
RTI vii
-------
-------
TABLES
TABLES
Number Page
3-1 Sampling Design ,. '. ........; 3-6
3-2 The Ideal Evaluation of a Risk Communication Campaign 3-8
4-1 Response Rates for Maryland Baseline Survey ;..' 4-3
4-2 Response Rates for Maryland Independent Followup Survey 4-4
4-3 . Demographic Profile of Respondents .; 4-6
4-4 Demographic Variables in Percent I „ 4-7
5-1 Changes in Awareness': Statistical Test Results ..' 5-4'
5-2 . Recursive Awareness Models '. i •.... . 5-6
5-3 Changes in Attitudes About Radon: Statistical Test Results ..........: 5-11
6-1 Knowledge About Radon by Town , -. ::....:. ........-....:... 6-2
6-2 Knowledge Models Followup Surveys '. ;.. 6-8
8:1 " Evaluations of Public information Programs ......;. , 8-5
RTI ix
-------
-------
Executive Summary
/RTI
-------
-------
EXECUTIVE SUMMARY
EXECUTIVE
SUMMARY
Research Design
The U.S. Environmental Protection Agency "(EPAI has become increasingly
concerned about the threat to health posed by'radon gas in homes. New evidence has
emerged that shows radon to be more pervasive than originally thought, with high lewis
found in North Dakota, Minnesota, and several midwestem stated As a result of this new
evidence, the EPA and the Public Health Service recently issued a joint announcement
.that included a health advisory from the Public Health Service about radon. In addition,
EPA has developed a risk communication program to inform people of the risks from
radon and to help them make decisions aboutithe necessity of mitigation.
As part of the radon risk communication program, EPA's Office of Policy, Planning,
and Evaluation; the EPA Region ffl; the Maryland Department of the Environment; and
Research Triangle Institute jointly conducted trus study which developed and evaluated
practical risk communication methods that could be adopted by other regions and states.
The .community-based program, which emphasized local communication activities—
radon awareness week, posters, and presentations—supplemented by media, showed
increased awareness, knowledge, and attitudes that are more likely to lead,'to future
. testing. The level of testing increased about 10! percentage.points from 5 to 15 percent
of the target population, but a 15 percent overall level of testing is still relatively low. From
a marketing perspective, these changes are quite large given the modest level of
communication resources and the 3-month time frame. However, from a public health
viewpoint, the increase in testing is not commensurate with the risk.
This study used a three-community design to test and evaluate the risk
communication materials. The three Maryland communities chosen for the study all
had high reported levels of radon and were similar in socioeconomic terms. Hagerstown
received an integrated but modest media campaign—radio and print public service
announcements and a utility bill insert; Frederick received the same media campaign
plus a community outreach program that included presentations, posters, and related
activities. The project team designed, developed, and produced the communication
materials after conducting a series of focus groups in Maryland. Randallstown served as
the comparison community and received no special radon information.
' We conducted a baseline survey of homeowners in each city during December 19S"7
to gauge initial attitudes, knowledge, and levels of testing. The telephone survey used
a random digit dial technique. Response rates ranged from 48 to 64 percent. In April 1988
another telephone survey followed after the campaign ended in each city. We interviewed
respondents from the baseline survey and an independent sample of homeowners in each
city. Response rates for the panel sample followup were approximately 80 percent, while
those for the independent followup- ranged 'from about 55 to 70 percent. While the
response rates are below what is often desired for developing accurate generalizations
RTl
-------
EXECUTIVE SUMMARY
for each city, the consistency- between surveys and comparisons of demographic variables'
with census data suggest that our respondents are typical of those described by the 19SO
census.
Our research design offers several advantages. The panel component includes "before"
and "after" observations for the same household and therefore permits control for
interpersonal differences in analyzing responses to the program. This format allows the
changes in the risk communication effectiveness indicators to be measured with greater
statistical power. The independent component of the study, however, enables us to test
for sensitization bias resulting from, reinterviewing the same person who may have
responded more to the program message because of being interviewed for the baseline.
With any research design, potentially confounding factors may occur. During this
evaluation, WJLA, an area television station, ran an extensive television campaign to
inform the public about radon that also included radon monitors being made available
at reduced prices through a major grocery store chain. As noted in U.S. EPA [1988], this
campaign was most influential in Frederick. While this campaign complicates our ability
to evaluate the effectiveness of the community approach, it was, after all, a risk
communication exercise. The WJLA campaign unintentionally augmented the
community approach that was used in Frederick. Thus, the WJLA campaign complicates
what can be said about the practical community campaign that was designed, but it does
not prevent us from drawing the more general risk-communication inferences.
Awareness
A risk communication program cannot be effective unless it increases the target
population's awareness of the risk. Frederick, which received both the media campaign
• and the community outreach, had the highest level of radon awareness in the followup
survey, at 90 percent—an increase of almost 20 percentage points. This level of awareness
was approximately 10 percentage points higher than in Hagerstown and 20 percentage
points higher than in Randallstown. Frederick also had the lowest percentage, 7 percent.
of respondents who had not seen or heard about radon in either survey. Even after control-
ling for differing baseline levels of awareness, Frederick still showed a significantly higher
level of awareness in the followup survey. • •
• Respondents who were aware of radon mentioned newspapers and television most
often as sources for radon information. About 75 percent of these people in the baseline
and both followup surveys from Frederick and Hagerstown cited newspapers as an
information source, while approximately 65 percent of those in Randallstown did.
Television showed.the greatest increase as a source of radon information between the
baseline and followup surveys. References to television in the two treatment communities
increased from about 55 percent in the baseline survey to around 75 percent in the
independent followup and 68 percent in the panel followup. The public information
campaign conducted by WJLA-TV at the same time as the EPA project is largely
responsible for the increase in the citations of television as a source of information. WJLA's
campaign consisted of public service announcements, promotion of their intensive
coverage of radon on newscasts and a prime time special, and announcements about the
availability of radon test kits at a reduced price through local supermarkets. Because of
• the coincident timing, we have attempted to control for the WJLA effects. We also consider
the WJLA efforts as pan of a broader risk communication effort in evaluating overall
effectiveness.
RTI
-------
EXECUTIVE SUMMARY
• • .'. Features of EPA's program included utility bill inserts, posters, and promotion of the ;
I Maryland State Radon Hotline. Approximately 25 percent of panel sample respondents
in the 'two treatment"communities recalled seeing a utility bill insert, while 19 percent
of the independent followup respondents in Frederick did and 14 percent in Hagerstown
I did. There was also a relatively large number of respondents in the control community.
. , who incorrectly reported that they had seen a utility bill insert—11 percent in the
' independent followup and 18 percent in the panel foljlowup. The respondents were either
j . mistaken or had seen 'a utility bill insert in another community because the power
) company serving RandaUstown was not used m the ;study. However, less than 10 percent
of the people recalled seeing posters.or said they called the state toll-free number.
' Statistical tests for sensitization indicated that there was no substantial difference in
" awareness between the panel and independent saniples.
Attitudes '.'•';'
« Three questions dealing with attitudes were included in the survey questionnaire:
• 'It is important to test my home to find out if I have a radon problem."
• "If I had a radon problem, it would be costly to fix." '
• • "Even if a radon problem was fixed, my home would still be worth
\' a lot less" | . '
Respondents were asked to what extent thpy agreed or disagreed with these
- ' statements. These three themes correspond to the risk communicationi messages which
emphasize the importance of testing, that mitigation need not be costly, and that.
mitigation can be effective. If the risk communication messages were processed correctly,
' the proportion agreeing with the first would increase, and the proportion disagreeing with
the second and third would increase as well. ' • \. ,
Statistical tests of differences between the baseline and followup responses for the
independent sample indicate that Frederick respondents ,in the independent followup
sample exhibited more informed attitudes for all| three questions than did Frederick
respondents in the baseline. Hagerstown respondents in the independent followup were
more informed with regard to cost of mitigation and its effectiveness than their baseline
counterparts, while Randallstown showed no statistical difference between the baseline
and independent followup surveys. The panel sample also showed that people in both
Frederick and Hagerstown who have children were more likely to change their attitudes
' • - about the importance of testing compared to people in Randallstown.
What are the general attitudes about radon after the program? A sizable fraction |at,
least 75 percent1, of both testers and nontesters in «ach town felt that it Was important
to test their homes for radon. However, many believe that radon mitigation will be
expensive. Despite improvements in both treatment communities, almost half of the
• -testers and more than half of the nontesters still felt; mitigation would be expensive. Both
testers and nontesters in Frederick and Hagerstowri understood the message concerning
mitigation effectiveness while there was essentially no change in Randallstown. A sizable
maionty in Randallstown still felt that even if a radon problem were fixed, their home
would still be' worth less. Even in the treatment; communities,,35 to 40 percent of
. 'respondents either have not understood the mitigation message or do not agree with it.
RTI -
-------
EXECUTIVE SUMMARY
Knowledge
Testing
Both baseline.and followup questionnaires included a seven-question radpn "quiz"
presented in a multiple-choice format. The improved performance on questions covering
general knowledge about testing, health risks, and mitigation in the followup surveys
serves as an indicator of program effectiveness. Frederick had much higher levels of
knowledge in the followup surveys than did Randallstown, with levels averaging from
15 to 20 percentage points higher. Hagerstown also performed better than Randallstown,
but not as well as Frederick. The biggest improvement in knowledge came with the
question "How can one test for radon?" This was one of the main messages EPA tried
to communicate in its information program as well as the WJLA campaign. Frederick
showed increases of around 40 percentage points in both followup survey's for this
question, while Hagerstown and Randallstown both showed increases of about 20
percentage points in the independent followup and 30 percentage points in the dependent
followup. Baseline performance on this question was roughly the same in the three cities.
Testers did 15 to 20 percentage points better on knowledge questions than did
nontesters. However, we are not able to prove causation in either direction. Higher
knowledge could have led people to test, or they could have gained more knowledge as
a result of testing their homes. One conclusion that we can draw is that the information
materials conveyed general facts better than they did more technical information.
Between 70 and 90 percent of Frederick respondents in the followup surveys correctly
answered general-knowledge questions, with answers saying that radon occurs naturally
and has no odor. However, only 40 to 45 percent of the respondents in Frederick in the
followup survey correctly answered questions dealing with more technical issues, such
as what affects the level of radon measured. Hagerstown knowledge levels were about
lO.percentage points lower for these indicators, and levels in Randallstown were on the
order of 20 percentage points lower.
. We also modeled knowledge as a function of program components, socioeconomic
factors, and other elements such as WJLA's radon campaign. Frederick and Hagerstown
homeowners both performed better on the knowledge questions than did Randallstown
homeowners, even after controlling for exposure to the WJLA campaign and various
socioeconomic factors. People who had called the State Radon, Hotline, had talked to
someone about radon, or had recalled seeing radon on television scored higher on the
knowledge quiz than those who had not. With respect to demographics, whites and males
scored better than blacks and females. Also, the better educated and higher income
respondents knew more about radon than those with less education or income.
We did detect some sensitization in the area of knowledge for the panel sample in
Hagerstown and Randallstown. It appears that the baseline interview may have
preconditioned people in those two cities. Frederick's higher initial levels of knowledge ,
may explain why there was no sensitiration detected in that city.
One could gauge the ultimate success of the EPA information program by seeing
a large increase in the number of people testing their homes for radon. Likewise, one could"
argue that the program was a success if people made an informed decision not to test
; In our study we attempted to look at both aspects' of the testing situation. Testing levels
tripled in Frederick. Because of.low initial fractions of the population, this increase was'
RTI '. .
-------
EXECUTIVE SUMMARY
about 10 percent'of the households. We detected no change in Hagerstown or
Randallstown. Several points need to be made witla respect to these levels. First, the time-
frame of this study was only 3 months, which sugge<;ts that respondents may not have
had time to complete the decision cycle, especially since it involved a relatively unfamiliar
risk. Second, marketing campaigns are often deemed successful if they increase market
share by fewe'r than 10 percentage points. Thus, the increase in testing in Frederick points
to the effectiveness of an integrated media and community outreach campaign.
Whether a respondent had talked with, others; about radon emerged as a key variable
in the testing models. Over three-fourths of the testers in Frederick and Hagerstown had
talked to someone about radon. In contrast, only 49 percent and 38 percent, respectively.
of nontesters in Frederick and Hagertown had. This finding reinforces the importance
of informal communication channels in promoting radon testing. The program in
Frederick contained community outreach that made use of these informal channels and
thus helped lead to increased levels of testing in that community. Neither income, age,"
nor smoking emerged as significant in the testing decision.
We also examined the reasons people gave for not testing their homes. Using the
survey responses, we placed the reasons people'gave into categories—procrastinator.
uninformed, not in my home, not in my area, antii:ester, and fatalist—and modeled these
categories on various factors. As shown in"Figure 1, our findings show that
• Overall, the decreases in the "not in area" numbers in both Frederick (14 per-
centage pointsl and Hagerstown [7 percentage points! suggest that the programs
helped some people personalize the risk'from radon on the community level.
However, they have not personalized the; risk completely, as evidenced by the
• increase of about 4 percentage points in the category "not in my home.",
• In Frederick the "not in area" category depressed by 14 percentage points" This
• compares to only 6 percentage points in'both Randallstown and Hagerstown.
Proportions were similar between the independent and dependent samples.
• The number of Frederick respondents in the procrastinator category increased
by 12 percentage points in the panel sample and by 18 percentage points in the
independent sample. In Hagerstown, increases were 19 percentage points for the
panel sample and 13 percentagexpoints for the independent sample. Randallstown
showed increases of 13 percentage points and 6 percentage points in the panel and
independent samples, respectively. ' ' ' '
• There were no large differences among 'the percentages 'of nontesters in the
uninformed category—each city showed decreases of about 10 percentage points.
However, Frederick had a lower initial proportion of uninformed nontesters in the
baseline-—6 percentage points lower than either Hagerstown or Randallstown.
i
• Recall of WJLA's campaign and a person's educational level helped explain
movement from uninformed to procrastinator (the person said they lust had not
gotten around to testingl Living in Hagerstown relative to Randallstown and
talking to someone, about radon influenobd the mow from not in home area to •
, procrastinator.
There appears to have been sensitization of the panel sample in Hagerstown and
n with regard to reasons for not testing. The proportion of procrastinators
RT1
-------
EXECUTIVE SUMMARY
WHY I HAVEN'T TESTED: CHANGES BETWEEN
BASELINE AND FOLLOWUP SURVEYS
FREDERICK
Unmfonned
Not In Ana
Main HOCK
Proeutiauon
HAGERSTQWN
Uninfonned
NtxtoAret
Not In Home
Frecmiiauon
o IASEUKISLTVEY Q jiim.stt.yr HXLOWLT •
Figure 1. Reasons for not testing.
RTI -.
-------
EXECUTIVE SUMMARY
in Hagerstown increased by 19 percentage points in the.-panel sample, but -by. only 13.
percentage points in the independent sample. Likewise, the proportion of procrastiriators
in Randallstown increased by 13 percentage points in the panel sample, but by only 6
percentage points in-the independent follow-up. These differences are statistically
significant. ' . /. , : .
However, for the testing decision itself, we did not find any sensitiration among
members of the panel sample. In fact, the testing increases in Frederick were slightly
higher for the independent sample. . j
The WJLA factor
To evaluate the effectiveness of the EPA program, we tried to isolate the EPA program
effects from other, activities. The foremost of these was WILA-TV's independent radon
awareness campaign, which ran concurrently with our study. We used statistical models
to simulate what changes would have occurred with and without, the EPA program, as
well as with and without the WTLA campaign. Ws also used a high and low measure for
the campaign. Our results from these simulations are suggestive, but not definitive. The
high and low measures for the WJLA campaign dip change its relative influence. At this
time, our results suggest that the WJLA campaign probably affected homeowners'
awareness, knowledge, and attitudes. Our simulation models suggest that the WTLA
campaign had fewerdirect effects on the level ofjradon testing. . ' • . '
The effects of the EPA program and those of the WJLA campaign may have worked
together to produce effects greater than either would have alone. We have already
mentioned that television and community outreach campaigns are complementary.
Thus, there may have been some synergism between the WTLA program and the EPA
program. ! •' ' .
Overall Effectiveness
For an overall evaluation of effectiveness, Figure 2 highlights the changes in.the four
main indicators. The figure compares each treatment community with the comparison
community, Randallstown. The results focus only on the independent sample, but the
general pattern holds true for the dependent sample also. Frederick, which received both
the media campaign and the community outreach, performs better than Randallstown
in terms of increased awareness ;6 percentage points higher^, attitudinal changes ,15
percentage points higher1, knowledge ;15 percentage points higher', and testing ;8
percen,tage points higher1. The changes in indicators for Hagerstown, which received only
the media program, were much more limited with attitudes the only measure to display
a significant improvement [8 percentage points! qbmpared to Randallstown. There were
no significant changes in the summary measures of awareness, knowledge, or testing
for Hagerstown relative to Randallstown. • • : j :
"In addition to these specific results related to the relative effectiveness of the two
variations on the EPA program, we also observed some general patterns about radon
communication issues. There were high lewis of awareness of radon in each city in the
followups. with over."0 percent of respondents saying that they had seen or heard about
RTI
-------
EXECUTIVE SUMMARY
Ji-
ll 30
£ =
II »t
D-2 10-
>£ ~
40-
30
x =
it20
II 10
OVERALL PERFORMANCE:
FREDERICK VERSUS RANDALLSTOWN
37
18
12
17
0.8
Awareness
Altitudes
Knowledge
Testing
OVERALL PERFORMANCE:
HAGERSTOWN VERSUS RANDALLSTOWN
13
10
0.5
0.8
Awareness
Attitudes
Knowledge
D Hagerstown B Randallsiown
Testing '
Figure 2. Overall performance.
radon. People also apparently realize that it is important to test. They see mitigation as
expensive, but a majority feel that it would lead to restored property values. Nonetheless.
there remains a sizable percentage of people. between 30 and 40 percent in Frederick and
Hagerstown. who feel that their property would be worth less even if a radon problem
. were fixed. ' . ' • '' ' '. ' .
The program was effective in conveying general knowledge about radon, such as what
it is and how to test one's home. These were themes stressed in the risk messages. In
addition, our results show that the respondents in all three, towns did not know what
affects the levels of radon measured and what they can do to mitigate their risk from
radon. Neither of these themes was addressed in the risk messages.
During the 3 months of the programs, testing levels reached about 15 percent in
Frederick and remained at about 5 percent in Hagerstown and in the comparison
community of Randallstown. While these levels are not as high as one might hope from
a public health perspective they are not unreasonable for a practical short-tern:
RTI
-------
EXECUTIVE SUMMARY
''"'''.'. communication program: They are also not out of line with testing levels elsewhere. Even-
| in New jersey, where radon has been a maior concern for several years, only 10 to lb<
percent of the "homeowners have tested lor radon. :j
.' . Changes in the reasons why people have not tested for radon also show some program
j . effects The proportion of procrastinators has, risen significantly, while the proportion
, ' 0fimiriformednontesteishasdecreased.ThenumterofpedplewhostillbeHevethatthey
' " . do not need to test because radon is not problem m thf?ir area has decreased. But about
; - 10 percent of Homesters in the treatment communities still feel that radon is not a
1 ' ' . ., problem in their area. Since the risk messages stressed personalization at the city level,
these people could be regarded as still possessing anjuninformed attitude; Approximately
| one-fourth of the nomesters still say that radon is not a problem in their home. Such a
large number indicates that influencing people to personalize the risk to that last level,
• the home, will be very, difficult and will require even more innovative means of risk -
communication, or some other type of initiative.;
Implications i
From an overall risk communication perspective, one can view the Frederick
Program as including community outreach and. presentations, radio public service
announcements, utility bill inserts, and the extensive WJLA campaign. After all of this
• . risk communication, testing in Frederick only increased by about 10 percentage points.
At the same time, however, the combined Frederick program may have improved
conditions for increased testing later. The increasesin the number of people feeling.that
. . h is important to test and those who have procrastinated about testing, and inthe
- - • proportion of respondents having more knowledge about radon, all support this
• • ' 'hypothesis. The increased levels of people in Frederick talking to each other also had a
positive influence on testing and more informedatititudes. However, reaching high lewis
of testmg with a modest risk communication progi-am is unlikely. Even .if all of the
procrastinators in Frederick eventually did test, testing levels would only increase to
around35 percent. . : , , , i. i -i
Based on the results in Hagerstown, there appears to be relatively little payort to a
' smaltscale media effort alone. There was less exposure to WJLA in this town.' Radio and
••••-'• newspaper public service announcements and utility bill inserts served as the main nsk
communication vehicles. There were modest improvements from baseline levels in the
areas of awareness, knowledge, and more informed attitudes; however, the differences
were not very much larger than in Randallstowni, which received no treatment.
Our study shows that risk communicators still do not know how to convince people
to test for radon. People are uncomfortable with mitigation and its potential effectiveness.
Reducing this concern is likely to-require more ^idespread certification of mitigation
'• " contractors and more communication about effectiveness. The community-based
programs.suchastheoneinFrederick,offeragcxxl;whiclefor''vTOrd<)f-mouth''rnitigation
success stories to be passed along. Radon mitigation needs to be perceived as being in
the .mainstream of normal home repairs. The formal communication messages alone-
such as the recent press announcements—are unlikely to achieve this goal.
•\ topic only touched on in our study is the is'sue of property value and how it relates
-'••-. . ' to the testing decision. Perhaps it will require sounding the alarm to homeowners that
RTI
-------
EXECUTIVE SUMMARY
testing now could make it easier to sell a house later. A communications message might
emphasize that radon tests may become a required pan of all real estate transactions.
Such issues were not addressed in our study because the Agency felt that it was unclear
whether it was within their jurisdiction to be providing information or raising concerns
in these areas. This in turn implies that states may need to rethink the boundaries they
place on their radon programs with regard to what types of information these agencies
can provide.
Finally, is the recent Public Health Service Advisory strong enough to increase testing
on its own? Testing levels could be influenced more by a public health message that may
overstate the risk to some in hopes of erring on the safe side. However, based on our results,
we think the answer is no. Even in Frederick, where residents had community
presentations, unsolicited mailings, and coverage on both television and radio, testing
levels increased to rather modest levels. The messages heard in Frederick emphasized
the health risk from radon and that high levels had indeed been found in that city. Still.
85 percent have not tested. This suggests that the Advisory may lead to a short-lived
increase in testing, but that its effects will soon fade away'without a comprehensive
communications strategy for radon. , ' '
Because of the difficulty in getting people to test their homes for radon, we
recommend that a radon risk reduction strategy include the following elements:
• Legislative action by states to require radon:testing for all property
transactions . • . '
. • Building code changes at the state and local levels to make new construction
"radon free" ...
• Increased certification of radon mitigation contractors : - - - -
• A comprehensive risk communication program that identifies national concerns
about radon testing and mitigation and provides communications assistance to
•state and local governments for organizing community-based communications
efforts.
A comprehensive strategy is needed because none of these elements alone will lead
to radon testing levels and mitigation that are commensurate with the risk. Increased
legislative actions to require more testing will only work if people know how to test, where
to get test kits, how to interpret the risks implied by their test results, and how to make
sound mitigation decisions. A comprehensive risk communication is essential to making
these actions happen with a minimum of hassle and anxiety. At the present time, we
see only small pans of such a plan in EPA's current risk communication activities.
Such a comprehensive program would require substantial commitment and
resources. However, it would be much'less expensive than a direct regulatory approach.
Given the widespread concern about the risks from radon, such a program seems
appropriate.
RTI
-------
Section 1
Section 2
Section 3
Section 4
Section 5
Section 6
Section 7
Section 8
Section 9
Introduction
Objectives
Study Design
Data Collection
i
Radon Awareness
and Attitudes
Knowledge About Radon
Radon Testing, !
Measuring Effectiveness
References
/RTI
-------
-------
1. INTRODUCTION
SECTION 1
INTRODUCTION
Environmental Protection Agency (EPA) Administrator Lee Thomas has made
radon a main target of the Agency's efforts io manage environmental risks.
Radon's physical characteristics, however, complicate the development of an.
effective policy to manage its risks. Radon is 'a radioactive gas caused by the
natural breakdown of uranium in the underlying ground. It is not perceptible
to the senses—you cannot see, smell, or taste it. Because radon occurs nat-
urally and people experience its risks primarily in their homes (or buildings),
EPA's conventional regulatory approaches are unlikely to be appropriate The
Office of Radiation Programs is charged with implementing EPA's radon
policy. Its Director, Richard Guimond, has described the Agency's role as
facilitating the flow of information about radon. (See Coyle and Drachler
[1985].) EPA's efforts include providing information and technical assistance
to the States, setting guidelines to assist homeowners in determining whether
mitigation actions are needed, and conducting engineering research to find
ways to effectively reduce radon concentrations in homes Although EPA has
maae some'progress in managing the risk from radon, a recent Agency eval-
uation (US Environmental Protection Agency [1987]) rated radon as a high
risk/low effort area •'!-.' • '
I • •• , •• t •
The lack of clear regulatory authority accouHts for some of the low effort on
radon .Limited funding for radon activities Has also hindered the Agency's
progress (Knmsky and Plough [1988]). Finally, the Agency may not.have
anticipated some of the difficulties in using an information-based approach to
radon risk management Such an approach differs significantly from the
majority of EPA's regulatory efforts, which u'sually set maximum emission
levels from some type of pollution source. !A comprehensive information
'program for radon would involve motivating people to test their homes, pro-
viding .information about the reliability of tests, helping people understand
their risks, developing and evaluating mitigsition alternatives, and certifying
contractors who could perform mitigation effectively ; .
-In this study, we address one important aspect of such an overall program:
How car EPA or .other agencies motivate people to test their homes .for
radon0 Given the magnitude of the risk from Yadon exposure, the answer to
this question should be relatively simple. However, some early studies
indicate that relatively few homeowners halve tested their homes In the
Reacrng Prong area o* New Jersey, only 10 to 15 percent of residents had
RTI 1-1
-------
1. INTRODUCTION
tested their homes (Newsweek [1988]). This low testing occurred in spite of
high radon levels being found several years ago.
This study evaluates the effectiveness of two alternatives for motivating
people to test for radon:
• A media program that also includes a utility bill insert mailing;
• The same media program reinforced with a community-based outreach
effort—community presentations, slide shows, and a radon awareness
week.
Both alternatives involve relatively modest communication expenditures arid.
were designed to be adapted for use in other states. Nevertheless, the com-
munity-based EPA program did require more staff time for logistics and other
activities than the media campaign. The issue we consider is whether the
communication alternatives yield additional benefits in terms of increased
levels of radon testing. We also consider other indicators of effectiveness
including changes in awareness, knowledge, and attitudes •
The remainder of this report is organized into eight sections and six
appendixes:
•*
• Section 2 describes the study objectives and the rationale underlying
them; . • .
• Section 3 highlights the study design issues: ...
• • Section 4 summarizes the data collection activities along with the
demographic profile of respondents;
• Section 5 presents the empirical results for changes in radon awareness
and attitudes:
• Section 6 describes the empirical results for changes in knowledge about
radon:
• Section 7 highlights the findings for radon testing and reasons people
gave for not testing their homes:
• Section 8 examines the overall effectiveness of the risk communication
program and develops the public policy implications from the study:
• Section 9 contains references cited in this report: .
'• Appendix A describes statistical modeling; • ,
• Appendix B discusses interpreting selection effects:
• Appendix C gives a description of the coefficient of determination, more
commonly known as R2, for models involving qualitative dependent
variables.
RTI 1-2
-------
1. INTRODUCTION
Appendix D show.s'the'results of testsjfor differences, in proportions-;
between the baseline and independent fpllowup surveys in the areas of
awareness of and changes in attitudes abjout radon;
Appendix E discusses tests for sensitization effects: and •
Appendix F contains the baseline screen'er sind the three questionnaires
used in this study: Baseline, independent Followup, and Dependent
Followup. • . ',
/ '••.• '
Appendix G highlights some of the risk communication materials that
were developed as part of this study. , '
RTI 1.-3
-------
-------
2. STUDY OBJECTIVES
SECTION 2
STUDY OBJECTIVES
The overall objective is an important guidepost for any study. The primary
objective of this study is |
; To identify and analyze the effectiveness of selected outreach
methods in communicating the need to test for radon in
communities with high radon levels, i .
-'" 1 ' '
This objective implies .that our activities must address many facets of risk
communication. Among the most important are
• Developing messages to communicate the risk from radon exposure,
• Adapting outreach methods to communicate messages,
• Designing a plan for evaluating risk communication effectiveness,
• Implementing the effectiveness evaluation, and
• Developing the' implications for future radon risk communication activities.
To be relevant to other states and communities that want to implement a risk
communication program, our study has focused on outreach methods and-
strategies that are practical. For example,! we did not pursue an expensive
television campaign because it is unlikely that states or communities could af-
ford to implement such a program on their own. ,
Our study also has had to reflect the time constraints imposed by the policy
process For example, because of the desiVe of the Region III office to have
results before the 1988-1989 winter heating season, we measured the
changes in radon testing rates before some^of the people exposed to the risk
communication messages could be expecte1 d 'to have made a decision about
testing Fortunately, radon tests are inexpensive and relatively easy to
perform, which may make it easier to observe changes. Additionally, the
mes'sages were disseminated at one of the1 best times of the year to test. A
longer term appraisal would be necessary to observe the full effects of the risk
communication efforts. For now, we recogriize that we will'be taking an early
snapshot of what is a long-term dynamic proceiss.
A further implication of our primary objective is to develop supplemental
indicators of effectiveness that are easy to measure in a short time interval.
RTI 2-1
-------
2. STUDY OBJECTIVES
For example, changes in awareness or intentions to test reflect people's eval-
uation of the risks from radon exposure. Other supplemental indicators used
to measure effectiveness include changes in knowledge and attitudes about"
radon. ;
Four supplemental objectives follow directly from our primary objective.
These objectives are to ,
*" ,
• Identify who is now testing and factors that motivated testing,
• Identify a few effective and practical radon risk communication methods,
• Develop a limited number of prototype communication materials, and
• Evaluate effectiveness of prototype materials and methods.
The remainder of this report describes the study design and findings that ad-'
dress all the objectives.
RTI 2-2
-------
3. STUDY DESIGN
SECTION 3
STUDY DESIGN
3.1 Introduction
This study has involved a wide range of risk communication activi-
ties—developing risk messages, designing arid implementing communication
strategies and evaluating their effectiveness!, The study design provides a
basic framework for organizing and integrating these diverse activities.'
Because the activities are described in some;detail in a companion study by
EPA's Program Evaluation Division (U.S. Environmental Protection Agency
[1988]), our report provides a broader overview. .
Our study design includes,two types of risk communication evaluations:
qualitative and quantitative. The qualitative .evaluation involved a series of
focus groups dealing with radon risk communication., The quantitative evalua-
tion involves the use of survey methods to appraise effectiveness. Although
we have drawn oh some findings from the focus groups to help illustrate, our
quantitative findings, this report focuses mainly on the quantitative evaluation.
The study design highlights four main parts of the risk communication
evaluation: the use of focus groups, the three community design, the risk
communication program itself, and the. sampling plan. Each of these is
discussed below. ' ! ' "'•--.
3.2 Focus Groups
Focus groups, or small group discussions led by an experienced moderator,
played several key .roles in the development of risk communication materials.'
We conducted four focus groups in two Maryland communities—Westminster
and Hagerstown—-in the fall of 1987.* ;
The first two sessions held in Westminster, Maryland, compared a group of
people who had tested their homes for radon with a group of nontesters
These sessions examined factors that influenced people to test, potential
barriers to testing, and other related factors.; The most important factor that
emerged from these groups was the need to help people personalize the risk
from "radon The second group of sessions in Hagerstown focused on
'Summaries of these sessions are found in Desvousges and Rink [i9B7a,b].
RT! 3-1
-------
3. bJUU.Y DcjjIlaN
evaluating preliminary risk communication materials that included a leaflet,
posters, and radio public service announcements, -
3.3 Three-Community Design
The second part of the study involved the design for evaluating effectiveness.
We use a three-community design to measure the effectiveness of our radon
risk communication efforts Figure 3^1. illustrates the basic design. Hagers-
town, Maryland received an integrated radon risks/testing media campaign
during the winter heating season. Frederick, Maryland received the same
media campaign plus a community outreach program that included
presentations, posters, and related types of activities (The treatments are
discussed more fully in the next section.) Randallstown, Maryland served as
the comparison community and received no special radon risk information.
These three communities were chosen because they are high-risk areas and
the comparison community has media markets that are separate from the
experimental communities. Randallstown has two features that make it an
effective comparison community:.
• Its population size and general character are similar to the treatment
communities. ,
• It is more likely to receive the normal flow of radon information than an
isolated community. , ,
However, there are some important differences between Randallstown and
the treatment communities, especially Hagerstown. On average, Randalls-
town's residents are wealthier, are better educated, and live in more
expensive homes Because our study design incorporates both panel and
independent sample features, each community can serve as its own reference
Communicating Radon Risk Effectively:
Radon Testing and Related Measures
in Three Communities
Hagerstown, MD
Media only
Frederick, MD
Media plus
community program
Randallstown, MD
No experimental
treatment (comparison)
Figure 3-1. Three-community design.
RTI 3-2
-------
3. STUDY DESIGN
group' Therefore, the higher income and education levels should not affect
our ability to measure the treatment effects For example, changes in
awareness of radon, or the levelpf testing, will be measured using three
samples for each community. In this w.ay each community serves as the
reference point for measuring changes. The Relative size of treatment effects
can be measured by comparing the changes that occur across the three
communities. • !
The randomization precepts underlying the theory of experimental design
cannot be achieved in most social settings! This study is no exception. In
such studies, where equivalent comparison! groups cannot be achieved by
randomization, substitute methods are ofteVi employed to approximate the
randomization. These, methods are referred to as quasi-experimentation in
which attempts are made to isolate treatment effects. Potential confounding-
factors that may complicate these attempts include
• The change in awareness, attitudes, and behavior that would occur during
the study period without the treatment being implemented.
• The sensitization effect of the baseline survey interview for the panel
sample •' . . . ' :i' '""",'•. :
,...•" The study design addresses these potential effects by comparing the
,' . treatment samples with a closely matched comparison, sample to isolate
changes in awareness that would have occurred without our special actions
-. Additionally, the desigri~uses separate, independent samples for the "before
; . and after measures" to eliminate the preconditioning bias. Finally, the two
treatment effects are compared between communities that share media areas
in order to estimate the incremental effedt of the community programs
component. . • . v •!.''.' . .
3.4 Risk Communication Design i
Developing communication messages that will motivate people to test their
homes-for radon is an important part of our research activities Our activities
in this area included conducting focus groups to evaluate factors that
influence testing (Desvousges and Rink [I987a,bj). developing and pretesting
new communication materials, developing'a media and community-based
communication plan, and implementing thdse plans. These materials are
intended to serve as pilots for adaptation to other states and Communities if
they proved effective in Maryland. ;-.. " " •
- '. The heart of risk communication is the messiages>that are communicated to
people The message content must motivate and overcome resistance to
change (Manoff [1982]) An effective message must tackle four basic
Ccnte-
it—W^at -s t^e target audience7 What are their resistance points9
W-.a: are tne a'esirec actions9 I.'."'
RTI 3-3
-------
3. STUDY DESIGN
• Design—What is the basic single idea? What is the best style? Can
people identify with the message?
• Persuasion—Will the message arouse concern? empathy? Is it believa-
ble?
• Memorability—Is the message reinforced? Are distractions minimized? Is
• it repeated?
. (Manoff[1982],p. 188)
Experiences in other publi.c service advertising campaigns suggest that
personal contacts with media, high-quality materials, donated air time, and
positive messages are key ingredients in an effective campaign.
Figure 3-2 highlights the message theme: "Test now and be sure." The
theme is short (five words), emphasizes desired actions, and provides a"
positive message. Figure 3-2 also identifies the messages, media, and target
audiences for encouraging radon testing. The plan shows print and radio as
the primary media. Ideally, television would have been used to reinforce the
messages as many ways as possible. Regardless of the media chosen, the
messages should be consistent and take advantage of the usefulness of the
panjcular media.
The second part of the communication design involved a community-oriented
communications strategy. The strategy included presentations to community
organizations, poster and brochure placement at various Community
.locations, and a focal point for the campaign: a radon awareness week. The
• design calls for the same basic messages as in the media plan to be used in
the.community-based efforts. Appendix G highlights these efforts. For a
more indepth discussion, see U.S. EPA [1988].
3.5 Sample Design
The sample design plays two important parts in the study design. First, it
must provide for a sample of households who are representative of the target
population. Second, it must enable effectiveness to be measured with a rea-
sonable degree of precision. Many evaluations of media effectiveness have
failed because of problems with insufficient,experimental controls, limited
statistical power, or nonrepresentative samples (Lau et al. [1980]). Wallack
[1981] suggests that inadequate sample design dooms an evaluation effort to
failure before it e'ven begins. , ,
The sampling design for this study is shown in Table 3-1. The "O" indicates
points in time when'observation (interviewing) occurs and the "X" indicates
when the various treatments are introduced. Dashed lines are used to show .
that the treatment occurred between interviews. .
RT1 3-4
-------
3. STUDY DESIGN
I. MESSAGES
Message 1:
Radon is a serious
health risk:
You may be at risk
Message 2:
Radon testing is
easy/inexpensive
Message 3:
Radon problems
Can be fixed
Message 4:
The State of
Maryland has a
toll-free number to
provide testing
information
II. MEDIA
Print
Radio
III. COMMUNITY
OUTREACH
Posters
Presentations
Community
Activities
IV. TARGET
AUDIENCES
Parents
Messages: 1,3
Women
Messages: 1,2,3,4
Men
Messages: 1,2,3,4
Homeowners
Messages: 2,3,4
Figure 3-2. Radon message theme: Test now and be sure.
RTI.3-5
-------
3. STUDY DESIGN
Table 3-1. Sampling Design
Sample Design
Frederick sample 1 O X- O 2
Frederick sample2 ~~~ I "X" ~~0~
I 11 2
Hagerstown sample 1 o,' X2,' O2
Hagerstown sample 2 " I "X" ~I~0~
, - I ' 21 2
Randallstown sample 1 O-..
Randallstown sample 2 . I ~~T~0~
• - I I 2
Time •».
Key: p, and O2 are measurements (surveys) at two points ..' '
in time and X1 and X2 are treatments.
The design includes both a panel component and independent sampling
components. The panel component, which includes before and after
observations for the same household, allows for each household to serve-as
its own control. This allows changes in the risk communication effectiveness
indicators to be measured with greater statistical power than can be achieved
with two independent samples. To see why this is true, we can compare the
variance equations for differences under the two approaches: •
independent samples— " •
Var (xn - x.) = Var (x,) - Var (x2) and for : - .
paired observations— ,
Var (x, - x2) = Var (x,) - Var (x2) - 2Cov(xvx2)
•I '
If the covariance, Cov(x..x2). is large and positive, we see that the variance of
an estimated difference can be substantially smaller for the paired observation
sample (see Wolter [1979]) • . ' .
For such" items as knowledge and attitudes about radon, the covariance for
the same individuals at two points in time could .well be large and positive
Most major continuing or longitudinal surveys that are designed to measure
RTI 3-6
-------
3. STUDY DESIGN
change over time employ paired .observations from some form of totally or
partially overlapping samples (samples containing largely or totally the same
individuals at different points in time). Some examples of the major national
surveys that.exploit paired-observation design's include
• The major crop and livestock survey of the' National Agricultural Statistics
Service. This probability-based survey, retains approximately 80 percent
overlap of its area segments from one year to the next (U.S. Department
of Agriculture [1977]). j .
• • • . • . ' i .•••
• The Current Population Survey is a major, multipurpose survey that Has as
one of its primary objectives the estimation of change in labor force and
other demographic characteristics. According to the Bureau of Census,.
75 percent of the sample households are retained in the sample.for each
succeeding, month to improve reliability of estimates of.month-to-month
change (U.S. Department of Commerce [1;978]).
• The Consumer Expenditure Survey uses! approximately an 80-percent
overlap in their "panel" of sample households to ". . .improve efficiency in
capturing changes in expenditure patterns" (U.S. Department of Labor
[1986]). .. .. .;..•'••
• ' (- • •
However, preconditioning bias is a potential concern with the panel sample.
This bias occurs when the baseline interview sensitizes or conditions the"
panelmembers to the program messages. To test for such bias, our design
also includes an interview after the treatment with an independent sample" of
households. , ••.'.'. :
the target population for the study includes homeowners in each town who
have a working residential telephone exchange We excluded renters from
•the target population because of potential interpretation difficulties for this
group. For example, we expect that renters'and landlords could disagree
over whose responsibility it is to test for radon. .By including only
homeowners, we focus on the target population that has a clearer property
right. However, this means we cannot address potential differences between
homeowners and renters in their attitudes, knowledge, and testing decisions
for radon. - i ' •
The sampling design for each town is very nearly a simple random sample.
The use of active hundred-blocks that represent potential working residential
(8-digit)* telephone numbers results in some'clustering of sample numbers,
but the clusters will be small, say five or six, 'and the intracluster correlation
should be negligible because of the geographic dispersion that is represented
by the hundred-blocks. Stratified sampling was not used because little is
known about factors that may have served as ;a basis for stratification.
'The 3 digits include the area code/the, exchange, and the first two digits of
the individual number. These last two digits are used in the randomization
process '.-.••
RTI 3-7
-------
3. STUDY DESIGN
The evaluation design is an essential aspect of measuring risk communica-
tion effectiveness. Table 3-2 summarizes, the features of the Maryland risk
communication design and compares them with an ideal evaluation design
proposed by Lau et al. [1980]. The table shows that the Maryland design
includes most of the proposed features.
Table 3-2. The Ideal Evaluation of a Risk Communication Campaign
Steps
1. Selecting a Health Problem
2. Breadth of Stimulus
3. Dose and Duration
4. Sampling
5. Design Considerations
6. Data-Gathering Technique
7. Measurement of Dependent
Variable
Critical Elements
Seriousness, prevalence, effectiveness
of remedial or preventive actions
Single health topic
The larger and longer the better
Representative survey, perhaps
oversampling specific target groups
Expenmental or quasi-experimental
. design
a. Nonequivalent control-group design,
if small communities with separate
media are of interest
b. Inducements for watching if single
large community is being studied:
include techniques for minimizing
, and measuring reactivity of prompt
c. Incorporate Sotoman four-group
design into a or b; pretest-posttest
more powerful in detecting small
effects
Probably telephone survey, possibly with
.mailed followup
Self-report, with direct observation of
subsample for validation
Maryland Radon Study
Exposure to radon accounts for &000 to 20.000 lung
cancer deaths per year, targeted to three high-risk
cities, the campaign stressed the importance of testing
. and that mitigation need not be difficult or expensive.
Radon exposure
Three-month media campaign and community
outreach. <
Random digit dial telephone survey with random choice
- of respondent .
Quast-expenmental design
a. Randallstown. with a separate media market, was
comparison community; treatments administered in
two comparable communities
b. Not applicable '
c Used Soloman four-group .design: conducted
baseline and followup interviews with a panel
sample to control for interpersonal differences: *
conducted, a followup interview with an independent
sample to test for sensitization.
Telephone survey . •-
Self-report -.
Source: Lau et al. [1980]
3.6 WJLA Campaign
Independent of the EPA project, a Washington, D.C. television station, WJLA,
conducted a month-long campaign to encourage people to test their homes,
for radon during January/February of our study period. The campaign had
multiple components, with coordination by television and'newspapers, and
availability of radon test kits at a reduced price at Safeway Supermarkets.
The campaign began on January 12, including a three-part news series dur-
ing the week of January 18, and ended with another three-part series during
the week of February 15-19 Advertisements for the effort appeared in the
Washington Post and on television. 100,000 radon test kits were purchased
with 70,000 returned for analysis. Television ratings indicated an audience of
RTI 3-8
-------
3. STUDY DESIGN
76,000 viewers. The television programs
ing area. Residents of Frederick were most
the programs. Hager.stown residents eould
without cable, Randallstown is outside the
targeted the Washington, D.C/view-
ikely to' watch the station airing
vew-WJLA, but reception is poor
area.
viewing
In Frederick, more than 65,0 tests were returned
surveys mentioned either Safeway or WJLA
earlier, the WJLVcampaign makes'it more
the EPA program in Frederick. We examine
Section 8.
but none, of the testers in our
during the interviews! As noted
c ifficult to -interpret the effects of
this question in more detail in
RTI 3-9
-------
-------
4. DATA COLLECTION
SECTION 4
DATA COLLECTION
4.1 Overview
The data used in this study were obtained through three telephone surveys
using a random digit dial technique. The baseline survey was conducted dur-
ing December 1987, and the two followup surveys were conducted concur-
rently during April 1988. The baseline survey provided information on
people's awareness, knowledge, and attitudes about radon, and whether they
had tested for radon before the EPA program was implemented during the
winter heating season. The followup surveys gauged these factors,shortly
after the EPA program's completion. One of the followup surveys
reinterviewed the same respondents from the baseline survey; the other
followup survey interviewed a new set of respondents in each of the three'
cities. The followup survey timing corresponded closely to the end of the risk
communication treatments'in both Frederick jknd Hagerstown.
i ..-
The telephone interviews were conducted in RTI's Telephone Survey Unit by
experienced professional interviewers who had been extensively trained by
the study team. "Interviewers were provided'with a comprehensive training
manual for each'survey and copies of the questionnaires before beginning the
actual survey. The interviewers used thes;e to conduct mock interviews
among themselves before calling survey respondents. , \
Throughout the data collection the study t;earn, as well as the telephone
supervisory staff, monitored interviewers and tracked interview completions
on a daily basis. Quality control remained an'important focus throughout the
d.ata collection and data preparation. All questionnaires were edited and
coded by personnel trained specifically: for this survey The coded-
questionnaires were then entered by djata entry personnel using a
computerized data entry program that inclucled range checks for each field.
as well as other features to reduce the number of invalid entries. The comput-
er programmer who wrote the tfata entry program, as well as project"
personnel, tested the programs before they;were released to the data entry
personnel for use Cooperation between thej programmer and a member of
the project team during the writing of this program guaranteed that the data
would emerge from this process in a readily usable form. The data entry
• process included 100 percent verification of a|ll keyed data.
RTI 4-1
-------
4 DATA COLLECTION
4.2 Response Rates
Response rates for all three surveys attest to-efforts to obtain completed
interviews .with sample respondents. As shown in Tables 4-1 and 4-2
screener completion rates for all three cities in both the baseline and
independent followup surveys are well above 90 percent. (Since the
.dependent followup survey went back to respondents interviewed in the
baseline survey, no screening was necessary for it.)
Response rates for the baseline and independent followup surveys are also
.presented in Tables 4-1 and 4-2. Total response rates for the baseline survey
ranged from around 48 to 64 percent, and those for the independent followup
survey ranged from 55 to 68 percent. Several external factors might have
contributed to the lower response rates. First, the baseline survey was
conducted in the middle of the Christmas season during which telephone
response rates typically are lower than average. The more urbanized cities,
Frederick and Randallstown, had lower response rates in these two surveys
. which perhaps reflects the fact that it is harder to obtain participation from
people in urban areas where they^are more likely to be inundated with sales
calls. Most refusals indicated that the individuals involved simply did hot want
to be bothered.
The method used to compute the response rate for each city is given by the
.•following formula: •
/ Screener j J Intervfew J _ {Total response)
Icompletion rate] Icompletion rate/ ~ V rate /
Screener Total eligible tele- . Total ineligible tele-
comoletion - pnone numbers identified phone numbers identified
rate Total telephone numbers called
Interview T . . , - •.••.•
completion = Total number of completed interviews
rate Total eligible telephone numbers identified
Response rates for Maryland dependent followup survey: ".. .
Frederick -s^ = 82.60JJ
Hagerstown ~ = 81.66J{
•307
Randallstown --7—= 78. 1855
495
RTI 4-2
-------
V 4. DATA COLLECTION
Table 4-1. Response Rates for Maryland Baseline Survey !
] Frederick
Total Eligible Telephone
j Numbers Identified
. • Physically 'Mentally Incompetent 11
.* ' . ' Language Barrier , '2
Interview Completed 523
j Partial Data 12
'• . Final Interview Refusal 305
Other 1
* .. ' , Tbtal 854
Screener
Completion Rate
, 1.921
2.095
(.92)
Total Eligible Telephone
Numbers Identified
Physically Mentally Incompetent 22
Language Barrier 0
• Interview Completed 52P
Partial Data 6
Final Interview Refusal 216
Other ' . 2
.Total 7S2
Screener
Completion Rate
1.973 '
2.090
(.94)
Total Ineligible Telephone
Numbers Identified
Not Residential Number
Not Homeowner
Nonworking Number
Double Wrong Connection
No Result from Dial
Fast Busy/Computer Tone
total
Interview
x Completion Rate
523
854 '
(.61)
Hagerstovvn
Total Ineligible Telephone
Numbers' Identified
Not Residential Number
Not Homeowner
Nonworking Number
Double Wrong Connection
No Result from Dial
Fast Busy Computer Tone
Total
Interview
'• Completion Rate
. 529
(.65)
I Indeterminable
i . *• ••
360 ! Ring. No Answer 137 .
385 Answering Machine Service 22
251 . Unable to Contact 15
40 Total 174
2 i . . . . .
ll)67 j '.. " .. ' '
total
Response Rate
i .
• . *
i . .
•i . .. - •-•_..-
I Indeterminable
317 |. Ring. No Answer • 98 ,
39(v j. Answering Maclune Sen-ice IS
400 ! Unable to Contact 7-
"51 1 .. Total 123
2li. t ' , • '. '
1.191 ' ' '•-..''-.•
iTota!
Respionse Rute >
"• , V*---. . - '.•••. ,.•-....,
(ii:3.80'.!o v.
• ' i ' ' ' ,'....-
Randallstown
Total Eligible Telephone
, . • Numbers Identified
Phvsically Mentally Incompetent 1 1 '
Lancuace Earner S
Interview Completed 4f»o
Partial Data 19
Final Interview Refusal 446
• Other .: • , ' 5 •
, Total 9S4
Screener
• • Completion Rate
l.HPS
2.IJP7
•..Po'
Total Ineligible Telephone
Numbers Identified
Not Residential Number
Not Homeowner
Nonworking Numt>er
Double Wrong Connection
No Result from Dial
Fast Busy Computer Tone
Total
Interview
x Completion Rate
49S
9S4~
- (.50.!
i Indeterminable
194 1 Ring. No Answer 54
450 ! ' Answering Machine Service 22
320 Unable to Contact 20
37 ' Total ' 1U2
o • r • - . . • ,• .
10 ."j' .
1.011 1" , '. , '•' v •.
jtotal .
• Resixinse Rate
RTI 4-3
i
' l" , . . • •
-------
4. DATA COLLECTION
Table 4-2. Response Rates for Maryland Independent Followup Survey
Frederick
TotaJ Eligible Telephone
Numbers Identified
Physically Mentally Incompetent
Language Earner
Interview Completed
Partial Data
Final Interview Refusal
Other
Total
1-2
4
481
11
284
8
800
Screener
Completion Rate
1.817
1,937
(.94)
Total Ineligible Telephone
Numbers Identified
Not Residential Number
Not Homeowner '
Nonworking Number
Double Wrong Connection
No Result from Dial
Fast Busy:Computer1bne
Total
x Interview
Completion Rate
K 481
800
(.60) l
382
352
235
28
2
18
1,017
*> -
!m
-
_ Indeterminable
Ring. No Answer
Answering Machine Service
Unable to Contact
Total
Total
Response Rate
;
56.40"o
99
14
7
"l20
Hagerstown
Total Eligible Telephone
Numbers Identified
Physically Mentally Incompetent 13
Language Earner 4
Interview Completed 534
Partial Data 7
Final Interview Rr:"u>:ii 19s
Other
•
. 4
Total 760
Screener
Completion Rate
_U"!T_
1.9SS "
(.90)
Total Ineligible Telephone
Numbers Identified
Not Residential Number
Not Homeowner
Nonworking Number
Double Wrong Connection
No Result from Dial
Fast Busy Computer Tone
Total
Interview
Completion Rate
• ML
7(50
(.70)
• 267
436
412
9
2
21
1.147
*
-
Indeterminable
Ring, No Answer
Answering Machine Sen-ice
Unable to Contact
Total
Total
Response Rate
' 67.40'.'.,
' 65
*
9
SI
RandaUstown
Total Eligible Telephone
Numbers Identified
Phy.Mi-ally Mentally Incompetent
Lannuail" Barrier
Interview Completed
Partial Data
Final Interview 14-fii>al
Other
Total
' S.
Comp
ID
Hi
(i
.S7.S
Teener
lenon Itite
1 SIX)
UI1J
'.'.4.
Total Ineligible Telephone
Numbers Identified
Not Residential Number
Not Homeowner
Nonwnrkmg Numlier
Double Wrong Connection
No Result from Diai
Fast Busy Computer Tone
Total
Interview
Completion Hate
SIfJ
, S7h
•( ~>*>
385
422
290
16
0
9
'9±>
m
-
Indeterminable
Ring. No Answer ;
Answering Machine Service
Unable to Contact
Total
Total
Response Rate
55.0", • - . ,
74
31
RTI 4-
-------
4. DATA COLLECTION
Another potential influence on response'rates was the randomization
technique used to.choose the respondent once the observation had been
screened. To avoid the potential bias introduced from using decisionmakers
who are most accessible to the telephone, first names of all adult
decisionmakers in the household were obtained land then one was randomly
chosen for interviewing. Such a technique resulted in an increased number of
callbacks and several cases in which the person chosen in the random
selection process was not willing to be interviewed. Thus, response rates
were potentially lower than they would have been without choosing the
respondent by this random selection process: Nonetheless, randomization of
the household, respondent was judged a;s important consideration in
designing the protocol to be used in composing the sample.
Considerable efforts were made to convert! refusals. Cases classified as
refusals were reassigned to the most experienced interviewers. These
interviewers recontacted the households in question at least two additional
times trying to obtain the interview. We decided that further attempts to
obtain the interview could.be construed as harassment.
, ' t ' - ' ' - '•
To evaluate whether our respondents are representative of the target
population, Tables 4-3,and 4-4 compare deVnographic information for our
respondents with those from the 1980 census. The table shows that our
respondents have similar demographic profile, s. The main variable displaying
any significant differences is the median value of the house. Much of this
difference probably can be traced to the significant increases in housing'
.values that have occurred during the pa|st 10 years In the Baltimore-
Washington Standard Metropolitan Statistical Area. The ratio of males to
females in our study is'slightly different than that in the census. This may also
be a result of the randomization process meritioried above.
An additional consideration in evaluating the response rates for our survey is.
the general consistency between the basejine and followup survey in the
demographic characteristics of respondents. \ For example, the response rate
for Randallstown in the followup survey was ailmost 10 percent higher than the
baseline but there were almost no differences in the demographic
characteristics of respondents in the two surveys.
Finally, our response rate calculations are based on the most conservative in-
terpretation of indeterminate calls. Some studies; assume that these calls are
ineligible numbers which would raise the overall .response rate to 61, 64, and
50 percent for'the baseline interviews for Frederick, ^Hagerstown, and
Randallstown, respectively. • i ; '
RTI 4-5
-------
4. DATA COLLECTION
in
I
C«
CM
CO
CO
8'
TO
TO
01
^i
O
c
o
Jfi
75
1
tr
z
• e
,0
OJ
•o
TO TT O
eccsoo
w ^ TO
OCO CM
•c — o>
TO ^ CM
— mm
o o eo
CM TO 1
mcoin
CM CM CM
N. O —
OCM CM
T T TO
000
OOO
CD CO CM
co "-"CD"
«^*>v*
» CM in
TO m ^>
w TO
ooo
0.0.0
f^ ^r" o"'
5m m
*•><*>
O">CM TO
CO OCO
•
O) O (O
in o> n
CD — o
co co co
ooo
g
en —CD
*T — TO
222 ^™S! JMCMCM
§(0
en
> —*»
TOCOTO
m OSD'
ooo
a>
•o
I
(0
a>
OC
*»•
o
0)
D.
O
ic
o.
s
O)
o
0)
a
s
CD
IT
mm 7
e • .
.S in — en
^ *» r^ r.-
g
E
o>
U
I
"
CM
CD
— com mocM
-------
1
1
1.
i " ''•''••'.
• _ « '
/ . , ' .
1* ! ,
-
' '
'
^
•
. .
^—
C
O
o
' OJ
a.
. . . _c
w
' ' S
' • ' - .2
ro
•. o
•'''•• !c
Q.
ra
ra
i O
•- ' E
.0
'••. , . 3
_a>
ra
. " ... .1-
, c
i
2
•
g
CD
QC
c
S
• fc
BJ
'
• ij
1
U.
01 o
a=
a> o
O-
f
c
•o §"
i *
!•=
£.*"
— 8"
°:
|
c
Q)
U
IB-
TJ 2
C 5
O O
0.=
O *"
s
* o.
si
°-=
s ^
-"
c ^
S 2
01 5
O M
CD
3
CD
C
a>
U
ca
(j, a.
II
o —
la
? =
S
1 1
^
e>
l>
g 3
CD
CD Cu
00 —
f«» e*>
CO-
CO CM
CD •-
»".
Jo"
a> —
CDCM
CO CM
CD CM
'
CO CM
CDCM
^ to
(D n
, !K8
'28
i
SS
: si
Type of reslden
Sin<|lo -family 1
Olhar
: 4. DATA COLLECTION
;i • "• ' .
CDf^O <<• <.<. tC^1 ffJO.
®" .' ZZ ; ZZ '. °=~ •'.««_
•mtDo^ > N» n • e CM N.CO CTJ —
,p*«^ CD — cur** too cote
• i -
• • - • ;
•'. ' ' i •" •
' .- • • •- I
f**CNJO N« W • ift lO ' ^2^ tf) ifl •'
l^CNi ^CD- j WN. t£)CO Vin,
! . . - • •
K cy CD •»- w r- CD eo ^ to
1 . :
1 ' ' ,
SWC*> ™*
OJ ITS ^* CD^.,CNjr*- Ol P3(D_
. - . '
1 .
' - • ' .•."'••.. v
.
- 1 '
^ (N ID (D ^ ' W ^* ^ W • tD ^T1
W tf3 ^ CD ^ , CN r*« O) ' CO CD
-••"'••" • : ! * " •
" •' . ' i " >•'
•'•'':.-• ' ' ' ' ' "5 . ;
h*. to to 'f*-co irtin t^co CDCVJ -£
CMlO*- CD — • ! CSJ f*. C) COCO O •
i " Q
• • -S
. ; ••-.•• • f .. - .'
; . .- • - re
^coto << <^ ^ *o~ *n o "5
Z Z s Z Z - Q
CL
'- "*5
tw
1
^^^ «^ O> !• tt1 ^ *ff *~ rt
CyC*3^J" 'CD^ Cyf*» Ol C*>(C ^
• - ' •& • '
i '•• • L ' ,: - C. '
•5
-!.._*' £ '
' E • •
2S?i? SS "t en" SS |
' ; ' ' 8 . ' '
£
'.'•-•. ' \ • .' -s. • -
o»-— ocj) JD« g>n =ffi'^ "8
= i a> "G •
1 R g
*.'".'' ' 8 1 ,
« • c ' "'? ^ ° o
liiss-ii •! Is^-I • .
go,- e» Si 5 S .§ .- . e-oS?iC-=
fHit -«r -g. IS .«! i*22 1
il*i5' i|^i ;|^i |i= ^i^ | 1
5£ oS /o tc w ' ? ^-
RTI 4-7
-------
••*
-------
5. RADON AWARENESS AND ATTITUDES
SECTION 5
RADON AWARENESS AND ATTITUDES
5,1 Introduction
This section addresses two indicators of risk communication effectiveness:
changes in people's awareness of radon and'changes in their attitudes about .
radon. :.
The awareness indicator is straightforward: Are people more aware of radon
as a result of the risk communication program? Awareness is a basic
precondition for any type of decision involving radon.- In this section we
discuss the results for awareness changes, allong with the important sources
Of awareness. These sources include'both formal communication
channels—radio, television, and newspaper—and informal ones—-friends,
relatives, or coworkers.
j .
Attitudina! change is a more subtle indijcator of risk communication
effectiveness. The subtleties stem from several sources. First, it is unclear
which attitudes are really preconditions to tesjting. Second, the interpretation,
of .'more informed" attitudes can be ambiguous. Finally, we may have only a
partial picture of people's attitudes toward radon testing. Despite these
subtle, and potentially ambiguous, dimensions of attitudinal change, the
results presented in this section do shed some light on people's attitudes
toward radon. These results are important for both interpreting subsequent
testing decisions, as well as.offering insight into potential obstacles that
remain in increasing the level of radon testing.
This section combines information from bpth the panel and independent
surveys. This combination allows us to take advantage of the relative
strengths of each part of the overall design. It also considers whether or not
awareness and-:attitude of the panel sample i
effects resulting from the initial telephone int«
espondents reflect sensitization
5.2 Radon Awareness
Increased awareness is a basic indicator
program. Awareness of a risk is a necessary
behavioral actions to reduce the risk.
rview.
fdr any risk communication
condition for any subsequent
However ..increased awareness-does
RTI 5-1
-------
not guarantee that the desired behaviors will occur. It is a starting point or
building block that underlies almost every model of behavioral change (see
McGuire [1985]). - .
In this study, the measure of awareness was straightforward: Had
respondents read or heard about radon recently? This measure requires
people to recall some previous exposure to radon information. The source of
the information is open ended—it could be from either formal or informal
sources. No information was given to respondents that would have indicated
• our survey was about radon. The introduction focused on general
. environmental issues in the community..
Awareness can be appraised from several perspectives: '
• The levels in the followup survey in each community,
• The change in awareness in each community between the baseline and
followup survey, and .-,.••
• The change in awareness in the experimental communities compared to
the control community.
Each of the perspectives provides somewhat different insights into the
effectiveness of the risk communication programs in;Frederick;and
Hagerstown. Each can also be evaluated for the panel sample and the
independent sample.
As'shown in Figure 5-1, Frederick had the highest level of awareness in the
followup survey with 9 out of 10 people interviewed having heard about radon
recently Frederick received both the media and communjty outreach
programs. In Hagerstown, which received only the media campaign, approxi-
mately 8 out of 10 people interviewed had heard about radon recently In the
comparison community, Randallstown, about 7 out of 10 people interviewed
were aware of radon. The differences in awareness levels between the panel
and independent sample are insignificant. Overall, these levels suggest that a
clear majority in all three communities was aware of radon by the time of the
followup survey.
Figure 5-1 also shows the changes in awareness between surveys. For any
one community this change could reflect other factors besides the
experimental programs. However, by comparing the changes between the
experimental communities and the comparison, we obtain a ^better gauge of
the risk communication program's effectiveness. As we discuss later, this is
still an overestimate because of the WJLA campaign. The simplest evaluation
can be performed with the independent sample. In Frederick, awareness
increased by 18 percent—from about 72 percent in the baseline survey to
90 percent in the independent followup, The increases in Hagerstown and
Randallstown were similar to each other-—about 13 percent and 12.5 percent,
RTI 5-2
-------
5. RADON AWARENESS AND ATTITUDES
RADON AWARENESS BY TOWN AND SURVEY
PERCENTAGE WHO HAD SEEN OR HEARD ABOUT RADON
Figure 5-1. Radon awareness.
respectively. As shown in Table 5-1, all of thes
nificant at the standard .05 level of significance
Figure 5-2 uses the panel to .develop some further
awareness for the panel sample. Because the
for the same homeowner across surveys, we
showing how the awareness of each person
states include
• Aware baseline, unaware followup,
• Unaware both surveys,
• Aware both surveys, and
• Unaware baseline, aware followup.
Again, we see that Frederick had the highest
70 percent, who were aware in both surveys.
the comparison community Randallstown.
percentage, 7 percent, who were unaware
12 pe-cent were'unaware in Hager.stown in
Randallstown • •
e changes are statistically sig-
insic-hts into the changes in
panel matches observations
can measure the changes .by
changes. The four awareness
percentage of homeowners.
This is 21 percent higher than.
Frederick also had the lowest
in both surveys. In contrast,
both surveys and 15 percent in
RTI 5-3
-------
5. RADON AWARENESS AND ATTITUDES
Table 5-1. Changes in Awareness: Statistical Test Results
Have you seen or heard anything about radon in
the past few months?
(Yes/Total - "Don't Know")
Frederick
Hagerstown
Randallstown
Baseline
proportions
388/522"
350/526 '
291/493
Independent
followup
proportions
433/480
481/531
357/509
Z Statistic
-6.68-
-9.95-
-3.70-
*Z statistics for sample proportions that are statistically different.
PERCENTAGE BY TOWN OF THE PEOPLE WHO HAD SEEN
OR HEARD ABOUT RADON ON THE BASELINE AND
FOLLOWUP SURVEYS
PANEL SAMPLE
BASELINE, NOT ON
FOLLOWUP
NOT ON EITHER
ON BOTH
NOT ON BASELINE, BUT
ON FOLLOWUP
23 FREDERICK
D HAGERSTOWN
RANDALLSTOWN
Figure 5-2. Panel sample radon awareness.
RTI 5-4
-------
5. RADON AWARENESS AND ATTITUDES
The pattern of changes With people moving from unaware to aware between
towns is more amb.igupus The comparisbn community has'about a,
6 percent larger increase in this category than l|he two towns that received the
risk communication treatments This may reflect either of two factors—that
the higher initial 3v.areness levels left le'ss room for change or the
Randallstown panel respondents were monb sensitized by the baseline
interview than their counterparts in Frederick or Hagerstown.
* ' *•!
To evaluate which of these two explanations was correct, we performed two
additional evaluations which are described in more detail in Appendix A. Our
first analysis estimated an awareness model wtiich used a homeowner's initial
level of awareness in the description of changes in awareness. The results of
several models are shown in Table 5-2. As shown in the models, people in
Frederick and Hagerstown were more likely to be aware of radon in the
followup than people in Randallstown. This held true even when the prior
levels of awareness are controlled for in the model. Not surprisingly,' people
with higher education were more likely to be avi/are in the followup.than those
with lower levels. The same was true for wriites and people who lived in
homes built between 1940 and 1976. These results are not definitive because
the analysis used ordinary least squares (regression) analysis in this phase of
the evaluation. We have extended these models to adequately re^"ct the
type .of response that must be modeled ajnd report more appropriate
statistical methods below. , ' . j
Our second analysis tested for sensitizatipn ih each town using a standard
statistical test. There was no evidence that thb baseline interview prompted
greater awareness in the followup interview among the homeowners in the
panel sample compared to those in the independent followup. This is
consistent with Figure 5-1 which showed little, if any, differences in awareness
between the panel and independent followups | . •
i- . ' • *, '
! • N
The sources of awareness about radon are also helpful in evaluating risk
communication issues. They identify the charinels people use to obtain risk
communication messages. Figure 5-3 showsj.the sources of information for
homeowners who were aware of radon, not the full sample Homeowners
also may have indicated more than one source. 'As shown in Figure 5-3.
newspapers and television are the two main gources of information about
radon The results indicate that television as a source increased substantially
between the baseline and followup surveys with increases of approximately
25 percent in both Frederick and Hagerstown and 15 percent in
Randallstown The substantial increase rejects the greater coverage for
radon issues.in Maryland, especially the information campaign launched by
WJLA s Roberta Baskm This effort is a potential confounding factor for
interpreting changes m our indicators. The WJLA campaign was available in
ali three towns but the exposure seems to have been somewhat higher in
FreaencK than m the other two towns . About 15 percent of Frederick
hcmeov.-ners mterviewec recalled-seeing Fioberta Baskin compared to
RTI 5-5
-------
... 5. RADON AWARENESS AND ATTITUDES |
Table 5-2. Recursive Awareness Models
Coefficient (t-ratio)
Independent
variables Model 1 Model 2
INTERCEPT ' 0.4052
BASELINE • 0.1610
AWARENESS.
FOLLOWUP 0.0045
PERCEIVED RISK
EDUCATION 0.0169
AGE 0.0013
ASK DOCTOR -0.0555
YOUSMOKE 0.0114
FREDERICK
HAGERSTOWN
• ' SEX
RACE
• HOMESELL
FOLKS
USE BASEMENT
MOVESOON
YR40_76 -
YR76
YEARS AT ADDRESS
n 1,062
Adj R2 0.0573
F 11 .750
(4.532) 0.2015
(6.308) 0.1257
(1.243) 0.0048
(3.609) 0.0222
(1 .546) 0.0008
(-2.263) -0.0538
(0.429) 0.0241
0.1269
0:0702
-0.0174
0.1230
-
1,037
0.0888
11.106
(2.078)
(4.851)
(1.332)
(4.603)
(1 .003)
(-2,176)
(0,900)
(4.288)
(2.286)
(-0.743)
(3.647)
.... ... - . . .
Models
0.0723
0.1167
0.0044
0.0240
0.0002
-0.0477
0.0253
0 1741
0.1125
-0.0151
0.1180
4.1702E-8
00033
00513
0,0167
6.0939
0.0388
0.0010
945
0.0964
6.931
(0.625)
(4.242)
*
(1.153)
(4.674)
(0.160)
(-1.817)
(0.896) •
(5.208)
<3.353)
(.-0.610)
(3.365)
(0.863)
(0.337)
(1.951)
(0,436)
(2.733)
(0.971)
(0.666)
X ' " • ' • - . = •
RTI 5-6
-------
5. RADON AWARENESS AND ATHTUQES
SOURCES OF AWARENESS FOR THOSE
HEARING ABOUT RADON
Tc
'100 T
90{
y
60
50
30-r
NEWSPAPER
RADIO
HAGERSTOWN
° 73.5 77'2'
33.5
2-J.9
29.4
76.S
56.5'
NEWSPAPER
RADIO
NE'A'S PAPER
RADIO
C DASG.Iti: SLTRST.Y , D INDEPHNDnsT RDU.OV.T^> • DlTESIirST I-OIJ.OWLT
Figijre 5-3. Awareness sources.
RTl 5-7
-------
5. RADON AWARENEfie: AND ATTITUDES
14 percent in Hagerstown and only about.4 percent in Randallstown. We
have tried to isolate the effects of our risk communication program by using
. statistical procedures that account for the influence of factors such as the
WJLA campaign. A summary of these analyses is presented in Appendix A.
" i
5.3 Exposure to EPA Risk Communication Programs
Exposure to EPA's risk communication program is a second indicator of
effectiveness. To measure exposure, we asked people whether they recalled
seeing or hearing about radon from various sources such as the newspaper
or television. We then used an open-ended probe to learn about details that
people recalled from their exposure.
Figure 5-4 presents three elements that can be associated with the EPA risk
communication program—whether people recalled seeing a utility bill insert,
recalled seeing a poster, or called the Maryland toll-free number which was
publicized in all the risk communication messages. These three elements are
ifkely to understate exposure to the program, especially in Frederick. If
homeowners recalled reading about a radon awareness week activity in the
newspaper, this would not have been counted as a program element—it was
classified in the survey as only learning about radon from a newspaper.
As shown in Figure 5-4, the utility bill insert had the highest recall among
homeowners in Frederick and Hagerstown. The panel sample recall was
• about 25 percent in each community while between 14 and 19 percent of the
independent sample recalled seeing the insert. Between 7 and 10 percent of
the homeowners in Frederick recalled seeing the radon posters. The findings
on exposure should be interpreted with some care because a sizable
percentage (as high as 18 percent) of respondents in the comparison
community stated that they saw a utility bill insert. These respondents either
incorrectly recalled what they saw or actually saw an insert about radon from
some other source.
The exposure indicator of effectiveness can be difficult to' interpret. For
example, the measure is based on people's recall of the source of particular
messages. Given the 3-month time duration for the exposure to the
messages, people probably forgot where they heard the message.' If
exposure is the only indicator of effectiveness, then we would :have designed
our evaluation differently to account for recall problems—e.g.,; the number of
people picking up leaflets from the posters could have been counted. Our
concern, however, is with more robust indicators, such as; the change in
knowledge or attitudes on testing, which required us to wait a sufficient period
of time for changes to occur and to use a broader population based
evaluation. Because of these measurement difficulties, compared to the other
indicators, we place less weight on these indexes jDf exposure..
RTI 5-8
-------
5. RADON AWARENESS AND ATTITUDES
EXPOSURE TO EPA PROGRAM
Utility Bill Insert
Poster
Called Toll Free-Number
HAGERSTOWN
10 ?,
I'tili.tv Bill Insert
Poster
Called toll Free Number
30
RANDALLSTOWN
2.S
3.2
Y///////S
Utility Bill Insen
Poster
'! D INDEPENDENT FOLLOWX.T • DOPLNDENTRXLOWUP
Figure 5-4. EPA program.
CalU d ToU Free Number
J
RTI 5-9
-------
5.4 Attitudes Toward Radon
Attitudes are an important indicator of risk communication effectiveness
Aizen and Fishbein [1977] argue that attitudinal change is an important
condition for behavioral change. They also argue that attitudes that are close-
ly linked to the behavioral patterns under investigation can also help to
predict changes in that behavior. Most experts tend to agree that attitudinal
measures are an important part of the evaluation of communication
effectiveness. There is far less agreement over the ability of attitudinal
measures as predictors of behavior (McGuire [1985]),
For this study, we will use attitudinal indicators, as an additional gauge of
whether people received and responded to the risk communication messages
in the treatment communities. Our attitudinal measures are more narrowly •
defined than the cultural and social attitudes suggested in Krimsky and
Plough [1988], but they do offer some useful insights into how people think
about radon. Our attitudinal measures, were also constrained by the need to
cover a wide range of topics within the confines of a relatively short—10 to 15
minutes—telephone interview. '-...-.
The survey questionnaire included three attitudinal measures that were asked
in a strongly agree, agree, disagree, and strongly disagree format The three
statements were . .
• "It is important to test my home to find out if I have a radon problem":
» "If I had a radon problem, it would be costly to fix"; and
• "Even if a radon problem was fixed, my home would still be worth a lot
less • ' . .
These three items correspond closely to toe risk communication messages
which emphasized that testing is important, that remediation need riot be
expensive, and that remediation can be effective. -If the messages were-
received and processed, the proportion agreeing to the first would increase,
and the proportions disagreeing to the second and third would increase. ' ;
Table 5-3 summarizes the results from statistical tests to determine whether
the proportions of people having these .three attitudes changed between the
baseline and followup survey. For illustrative purposes, we have focused on
the baseline sample and the independent followup sample. The results indi-
cate that the risk communication messages were received and accepted by a
significantly highe/ proportion of respondents in the towns that received the '•"
communication treatments. In Frederick, the proportions increased in the
desired direction for all three questions, while in Hagerstown significant '
chances'were measured in the second and third case. There were no signifi-
cant differences in the'comparison community , . -'';:•,.•
RTI 5-10
-------
5. RADON AWARENESS AND ATTITUDES
Table 5-3. Changes in Attitudes About Radon: Statistical Test Results
! ' \
"It is important to test my home to find out if I have a radon problem."
(Strongly Agree or Agree/Total - "Don't Know")'
Frederick
Hagerstown ,
Randallstowri
Baseline
proportions
342/467 ,
333/452
-326/428
%
73
74
76
Independent ,
followiip
. proportions
i
366/445
. 351/460.
346/456
%
82
76
76
2 Statistic
-2.67*
-0.78 ,
0.13
"If I'had'a radon problem, it would be cjostly to fix."
(Disagree or Strongly Disagree/Total- "Don't Know")
Baseline
proportions
%
Independent
followiip
proportions
%
Frederick
Hagerstown .
'Randallstown
115/392 . 29
92/363 25
99/311 32
177/358 49
144/353 41
. 123/325 38
"Even if a radon problem were fixed, my home woulc
(Disagree or Strongly Disagree/Total
still be worth a lot less."
Don't Know")
Frederick.
Hagerstown
Randallstown •
Baseline
proportions %
212'442 48
211'433 49
233:394 - • 59
Indepenc ent
followup
proportions
/427
291/•
250/
238/400
/419
Z Statistic
-5.44*
-4.33*
,1.44
68
60
60
Z Statistic
-5.43'
-2.92'
-004
Note- The sample prcoo-nons were tested at the 95% confident e ievelr(2 Stat'Stic ot-'.96 < 2
*Z Statistics for sample proportions that are statistically different.
RTI 5-11
-------
5. RADON AWARENESS AND ATTITUDES
Figures 5-5, 5-6, and 5-7 provide supplemental information on the three atti-
tude questions by comparing tester and nontesters responses. In these
figures the responses of both the panel and independent samples are pooled
for the followup survey. We have pooled the responses because there are no
substantial differences between the two followup surveys in the area of
attitudes about radon. . .'•
Figure 5-5 suggests that the patterns in the followup surveys are relatively
similar across all three towns—a very high percentage of testers and approxi-
mately 75 percent of nontesters agreed that it was important to test for radon.
This message seems to have been communicated about the same in all three
towns.. .
Figure 5-6 shows that the second message about the expense of radon miti-
gation had effects in both Frederick and Hagerstown but not in Randallstown.
In Frederick, about 15 percent more testers and nontesters disagreed that
mitigation would be expensive ' In Hagerstown, 22 percent more testers
disagreed with this statement and 16 percent more nontesters. In
Randallstown, the comparison community, 2 percent fewer testers disagreed
and 6 percent more nontesters disagreed. Across the communities,
Hagerstown had the largest increase and also had the largest differential
between testers and nontesters—13 percent in the,followup survey. 'Despite
the improvements, almost half the testers, and more than half the nontesters,
still felt that radon mitigation is an expensive undertaking. This view could be
an important reason why radon testing levels are less than expected in some
'areas. People are concerned that mitigation will be a major investment.
•\
Figure 5-7 presents the results for the homeowners' opinion about the effec-
tiveness of mitigation. In this case the results are far clearer—a significant
majority of both testers and nontesters in all three towns disagreed that their
housing values would be worth less after mitigation. The percentages, of
nontesters in both Frederick and Hagerstown holding this opinion increased
substantially between the baseline and followup surveys—17 percent more in
Frederick and 14 percent more in Hagerstown. The percentage of testers in
Hagerstown having this opinion increased by 23 percent while increasing only
8 percent in,Frederick and 9 percent in Randallstown/ It appears that the
message.about mitigation effectiveness came through reasonably clearly in
both Frederick and Hagerstown, while there was relatively little change in
Randallstown Because of time constraints, we have not modeled the
attitudinal changes as we have for other indicators Thus, the changes
observed between the baseline and followup surveys probably reflect both the
WJLA campaign and the EPA program This is an area where further investi-
gation may be warranted to isolate the effects of our program. '
RTI 5/12
-------
5.. RADON AWARENESS AND ATTITUDES
DISAGREEING THEIR HOMES
EVEN IF A RADON
RESPONDENTS
WOULD BE WORTH LESS
PROBLEM WERE FIXED
BASELINE
54
60
59
FREDERICK
HAGERSTOWN
RANDALLSTOWN
PANEL AND INDEPENDENT FOLLOWUP SURVEYS
FREDERICK
HACERSTOWN
RANDALLSTOWN
E TESTERS
NONTESTERS
Figure 5-7. Attitudes^mitigation effectiveness.
RTI 5-: 5
-------
Attitudes toward radon mitigation appear to be a fruitful area for further inqui-
ry. Even after the experimental program, participants in our study were con-
cerned that mitigation would be expensive. An important area we have not
explored is whether people feel that they could obtain competent mitigation
contractors in their area or whether they are reluctant to undertake mitigation
because they think more effective methods will be found relatively soon as
more public attention is given to radon.
RTI 5-16
-------
6. KNOWLEDGE ABOUT RADON
SECTION 6
KNOWLEDGE ABOUT RADON
6.1 Introduction
The fourth indicator of program effectiveness is the increase in knowledge, or
learning Did homeowners increase their knowledge about radon as a result
of being exposed to the EPA Region Ill/Maryland program? Like attitude
changes, knowledge can be viewed a,s both ail endpoint and a precondition
for some desired behavioral action such as testing for radon. As an endpoint,
we are interested in measuring whether our risk communication program
transferred information to homeowners about radon. As a precondition for
behavior, we are interested in evaluating whether the transfer of certain types
cf information has an effect on the level of radon testing. This question is
explored more fully in the next section. For now, we focus on knowledge as
an endpoint or effectiveness indicator | ,
6.2 Changes in Knowledge About Radon
To measure homeowners' knowledge about radon we administered a seven
question "radon quiz" in both'the baseline and| followup surveys Such quiz-
zes have been used by Weinstein, Sandman', arid Klotz [1987] and Smith
et'al [1988] For this study, the quiz was multiple choice with three answer
choices with the same questions used in both surveys. The advantage of .this
strategy is that we can compare improved performance-tor each question.
The only potential disadvantage is that the strategy could alienate some
members of the panel sample who had answered the same questidns three
months earlier. Fortunately, although a few homeowners expressed
complaints, no interviews were terminated because of the administration of an
identical second quiz ... .ix.:1-•••„'•/ - .. ", . ....«,
Table 6-1 summarizes the performance on each question of the radon quiz
The results are presented for the baseline survey, the independent followup
survey, ana the dependent (panel) followup survey. Several important
patterns can be seen in these results First, Frederick, which received both
the media program and community outreach, has substantially higher levels
of knowledge in the followup surveys than the comparison community.
Randalistown /Knowledge about radon is generally higher in Hagerstown
than in the comparison community but the differences are smaller than those
for Frederick i" :. •.; '
RTI 6-1
-------
IMMUVVICUUC AtiUU I 'HAUUN
' i
|
?
t
O
O
o>
(0
C
0)
0)
0^
«
1
i
(C
M
i
m
dL
If
c g
l|
c
£ Q.
V 3
II
«S
£
if
s2
•f
T3 |
Ji
i.
"O 3
u
x>£
£*"
c ^
= »
g-
• '
i !
o
1-
a
^^
C
o
•o
(Q
C
3,
O
5
^
0)
0)
•o
0)
Table 6-1. Knowl
Q
i
£
il
00 CO •*• (0 ifl ! CD
'".,.'.; "
;
0) K ^ CD ^ ^ flD
* w w !» o ?
&j 52 • "* M ••
» te »- ca o v
S — »- h- n CD
CD n in ^ u
'in iv e\j ^ Jo : • «
( •
0) (0 0) ^^ *" ' CO
• *
1 • ' -
r
— v o m wo
N • (O *T «0 , 1 N
A) > 1 •• w m CD • co • i ^
V) 2 ' ^ ^ C^ ^ ^ ^
m *
MM
Questions
\
"
sr ' M
- n . i? E£ ; '
•S? S '-gS 1 IS £|
II IV 111. 1|I_
w —£ ^s *w So ^5^
Is li'lf |s Ifffs
£8 f £ 2c S«>c ?2 fl^
«p ng re o $o 0) c
-------
61 KNOWLEDGE ABOUT RADON
Second, the largest improvement in performance between surveys occurs for
Question 15, "How, can one test for radon?': This was one of the main
messages of the EPA communication prograhr In Frederick, 37 percent
more people answered this question correctly! in the independent foliowup
survey, while 44 percent more people in the p'anel sample answered it cor-
rectly in the foliowup survey. Both Hagerstown and Randallstown showed
increases in the numoer of people with correct answers in the range of
20 percent. ' - [. -
Third, people showed more knowledge of general characteristics—it occurs
naturally and has no odor—than more technical issues—what affects the
level measured and when do health problems! usually occur. Quiz perfor-
mance was at least in the range of 70 to 80 percent—correct for the general
knowledge questions but it declined to 40 to 45ipercent for the more tecnnical
questions. This suggests that there remains more to be done in
communicating information about measurement factors and the timing 'of
health problems. . ' ! . '-
Fourth, Frederick and Hagerstown homeowners in,both samples also
improved significantly on being able to identify the health problems caused by
raaon (Q17), and the question associated with knowing what can be done to
reduce radon levels (Q18). Hagerstown pianel members had 20 and
26 percentage points increases, respectively; Frederick panel members
increased by 24 percentage points and 18 percentage points, respectively.
Randallstown panel members improved slightly; more, with 1,7 percentage and
20 percentage'points, respectively. \,<.v;i;|>-
Fifth, the pattern of responses suggests that the baseline interview may have
served as a prompt for learning in the panel sample. Appendix E presents the
results of. statistical- tests for learning. Our findings indicate that there was no,
sensitization in Frederick, but there was sensittiation in both Hagerstown and
Randallstown. One reason why Frederick was less affected may be,the
higher initial levels of knowledge in the baseline survey. In Frederick, the
baseline interview communicated less new informeition.
Finally, the.finding of knowledge sensitization in Randallstown, but not in,
Frederick, is particularly significant. Because our comparisons are made
between the treatment community, Frederick, and the comparison
community, Randallstown, sensitization implies that we are understating the
changes,in knowledge attributable to EPA's experimental programs when we
use the panel sample for comparison. The knowledge indicator was the only
one which showed evidence of any significant sensitization of panel
members. . • .;•-.-•,.! •.;•:' •"•' • -v->. ••
Figure 6-1 provides some additional insight into the knowledge indicator by
comparing the performance of testers and nontesters on four key questions in.
the'foliowup survey For this figure, we have pooled the panel and
independent foliowup-survey responses. The figure shows a clear pattern
RTI 6-3
-------
6. KNOWLEDGE ABOUT RADON
RELATIVE PERFORMANCE OF TESTERS VERSUS
NONTESTERS ON KEY KNOWLEDGE
QUESTIONS IN THE FOLLOWUP SURVEYS
FREDERICK
Question 12 Question 15 Question 17 Question 18
HAGERSTOWN
Question 12 Question 15 Question 17 Question 18
RANDALLSTOWN
Figure 6-1. Key knowledge questions.
RTI 6-4
-------
6. KNOWLEDGE ABOUT RADON
that people who had tested their homes had (Significantly higher knowledge of
these,key questions. -The differences are oh the order of 15 to 20 percent.
Unfortunately, we cannot answer whether higher knowledge led them to test,
or they acquired more knowledge as a resujlt of testing. Either explanation
would be supported by the data. A conclusive answer would have required
us to have a knowledge measure at the time when people decided to test.
Such measures are usually impractical i'n a standard survey context.
However/some type of'evaluation might be; developed in conjunction with
testing companies to acquire an answer to this important communication
issue. '!"'.-• •
Figures 6-2(a) through 6-2(d) profile learning patterns of the panel sample for
the same four key questions. Learning patterns are similar in allthree tcwns.
By linking responses for the same individuals v/e can develop more refined
learning measures than are possible with the independent samples. Figure
6-2(a) shows that homeowners in Frederick and Hagerstown were more likely
to know that radon is a natural source of radiation in both surveys than
homeowners in Randallstown. ' j ' , •
'' • - ' !' •
Figure 6-2(b) shows the substantial improvement in homeowners' knowledge
of now to test for radon—the main messaige of the EPA communication
program. Learning is 12 points,higher in Frederick than in Randallstown,
while Hagerstown homeowners improved only 3 points more than those in
Randallstown.. Frederick also had about 15 points fewer homeowners answer
the question incorrectly in both surveys. . !.' .
i-
* , ,•-, i .'''•' •
.Figures 6-2(c) and 6-2(d) follow a pattern that is similar to Figure 6-2(b). More
Frederick homeowners answered the questions correctly .in both surveys than
in-Randallstown. Hagerstown homeowners a|so did better than Randallstown
homeowners on both questions in this respeclt, but the differences are usually
smaller. Learning is very similar across this towns, which may reflect the
sensitization caused by the baseline interview .in.both Hagerstown and
Randallstown. I " - . .
Table 6-2(a) shows=the results from statistical janalyses of the performance on
the radon quiz. To estimate these models, we used the standard regression
analysis to explain performance on the|quiz as a function of EPA's
experimental program, various socioeccinomic factors, and potential
confounding factors such as television and '^he Roberta Baskin campaign.
We present the results from the standard arialysis for both the independent
and panel samples For illustration purposes jwe also used the more complex
Poisson regression technique and found very! similar results. -•
- * ' i " " | -.
The R2 summary statistics indicate that the model explain about 30 percent of
the variation in knowledge 'R2 measures how well the model fits the data.
Appendix C provides more discussion on theluse of R2 in evaluating a model,
It is generally lower for cross section-than th'e time series data. The results
RTI 6-5
-------
_£iANUWLED:GE ABOUT RADON
80
70
604
50
40-
30
20-
10-
0
"WHERE DOES MOST RADON IN HOMES COME
FROM?"
PANEL SAMPLE PERFORMANCE BY CITY
57
(a)
35
25
18
6 7 8
46,
35
20
22 23
UNLEARN . WRONG ON BOTH RIGHT ON BOTH
LEARN
80-
70"
60-
50"
40-
30"
20-
10-
0--
"HOW CAN YOU TEST YOUR HOME FOR RADON?"
PANEL SAMPLE PERFORMANCE BY CITY
(b)
4 3
UNLEARN
WRONG ON BOTH RIGHT ON BOTH
LEARN
FREDERICK D HAGERSTOWN • RANDAIlSTOWN^
Figure 6-2. Panel sample radon knowledge.
RTI 6-6
-------
"WHAT KIND OF HEALTH PROBLEMS
OF RADON EXPOSURE LIKELY
PANEL SAMPLE PERFORMANCE BY CITY
ARE HIGH LEVELS
TO CAUSE?"
UNLEARN
WRONG ON BOTH
RIGHT ON BOTH
LEARN
80"
70-
•60"
50-
40"
30-
20-
10-
"WHAT CAN HOMEOWNERS DO
RADON LEVELS IN THEIR
TO REDUCE HIGH
HOMES?"
PANEL SAMPLE PERFORMANCE BV CITY
48
25
28
17
10 11
6
38
UNLEARN
WRONG ON BOTH
RIGHT ON BOTH
Q FREDERICK D HAGERSTOWN
Figure 6-2. (con.)
(d)
32 31
26
%
LEARN
RANDALLSTOWN
RTI 6-7
-------
6. KNOWLEDGE ABOUT RADON
V)
0) ^
& "^
w 2.
o
£6
B)
TJ
_Q
I
£)
^)
TJ
O
S
0)
O)
0)
Table 6-2. Kn
(Number Corre
S N
0)
?SHIillIIsIIliIii1iil.il8
p 0000000066^66 66 -'6666^6-3-
g w
c
0)
TJ
I
d>
O
TJ
2 t^N.'o>^>O)o>u?'coco«cfl
3 S 8 8 S 8 §8 8 =•« I I 8 3 8 S-'8 8
T T9'r'9CNqT);ir>cr)Cvj-qcqw^;o^Kl^
0000660666^6666 <- 666 , _ ..,
w
Sf £T ^ ^'J^^S'ParPPorPo^C'rS' P in" n m
Vii22l'5t^tT3CNJT3''^l^iOCONiOr*.rt-rfS?tom
CNJ cq p 05 in ro ^-; T-• <}• in CD ^- •»- oo CD in eo. N- ro o
_.N^ ^.^^ _ —'>-^.>_-^_^^,
O «
fl.I
(0
Qj
3 00009 ooooocn oooo-oooo OID
| • . «
o
TJ
Q>
DJ
0)
TJ
0)
TJ
O
S -
• CD
8
g
8
§
00000006 66 oi 6 6 6 6 -r- 6
6o
ocb
Variables
LU
co or
to tu
LU CO
£1
111
O
D
H 2"
O
D
8B
LLQ
i-i-CE
O
O
'
v
II
UJCC
SLU
S 0
,
D:
UJ
CD.
O
cv
CC
.O
1
0>
o
CO
(0
o
w
o
I'
to
Q.
.c
s
o
£
p
5
RTI 6-8
-------
KNOWLEDGE ABOUT RADON
I , •
suggest that other unexplained factors, are affecting knowledge about radon
This is not surprising since we only used a 'snapshot at one point in time.
There may be room for exploring this further with the panel sample and using
the matching observations overtime to get a .better depiction of knowledge.
An important aspect of model evaluation is the significance of coefficients for
the variables in the moael and their stability across various specifications.- On
these criteria, the mode!'performs well with relatively stable coefficients and.
many significant variables. \ '
Among the implications that can be, drawn from the model are
j • '
• Frederick and Hagerstown homeowners both have higher levels of radon
knowledge than Randallstown, even after! controlling for exposure to the
Roberta Baskin campaign and various socioeconomic factors'
1 ' . • ' !' . ' • '
• People who called the State of Maryland's'toll-free number also performed
better on the quiz than those who djd not:! '• .
• People who talked about radon between trie surveys/had much hiaher ra-
don knowledge: , ' ; . .
• People who recalled seeing radon on televisionrhad higher knowledge
.. than those who did not recall seeing radon1 on television; •
• Whites and males had higher knowledge!than nonwhites a.nd females-
and , , j
•. Not surprisingly, homeowners with higher! education and higher income
. levels performed better than those with low;er levels.
j -
These findings are not definitive because other' factors are at work. But, they
do suggest that the EPA program has led to higher knowledge about radon in
both Frederick and-Hagerstown. They also suggest that both informal
channels—talking with other people about radon—and formal chan-
nels—television—contribute to higher levels pfj knowledge about radon. .-.
RTI 5-9
-------
-.••'**•'
-------
7. RADON TESTING
SECTION?
RADON TESTING
7.1 Introduction
This section examines changes in the-level of radon testing. Evaluations
often place the most weight on this indicator because it represents the
desired behavioral outcome of the risk communication program—more
homeowners testing their homes for'radon. It is also the most difficult indica-
tor to influence because it requires people 'to receive the risk message,
process it, and undertake an action. The literature on the success'of public
education, or risk communication, in encouraging behavioral changes
emphasizes how difficult it is to achieve this goal.
• • • I ' •"
This section provides some perspective on the nature of the radon testing
decision. It considers the characteristics o'f the decision, as well as the
potential consequences homeowners might-experience. It also discusses the
factors that influence the level of radon testing), especially the importance.of
informal communication channels. As part of pur analysis, we also examine
the factors that helped encourage informal-Tadon communication.
Finally, this section evaluates the reasons people gave for not testing their
homes We suggest that som.e nontesters may have gotten the risk
communication message and simply not yet gotten around to testing While
the ultimate evaluation is whether these proci;astinators subsequently test,
changes in their reasons for not testing cari also be associated with the
effectiveness of EPA's risk communication procjram'. • '
7.2 Characteristics of Radon Testing Decision
On the surface, the decision to test your home for radon appears relatively
simple The tests are relatively inexpensive! and not terribly difficult to
perform Unlike some actions (e g , smoking cessation) which have been the
focus of public information campaigns, the radon testing decision does not
require people 10 break some well-establishec, or even habitual, behavioral
pattern People need to buy the monitor, place [it in their homes, and return it
for analysis These characteristics of the rajdon testing decision would
suggest that increasing testing should not be tob difficult
RTI 7-1
-------
Below the surface, however, the radon testing decision is more complicated
For example, in our focus groups we found that nontesters found it easier to
put off testing because the radioactive gas is imperceptible to the
senses—there are^no sensory reminders of the'problern.to nag or annoy
people (Desvousges and Rink [1987a]). The implementation'of the'tests may
also put some people off—they might prefer having a testing company per-
form the monitoring or might be concerned about making mistakes in testing
their homes;
Svenson and Fischhoff [1985] suggest that the radon testing decision
involves four significant consequences: the radiation health risk; the immedi-
ate monetary costs; long-term costs; and worry. A low radon test result leads
to low levels of all four consequences. A moderate or high result can lead to
a more complicated decision situation. Mitigation could require significant
monetary outlays which might discourage people from testing—they want to
avoid the potential bad news. If people are further concerned about the avail-
ability or effectiveness of mitigation, this concern could also inhibit radon test-
ing ,- ' ;
An additionalcomplication is that radon testing may be relatively unfamiliar to
psoole. Unlike diagnostic tests for cholesterol in people, which have been the
subject of media attention for years, radon is a new risk. The availability of in-
formation about cholesterol, the reliability of tests for it, and what to do about
it are far greater as well. ;
.A final consideration in the decision to* test for radon is how the property
values of their homes might be affected if high radon levels are found. The
property value aspect appears to be a two-edged sword: some people
preparing to sell their, home may decide to test to avoid complications af the
time of sale; others may choose not to test in the hope that buyers will not
ask for a radon test. . ' . ' • ..
. - i - ' \
The bottom line on the characteristics of the radon testing decision seems
unclear The test is relatively inexpensive and involves no habitual behavior.
On the other hand, concern about mitigation expense or effectiveness could
be a significant deterrent. ' • • "''.-.
7.3 Changes in Radon Testing Levels
To measure the level of radon testing, we used the self-reports from the study
participants Self-reported responses would seem relatively reliable measures
of the actual levels of testing. It is unlikely that the homeowrier would have
forgotten about performing a test. It is an action that is likely "to stand out and
;was probably performed in the recent past, To minimize potential over-
reporting of responses.that are perceived to be socially desirable, the survey
questionnaire did not indicate or imply that increased testing was a study
objective If anything, self reports may understate actual-levels if the
RTI 7-2
-------
/. RADON TESTMfi
household respondent were unaware that another decisionmaker had tested
It .is also unclear that there is a feasible alternative to self-reported data.for'ra-
don tests, •• .•-•..-.. i
•• ' • • '"•••" ' . \ '•-'"••
Figure 7-1 highlights the percentage of homeowners who tested in the
baseline and followup surveys for the three! communities. The figure clearly
shows that testing increased by approximately 10 percentage points in
Frederick while.remaining constant in both Hagerstqwn and Randallstown
This difference was statistically significant'using standard tests and signifi-
cance criteria. None of the increase can be attributed to sensitization of
members in the panel sample-the level df testing is actually higher in the
independent followup than for panel households. The finding is also
supported by statistical tests for sensitization described in Appendix E.
;To further evaluate the factors that influenceb homeowners' decisions to test"
for radon, we performed a probit statisticallanalysis of the individual testina
decisions. Described in detail in Appendix A (see Table A-2), this analysis
identified three factors that increased the livelihood of a homeowner testing
his home before the EPA program was implemented: the level of knowledge
about radon, a qualitative variable designed as a health indicator and mea-
sured as a response to whether he asks his physician a lot of questions about
his health and, to a lesser extent, his education level. No negative influences
were statistically significant.; | - .•••••..
i
After the program was implemented, qurarialysis identified two factors that
. were significant factors in influencing testing in the panel followup survey: - :
" ' ' ' I • '
• Whether the homeowner lived in Frederick compared to Randallstown
• Whether the homeowner had talked with someone about radon/
• ' • . i . ' .
Among the factors that were not significant! were'income, age, whether the
homeowner recalled seeing the WJLA campaign, or whether the respondent
was a smoker. The Frederick variable indicates that the EPA program did
help to increase testing in that community. 'This variable was consistently
positive and significant for all versions of the model. Our model does explain
only a small amount bf the variation in the letting relationship- but this is not
surprising m cross-section data. Measure* of the explanatory power of
models estimated for qualitative dependent variables do not have the same
type of interpretation as they do with simpler models. They are simply'a -
crude index of models'fit to a particular set of data. '
1 ' .
Figure 7-2 sheds some light on the one factor" that differentiates testers from
nontesters—Whether they had talked with alfriend, neighbor/or coworker
about radon: Almost 80 percent of testers-in (Frederick and Hagerstown had
talked with someone about radon while only 49 percent and,38 percent of
RTI 7-3
-------
•7. RADQ.N TESTING
PERCENTAGE OF PEOPLE IN BASFTTvr Avn
INDEPENDENT FOLLOWUP WHOHAE[ TESTED
FOR RADON
FREDERICK
0 BASELINE SURVEY Q WDEPENDENr'FDLLOWUP
PERCENTAGE OF PEOPLE IN BASFI TXT
DEPENDENT FOLLOWUP WH(>HAEf TESTCb FOR
RADON
FREDERICK
HAGERSTOWN
RANDALLSTOWN
Figure 7-1. Radon testing.
RTI ?--
-------
7. RADON TESTING
PERCENTAGE OF TESTERS AND
THE FOLLOWUP SURVEYS WHO
SOMEONE ABOUT RADON
NONTESTERS IN
BAD TALKED TO
100
60
40
20
0
79
79
' 49
38
58
26
FREDERICK
HAGERSTOWN
RANDALLSTOWN
0 TESTERS
NONTESTEI S
Figure 7-2. Informal communication channels.
nontesters had discussed radon. In the
town, the levels are 20 percentage points
supporting information discussed below
communication channels play an important r
of radon testing
comparison community, Randalls-
bwer. From this figure, and
it seems clear that informal.
ro e in leading to increased levels
We also estimated a probit model for whether
friend, neighbor, or coworker about radon
the significant positive influences on talking
Frederick or Hagerstown (compared to
seeing a television program about radon:
were more likely than nonwhites to have
The only significant negative influence was
were less likely to have engaged in the inform.
These results provide some insights into one
risk communication (Krimsky and Plough
important: it was enhanced by EPA's
Hagerstown and was greater among
pODulaiipn—whites, younger homeowners,
levels As noted in an earlier study, olde
special zee commumcaticn channels (see
someone had talked with 3
the baseline survey, Among
about radon were living in
Randallstown). education: income.
the WJLA campaign. Whites
talked with someone about radon.
person's age—older people
ial communication abput radon.
aspect in the social process of
Informal communication is
in both Frederick and
several segments of. the target
those with higher education
r people may require more
etal. [1987j)
[1988])
programs
and
Smth
RTI 7-5
-------
7. RADON TESTING
7.4 Reasons for Not Testing
We also examined the reasons people gave for not testing their homes.
Using the panel sample survey responses, we placed the reasons people
gave into categories—-procrastinator, uninformed, not in my Ihome, not in my
area, antitester, and fatalisi—and modeled these categories on various
factors. Figure 7-3 presents the results for the more important categories. In
particular, this figure shows that
• Overall, the decreases in the "not in area" numbers in both Frederick
(14 percentage points) and Hagerstown (7 percentage points) suggest
that EPA's programs helped some people personalize the risk from radon
on the community level. However, they have not personalized the risk
completely, as evidenced by the increase of about 4 percentage points in
the category "not in my home." - ' «
• In Frederick the "not in area" category decreased by 14 percentage
, points. This compares to only 6 points in.Randallstown and 7 points in
Hagerstown. Proportions were similar between the independent and
dependent samples. ' :
• The number of Frederick respondents in the procrastinator category
increased by 12 percentage points in the panel sample and by
18 percentage points in the independent sample. In Hagerstown,
increases were 19 percentage points for the panel 'sample and 13
percentage points for the independent sample. Randailstown showed
increases of 13 percentage points and 6 percentage points in the panel
and independent samples, respectively. ' " '
•' There were no large differences among the percentage of nontesters in
the uninformed category—each city showed decreases of about
10 percentage points. However, Frederick had a lower initial proportion of
uninformed nontesters in the baseline—6 percentage points lower than
either Hagerstown or Randailstown. " • , • •
• Recall of WJLA's campaign and a person's educational level helped
explain-movement from uninformed to procrastinator (the; person said he
just had not gotten around to testing). Living in Hagerstown relative to
Randailstown and talking to someone about radon influenced the move
from not in home/area to procrastinator. '
There appears to have been sensitization of the panei sample in Hagerstown '
and Randailstown in regard to reasons for not testing. The proportion of
procrastinators in Hagerstown increased by 19 percentage, points in the panel
sample, and by 13 percentage points in the independent sample. Likewise,
the proportion of procrastinators in Randailstown increased by 13 percentage
points in the panel sample, but by only 6 percentage points in the
independent followup. These differences are statistically significant, but they
are net based on the models used to evaluate sensitization in other areas.
However, for the testing decision itself, we did not find any sensitization
among members of the panel sample. In fact, the testing increases in
Frederick were slightly higher for the independent sample, '[ •
BTI 7-6
-------
7. RADON TESTING
WHY I HAVEN'T TESTED: CHANGES BETWEEN
BASELINE AND FOLLOWUP SURVEYS
FREDERICK
10-
Uninformed
Not In Area
Not In Home
Uninfo'rmed
Not In Area ...
Not In Home
a BASELINE SURVEY D ivDereNDEhTFOLLowup • DEPENDENT FOLLOWUP
Figure 7-3. Reasons for not testing.
• Procrasiinators
Procrastinators
RTI 7-7
-------
-------
SECTIONS
MEASURING EFFECTIVENESS
8.1 Introduction
Measuring the effectiveness of a radon risk information program involves sub-
tle considerations. For example, Viscusi, Magat, and Huber [1986] described
effectiveness in terms of exercising a "soundludgment." While this criterion
is sensible, it is difficult to implement because! it requires the analyst to define
what constitutes a sound judgment. Or, we might ask whether individuals test
their homes and then whether they tak| e the "proper'" amount of
precautionary behavior after receiving the results. The process of developing
such evaluations is easiest for those at the very lowest and the very highest
for different reasons. In the middle, there'is inevitable ambiguity in the
benchmark. !' ,
•I ' '
What is clear, however, is that perceptual/behavior measures of effectiveness
are necessary because simply asking people, to evaluate effectiveness can be
very misleading. Fosr example, Smith etal. [1987] found almost 90 percent of
homeowners receiving a radon fact sheet! considered it very effective.
.Perceptual/behavior measures of effectiveness showed these same home-
owners to have less understanding of key radon concepts and a greater
divergence between their perceived and technical risks compared to similar
homeowners who received experimental brochures
'• •' I '•''-..
This section considers three questions'relatedjto effectiveness. First, what do
the various indicators show about the overall ejffectiveness of the.risk commu-
nication program'' Second, how do these findings compare with other public
information efforts to improve public health? Third, can we isolate the effects
of EPA's experimental risk communication program from the WJLA cam-
paign? . " . ' ! ' .
8.2 Overall Effectiveness
i , i
For an overall evaluation of effectiveness. Figure 8:1 highlights the changes in'
the four main indicators This figure compares each treatment community
with the comparison community, Randallstown. The results focus only on the
independent sample, but the general, patternjholds true for the dependent
sample Frederick, which received both EPA's media campaign and the
RTI 8-1
-------
6. MEASURING EFFECTIVENESS
FREDERICK VERSUS RANDALLSTOWX
HAGERSTOWN VERSUS RANDALLSTOWX
Awareness
Attitudes
Knowledge
D Hagerstown
Randallstown
Figure 8-1. Overall performance.
Testing
RTi 5-2
-------
8. MEASURING EFFECTIVENESS
community outreach, performs'better than RanjdaHstown in terms of increased
awareness (6 percentage points higher), attitujdinal changes (15 percentage '.
points-higher), knowledge (15 percentage,' points higher), and testing,'
(8 percentage points higher). The changes-in- indicators for Hagerstown,
which received only the media program* were much more limited with
attitudes the only measure to display a significant improvement (8 percentage
points) compared to Randallstown. There were no significant changes in the
summary measures of awareness, knowledge/or testing for Hagerstown
relative to Randallstown. AS noted earlier, the WJLA campaign may account
for some of the differences between Frederick iand Randallstown.
In addition to these specific results related to the relative effectiveness of the
two variations on the EPA program, we also observed some general patterns
about radon communication issues. There we|re high levels of awareness of
radon in each city in the followups with over 70| percent of respondents saying
that they had seen or heard about radon. People also, apparently realize .that -
it is important to test. They see mitigation as expensive, but a majority feel
that it,would lead to restored property values, j Nonetheless, there remains a
sizable percentage of people, between 30 arid 40 percent in Frederick and
Hagerstown, who feel that their property would be worth less even if a radon
problem were fixed, .• . | •' . , •
[
I * -
the program was effective in conveying general knowledge about radon. Our
results show that the respondents in all three towns did.hot know what factors
influence the radon measurement levels.andjwhat they can do to mitigate
their risk from radon. Neither of these themes was addressed in the risk
messages. ' . ' '-'•!.-. ' ' .
• . . f; - . • .
During the 3 months of the EPA program, testing levels reached about
15 percent in Frederick and remained at .about 5 percent in Hagerstown .and
in the comparison cdmmunity of Randallstown! While these levels are not as
high as one might hope from a public health perspective, they are quite rea-
sonable for a practical short-term communication program. They are also
consistent with testing levels elsewhere. In the Reading Prong area of New
Jersey where radon has been a major concern for several years, only 10 to
15 percent of the homeowners have tested for radon (Newsweek [1988]).
f • ' . •' •
Changes in the reasons why people have not! tested for radon also indicate
that the program has had effects. We argued'jearlier that changing people's
attitudes so that they acknowledge radon as a.problem.even if they had put
off testing should be regarded as making a mojdest positive contribution. 0,n
this basis, we examined the effects on those having attitudes consistent with
what we have designated procrastinators. The.proportion of procrastinators
has risen significantly, while the proportion of uninformed nontesters has '
decreased The number of people who still believe that they do not need to
test because radon is not a problem in their afea has decreased. But about.
10 percent of nontesters in the treatment conlmunities still feel that radon is
RTI 8-3
-------
not a problem in their area. Since the risk messages stressed personalization
at the city level, these people could be regarded as still possessing an
uninformed attitude. Approximately one-fourth of the nontesters still say that
radon is not a problem in their home. Such a large number indicates that get-
ting people to personalize the risk to the level of their home remains difficult.
For perspective, the results for our indicators can be'viewed against other
evaluations of public information programs. Table 8-1 presents a summary of
studies that have evaluated various public information programs The
summary updates the Lau et al. [1980] and Wallack [1981] reviews. Several
interesting points can be drawn from this table. First, as both'Lau et al. [1980]
and Wallack [1981] persuasively argue, the designs underlying many of the
evaluations have serious flaws, most notably the lack of control groups and
inadequate baseline information. In contrast, the design of the Maryland
study mirrors the features of the design proposed by Lau let.al.'[1980] as
shown in Table 3-2 (page 3-8). The major weakness in our evaluation is the
response rate for the telephone surveys. Nevertheless, these fates are based
on a random sample and are comparable to those for other environmental
risk studies that have used mail techniques or mall intercept surveys They
are also comparable to, or higher than, the response rates :for the political
poi!s (seeMorin [1988]). ;
Second, the magnitudes of our indicators measure up reasonably well when
compared to the other studies. For example, the Stanford, California
community-based risk information program for preventing heart disease
found that knowledge of cardiovascular risk factors increased 50 percent in
' the control city, 250 percent in the media-only city, and 400 percent in a
group which received intensive individual instruction (Lau et al [1980]). This
campaign was far more extensive than what our budget allowed, but we
achieved a 132 percent increase in knowledge from 28 to 65 percent correct
answers to questions about radon testing in Frederick and a 100 percent .in-
crease from 22 to 44 percent in Hagerstowrf. We also found1 significant atti-
tude changes and an increase of 136 percent (from 6.9 to 16.3 percent) of
homeowners testing their homes for radon. :
- '" ' - ' t •
Two recent studies, however,-offer some additional promise that risk commu-
nication can be effective. An evaluation in Finland shows that a weekly televi-
sion series on health topics resulted in changes in health-related behaviors
including smoking cessation and weight loss (Wiio, Puskai'and McAlister
[1986]). Flay et al. [1986] also found that integrating mass media viewing into -
the school resulted in significantly higher exposure and positive changes in •
knowledge, attitude, and behavior. • ' ^ '
RTI 8-4
-------
8. MEASURING EFFECTIVENESS
Table 8-1. Evaluations of Public Information Programs
,C
Study
Farquhar et al
|i977]a
MaccoDy et al
[I977)a
Nash and
Farquhar
11978]"
Stem et a!
(19761"
Fieisner [1973]
Freimuth and
Marron (1978]°
Greenberg and
Gantz |197S]a
Menaeisohn
[1973!3
O'Keefe [I97i]a
Robertson
[1974!C
Salzar.
Marsnali, ana
Glazer |1977]a
Schmelmg and
Wotnng (I976!a
Simons-Morton.
Bnnk. and
Baiesil987]
Uar>' |l972;a
Warner [1977]a
Wuo. Koskeia.
ana McAiister
|19851
Yarneli J1976ja
I
Type of Behavior
RISK of cardiovascular diseas
Increasing seat bell use
Increasing demand for health
information
Increasing knowledge about
veneral disease
Instill pnde and order m Los
Angeles hispanic community
Stopping smoking
Increasing seat belt use
Encouraging free medical .
screening among the eloerty
Preventing prescnption drug
abuse among women
Increasing seat belt use
Increasing use of
contraceptives
Stopping smoking
i
Health behaviors such as
changing diet and stopping
smoking
Encouraging riotners to
vaccinate chitaren agamst
measies
Communication
Program
Multi-media
campaign and
personal instruction
Media campaign in
two cities; one control
city
Radio and television
announcements
"Television show
Television program'
Television
commercials
Television campaign
Cable television
commercials in a
New Vbrk highnse
Television
commercials
Vanous means.
including Persuasive
Communication (PC)
only, PC plus
monrtonng, PC plus
incentives, and PC
plus incentives and
prompts, study
included control
group as well
Media campaign with
treatment and control
cities oesign
Surgeon General's
report, radio, and TV
commercials
Weekly television
series
Maileo an invitation to
clinic
Effectlveneu Meaiure
Measured knowledge changes and
physical changes at 1 year intervals
No major change in attitudes about
Mat bete among ttni tnreo cities '
due to campaign !
10% of sample requested
additional information
Viewers 10-25% nxxe
knowledgeable than nonviwwrs
6% of viewers-reported joming a
community organization— (he goal
of the campaign i
1% of nonsmokers aid exsmokers
said commercials had an effect in
.. helping them stop smokinji
No effect on use of Mat belts
3-14% of the residents participated
in the screenings
No change between lire- and post-
attitudes; no different in altitudes
between viewers and nonviewers
Communication only and
communication plus monitoring
produced changes ofJT-IWb: PC
with incentives and incentives with
prompts showed increases of well
over 100%
}
[
No difference in saies>of
contraceptives between moment
and control communities
3-6% annual decrees*) in por
capita cigarette consumption
4-7% of viewers lost weight;
12-24% decreased fat 'intake: and
8% tried to stop smoking
Knowleage of campaign did not
distinguish atienders from
nonattenoers ;
Potential Limitations
Comparability of different
sites; independent event
could happen in one site
and not the others; cost
Nonequivaient control group
Things other than program
could have influenced
change
No baseline survey; no
control group; correlational
only
No control group; no
baseline survey;
correlational only
No control group: only
interviewee program
viewers: no baseline survey
Things other than program
could have influenced
change
No baseline survey; no
control group: correlational
only; potential bias in
sample
No control group:
correlational only
Small sample size: 'selective
population
Nonequivaient control group
Time-series design with no
comrol.over other forces
which could have
nfluenced change ,
Influence of otner factors on
behavior: self-reporting with .
small number of validation
interviews
No control group: no
baseline survey:
correlational only
aCitec in uu e: a: |1980!
3Ct,teo m Mammono. Freimuth. and Momson [19S7]
:C;tecir Waiiack[i98-;
RTI 8-5
-------
MEASURING EFFECTIVENESS
Another .mportant consideration to keep in mind when evaluating the effec-
tiveness of a media campaign is the role of the gatekeeper. The gatekeeper
refers to the screening operative, be it a programming director or public
affairs officer, who decides when and how often a particular public service an-
nouncement (PSA) is broadcast Hammond, Freimuth, and Morrison [1987]
found that the gatekeeping p-ocess effectively narrowed the potential target
audience for cancer prevention PSAs to 14 percent of the original target
population _Sueh a reduction in the potential audience coupled with the fact
that only a given percentage of the remaining audience will, see the message,
much less take action as a result of it, greatly reduces the chances that a
communication campaign will produce any behavioral change at all The
Frederick program directly involved the'community leaders who also are im-
portant gatekeepers for public opinion. The presentations to key community
organizations reached an even broadef group of potential gatekeepers In -
contrast, the Hagerstown media effort had to rely on the programminq '
directors at the radio'stations and the newspaper editor to disseminate the
message -. ' . ;. •
Finally evaluations of actual behavior changes are the exception rather than
.he rule in public, information campaigns. Our 10 percentage point increase in
test,ng involves the final stage of the decision process-Hhe actual behavior
Many of the campaigns in Table 8-1 focus only on indicators such as
awareness, knowledge, or attitudes. In the area of product advertising, when
an evaluation is actually performed, companies, are often content to see a 1 to
2 percentage point increase in market share (Lau et al [1980]) Wallack
[1981] also cites a 1 percentage point increase in market share as significant
Comparisons with advertising effectiveness are not completely correct
because a company's advertising has to compete with products substitutes
and their advertising. Nonetheless, a risk communication, program has to
compete with other health-related advertising and information that people
receive from various sources. '
8.3 Isolating the WJLA Factor
Another important consideration in evaluating effectiveness for this study is
the need to isolate the EPA program effects from other confounding activities '
Most prominent among these is the WJLA TV radon campaign that could be
seen in all three communities almost simultaneously with the EPA program.
To examine effectiveness we-haye-used .our statistical models to simulate
what would have occurred with and without the EPA .communication program •
(See Apoendix A for a more .detailed discussion of this simulation) The basic
components of the simulation include'•
RTI 6-6
-------
_8. MEASURING EFFECTIVENESS
•_ The coefficients from the statistical modejls; . ' •
•. The demographic characteristics of a typical person in each town—we
used arithmetic averages for each socioe;coriomic or attitudinal variable;
• The effect of the EPA program—estimated from the statistical models'
. and .••]•-•..
• The effect of the WJLA campaign—also estimated from the statistical
models. ' j- .
To measure the effect of the WJLA campaign, we have used a high and low
range. Our low end of the range is based on the variable that measures
whether or not the respondents recalled seeing a WJLA campaign. The high
end of the range includes both the low end measure and a measure based on
respondents who said they saw or heard about radon from any television
source. . • | '
In the simulation, we focus on the^predicted/ effect of the EPA program on
awareness, knowledge, testing, and reasonis nontesters did not test. Our
'simulations consider the following cases: •'[ '.".'•••
'. [••"'•'.'
• What effect would the EPA program have had on the various indicators for
Frederick?, . • j - • -•'•„•
• What would have happened without the WJLA campaign using the high
and low range.for the effect? -. |
• What effect would the EPA program that v/as used in Frederick have had if
it had been implemented in Hagerstown arid Randallstown?
• What would have happened with the EPAi program but .without the WJLA
campaign in Hagerstown and Frederick? i ,
Figure 8-2 presents the simulation results for the typical person in each town
being aware of radon in both surveys. The simulation results reported in
these figures.show that in all situations injthe two treatment towns, the
probability of the typical person being aware of radon in both surveys
exceeds 80 percen-t. The simulation also suggests that-the EPA program
does not have an appreciable effect on the probability of being aware in both
surveys In this case, the WJLA campaign has a larger predicted effect than
the EPA program—ranging from about 6 to 53 percentage points. Clearly,
television had the largest single effect.on the pirobability that people would be
aware of radon in both surveys. However, as fioted in Section 5.2, when prior
levels of awareness are accounted for in the statistical analysis, there was a
Significant effect of the EPA program in Frederick, even after the television var-
iable is controlled for in the model f
RTI 8-7
-------
_s. MEASURING' EFFECTIVENESS
PROBABILITY OF HAVING SEEN OR HEARD
ABOUT RADON IN BOTH SURVEYS
FREDERICK
99.1
93.3
45.5
100T
Overall
99.1
Without EPA Program Without WJLA
(low)
HAGERSTOWN
98.3
913
Without W.JLA.
(high) ,
48.9
Overall Without EPA Program Without WJLA
(low)
RANDALLSTOWN
78.8
Figure 8-2. Awareness of radon in both surveys.
Without WJLA
(high)
44.7
Without EPA Program Without WJLA Without WJLA.
(low)
RTI 8-8
-------
8, MEASURING EFFECTIVENESS
Figure 8-3 presents the results, of our simulation-.of program effectiveness on
the probability of whether atypical person kn|ew the correct answer to the
question;,"How can you test your home for radon?" The simulation
considers two outcomes/ whether the typical person answered the question
correctly on both surveys or whether the typical person learrfed the correct
answer between the baseline and followup surveys. The figures show that the
EPA program has little effect on the probability of both correct, but increases
the probability of learning for the typical persoip by approximately 9 percent.
This predicted increase is similar across all cjommunities. The simulation
suggests that the range for the WJLA campaign is quite wide; it ranges from'
almost no effect on either outcome in any of the communities to
15 percentage points for learning in both Frederick and Hagerstown. Both
the EPA program and the high measure for thejWJLA campaign seem to have
affected learning. ! .'•''! ' '
Figure 8-4 presents analogous results fojr the question, "What can
homeowners do to reduce high radon levels in their homes?'' These
simulations suggest that the EPA program has little or no effect on the
probability of learning, but increased the probability of the typical person hav-
ing correct answers on both surveys by approximately 10 percentage points
This pattern is similar for all three towns. Agairji, the WJLA campaign ranges
from no effect to approximately 12 percentage [points, - , '
Figure 8-5 presents the simulation results forjthe effectiveness of the EPA
program on the probability of the typical person testing their home for radon.
The simulation suggests that the program.increases the probability of some-
one testing by approximately 5 percent. Thejincrease is similar across all
three towns. The simulation results suggest that WJLA campaign has almost
no effect on the probability of the typical perslon testing in any of the three
communities. Only -the low WJLA variable is 'Included because whether a
person recalled seeing something about radon on television in the followup
survey was not a significant determinant of whether they tested their homes
~for radon. Figure 8-6 shows the simulation .results for the probability of some-
one having talked to a friend, neighbor, or colworker about radon as mea-
sured in the followup survey. The EPA program raises the probability of infor-
mal communication taking place by 16 percentage points.' The WJLA effect
ranges from 2 to 21 percentage points depending on the measure chosen. -
I • '
In summary, the simulation results suggest that the EPA risk communication
program has increased the chances that the tyJDical homeowner would know
more about radon testing and would test his dr. her home for radon. After
other factors are controlled for in the simulation, the EPA program effects
generally increase from about 5 to 10 percentage points for the knowjedge
and testing indicators The simulation also suggests that the effect of the
WJLA campaign depends on how the campaign i's measured In the low end
of the range, the simulation shows only a modest increase in awareness. On
the high end, the campaign has effects that a!re somewhat larger than the
EPA program for awareness, knowledge, and informal radon communication
'. . - - i • ' "-....
___^_ - RTI 8-9
-------
PROBABILITY OF A CORRECT ANSWER TO
HOW CAN YOU TEST YOUR HOME FOR RADON?"
FREDERICK
Overall
Without EPA Program Without WJLA Without WJLA
. flow) " (high),,
HAGERSTOWN 1
Overall Without EPA Program Without WJLA
(low)
Without WJLA
' (high) :
RANDALLSTOWN
Without EPA Program Without WJLA
(low)
G2 RijhionBoih • Lcam
Without WJLA
(high)
Figure 8-3. Correct answer probability-testing home for radons
RTI 8-ID
-------
ADMEASURING EFFECTIVENESS
WHAT CAN HOMEOWNERS DO
HIGH RADON LEVELS IN
TO REDUCE
THEIR HOMES?
FREDERICK
Overall
Without EPA Program Without WJLA
(low)
HAGERSTOWN
Overall
Without EPA Program Without WJLA
(low)
RANDALLSTOWN
34.9
Overall
Without EPA Program Without WJLA
(low)
0 Righi on Boih B Lcam
Figure 8-4, Probability of,correct answer-mitigation knowledge.
Without WJLA
(high) '
Without WJLA
(high i
20.7
Without WJLA
(high)
RTI 6-11
-------
PROGRAM EFFECTIVENESS: PROBABILITY OF
TESTING FOR RADON
"(PANEL-SAMPLE)
Overall
4'•
8.6
Overall
4 •
6.1
i Kerall
FREDERICK
Without EPA Program ' Without WJLA (low)
HAGERSTOWN
8.4
3.4
Without EPA Program
RAND'ALLSTOWN
Without WJLA (low)
6.0
.2.3
Without EPA Program Without WJLA i low i
Figure 8-5. Testing probability.
RTI 8-12
-------
PROBABILITY OF TALKING ABOUT RADON-
PANEL SAMPLE FOLLOWUP
60%"
50%"
V
40%-'
30%-
20% •
10% •
Overall
58%
FREDERICK
49%
35%
Without EPA Program Without WJLA j Without WJLA
(high)
HAGERSTOWN
42%
0%-erall Without EPA Program Without WJLA j
(low) j
RANDALLSTQWN i
Without WJLA
(high)
43%
.'27%
Figure 8-6. Discussing radon in followup.
30%
Overall Without EPA Program Without WJLA j Without WJLA
How) (high)
RTI 8-13
-------
The simulation also suggests that there was no measured efffect of the WJLA -
campaign on testing decisions. This: may be somewhat misleading because
of the increased availability of test kits,in Frederick as part of the WJLA
campaign.. The relative importance of this factor is unclear, While availability
of test kits probably increased testing in Frederick, none of the sample
members who had tested mentioned the WJLA campaign :in open-ended
comments. -
8.4 Implications
From an overall risk communication perspective, one can view the Frederick
program as including community outreach and presentations, radio public
service announcements, utility bill inserts, and the extensive WJLA campaign.
After all of this risk communication, testing in Frederick only increased by
about 10 percentage points. At the same time, however, the combined
Frederick program may have improved conditions for increased testing later.
The increases in the number of people feeling that it is important to test and
those who have procrastinated about testing, and in the proportion of
respondents having more knowledge about radpn, all support this
hypothesis The increased levels of people in Frederick talking to each other
also had a positive influence on testing and more informed attitudes.
However, reaching high levels of testing with a modest risk communication
program is unlikely. Even if all of the procrastinators in Frederick'eventually
did test, testing levels would only increase to around 35 percent.
• Based on the results in Hagerstown, there appears to be relatively little payoff
to a small-scale media effort alone. There was less exposure to WJLA in this
town. Radio and newspaper public service announcements'and utility bill
inserts served as the main risk communication vehicles. There were modest
improvements from baseline levels in the areas of awareness, knowledge;
and more informed attitudes; however, the differences were not very much
larger than in Randallstown, which received no treatment -"'.-'
Our study shows that risk communicators still do not know how to convince
people to test for .radon.. People are uncomfortable with mitigation and its
potential effectiveness. Reducing this concern is likely to require more
widespread certification of mitigation contractors and more communication
about effectiveness The community-based programs, such as the one in
Frederick, offer a good vehicle for "word of mouth" mitigation success stories
to be passed along. Radon mitigation needs.to be perceived as being in the
mainstream of normal home repairs. The formal communication messages
alone—such as the recent press announcements—are unlikely to achieve
this goal .
A topic only touched on in our study is the issue of property values and how it
relates to the testing decision. Perhaps it will require sounding the alarm to
homeowners that testing now could make it easier to sell a house later., A
____ RTJ'B-U
-------
I ' - '
communications message might emphasize tfisat radon tests may become a
required part of all. real estate transactions. Such issues were not addressed
in our study because the/Agency felt that it v\[as unclear whether it was within
their jurisdiction to be providing information or raising concerns in these
areas. This in turn implies' that states mayneed to rethink the boundaries
they place on their radon programs with reciard to what types of information
these agencies can provide. .
•. • -i ..-•/-.
Finally, is the recent Public Health Service Advisory strong enough to increase:
testing on its own? Testing levels could be influenced more by,a public
health message that may overstate the risk to some in hopes of erring on the
safe side. However, based on our results, we think the answer is no. 'Even in
Frederick, where residents had community presentations, unsolicited
mailings,"and coverage on both television and radio, testing levels increased
to rather modesHevels. The messages heiard in Frederick emphasized the
health risk from radon and that high levels ha!d indeed been found in that city.'
Still, 85 percent have not tested. This suggests that the Advisory may lead to
a short-lived increase in testing, but that its effects will soon fade away without
a comprehensive .communications strategy for radon
Because of the difficulty in getting people t4p. test their homes for radon, we
recommend that a radon risk reduction strategy include the following
elements: . . . . j.. • . -,-•.,
'! • '
• Legislative action by states to require! radon testing for all property
transactions . I
•' Building code changes at the state and local levels to make new construc-
tion "radon free" ' j "'•'"." ' ' ' . • •
i . - '
> Increased certification of radon mitigation! contractors - '
• A comprehensive risk communication program that identifies national
concerns about radon testing and mitigation and provides communica-
tions assistance to state'and local .governments for organizing
•community-based communications efforts. •
;i " *
A comprehensive strategy is needed because none of these elements alone
will lead to radon testing levels and mitigation .that are commensurate with the
risk. Increased legislative actions to require'rhore testing will only work if peo-
ple know how to test, where to get test kits, h;ow to interpret the risks implied
by their test results, and how to make sound mitigation decisions. A compre-
hensive risk communication program is essential to making these actions
happen with a minimum of hassle and anxiety. At/the present time, we see
only small parts of such a plan in EPA's current risk communication activities.
'••"'. ' i ' • '. .. •'
i . . . . ••
Such a comprehensive program will requirfe substantial commitment and
resources However, it would be much iless expensive than a Direct
regulatory approach Given the widespread concern about the, riskS'frOm
radon, such a program seems appropriate j "-'.--
' 4 ' i
— . • " : ~~~i ^ RTI 8-15
-------
• r
-------
s.
SECTION 9
REFERENCES
Aizen, I., and M. Fishbien, 1977, "Attitude-Behavior Relations; A Theoretical
Analysis and Review of Empirical Research," Psychological Bulletin, Vol
84, pp. 888-918. , i ,
Coyle, Marcia, and Stephen Drachler, 1986, "Officials Tread Lightly in Radon
Minefield," The Morning Call, October28,!pp. 1-3.
Desvousges, William H., and Hillery H, Rink, HI, 1987a, "Maryland Radon Risk
Study: Focus Group Findings," prepared for Office of Policy, Planning;
..and Evaluation, U.S. Environmental Protection Agency, Washington, D C ,
Research Triangle Institute, Research triangle Park, North Carolina,
September.. . ' .; j , • •. , • . . •
f ' '
! " . . -" '
Desvousges, William H., and Hillery H. Rink 111, 1987b, "Communicating Ra-
don Risk Effectively-Radon Testing in Maryland: Hagerstown Focus
Group Summary," Prepared for Office of policy, Planning, and Evaluation,
U.S Environmental Protection Agency-, Wiashington, D.C., Research Trian-
gle Institute, Research Triangle Park, North Carolina, November.
Farquhar, J. W., et al., 1977, Community (Educationlor Cardiovascular
Health," Lancer, Vol: 1, pp. 1192-98.
Flay, B. R...M. A Pentz, C. A. Johnson, S. Susisman, J. Mestell, L. Scheier, L
M, Collins, and W. B. Hansen, 1986, "Reaching Children with Mass Media
Health Promotion Programs: The Relative! Effectiveness of an Advertising
Campaign, a Community-Based Program, and a School-Based Program,"
in D. S. Leathar, G. B. Hastings, K. O'Reilly, and J. K. Davies, eds., Health
Education and the Media II, Oxford, U.K.: Pergamon Press, pp. 149-54.
Fleisher, G. A., 1973, "A study of a Radio/TV Campaign on Safety Belt Use,"
J. Safety ftes., Vol 5, pp 3-11. j
Freimuth, V. S.. and T. Marron,11978, MThe Public's Use of Health
Information," Health Education, July-August, pp. 18-20.
Greenberg, B , and W Gantz, 1976, "PublicJTelevision and Taboo Topics:
The Impact of 'VD Blues'," Public Telecommunications Rev., Vol. 4, pp
> '' ' !' ' ' " •
Hammond, Sharon Lee, Vicki S Freimuth, and William Morrison, 1987, "The
Gatekeeping Funnel: Tracking a Major PSA Campaign from Distribution
through Gatekeepers to Target Audience1," Health Education Quarterly,
Vol 14, No 2. Summer, John Wiley and Sons. ' • •
RTI 9-1
-------
9. REFERENCES
Krimsky, Sheldon, and Alonzo Plough, 1988, Environmental Hazards:
Communicating Risks as a, 'Social Process, Dover, Massachusetts'
A • • " .
h.aS R°b,ert uan?' Sandra Berry< John Ware,' and Douglas Roy,
, Channeling Health: A Review of the Evaluation of Televised Health
Campaigns, Health Education Quarterly, Vol. 7, No. 1-, Spring, pp. 56-89.
Maccoby, N., et al, 1977, "Reducing the Risk of Cardiovascular Disease," J.
Community Health, Vol. 3, pp. 100-14.
•'Man°^c^Td^Q^ Social Marketing: A New Imperative for Public
Health, New York: Praeger. • . .
; i .
McGuire, William J., 1985, "Attitudes and Attitude Chancje/' in Gardner
' Tu-J?^ E llot Aronson' eds- Handbook of Social Psychology, Vol II
Third Edition, New York: Random House, pp. 233-304.
Mendelsohn HA., 1973, "Some Reasons Why Information Campaigns Can
Succeed, Public Opinion, Vol. 37, pp. 50-61.
Morin, Richard, 1988, "Behind the Numbers: Confessions of a Pollster "• The
Washington Post, October 6. ' : '
Nas*h/ i-9 ' and J W Farquahar, 1978, "Community Approaches to Dietary
Modification and Obesity," in A J. Stunkard (ed.); Psychiatiric Clinics of
North America: Obesity, Philadelphia: Saunders. • .'
Newsweek, 1988, "The Risk from Radon," September 26 , •
O'Keefe, M T.-, 1971, "The Anti-Smoking Commercial's A study "of
Television s Impact on Behavior," Public Opinion, Vol 35, pp 242-48
Robertson, L S., et a!., 1974, ;'A Controlled Study of the Effects of Television
Messages on Safety Belt Use," American Journal of Public Health Vol
64, pp 1071-80. ••••-• ; '
Salzar, J E., C. L Marshall, and E. R. Glazer, 1977, "The: Use of Cable '"
Television as a Tool in. Health Education of the Elderly: Screening "
. Health Educ. Monogr., Vol. 5, pp. 363-78.
Schmeling, D., and E Wotring, .1976, "Agenda Setting Effects of Drug Abuse
Public Service Ads, "Jouma//sm, Vol. 53, pp. 743-746. . ;.
Simons-Morton, , Bruce G ... Susan Brink, and Deana Bates 1987
•Effectiveness and Cost Effectiveness of Persuasive Communications and
Incentives in Increasing Safety .Belt Use," Health Education Quarterly Vol
14, No. 2, Summer, John Wiley and Sons •
Smi!t2- V Kerry, WHIiam H Desvousges, Ann Fisher, and F. Reed Johnson
1987, Communicating Radon, Risk Effectively: A Mid-Course Evaluation
prepared for the Office of Policy Analysis, U S Environmental Protection
Agency, under Cooperative Agreement No CR-811075, by Vanderbilt""
University, Nashville, Tennessee, and Research Triangle Institute'
Research Triangle Park, North Carolina
RTI 9-2
-------
9. REFERENCES
Sm'!n'oo- Kfrry' Willia*m H: Desvous9es< Ann Fisher, and F. Reed Johnson,
, ••Learning About Radon's Risk," 'Journal of Risk and Uncertainty,
.._
. Stern, M. P., et al., 1976, "Results of a Two-Year Health Education Campaign
of Dietary Behavior," Circulation, Vol. 54 j pp. 826-33.
Svenson, Ola, and Baruch Fischhoff, 1985, "Levels of Environmental
Decisions, Journal of Environmental Psychology," Vol. 5, pp. 55-67.
Telecon . . -' . - [: -'.-.'.
Crellin, Glenn, 1988, National Association of Realtors, with Hillery Rink,
Research Triangle Institute, December 1-
P •••.'"•
Hockensmith, Dee, 1988, Greater Hagerstown Board of Realtors with
. Hillery Rink, Research Triangle Insitut'e, December 1.
!
Udry, J R., et al., 1972, "Can Mass Media Advertising Increase Contraceptive
Use7" Fam. Planning Perspect, Vol. 4. pip. 37-44.
U.S. Department of Agriculture, 1977, Scdp;e and Methods of the Statistical
Reporting Service, Misc. Pub. 1308, Washinqton, D.C : U S Government
Printing Office, pp. 6-8, , -j - * -• •
i ' • ' ...
U.S Department of Commerce, 1978,,Burejau of the Census, The Current
Population Survey: Design and Methodology, Technical Paper 40
Washington, DC.: U.S., Government Printing Office. '
U.S. Department of Labor, T986, Consume/; Expend/tore Survey: Interview
Survey 1982-83, Bulletin 2246, Washington, D.C. : U S Government
• Printing Office,, pp. 136-39 | • , , . '
U.S. Environmental Protection Agency. 1987, •'Unfinished Business: A
Comparative Assessment of Environmental Problems," Washington DC-
U.S. Environmental Protection Agency, February. '"
| ,
U.S. Environmental Protection Agency, 1988, "Region III/OPPE Radon Risk
Communication Project," Final Report! Program Evaluation Division,
Office of.Management Systems and Evaluation, Washinqton DC Au-
gust -' j .'•''•.• ;>
Viscusi, W Kip, W A Magat, and Joel Huberj 1986, "Informational Regulation
of Consumer Health Risks: An Empirical Evaluation of Hazard Warnings "
Rand Journal of Economics, Vol. 17, Autumn, pp. 351-65 '.
i • ' .
Wallack, 'Lawrence M,, 1981, "Mass Media Campaigns: The Odds Against
Finding Behavior Change," Health Education Quarterly, Vol.' 8, No. 3, pp
Warner. K E., 1977, The Effects of the Anti-Smoking Campaign on Cigarette
Consumption," Am. J. Public Health, Vol. 67, pp 645-50,
Weinstem, Neil D , Peter M Sandman, a|nd M. L. Klotz, 1987. Public
Response to the Risk from Radon. 1986. -final -/epo'rt to Division OT"
Environmental Quality. New Jersey Department of Environmental
. Protection. Rutgers .University, January ' •.'. ' '
RTI 9-3
-------
'•• st. nc.rc.ncwoc,o
WHO J P. Puska, K. Koskela, and A. McAlister, 1986, "Television in National
Health Promotion in Finland: The 'Keys to Health1 TV Program in 1982 "
in D s. Leathar, G. B. Hastings, K. O'Reilly, and J. K. Davies, eds:
Wolter, Kirk, M., 1979, "Composite Estimation in Finite Populations," Journal
of the American Statistical Association, Vol. 74, No. 367, pp 604-10.
Yarnell, J., 1976, "Evaluation of Health Education: The,Use of a Model of
Preventive Health Behavior," Soc. Sci. Med., Vol. 10, pp. 393-98..
RTI 9-4
-------
A. OIM I is i luAL MODELING
APPENDIX A
STATISTICAL MODELING
The analyses described in this appendix focibon the panel dimension of the
Maryland surveys. That is, we have considered the responses of those
households in each town who were contacted twice—in the baseline survey
prior to the introduction of information programs in Hagerstown and'
Frederick, and then after the completion df these programs. This design-
offers the opportunity to observe how the siame individuals have changed
their behavior, awareness, knowledge, and attitudes in response to the
programs These areas will be the focus of bur analyses
[• . • .
i.
Because the survey might s'ensitize respondents, raising their awareness to
radon and influencing the information and actions they have taken, tests for
sensitization were,included as an explicit part of our design. This comparison-
is developed in a separate section of this report; however, pur basic findings
suggest that sensitization has not had a dramatic effect on the households in.
our panel The e'ffects have bee:n quite modest and do not appear to be
responsible for the results we have measured as households' responses to
the information program. This finding is important because it allows us to
consider a wider range of responses to the! information program using the
data available on the panel, The potential for. ev?'uation with a design that
focused exclusively on independent samples! wou ; nave been more limited!
Table A-1 reports the total number of households included in our panel
sample by town It indicates the number! originally interviewed for the
baseline survey and then those who continbe to participate in the fotlowup
survey It also identifies the number of households who had not tested at the
time of the baseline survey and those who had not tested by the foltowup
These two classifications are important becjause they represent the sample
definitions that will be used in our subsequent analysis.
The objective of our analysis is.to evaluate the effectiveness of the two infor-
mation programs described earlier. Each' was designed to increase the
general public's knowledge of radon and to e'ncourage testing for it in private
homes We'have separated these /objectives into three components:
enhancing households' awareness of and attitudes.toward radon: increasing
their knowledge of how to test'for radon and of the problems-it causes: and
encouraging them to monitor their homes for radon. Measuring the impact
RTI A-1
-------
Table A-.1. Description of the.Panel Sample
. Panel households
Town ^ Baseline
Frederick 523
Hagerstown .-. 529
Randallstown , 495
Nohtesters
Followup
—
432
432
387
Baseline Followup
475 ; , 365
488 ! . 405
i
456, 367
on these three areas ,s one way to evaluate the effectiveness of the two infor-
mation programs The primary consideration in evaluating the commun ica-
tion programs is their effects on the rate of radon testing. e communica
; ? CaUti0r\that the increased testin9 arising from the information .
s is not a perfect measure of effectiveness. Each survey is -simolv a
ares of theTm **< ^^ "^ have d°ne and Wha< ™ ~
are as of the time of that interview. Decisions to test for radon and to take
subsequent actions may.require more time for some households Th s
means that changes in attitudes toward testing and toward the seriousness o
he .adonprob^m should be regarded as movements in 'the "right direct on ! .
-in evaluating the information programs. Equally important the process of
aTuShfl0,f h^f UdSS tOW3rd teStin9 °nCe the' ^ve been frmedb/
a household should be considered, a positive outcome of an information
• program At this stage, we do not know whether the acquisition oTsuSh
f'uhl r/11 f d? ^ ^ enhanC6d kn°Wled9e tha< accompames a su cess
ful transfer of information about radon will ultimately lead to increased testing
t« r S 6h y' We PreS6nt reSU'tS f°r a" types of ^ings-hew' decisions to
test, and changes in awareness, enhanced understandina, and modifications •
m basic attitudes toward testing. • moamcanons
p measurable out^omes can inf'^nce possible analys.s
Perhaps most important .of these involves trying to exolain
herent'y discret^^.ng is a yes no dec.s'on that'm
Ce by 3 h°USeh0ld Statistical anal^sis is *l"W s,mpler
,s some continuous measure that can be used to gauge the
effectiveness or performance of any program being evaluated, Po'exampie
f we were interested in the effectiveness of an irrigation system or a part,cular
class of fert,i,zers for agricultural production, a natural measure of
pe rforrnance would be the changes in crop yieldsJn response to the 'irrigation
o the,fert,l,zer Yields take on a reasonably continuous set of values In our
situation, we are observing people s statements that describe changes- in their
oUone,;andt tne'r teStm9 deCiSi°nS Any attemP' to ^"^rm S
responses into some continuous scale.requires the analyst to impose a
juagrnent on the statistical analysis The statistical findings will not be neutral
U. me scanng prOCeaure used to translate attitude changes into continuous -
' • 1 1 c i u ' 6 S * ' ~ • ' . . -
RTI A-2
-------
Appendix A
Appendix B
Appendix 0
Appendix D
Appendix E
Appendix F
Appendix G
Statistical Modeling!
Interpreting Selection Effects
Background on R* j
Testing for Differences
i- •
in Proportions
Tests for Sensitization Effects
Questionnaires
Risk Communication Materials
/RTI
-------
-------
APPENDIX B.-
INTERPRETING SELECTION EFFECTS
r el was used to
issues (the variable designated as ASK ^ 3rd health
2- ' st
Prob(TF>0)
or
/
- o
- e>0) = Prob(e>-k;) ' ,' ;.
'= 1-Prob (£ |c -x^}
j - '«•'
th£ 6Xpresslbn fcr th Polity, of test as
RTI E--
-------
ETF.TB •
(B-t;
where TF = x/J - e is the moael for the propensity to test'after the
program . ' .
and
TB = Z7 - 'n is the model .describing the propensity
program (e!g., 27 -./* > 0 test and 27 - /» < 0 do
If fi and £ are bivariate normal, then we can expect that
to test before the
not test)
where
•(B-2).
0 = standard normal density
.. $ = normal distribution function
a , a = standard deviations for /* and e, respectively
/?- = correlation between /j and e '- . :
With this background, it is possible to. gauge the effects of an indepenaenf
variable in the mode! that describes testing before the EPA program for the
propensity to test after • • • ; •. !
5ErTF|TB
-------
5 will be between 0 and v Thus.to determine the liqn of a
pr-obit model for testing before the 9
• ,h
in the
• 3EfTF!TB<01
(B-4)
where, x.
= 2
de.e,m,ne ,he sign o, the variable on the propensity ,o test ato ffie p™ •'
RTI B-3
-------
-------
APPENDIX A. STATISTICAL MODELING
I
To overcome this problem, we have used two modeling frameworks for dis-
crete choice variables/ These models describ0.the. probability that a particular
individual will undertake an action or assume a ''state.',' "In this case the term
"state" refers to a change in a person's knowledge or attitudes. The actual
.decisions are used to estimate probability mpdels that describe how each
individual's characteristics and circumstances influence his choices. The
models assume tha' 'he factors influencing1 such c.hanges in functional
relationships can be jinked to conventional probability distributions. Thus, a
discrete outcome is replaced by a continuous one over a predefined interval.
This continuous measure is the probability of the change in testing or
attitudes occurring, '| . ' , ' •
• . / ' • ;
Our models vary depending upon the particular measure of effectiveness.
The first type of model/the probit model, isjused primarily in the case of
variables that have only two possible outcomes. An individual either -chooses
to test for radon or he does not. He either has: heard of radon or he has not.
To convert the discrete outcome of having tested or not at the time of the
interview into a continuous framework, the model assumes that there was an
unobservable "propensity to test," given in o'ur notation as S,. We assume
that propensity to test is related ,to the person's characteristics, such as
education, age, income, knowledge of radon; and the characteristics of his
home Moreover, we include measures of thej information program in these
potential 'determinants. These' measures are jlimited because the programs
were intended to.be applied uniformly to large jnurnbers of households.* Two
overall variations and a control were considered. Frederick and Hagerstown
were the two communities experiencing different .information programs. We
assume in our analyses that every household in these communities had
access to, and experienced the effects of, each town's program. Thus, resi-
dence in the town is used to reflect whether ah individual' experienced each
town -s respective information program. These are measured relative to the
control community, which received^ no specialized program. ';.-•'•-'
The analytical description is given by Equation '(A-1):
si = V?
j
In Equation (A-1), S, designates the unobservable propensity to test and the
x, variables designate the individual's characteristics as well as those of his
residence that might influence the propensity to test. In this notation, a0 and
the b; coefficients are the parameters estimatecl'in the model They describe
the effect of-each'factor on this unobserviable propensity to test U"
acknowledges that the analyst cannot perfect!1/ understand or describe.why
individuals/make these testing decision's.' Even under the best of
circumstances, there will be an error associated with :ahy description of these
cecis-ons U. represents that error and plays ari important role in the model.
RTI A-3
-------
Because .only the outcomes of the decision process are measurable, and net
test^t^ ne5l' 'n3 deC'Si0n rU'e 1S'needed to relate;the P^nsity to
test to the actual testing decisions: This is given tn Equation (A-2)
Vj = 1 when S. > ()• . (A-2)
. ^ i • '
. . - 0 when S- < 0 • .
Where S' exceeds °' the rule assumes they will test. In
he propens.ty ,s less than or equal to 0, they will not test By
taking Equa ions (A-1) and (A-2) together, it is possible to relate the model to
he probability of observing an individual testing: this is given in Equation
'
Prob(y. rl).-pPob(U. >-a0-Ebjxj.:) (A-3)
-1-F .(-..o-IbjXj.) ;
Once we assume a probability distribution to derive the errors then it is
possible to use that assumption to characterize the probability of testing In
Equation (A-3), F represents the probability'function used ^characterize the
errors .For the probit model the error is assumed to be normally distributed '
F is then the cumulative distribution function associated with.lhe normal error
distribution •
9lVen in Equati°n'(A'3): Jt ls Possible to use the maximum
principle to estimate how individuals' characteristics, as well as the-
information programs they received, influence the likelihood of-test-no We
used this logic together with-tiata on the panel households' Our analyse
uses only homeowners who, had not tested their homes in the; baseline survey
and responded to the followup survey to estimate the effects of the program
on testing The baseline testers were excluded because they had already
tested before the communication program went into effect :
Table A-2 reports the model results - The model includes three types of
variables. The first type describes the socioeconomic characteristics—e a
income, education, ana age and the attitude variables—whether the person
asks his physician questions about health. This is treated as a qualitative
variable and identified m Table A-2 with the label ASK DOCTOR if the
individual's response to the health.att.tude question indicates concern over
ncaitn as a very good or good description of his attitude, then this variable is '
hh,e f/°b^?rel require-s a normalization rule that scales eacn coefficient bv
the s.anuara error associated with the stochastic erro-.. U • Convention^
p ac;ice sets-the-stancard error to unity because there is no Sa? o;-
sc^ara.eiy estimating botn tne coefficient and the stancarc error '
RTI A-:
-------
M. o i M 1,13 i ILAL MODELING
Table A-2. Probit Model for Testing-Foltowup Panel Sample
Independent
variables
INTERCEPT , ;:
HAGERSTOWN .'
FREDERICK .
INCOME , . '
EDUCATION , '. . ' '
TALK ABOUT RADON
/
ROBERTA/WJLA
ASK DOCTOR
USE BASEMENT '
AGE .' • .
SELECTION EFFECT VARIABLE
LOG(L) . . •
Coefficient t-statistic*
•1.65;10 ,
0.0780
I
.0.4367
2.84B9E-6
! ' "
!' -
-0.0623
0.3454
0.0907
•0.00(36
0.002'
-2.168
•T82.41
955
These numbers are the ratios of the coefficient to their estimated •
asymptotic standard errors The critical t value for statistical siq-
nmcance is ±1.96. - ' • !
C9ded as a 1 and 0 otherwise. USE BASEMENT is another qualitative
variable indicating whether the individual hcfd a basement in his home and
his household used it as an active part of his riving space.
The second type of variables provides an indirect indication of the baseline
awareness of radon. TALK ABOUT RADON is a qualitative variable that
describes whether individuals spoke about radon to others prior to the:
baseline survey We used the baseline level tc> minimize interactions between
the followup measure of this variable and the EPA programs. The second
attempts to account for nonprogram effects as a potential influence for testinq
decisions. It is designated in Table A-2 by ROBERTA/WJLA This variable is
qualitative with a value of unity when individuals reported seeing the. WJLA
special campaign on radon. This campaign occurred during the'same time
period as the EPA information program. Hacjerstown and Frederick are the
qualitative variables identifying the.individualjs town of residence'. They are
usec to reflect the influence of the 'information program on the testina"
decisions • • > •' ••;,'-; :., s
RTI
-------
APPENDIX A. STATISTICAL MODELER
°' 'he 'ast variable in t^ model, it is important
They
no. tell us the contribution of each variable to the a y e ,
C°mplex ca"^
-------
I
I
I
Table A-3. Probit Wodeljor Talking Aboi| Rado'n-Followup
Independent'
•variables5;-1'-. ;
INTERCEPT, _
FREDERICK
. HAGERSTOWN
EDUCATION
AGE '-. * .
i
RACE '
ASK DOCTOR
USE BASEMENT , , . •
INCOME
RESIDENCE TYPE
ROBERTA -WJLA . '
TV— FOLLOWUP
LOG(L) >'.••:
n
a . - .
Coefficient
•2.2181
0.4213
0.3418
0.0673
-0.0067
. 0.4500
0.114'f
" 0.1049
, 5.0739E-6 .
•olpoa- ' •'
0.4003.
0.8563
' *— -. -715.54
•*— r 1,273
t-statistica
•6.931 .'
3.961
3 137
3.936 . '
•2.426 • '.
3.298 ./''.
1,379
.,.1.260 '..•'•
2.142
. -0.030 ., • " :
2.701
10459
. ^ ' '
''.'—+ V
'lthaenHe nHumbers are the ratios of the coefficient to their estimated asymptotic
standard errors. The critical t-value for statistical significance is ± 1.96.
' • • • -' •'•••• i •' • •>
To use the testing model or the model for talking about radon, we'-need to
develop a series of simple hypothetical experiments. For example with the
testing model, we considered-an average ^dividual in each of the three
towns The average individual was described by taking the arithmetic
averages of the socioeconomic and attitudinaj variables for households in that
town Using that average person we then, evaluated the effect of the EPA
information program in Frederick on the probability he would test his home
for radon. Obviously in the case of Frederick, we know which sample
individuals tested their homes for radon after the program. This was the basis
for estimating the statistical model. These calculations were undertaken to
illustrate the effects of personal characteristics and attitudes on the reported
testing rates. With them it is possible to consider how the program might'
-have fared if it had been undertaken in Hagerstown or Randallstown. We'did"
so by composing the average person in thpse towns and evaluating the
Fredenck program s effects using this "person's" characteristics Thi*
method allows us to undertake other types of (empirical experiments as welP '
WJLA s radon campaign with Roberta Baskio was undertaken without the
.prior knowledge of EPA. Yet. it may haveiinfluenced how some of the
households in the test communities responded to the EPA program
Fortunately, the WJLA.campaign'was not seen by all households in all
communities Consequently, it was possible to isolate (and some models did
RTI A-7
-------
APPENDIX A. STATISTICAL MODELING
n roh h n f H 3n °therS) the P0tential influence of the Campaign on the
p obabihty for radon testing. To evaluate "this influence, we consider our
SnPe£?n '" tOWn With the EPA pr°9ram but'without'the WJLA
campaign This permits us to control, in hypothetical terms, the EPA program
cent Snn9?TW?flth- ' thvf EPA Pr°gram W3S the most important factor in
contributing to testing or whether the WJLA.campaign was^ore important.
• wMhPPPA CalCU7lations' consider R9«re A-1, Our model predicts that
with the EPA program, 7.6 percent of the nontesters in Frederick should test
I hi ah PerCe?1 L 9»A than the m°del Predicted testin9 would nave
but e ^±^1 w n A Pr09ram- M°reOVer' if we retain the EPA
tua ,v no 1^ ^ C-a,T' 9n' the m°del impHeS that there wou|d be vir-
tuaiiy no effect on the level of testing. Testing would be about the same-7 4
percent. The same methodology applies to each of the other two towns The .
primary d.fferences in testing probabilities arise because of differences in the
soc.oeconom.c characteristics and attitudes of the average:person in each of
the two communities. For example, the Frederick program was predicted to
be somewhat more successful in Hagerstown than in Frederick, but to be less
successful m Randallstown. The better potential in Hafierstown "
cond ions that are likely to lead to more testing, especial.y Le pfep
ing talked about radon with,a:friend or neighbor before the EPA program.
Of course the predictions for the EPA program in Frederick lean be compared
wrth actual experience. The 7.6 percent is somewhat less' than the actual
. increase in testing that took place as a result of the program The actual level
of testing was increased from 5 to about 15 percent. There ar.e at least two
reasons why the model understates this record. First, our prediction was for
he average person. It does not reflect the inherent diversity of individuals in
he community. Second, the model is a summary that attempts to capture
the features mfluenc.ng decisions of individuals across all three communities
Since rt is an average, it offers an approximate description ;of what happens
and not a perfect one. How much these two considerations influence the
t°hUfltGmnHS,0HbSe+red in,theSe fJ9UreS W0uld be mcu" to Jud96' b^ " «s clear
that model identif.es a fairly consistent effect for the Frederick program This
±C1'S °ne that d°es influence testing. It is larger than ;the effect of the
WJLA campaign and is apparently also influenced by the households'
characteristics w uo
The second type of model arises from the same basic rationale as we
described for the probit analysis, However, in this case the model seeks to
evaluate situations where individuals may realize one of several different
states In- these models we use more specifically the advantages of the oanel
dimension of the sample By observing an individual's answers to the same
questions in the interviews undertaken.before and after EPA's information
program, there are opportunities to observe changes 1n learning and in
RTI A-8
-------
APPENDIX A. STATISTICAL MODELING
PROGRAM EFFECTIVENESS: PROBABILITY OF TESTING
FORRADON
(PANEL SAMPLE)
10 T
' Overall
FREDERICK
3.0
' 7.4
Without EPA Program Without WJLA (low)
10 T
8--
6--
HAGERSTOWN
«•+
10-
2--
Overall
6.1"
2.3
Without EPA Program Without WJLA (low)
RANDALLSTOWN
6.0
Overall
Without EPA Program Without WJLA (low)
Figure A-1. Probability of testing.'
RTI A-S
-------
MUUc'LINVJ
There are technical reasons why the probit formulation is much; more difficult
to implement in these cases « The probability model used in the case of
multiple outcomes is referred to as the multinominal logit model and is given
in Equation (A-4) below. . . .••.:••
exp (Z
Prob (state = i) = ~
4
t exp(E a, .X.)
k=l j kJ J
where' :
state = one of four outcomes: correct on both; incorrect on both-
incorrect on baseline and .correct on followup (i.e , learning) •
, and correct on baseline and incorrect on followup (unlearn
(subscripted by k in the denominator of (A-4)).
X = jth determinant of probability, constant across'states varyinq
• • across respondents.. -
Equation (A-4) treats each discrete outcome as a state. Thus, we are
interested in modeling the probability an individual will be classified as being
in a particular state. To provide a tangible illustration of what this, means,
consider the first of our indicators—awareness. Both the baseline and follow-
up surveys asked households if they had seen or heard anything, about radon
in the months preceding that interview. There are three feasible outcomes
. when we consider the answers to this question at the baseline and followup
interview times, They may not have heard about radon at the time of the
' baseline and not heard of it after the followup interview (from sources other
than the interview itself). They may not have heard of it at the baseline and
then at the time of the followup interview had seen or heard information about
radon. And, of course, they may have heard at both the baseline and
followup interview The nature of the questions precludes a situation in which
they would have heard about radon at the baseline and not at,the followup.
We have 'used the multinominal logit model to describe combinations of these
responses as states. Thus, we are, interested in characterizing whether.
individuals became aware of radon, which we describe as reporting that they
had not heard about it at the baseline and then reporting they had seen or
heard information about radon in the followup interview. We are also
interested in reporting that they had. heard about it in both situations. These
*TO describe the probability where there are multiple outcomes is a multiple
integ-ation problem for continuous random variables In the case of the
normal distribution, this multiple integration must be undertaken using nu-
men:ai approximations that are'quite expensive and time consuming
Consequently, it is commonplace in modeling multiple discrete outcomes to
use a formulation for the stochastic error that closely approximates the
norma: and yet coes not .require the numerical integration that is so costly
i h;s is the' primary rationale for the multinominal logit framework
RTI A-10
-------
APPENDIX A. STATISTIC!/MODELING
,two outcomes are consistentwith
problem. . ' ••. .
heightening the awareness of the radon
• The multinominal logit model, used in this
choice models in that a normalization rule ..
coefficients. The coefficients in Table A-4
variable on each awareness outcome. Thai
each particular variable on the likelihood .
particular state relative to a baseline situation
the Frederick variable in the state "unaware
case is similar to other discrete
rrust be applied in interpreting the
describe relative effects of each
is, they indicate the influence of
that the individual will be in the
For example, the coefficient of
at the baseline and aware in the
• .. . toiiowup interview' reflects the differential efl
the individual being classified as having th
comparison
with the sequence associated
- ' .program. The case of two, unaware answers
: analysis
Table A-4. Multinomial
Inde ri t
variables
• INTERCEPT
FREDERICK
HAGERSTOWN
EDUCATION
ASK DOCTOR
.SEX
RACE
AGE
• INCOME
YR40_76
YR40
USE BASEMENT
.TV— FOLLOWUP .
.. TALK ABOUT RADON
ROBERTA 'WJLA.
LOG(Li '.•-"•• ',
n:
Logit Models for Awareness-
., No baseline/
yesfollowup
Coefficient t-statistica
• -2.2994 -2.681
, 0.5714 ' 1.969
-0 1976 -0.690
0.0737 1.551
•02541 -1.143
•0.5451 -2.531
0 4459 1 .359
•0.0218 -2.899
-6'.6796E-6 -0.997
1.1366 . 4.183
0.8398 2.450
0.2458 1.097
5.0827 9.536
-18.8440 -0.012
17.7616 0.007
-*-- ' ' • fl
. o
dTnese numbers are the ratios of the coefficient to their es
errors The critical t-vaiue for statistical significance is ±1
ect of the information program on
e not aware/aware sequence in
with being totally unaware of the
is the reference category for this
•Panel Sample
' Yes baseline/
yes followup
Coefficient t-statistic
-5.8748 -8.273
0.6665 ' 2.929
0..0903 0404
0.1975 ..' 5.361
-0.5937 -3.279 _
-0.2646 -1.570
1..1777 .4,009
0.0134' 2.274
3.3982E-7 ' 0.067 .
0.90.62; 4,268.
0.5076 . 1.873
0.0860 0.479 ....
5.0692 , , 9.759
•1.0997 5.778
18.0216 0.007
iq'cc' ' '
,?73 ;' ... '„ '
timated asymptotic standard
96 .' , ' -
RTI A-11
-------
_bfATISTIC.AL MODE! iKft
attitudes, «
we
be
no, heard about"adon™ , so ° T e WJU^^^ T" * ^
no,
.RTI A-12
-------
CM M-I id I IL*ML MUUCUftu '
, • i
'answer incorrect during both interviews. Th'ere are two situations, Dues-.
tion 17 and Question 18, in which the WJJLA campaign influenced the
probability of getting correct answers. , |.
' " ' ., '
Using these models with an "average" personjcornposed in the same way as
described earlier for,each community, it is possible to predict the probabilities
that this individual.would answer the questions in both'interviews correctly or
would learn (i.e., go irom an incorrect to a'correct answer). These are
displayed by question and town in Figures 8-3 jand 8-4. Once again, the EPA
program influences learning for the question. The'probability of learning is 9
percentage points higher with the EPA prograrh than without it. Moreover, the
WJLA special does not dramatically influence this difference. This difference
is approximately constant across the communities of Frederick and
Hagerstown. the model also predicts that there would be an increase of
9 percentage points in learning about how tq test your home for radon had
the EPA program been applied in Randallstown.
• ' .! " • -
-This pattern of effects differs by question. For'the remaining three questions
analyzed, the EPA program had no apparent influence on learning
Trie last category of analysis involves modejing how household attitudes
toward radon were influenced by the information program. Table A-6 displays
the classification, based on respondents' answers, we-developed to describe
those households whO'reported'they had noli tested in the followup survey.
We are interested in understanding whether there was an influence of the
program on households who did not test."Our objective is to describe
changes in the households' reasons for no:t testing Thus, the model is
'' a
Table A-6. Change in Attitudes of Nontesters
Baseline state
Uninformed, -
anti-testers,
fatalists '•'
Not-m-area
Not-in-home
Procrastinators
-
Uninformed,
antMesters,
fatalists
0
0
0 . • '
. '6
i • .
Followup
i
Not-in-1
area
1-
o ;
! '
i '• •
2 '.; •
-,3 '
' '. 6 ' !• ,
i-
state
Not-in-
home .
0
3
!-
5
6 -
,
Procrasti-
nators
1
4
4
7
aNumbers in the table represent the seven different, states included in the
multinomial.logit model- . ' ;{' ' ".'' , '" .
RTI A-15
-------
relevant only to the nontesting subset of the-pane! households ' The
defin ble iS°'ate 9eneral types of resP°nses based on
n,c .
nd viduals explanations for not testing. The table identifies three changes
that we classify as being movements in "the right direction." All seven
changes were described simultaneously as separate states in a multinominal
ogit model We have selected the three of most direct interest and displayed
the parameter estimates for the logit model in Table A-7. The model specifi-
cation includes socioeconomic, attitude, information, and the residence '
variables. In this case, the overall size of the model required that we limit the
number of variables included in the specification. Nearly 80 parameters had
to be simultaneously estimated to develop this model. !
The overall conclusion of the analysis is that the EPA program did have some
influence on nontesters, primarily those in Hagerstown who stated in the
baseline survey that they felt radon that was not a problem in their home or
area. Following .the, program they changed their reasons for nontesting
acknowledging the potential problem but indicating they had not potter!
•around to testing This finding is reflected in the coefficient for Hagerstown
which represents the EPA program in the third row of the table.
It is also important to recognize that' the program did notmfluence those
nontesters we classified as uninformed. For these individuals the/coefficients
tor the EPA program variables were negative but statistically insignificant
Such negative coefficients imply that the program reduced the chances that
.they would become aware of radon's problems and act on!that awareness
Since the coefficients were not statistically significant, this outcome should
not be interpreted as a result of the EPA program. ;
Using the same basic structure as in our other experiments, analyzing the
effects of the EPA program on the probabilities of testing, learning and
awareness, we constructed probabilities for each, of the' three, towns usinq an
"average" nontester for each town, '-.-':
The cross-hatched bars in Figure A-2 identify the predicted probabilities of
moving from uninformed to procrastinator. In the case of Frederick it
appears that without the program individuals are more likely to move in the
correct direction than with the EPA program. This is a reflection of the
insignificant negative coefficient for the Frederick variable reported in this
column of Table A-7 Since the coefficient is not statistically significant this
probability change should not be interpreted as a serious possibility ' It is
reported here to illustrate the implications of the model For other initial '
attitudes on the part of nontesters-, we find that the Frederick information
program. does have an effect on moving in "the right direction" when the
nontester reports his reason for not testing was because he did not believe
radon posed a problem in his home or area. In Table. A-.7. the coefficient for
Frederick nas a positive sign but it is not statistically significant at the
RTI A-16
-------
T " •'.''•- -J'
Table A-7. Multinomial Logit Models for Reasons for Not Testing— Pane! Sample*
Independent
variables'
INTERCEPT'
HAGERSTOWN .
FREDERICK - .
EDUCATION
AGE • . . *
INCOME •
^
SEX •
RACE ,
TV— FOLLOWUP
TALK ABOUT RADON. '-
' ROBERTA Vv'JLA
LOGfL)
N
Uninformed to
procrastinators
-3.8618
. (-3.550)
-O.T244 •.
(-0.355)
.-0.4560
:(• 1.280)
' 0.1342 .
(2 190)
-00014'
(-0.137)
' 8.932E-6
' (1-027).
0.0440
(0.163) • . . •
0.7507
(1.743)
• 0.3231
(1.155)
0.2545
(0 840)
,09417.
(2.01.2),
Not in home/ Procrastinators
area to to
procrastinators procrastinators
• -5.836
(-4.582)
' 1.1616
(2.602)
0.7036
(1.562)
. 0.1338
;:.q.9i9) .
0.0163
(1.488)
CM 026
(1059)
0.1555
(Cj.518)
' (i.3200
(0.582)
d.298T
(0.939) .
1,1651
(2.642) ' -'.,
0794T
(1.624) ,( .
, , 1 nriri °
U I0
-6.8587 ,
(-4.211)
0.6947
(1.332)
.-0.0376
• (-0.070)
0 1206
(1.357)
0.0244
(1.679)'
' 1.7491E-5
(1-489)..
0.2087 '
(0,549)'
, . 0.2086
(0.331)
0 1327
(0.327)
21346'
.(4.985) '
• 09083
(1 499j •
3 '••••• ! ....
Tne numbers, in parentheses are the ratio's of-the coefficient to their estimated asymptotic
standard errors Tne critical t value for statistical significance is ±1.96
RTI A-17
-------
A' STATISTICAL MODELING
PROGRAM EFFECTIVENESS: PROBABILITY OF
IN NONTESTERS' RATIONALES
(PANEL SAMPLE)
Overall Without EPA Program Without W.JLA
(low)
Without W.JU
(high)
HAGERSTdWN
' "verall
19.7
Without EPA Pn.eram \Vithout W.JLA Without W.JLA
Oow) • | high:
RANDALLSTOWN
nverall Wi(h,,ut EPA Program Without WJLA Without W.JLA
How) . (high!
Usinfofrictf 10 Procrisur.r.o-
Noi a: ArcvTionc lo fro
Figure A-2. Probability of changes and reasons for not testing.
RTI A-18
-------
conventional .05 level of significance. .In
alone did have a positive and significant effect
radon was a problem in their home or area
significance in Frederick could be the group
'"not in area." We are exploring now whether
would change the response.
gerstown. the media campaign
on nontesters realizing that
One reason for the lack of
ing of both "not in home" and
r separating these categories
and
Overall, these results tell a consistent story
influenced testing, awareness, knowledge,
had the same effect had'it been applied in Ha'gerstowi
effects would be most like thpse of Frede
Hagerstown. The socioeconomic charac
Randallstown households are sufficiently different
smaller response on all of our measures of effectiveness
EPA's information program
attitudes. It would not have
rn or Randallstown. The
rick if it had been applied in
teristics and attitudes of the
in that we would expect a
RTI A-19
-------
-------
References
Goldberger, Arthur, S., .1373, "CorrelatL..
Probabilistic Predictions," Journal of the <
Vol. 6.8, p 84V.
ions
Between Binary Choices and
American Statistical Association,
Maddala, G. S., 1983, Limited Depender
Econometrics, New York:, Cambridge University
Morrison, Donald G., lS72,V:Upper Bounds
Outcomes and Probabilistic Predictions,'
: cal Association, Vol. 67, pp. 68-70
for Correlations Between Binary
Journal of the American Statistr-
and Qualitative Variables in
Press. '
RTI C-3
-------
-------
APPENDIX D
TESTING FOR DIFFERENCE IN PROPORTIONS
Differences in proportions between the
surveys were tested in the areas of
about radon. .
baseline and independent foilowup
awareness of and changes in attitudes
The null hypothesis that the two proportion
percent confidence level, with a corresponding
The formula used,to compute the sample Z-statistic is as follows:
/7(P1-P2)
where
pr and p2 are the twc proportions.to te tested expressed in decimal
form, and • -
(P1 - p2).is assumed to be 0 under the null hypothesis-
- P) =
The calculation of these variances (Var.p..
.fact that the p. are'ratio estimates. To
.responses on the questionnaire can usually
vanate: ' ,••'.•
and
see
X =
•1 .for negative responses
0 for neutral responses
1 for positive responses
.Letting-n0, nr and n^ be the counts for each
.respectively.,our estimated proportion, p, car
ratio cf twc random variables
are.equal was tested at the 95
critical Z-statistic of +1.96
(D-1).
(D-2)
Var p2) must account for the
this, note .that the'multiple
characterized as alrinomial
i of the three categories of X.,
be expressed in the form of a
fill D-i
-------
rwn Lyi
ir< rnurun'i l
P = n2/(n2-.n0) . ;
-This is the estimated proportion of the population who would give positive
responses if contacted, excluding those who would.give neutral responses
The variance of their statistic is expressed by:*
Var n
nQ)] ,=
E n,
/ar ru Var (ru + ru)
2 cov (n2,
n)
(D-3)
where Var ( )
Cov ( )
E ( )
= variance
= covariance
= expected value.
The variance of an estimated proportion as expressed in Equation (D-3) can
be estimated from the sample data as follows: • .' • •
War (p) =
2n,
where n = n1 -
'L . I.+
no "! "
n(n2 * nQ)'
2
(D-4)
'See for example. Kendall, M G , and A Stuart, 1969. The Advanced Theory
of Statistics. Vol 1,p 233, New York: Hafner Publishing Co ! -' -.
RTI D-2
-------
APPENDIX F
QUESTIONNAIRES
This appendix contains'copies of the screener
pendent.followup.and dependent followup
our study. ' •
, baseline survey, the inde-
<: uestionnaires that were used in
RTI F-1
-------
-------
Radon Information Effectiveness SJurvey:
Maryland Baseline Screener
Final Interview Code
01 ineligible, Not Residential Number
02 Ineligible, Not Homeowner
03 Ring, No Answer
04 Nonworking Number
05 Double Wrong Connection
Hello, my name is
06 Answering Machine/Service
07 No Result From Dial
08 Fast Busy/Computer Tone
09 Unable to Contact
10 Physically/Mentally Incompeti
I'm callinq from the Research Trii
'11 Language Barrier ;
12 Interview Completed
13 Partial, Data
14 Final Interview Refusal
snt 15 Other • "
mole Institute (RTI). in North
know and think about environmental
Carolina. We are conducting a study on what people
won't take much of your time and your answers will be kept strictly confidential
if necessary: Your cooperation is very important because we want 1o
knows about environmental issues. This is not a sales call. The study
Maryland.) . : .
1. lsthis_ - —-_- — ?
Yes '......... 01 -CONTINUE
No 02 - "Thank You," HANG UP
2. Does this number serve a: (READ ALL CHOICES, MARK ONE.)
Residence '01 -CONTINUE
Business/institution •... 02
Or something else 03
- "Thank You," HANG UP,
3.
Do you own your residence?
Yes ....:.'.:...... 01
No 02
- CONTINUE
'- TERMINATE
4.
As part of our study, I need to randomly choose an adult who makes
household decisions. Please tell me the first names of the adult
(IF RESPONDENT CAN'T ANSWER, ASK FOR ADULT, REPEAT INTRODUCTION.)
1. Woman's Name:
2. Man's Name
Third Decisionmaker:
Fourth Decisionmaker
(TO CHOOSE RESPONDENT, LOOK AT LABEL AND CHOOSE THE FIRST NAME
OR THE SECOND NAME IF ITS A "2". IF YOU HAVE TWO MEN OR TWO
JUST WRITE THE SECOND NAME IN THE MAN'S SPACE AND FOLLOW
RULE. IF MORE THAN TWO DECISIONMAKERS, THEN CONSULT YOUR
CIRCLE NUMBER OF PERSON SELECTED.) -
May I please speak to NAME OF SELECTED DECISIONMAKER. (IF NOT AVAILABLE, SCHEDULE A
CALLBACK.) ~
READ INTRODUCTION IF PERSON ANSWERING IS NOT THE RESPONDENT.
issues. It
. (Additional information,
find out what the general public
is sponsored by the State of .
or shares in important <
decisionmakers in your household.
IF THE NUMBER IS A "1"
WOMEN DECISIONMAKERS,
THE SAME CHOICE SELECTION
RANDOM SELECTION TABLE.
TERMINATION
Thank you very much for your cooperation. Our study involves
neec any more of your time. Thank you again for your help.
only homeowners, so I won't
-------
-------
Communicating Radon Risk; Effectively
Maryland Baseline Surrey
2.
3.
Telephoned
RTL ID #
1. Compared'to other issues the State of.Maryland faces, do vou
(READ LIST; CIRCLE ONE NUMBER.) X
a. MORE IMPORTANT
__ b. JUST AS IMPORTANT ..'....
c. LESS IMPORTANT.........
.d. DONTKNOWfDONTflE/ID,)
We re interested in finding out how. serious you think the risks
your community and to your household. On a scale from 1 to
SERIOUS and 10 meaning VERY SERIOUS, please tell me '
and tayour household
from some types of pollution are to
10, with 1 meaning NOT AT ALL
serious you think the risks from
-'. (READ LIST, SCALE, AND
how
Not at all
Serious
a. LEAD IN DRINKING WATER
b. HAZARDOUS WASTES IN LANDFILLS
c. RADON INCOMES
For the rest of the interview I'm going to ask questions mainly
mentioned-radon in homes. During the past few months, have
about radon? (CIRCLE ONE NUMBER)
a. YES ...
b. NO ..'.....,..'
c. DONT KNOW (DCW7 READ)
think environmental issues are:
01
02
.03
94
Very .
Serious
10
I Your
1 community
, .Your i
household r
about one of the sources I
you seen or heard anything
Skip, to Question 5
on page 3
-------
on 'the radio or TV aboTradon7 "*" ^"^ "' " """"^ °' ma«""ie;?r neara
a-YES ' ............... 01 —*• Continue
bl N0 ' = ..:..;....... 02 —-*» Skip to Question 4B
Was that in the newspaper or magazine, or was it on the radio or TV?
(CIRCLE ALL THAT APPLY.)
a. NEWSPAPER .. 01 c. RADIO ' 03 ' ''•''.
b. MAGAZINE ...'. 02 d. TV 04
\ " ' ••
4B. Have you seen or heard any public service aa's about radon in a newspaper or magazine or on
the radio or TV in the past few months?' , ...
a< YES • ,- • 01 —*• Continue
b> N0 • • • 02 —*» Skip to Question 4C
Was that in the newspaper or magazine, or was it on the radio or TV'
(CIRCLE ALL THAT APPLY.) '
a. NEWSPAPER 01 c. RADIO 03
b, MAGAZINE 02 d. TV 04 ' '
4C. Have you seen a poster, read a utility bill insert, or heard a presentation about radon in the oast
few months? . • ...-
a- YES , •...;...,.. 01 —^Continue . ':.•
b> NQ • • • •... 02 —-*- Skip to Question 4D
Was that a poster o'r utility bill insert? (CIRCLE ALL THAT APPLY,)
a. POSTER 01'
b.' UTILITY BILL INSERT 02 ' • • .
c. PRESENTATION 03 ' •
4D. Have you talked about radon with a friend, relative, or coworker in the past few months?
a-YES • • :....-..•...... 01—^Continue .'
b-N0 •••• , ...02—+ Skip!to Question 4E
Was that a friend, relative,- or coworker? (CIRCLE ALL THAT APPLY.)
a- FRIEND 01 c. COWORKER 03 ':
b. RELATIVE 02
4E. Have you called the State of Maryland toll-free number for radon information? !
a. YES .. 01 '
b. NO, ; 02
4F. In the past few months have you learned anything about radon in some other way?
a- YES 01 —*• How was that? '' " ' ' • ,
b, NO .... .' , 02 • •' - . • • \ ' ~ ~ ~~~ ~
2
-------
wuuiu yuu
.' (-u>o'ivwj ncAu t/o'/,' w/rrucc /nc
a. MARYLAND DEPARTMENT OF THE .ENVIRONMENT
b. MARYLAND DEPARTMENT OF HEALTH AND MENTAL HYGIENE
C. LOCAL HEALTH DEPARTMENT..".
d. U.S. ENVIRONMENTAL PROTECTION;AGENCY : '.
e. OTHER (SPECIFY) • •
f. DONTKNOW (DON7 READ)
If you wanted to le,arn. about radon-related health problems, wf ich
you trust the most to give you that information? (READ LIST; CIRCLE
a. MARYLAND DEPARTMENT OF THE ENVIRONMENT ,.........
b. MARYLAND DEPARTMENTOF HEALTH AND MENTAL HYGIENE
c. LOCAL HEALTH DEPARTMENT ..'.-.
d, U.S. ENVIRONMENTAL PROTECTION AGENCY
e. FAMILY DOCTOR ....,...,.
f. SOME OTHER SOURCE (SPECIFY) ;
g. NO ONE (DONT READ) '...,.
h. DON'T KNOWYDCWT READ)
7. People have different opinions about radon. How much do you
following statements are your opinion? (READ ANSWER CHOli
Strongly i ; Agree j
agree • ; i :
IT IS IMPORTANT TO TEST
MY HOME TO FIND OUT IF
I HAVE A RADON PROBLEM .
01
02.
b. IF I HAD A'RADON PROBLEM,
IT WOULD BE COSTLY TO FIX
01
02
c. EVEN IF A RADON PROBLEM
WAS FIXED. MY HOME WOULD
STILL BE WORTH A LOT LESS
• 01
02,
Of
02
03
04
05
94
of the following sources would
ONE NUMBER.)
01
02
03
04
05
-06
07
94
agree or disagree that the
ES AFTER FIRST STATEMENT.)
Disagree . > Strongly •
1 ' disagree
Don't
know
03
04.
03.
04.
03:
04.
94
94
94
-------
a-YES ..! 01 —&* Continue ... .
b. NO ;-.'...... 021 '
C. DONT KNOW (DONT RE4DI V '.... 94} ~~~** Sk'P to-Question 9
8B. When did you get your results? /
(If "don't know", enter "34/94") MONTH/YEAR
> . .
8C. Were the results over 4 picocuries per liter?
a. YES 01 —•*»* Continue ,
b. NO 02)
c. DONT KNOW (DON7READ) . J —* Skip to OUettion 11
8D. Did you do followup testing, anything to fix the problem, both, or nothing?
a. FOLLOWUP TESTING 01
b. FIX PROBLEM \ ..02
c. BOTH : 03 Skip to Question 11
d. NOTHING : 04
e. DONT KNOW (DONT READ) . , 94
9. People may have various reasons for deciding not to have their home tested for radon. What is
the main reason you have not had yours tested. (DON7 READ LIST; ALL THAT APPLY.)
a. NEVER THOUGHT ABOUT IT 01 i. WOULD RATHER NOT KNOW IF
b. DIDNT KNOW IT WAS POSSIBLE ... <02 THERE IS A PROBLEM • •> • 09
c. DONT THINK I HAVE A PROBLEM J- ^SSfiSX,^ ™* ABPUT .
IN MY HOME 03 . RADON ANYWAY ....;.. ..10
d. DIDNT KNOW HOW TO TEST ........ 04 k FIXING A PROBLEM IS TOO EXPENSIVE :... 11
e. THOUGHT TESTING WAS TOO - L CONCERNED ABOUT CONFIDENTIALITY 12
EXPENSIVE 05 ' m. JUST HAVENT GOTTEN AROUND TO IT 13
f. DONT THINK TESTS ARE RELIABLE , . 06 n.' OTHER (SPECIFY) 14
g. NOT INTERESTED 07 o. DONT KNOW (DONT READ) : . 94
h. DIDNT KNOW IT WAS A ' ;
PROBLEM IN THIS AREA 08 ' - • '
10.
11.
Suppose your local health department was offering a radon test for a one-time cost of $10, S25,
S50, S100. The cost would cover two radon detectors, the results, and a booklet about radon.
Would you take part in such a radon testing program? (CIRCLE ONE NUMBER.)
a. YES ..:,.. ; ,.,...; ,...... 01
b. NO :.; - ,'.;,._ ;; ....,....' 02^
C. DONT KNOW (DONT READ) '.'. 94
Suppose you are just moving to this area and you-want a home like the one you're in now. You
have narrowed the choice to two houses that are almost identical. The only difference is that
House 1 has radon levels 2. 5 times higher than the government's guidelines for action, while
House 2 has no radon but costs an additional $5,000, $10,000. $15,000, $20,000. Which house
would you buy? (CIRCLE ONE NUMBER.) '
a HOUSEi . , .
b, HOUSE 2
: DCN" KNOW (DONT READ)
01
02
94
-------
,, I al n
you three choices. Please tell me which answer you think is best. If "don't know",is your best answer,
then say that. • .. - " . , - - , ...
Record
Responses
12. Where does most radon in homes come from?
a. INDUSTRIAL POLLUTION :.....,......,
b. .NATURAL URANIUM IN SOIL .......:......
c. OR HOME APPLIANCES
d. DONT KNOW'(DONT READ) .... ,
13. Which of the following best describes radon? Radon has:
a. A SLIGHT ODOR
' b. A STRONG ODOR ..:.....
a OR NO ODOR AT ALL /
d. DONT KNOW (DONT READ)
14. When radon is measured in a home, which of the following will affect the level?
a. THE TIME OF YEAR IT'S MEASURED :
b. THE AMOUNT OF INDUSTRIAL POLLUTION AROUND THE HOME .
c. OR THE NUMBER OF APPLIANCES IN THE HOME ......"...
d. DONT KNOW (DONT READ) '. ., /
15. How can you test your home for radon?
a. YOU CAN DO IT WITH A HOME TEST
b. ONLY TRAINED PERSONNEL CAN TEST .,.
C. OR YOU CANNOT TEST FOR RADON
d. DONT KNOW (DON7READ)
16. When do health problems from being exposed to radon usually occur?
a. WITHIN A FEW WEEKS '. ...:,,.,......
b. IN A FEW YEARS ....'....:... ,.......
c. OR NOT FOR 10 TO 30 YEARS .. ".....
d. DONT KNOW (DONT READ)
17. What kind of health problems are high levels of radon exposure likely to cause? \ '
•a.' MINOR SKIN PROBLEMS .,............' '., ..,
b. EYE IRRITATIONS ., ".
c. OR LUNG CANCER !
d. DONT KNOW (DONT READ)
18. What can homeowners do to reduce high radon levels in their
a. REMOVE THE APPLIANCES CAUSING THE PROBLEM :..'.
b. HIRE A CONTRACTOR TO FIX THE PROBLEM . ., ,..'..
C. OR THERE IS NO WAY TO FIX THE.PROBLEM . :,
d. DONT KNOW (DONT READ) ... ;
homes?
01
02
03
94
01
02
0'3
94
01
02
03
94
01
02
03
94
01
02
•03
94
01
02
03
94
01
02
03
94
-------
j Always _"; Often . : Sometimes Never Don't
!. -! " • i !•• .'•'•• .. . • , know
a. I TRY TO FIX THINGS ' • :
AROUND THE HOUSE 01 02 ...;../.. 03 .'.... 04 94
b. I EXERCISE.AND/OR WATCH
WHAT I EAT TO PROTECT MY
HEALTH -•'....: • 01 ;.,02 03 i I.,. 04:/.-...•... 94''-
c. I ASK MY PHYSICIAN A LOT ' '
OF QUESTIONS ABOUT MY
HEALTH ••••• '• 01...;...-..-02 .-.....:. 03...;......04........94
d. I WAIT UNTIL I HAVE A LOT OF '.'•'•
INFORMATION BEFORE I DECIDE
TO BUY SOMETHING LIKE A NEW
APPLIANCE = 01 02 ......;..03 ,...04.. ..94 -,
e. I QUESTION INFORMATION FROM '
EXPERTS OR OTHER AUTHORITIES .01 02 " 'm ' '-' ' rv» "' «„ "
wfc wo U4.......; y4
20. Please tell me how active you are in each of the following types of organizations or activities
(READ LIST AND SCALE; CIRCLE ONE NUMBER FOR EACH STATEMENT)
| Very i ; Somewhat .'Not at all' :"DonT
• . . I active i i active i \ active , .know
a. CIVIC CLUB (KIWANIS. LEAGUE OF WOMEN VOTERS) .... 01 02 03 94
b, CHURCH OR RELIGIOUS ORGANIZATION., 01 .. 02 03 94
C, VOLUNTEER ACTIVITIES (RED CROSS. UNITED WAY) 01 ........ 02 ....... 03 94
-------
Now, we nave jusi a lew more general paeKgrouna questions.
21. About how many years have you lived at this address? L_ YEARS
22. Is your home a: (READ LIST; CIRCLE ONE NUMBER.)
,. a. SINGLE-FAMILY HOME 01 d. TOWNHOUSE .
b. MOBILE HOME ...: '....,.......'... 02 e. CONDOMINIUM
c. DUPLEX .".,..'.: '........ 03 f. DONTH
23. to the best of your knowledge was your home built: (READ
a. BEFORE 1940 01 " * c AFTER •
b. BETWEEN 1940 AND 1976 . 02 d. DONT
r; CIRCLE ONE NUMBER.)
'975............................ ,,03
K NOW (DON7READ) 94
24. , Are you planning-to move during the next year? (CIRCLE ONE NUMBER.)
a. YES .' 01
b. MAYBE ., .'. *.' 02 .
c. N0: ;....." 03 <
d. DONT KNOW (DONT READ) 94
25. Does your home have a basement? (CIRCLE ONE NUMBER.;
a. YES ...'...,
b. NO '.. '........
C. DONT KNOW (DON'T READ) ...
26. Is any part of your basement used as living space by you or your family?
(CIRCLE ONE NUMBER.)
a. YES ,. .'.:.... '.:
b. NO-
c. DONT KNOW (DONT READ)
04
05
NOW (DONT READ) 94
01
02
94
Skip to Question 27
on page 8
01
. 02
. 94
-------
28. How many children under 12?
29. Do you smoke cigarettes or other tobacco products? (CIRCLE ONE NUMBER.)
a-YES ^........... 01
b-NO ;. :.....,...,..: ,.,' '..:•.'......_ 02
30. D'oes anyone else in your household smoke? (CIRCLE ONE NUMBER.)
a. YES ". ....; _ ;_;.'_' 01
b. NO .....; ;............ 02^
' 31. What was the highest grade'of school that you completed? (CIRCLE ONE NUMBER.) • «.
a. NO SCHOOL..... 01 e. SOME COLLEGE (13-15) .....05
b. GRADE SCHOOL (1-8) .. 02 . f. COLLEGE GRADUATE (16) 06
c. SOME HIGH SCHOOL (9-11) 03 g.'POSTGRADUATE (17+) 07
d. HIGH SCHOOL GRADUATE (12) .... 04 ' ,
32. What is your age? YEARS
33. Is your racial or ethnic background (CIRCLE ONE NUMBER.) '
a. WHITE OR CAUCASIAN 01 d. ASIAN OR PACIFIC ISLANDER ..... 04
b. B.LACK OR NEGRO 02 ' e. NATIVE AMERICAN INDIAN ...05
c. HISPANIC 03
34. (ASK ONLY I'F UNCLEAR.) What is your sex? (CIRCLE ONE NUMBER.)
a. MALE ...'.. ., ..01 ,
b. FEMALE 02 *
35. I'm going to read a list of income categories for FAMILY income from all sources BEFORE taxes
during 1986. Please tell me to stop when I get to yours. (CIRCLE ONE NUMBER.)
a. S5.000 OR UNDER •... 01. e. 535,001 -550,000 05
b. S5.001 - 515.000 ' ., 02 f. 550,001 • 565,000 ........ .;.,06 ' •'
c, S15.001 - 325,000 03 • g. 565,001 - 580,000 07
d. S25.001 - 535.000 04 h. 580,001 AND OVER .;,:... .08,
36. If you had to sell your home today, what do you think your home and property would sell for?
S _ (PROBE FOR APPROXIMATE)
Thank you very much for your cooperation.
Your answers will be most helpful in this study.
-------
Expires a/eo
Communicating Radon Risk Effectively:.
Maryland Followup Survey— Independent Sample
Telephone #
RTI ID #•
1. Compared to other issues the State of Maryland faces, do you think; environmental issues are:
(READ LIST; CIRCLE ONE NUMBER.)
a. MORE IMPORTANT •:.... -. j .:...• •.. .. 01
b. JUST AS IMPORTANT .: ' ;! ..' ...'..... 02
C LESS IMPORTANT '....' ,; 03
;' ' - •
d. DONT KNOW/DOA/T READ) .'...,....'.: ......,......,:....,!. , 94
2. We're interested in finding out how serious you think the risks from some types of pollution are to
your community and to your household. On a scale from 1 to 10, with 1 meaning NOT AT ALL
SERIOUS and 10 meaning VERY SERIOUS, please tell me how serious you think the risks from.
each type of pollution are to your community and to your household. (READ LIST, SCALE, AND
CATEGORIES; PROBE FOR NUMBER.) i
Not.
Ser
at all- -. ' • jVe
ous ' .' • Sen
i
ii i i ii t i
1 2 3 ' 4 ' 5 6 7 8 • 9 -I '
'
• . ' . I - .'
ry
ous
10 ' , .
\
Your • ' . i ,Your
! community • ! household
a. LEAD IN DRINKING WATER ...'.
b, HAZARDOUS WASTES IN LANDFILLS.
c. RADON IN HOMES
3. For the rest of the interview I'm going to ask questions mainly
mentioned—radon in homes. During the past few months, have
about radon? (CIRCLE ONE NUMBER.)
about
one of the sources I
you seen or heard anything
a. YES
b. NO .,;...-
C, DONT KNOW (DONT READ)
0'
OS
Continue
Skip to Question
on page4
-------
' the radio or TV about radon?" "C" a"yU""a '" * n6WSpaper °^9az.ne or heard anytnmg on j
a. YES..... .....
b. NO. ' " ............. '""• ........... ••••;.. ...01 —^Continue
................. '""' ............ /••...-.... ..... oe —-*»• Skip to Question 4C
on page 3
4B. (1) Was that in a newspaper?
a YES , .
..,_"' ..... .......... ........... ••••• ............. '••• 01 • — ^Continue -
D NO
............. ' ' ............. ....... ....... ....... 02 — *• Skip to Question 4B, part
^^^
ANNOUNCEMENT/DO NOT READ LIST.) SERVICE
*
Public Service
BABY WITH CIGARETTES ........ '.-. ..... . .......... / 01 Y Announcement
HOUSE WITH ARROWS COMING FROM GROUND.... """"" ^ Was this a 01
HEALTH AD WITH FOOD, EXERCISE ETC.. ..... ....... ...... 03 Public Service . - "'
ZIP CODES OF AREAS WITH HIGH RADON LEVELS Q4 Announce-
ANNOUNCEMENT OF MEETING ABOUT RADON ...... '.'.'.'.'' ' .' ] « ' f ^dJcle to*' !
STATE RADON HOTLINE TELEPHONE NUMBER ..... 05
ARTICLE/STORY/EDITORIAL .......... ........... '.-.'."'•"" 07 ' ' •
OTHER (SPECIFY) _ ' " " nq I -" 01
; — - ; - ; - • -- . . 08 / . ; . Q1
(2) Was that in a magazine? ' , - .
a YES
. ' .._ ..... '. ..... ........... • • • • ........ • ..... • • ; ........ 01 — *• Continue
D. NO .......... «« k. MI . . A-
. ................. •_ • • .......... °2 — -** Skip to Question 4B, part (3
Public Service
Announcement
SPECIFY1— - - --- ,01- WM,hisa „ : • •
SPECIFY 2 _ _ . _ , M Public Service
~ ~ — : - - ••• 02 Announce- 01
ment?|fYES,
(3) Was that on the radio? Circle°1
b' ............. • .............. • • •• ............... ....... '01 — * Continue
................... ''•••••. .......... ........... ••'• • • °2 — * Skip-to Question 4B, part (4)
on page 3
SPECIFY 1
SPECIFY 2
SPECIFY 3
SPECIFY 4
A<
• -01
M
••• 02
.
.-..03
. 04
Was this a
Public Service
Announce-
ment?l(YES,
circleOI
Public Service
Announcement
01
01
01
-------
b. NO
1 9neral 'theme or
details about wna*
PUBLIC SERVICE AD WITH JOHN GLENN
LITTLE GIRL WITH GAS MASK
PUBLIC SERVICE AD WITH SENATOR WARNER OF VIRGINIA .
PUBLIC SERVICE AD WITH A FAMILY ...
ROBERTA BASKIN ON WJLA-TV/CHANNEL 7 RADON WATCH
OTHER (SPECIFY)
m N
n?
,*«•
03
05
06 t
. Public Service
Announcement
. was tnis a 01
Public Service „•«
. 01
Announce-
( ment? If YES, 01
circle 01 n4
ft*
™
/H
4C. Have you'seen a poster, read a utility bill insert, or heard a pre<
the past few months?
a. YES
b. NO.
Was that a poster, utility bill insert, or a presentation? (CIRCLE ALL
(1) POSTER ... 01
Please briefly describe the poster or posters you saw (\
READ LIST.)
BABY WITH CIGARETTES /
HOUSE. WITH ARROWS COMING FROM GROUND .......
• HEALTH AD WITH FOOD, EXERCISE, ETC .:.......,..
ZIP CODES OF AREAS WITH HIGH RADON LEVELS ......
' OTHER (SPECIFY) .
' (2) UTILITY BILL INSERT .: 02' ' , :
1 (3) PRESENTATION/SLIDE SHOW 03 *
What organization was responsible for this presentation, slide
4D. Did you see any other printed information about radon?
a: ' YES '......:..., •; _
. . b. NO ! s •'--,'• ^
Where did you get this printed information? (CIRCLE ALL
THE STATE OF MARYLAND. .,-_.. 01
THE US. ENVIRONMENTAL PROTECTION AGENCY ; 02
WJLA-TV CHANNEL 7/RADON WATCH , '.. 03
AMERICAN LUNG ASSOCIATION/AMERICAN CANCER SOCIETY . 04
AT A LOCAL.STORE OR,MERCHANT ..' Q5
AT THE LIBRARY ;.,;..... .....-..;•
A REAL ESTATE AGENT/TRANSACTION . 07
TESTING COMPANY ,
OTHER (SPECIFY) • ' • ' ___'
DONT KNOW . .'• '
06 -
08
09
94
Continue
Skip to Question 4C
saw on TV (PROBE FOR
'.DO
;enl:ation/slide show about radon in
Continue
Skip to Question 4D
THAT APPLY)
•IRCLEALL THAT APPLY. DO NOT
01
02
03
04
05
show?^
Continue
Skip to Question 4E
on page 4
THAT APPLY. DON7 READ LIST.)
Who published this
information?
Specify
' Specify
- Specify
Specify
Specify
-------
" "' '"u-;"c' »•" ^wui/\ci in me H<«H iew mornnsi'
b'N0 "" ' " ;• .•:••'• 01 .—^Continue
, 02 —* Skip'to Question 4F
Was that a friend, relative, or coworker? (CIRCLE ALL THAT APPLY.)
a< FRIEND '01 c. COWORKER ' 03 '•
b. RELATIVE ;.. 02 '
4R Have you called the State of Maryland toll-free number for radon information?
a-YES : .01 '
b. NO 02
4G. In the past few months have you learned anything about radon in some other:way?
a> YES • • • ;-.. 01 —*• How was that? - .
b. NO ' 02 . ~~ ; ~—•
L..
a. MARYLAND DEPARTMENT OF THE ENVIRONMENT '
............. , • .............. :• ---- ...... ...... 01
b. 'MARYLAND DEPARTMENT OF HEALTH AND MENTAL HYGIENE ...........
c. LOCAL HEALTH DEPARTMENT
• ' ........... ' ............................... •• .................... 03
d. U.S. ENVIRONMENTAL PROTECTION AGENCY
..... ' "' ......................... • ................. 04
e. OTHER (SPECIFY) _ _ __ • ' • . - '
f. DONT KNOW (DO/VT READ; ~. ,
............. ' ............. ' ........................... '. ....... 94
6' voyu°tmWsf ?fpd^5fn ab°Ut rad°n:re'ated health P^blerns, which of the following sources would
you trust the most to give you that information? (READ LIST; CIRCLE ONE NUMBER.)
a. MARYLAND DEPARTMENT OF THE ENVIRONMENT . : ' •'
• —'••••• • 01
b. MARYLAND DEPARTMENT OF HEALTH AND MENTAL HYGIENF
02
C. LOCAL HEALTH DEPARTMENT.... ' :' .
03
d. US. ENVIRONMENTAL PROTECTION AGENCY '.
e. FAMILY DOCTOR ; '"'""'"."."",' °*
f. SOME OTHER SOURCE (SPECIFY) - " °5
: ' :—; '—. .—_ ... 06
9 NO ONE (DON7READ) ••
' " :' '•••••... • '. . . ,-/- -Q7
h. DON'T KNOW (DON7 READ) ;
"' "" ': " ' •;•;•••••••••:•••'•• .V...1.....•:/.. •.... ,. 94
-------
7. People have different opinions about radon. How much do you
. . statements are your opinion? (READ ANSWER CHOICES AFTE
Strongly
,• ' . , _ agree - .
a. IT IS IMPORTANT TO. TEST MY HOME
TO FIND OUT IF 1 HAVE A RADON . •
PROBLEM 01
b. IF I HAD A RADON PROBLEM, IT '
WOULD BE COSTLY TO FIX 01 .....
' . • c. EVEN IF A RADON PROBLEM WAS .
FIXED, MY HOME WOULD STILL BE
WORTH A LOT LESS 01 . . . , .
Agree j
- -I ..
. .. 02... .
... 02 .....
... 02
agree or disagree that the following
R FIRST STATEMENT.)
I Disagree Strongly; j Don't i
| ' disagree ; ! know •
. . 03 04 94 ,
....03. ...04. ...94
.... 03 04.. ...'....94
8A. Have you had your home tested for radon? (CIRCLE ONE NUMBER.)
a. YES...:....,.... !oi .—
b. NO
c. DON T KNOW (DON7 READ)
02
94
When did you get your results?
(// "don't know", enter "94/94")
i
MONTH/YEAR
8C. Were the results over 4 picocuries per liter?
a. YES :. ' *.. ....
b. NO '....
' C. DONT KNOW (DON7READ)
8D. Did you do followup testing, anything to fix the problem, both,
a. FOLLOWUP TESTING .'....'.':
i b. FIX PROBLEM
c. BOTH... ••..... -..:..- ;
d. NOTHING -. ...;......
e. DONT MOW (DONT READ) ,
Continue
Skip to Question 9
on page 6
01
02
94
?r nothing?
01
02
03
Continue
Skip to Question 11
on page 6
04
94
Skip to Question 11
on page 6
-------
main reason you have not had yours tested. (DCW7 READ'UST; 'ciRCLE^LLTHAfAPPLY') '
a. NEVER THOUGHT ABOUT IT...: ......V 01 :: . ~
b. DIDNT KNOW IT WAS POSSIBLE ...!..,... :.... 02 : "
c. DONT THINK I HAVE A PROBLEM IN MY HOME ...03 ;
d. DIDNT KNOW HOW TO TEST .. 04
e. THOUGHT TESTING WAS TOO EXPENSIVE'..: ...'. 05
f. DONT THINK TESTS ARE RELIABLE .06
g. NOT INTERESTED 07 . - :
h. DIDNT KNOW IT WAS A PROBLEM'N THIS AREA, 08
i. WOULD RATHER NOT KNOW IF THERE IS A PROBLEM 09
j. NOTHING CAN BE DONE ABOUT RADON ANYWAY 10 ,
k. FIXING A PROBLEM IS TOO EXPENSIVE u - ' .
I. CONCERNED ABOUT CONFIDENTIALITY 12
m. JUST HAVENT GOTTEN AROUND TO IT 13
n. OTHER (SPECIFY) ' 14
o. DONT KNOW (DON7READ) 94 ;
10. Suppose your local health department was offering a radon test for a one-time cost of S10, $25,
$50, $100. The cost would cover two radon detectors, the results, and a booklet about radon
Would you take part in such a radon testing program? (CIRCLE ONE NUMBER.) , -:
a. YES .'. : • : 01 .'.;
b. NO ; ;...... 02 '- . •
C. DONT KNOW (DONT READ) 94
11. Suppose you are just moving to this area and you want a home like the one you're in now. You
have narrowed the choice to two houses that are almost identical. The only difference is that
House 1 has radon levels 2, 5 times higher than the government's guidelines for action, while
House 2 has no radon but costs an additional $5,000, $10,000, $15,000, $20,000. Which house
would you buy? (CIRCLE ONE NUMBER.)
a. HOUSE 1 (Cheaper House with Radon) 01
b. HOUSE 2 (More Expensive House without Radon) .,..-... 02
C. DONT KNOW (DON7READ) 94 - -' >
-------
Some people have heard a great deal about radon while others have
in learning how much people know about radon. For the next grou|D
you three choices. Please tell me which answer you think is best. I1
then say that. . ' .• • '
12. Where does most radon in homes come from?
a. INDUSTRIAL POLLUTION '
b. NATURAL URANIUM IN SOIL
c. OR HOME APPLIANCES .......:
• d. DONT KNOW (DON'T READ)
13. Which of the following best describes radon? Radon has:
a. A SLIGHT ODOR .'.
b. A STRONG ODOR •.....,.
c. OR NO ODOR AT ALL
d. DONT KNOW (DONT READ)
14. When radon is measured in a home, which of the following wi I affect the level the most?
' a. THE TIME OF YEAR ITS MEASURED
b. THE AMOUNT OF INDUSTRIAL POLLUTION AROUND THE HOME .
C. OR THE NUMBER OF APPLIANCES IN THE HOME
d. DONT KNOW (DONT READ) .: .'.
15. How can you test your home for radon?
a. YOU CAN DO IT WITH A HOME TEST ......
b'. ONLY TRAINED'PERSONNEL CAN TEST....
c. OR YOU CANNOT TEST FOR RADON '.....
,d. DONT KNOW (DONT READ) '.'....
16. When do health problems from being exposed to radon usual y occur?
a. WITHIN A FEW WEEKS '.-..,.' • .'.............
• b. INAFEWYEARS
c. OR NOT FOR 10 TO 30 YEARS ..:. '......
"d. DONT KNOW (DONT READ)
17. What kind of health problems are high levels of radon exposuie likely to cause?
a. MINOR SKIN PROBLEMS ...........
b. EYE IRRITATIONS
c. OR LUNG CANCER '..:
d! DONT KNOW (DONT READ)
What can homeowners do to reduce high radon levels in their
a. REMOVE THE APPL'ANCES CAUSING THE PROBLEM .'...'
b; HIRE A CONTRACTOR TO FIX THE PROBLEM
c. OR THERE IS NO WAY TO FIX THE PROBLEM .......... ....
d. • DONT KNOW (DONT READ) . . . .• .
heard very little.-We're interested-
of questions, I am going to read
"don't know" is your best answer,
Record
Responses
homes?
01
02
03
94
01
02
03
94
01
02
03
94
01
02
03
94
.."01
02
03
94
01
02
03
94
01
02
03t
94
-------
d8SCrbe themselves in various
each statement I rea please tell me
°r
Always
a. I TRY TO FIX THINGS
AROUND THE HOUSE
b. I EXERCISE AND/OR WATCH
WHAT I EAT TO PROTECT MY
' HEALTH
01.
C. I ASK MY PHYSICIAN A LOT
OF QUESTIONS ABOUT MY
HEALTH
01.
d. I WAIT UNTIL I HAVE A LOT OF
INFORMATION BEFORE I DECIDE
TO BUY SOMETHING LIKE A NEW
APPLIANCE ....'.'...........
e. I QUESTION INFORMATION FROM
EXPERTS OR OTHER AUTHORITIES
Often
.... 01 ...'. 02
02 .
02 .
01 02 .
.... 01. 02 ..
Sometimes
I Never | . Pont ,
i • I iknow !
03 ;. 04.......: 94
03 ,04 94
03 ; 04 94
03 04 ..94
03 04... 94
2°'
c?n nir^w^?,1^ fti?8 ln 6aCh °f ^ f0"OWin9 ^es of °W™*
(READ LIST AND SCALE; CIRCLE ONE NUMBER FOR EACH STATEMENT.)
or activities.
a, CIVIC CLUB (KIWANIS. LEAGUE OF WOMEN VOTERS) .
b. CHURCH OR RELIGIOUS ORGANIZATION ..;....
C. VOLUNTEER ACTIVITIES .(RED CROSS, UNITED WAY)
Very
active
.... 01 ...
. .... 01 ...
01 ...
Somewhat
active
......02 ,....,
02 .....
......02
Not at all -
active
.....03
....03,,...
....03......
Don't |
know |
...94
...94
.'..94
-------
Now, we have just a few more general background questions.
21. About how many years have you lived at this address?
22. Is your home a: (READ LIST; CIRCLE ONE NUMBER.)
a. SINGL&FAMILY HOME .........01
b. MOBILE HOME 02
: C. DUPLEX .,..'-. 03
d. TOWNHOUSE ./. 04
e. CONDOMINIUM ....,.'....'.;....'..'. 05
f. DONT KNOW (DONT READ) 94
23. To the best of your knowledge was your home built: (READ
a. BEFORE 1940 ......,., 01"
b. BETWEEN 1940 AND 1976 02
c. AFTER 1976 ". 03 ' .
d. DONT KNOW (DON7READ) .'.....!.. 94
24. Are you planning to move during the next year? (CIRCLE ONE NUMBER.)
a/YES,.•/...' ......... 01
b. MAYBE ,....; Q2
' c. NO '...., - ....03 •'
•'• d. DONT KNOW (DONT READ) 94
25. Does1 your home have a basement? (CIRCLE ONE NUMBER.)
a. YES
b. NO'...
c. DONT KNOW (DONT READ)
26. Is any part of your basement used as living space by you or your family"7
(CIRCLE ONE NUMBER.)
a. YES
b. NO
C. DONT KNOW (DONT READ)
27. How many people are there in your,household?
28. How many children under 12?
YEARS
1ST; CIRCLE ONE NUMBER.)
01
02
94
02 1
94 >
Skip to Question 27
.01
02
94
-------
29. Do you smoke cigarettes or other tobacco products? (CIRCLE ONE NUMBER.)
a-YES '..: .......... ...,.; ..: '.m
>• NO......;..... .: ... ........02
30. Does anyone else in your household smoke? (CIRCLE ONE NUMBER.)
a-YES '.: ...... :.... ;.'.• 01
b- NO ; .,;. 02
31. What was the highest grade of school that you completed? (CIRCLE ONE NUMBER.)
3.NOSCHOOL :...,..; 01 e. SOME COLLEGE (13-15) ......... 05
b. GRADE SCHOOL (1-8)....:....... 02 f. COLLEGE GRADUATE (16) 06 , *
c. SOME HIGH SCHOOL (9-11) 03 g. POSTGRADUATE (17+) 07
d. HIGH SCHOOL GRADUATE (12) .... 04
• x'
32. What is your age? YEARS , '
33. Is your racial or ethnic background (CIRCLE ONE NUMBER.)
a. WHITE OR CAUCASIAN 01. d. ASIAN OR PACIFIC ISLANDER .'...; 04
b. BLACK OR NEGRO 02 e. NATIVE AMERICAN INDIAN'........ 05
c. HISPANIC- '03 . '
34. (ASK ONLY IF UNCLEAR.) What is your sex? (CIRCLE ONE NUMBER.) .
a. MALE ..01 •
b. FEMALE,,' 02 ' . - • -
. \ •
35. I'm going to read a list of income categories for FAMILY income from all sources BEFORE taxes
during 1986. Please tell me to stop when I get to yours. (CIRCLE ONE NUMBER.)
a. $5.000 OR UNDER 01 e. $35,001 - $50,000 i...,05 v
b. S5.001 - $15.000 02 f. $50,001-$65,000..., .'.. / 06'.,
c. $15.001 - $25.000 ....03' g. $65,001 - $80,000 ,...07
d. $25,001 • S35.000 04 h. $80,001 AND OVER...! ..'.08
36. If you had to sell your home today, what do you think your home and property would sell for?
$— (PROBE FOR APPROXIMATE)
Thank you very much for your cooperation.
Your answers will be most helpful in this study.
10
-------
Communicating Radon Risk Effectively:
Maryland Followup Survey—Dependent Sample
(Label)
Final Interview Code
03 Ring, No Answer 09 Unable to Contact
05 Double Wrong Connection 12 Interview Completed
06 Answering Machine/Service-13 Partial Data
07 No Result from Dial 14'Final Interview Refusal
08 Fast Busy/Compyter Tone 15 Other __
Helio, my name is
.. I'm calling from the Research Triangle Institute in North Carolina.
I 1CHU, my name 10 .— -— 9 — _...._ w -
We called your home in December as'part of the State of Maryland's Situdy on what people know and
think about radon. We are now conducting the followup survey for the State of Maryland..! would like to
talk with (READ NAME ON LABEL), or the person we spoke with in December. (II person is
unavailable, schedule a callback.)
(WHEN YOU ARE TALKING WITH THE PERSON WHO
THAT NAME BELOW.)
WAS PREVIOUSLYINTERVIEWED, RECORD
Respondent's Name:
As I said, we are conducting the followup survey for the State
think about radon. This won't take much of your time, and your
of Maryland on what people know and
answers will be kept strictly confidential.
1 Compared to other issues the State of Maryland faces, do you thir k' environmental issues are:
(READ LIST; CIRCLE ONE NUMBER.)
a. MORE IMPORTANT
b. JUST AS IMPORTANT .......
c. LESS IMPORTANT
d. DONT KNOW (DON7 READ)
2.
with
We're interested in .finding out how serious you think the risks from
your community and to your household. On a scale from 1 to 10, v
SERIOUS and 10 meaning VERY SERIOUS, please tell me how s
each type of pollution are to your community and to your household
CATEGORIES; PROBE FOR NUMBER.)
a. LEAD IN
b.' HAZARDOUS
c. RADON IN HOMES
Serious
1 i i it i t
12345678
TPR
IN LANDFILLS
Seri
i
9.
> CO
01
02
.03
94
some types of pollution are to
1 meaning NOT AT ALL;
i erious you think the risks from
. (READ LIST, SCALE, AND..
Very
10
Your
community
Your i
household ;
-------
past few months, have you seen or heard anything about radon? (CIRCLE ONENUMBER.f
a- YES ...' 01—* Continue ,
b. NO 02 I
c. DONTKNOWfDONT/raoj..: •... 94 j—^ Skip to Question 5
on page 5
4A. In the past few months have you seen anything in a newspaper or magazine or heard anything on
the radio or TV about radon?
v
a. YES 01 —i* Continue
b. NO / .......;.... 02 —i* Skip to Question 4C
on page 4 *
4B. (1) Was that in a newspaper?
a. YES..... .....01 —-*> Continue
b. NO 02 —> Skip to Question 4B, part (2
Please tell me the general theme or any details about what you saw in the newspaper.
(PROBE FOR EACH SELECTION TO FIND OUT IF IT WAS A PUBLIC SERVICE
ANNOUNCEMENT. DO NOT READ LIST.) '
'
Public Service
Announcement
BABY WITH CIGARETTES "...'. 01 \ ' 01
HOUSE WITH ARROWS COMING FROM GROUND 02 Was this a 01
HEALTH AD WITH FOOD, EXERCISE. ETC 03 Public Service ' Q1
Announce-
ZIP CODES OF AREAS WITH HIGH RADON LEVELS 04 I ment? If YES, '01
ANNOUNCEMENT OF MEETING ABOUT RADON ....05 ' circle 01 ' 01
STATE RADON HOTLINE TELEPHONE NUMBER 06 '' -, 01
ARTICLeSTORY/EDITORIAL .....07 . 01
OTHER (SPECIFY) . . 08 / 01
(2) Was that in a magazine? . ,..':!-.
a. YES ...01 —^Continue
• b. NO 02 .—•» Skip to Question 4B, part (3
on page 3
!•'.'.
Please tell me the general theme or any details about what, you saw in the magazine. (PROBE
FOR EACH SELECTION TO FIND OUT IF IT WAS A PUBLIC SERVICE ANNOUNCEMENT.
DO NOT READ LIST.) . :
SPECIFY 1
SPECIFY 2
01
02
Was this a
Public Service
Announce-
ment? If YES,
circle 01
Public Service
Announcement
01
01
-------
a. YES.....
b. NO .....
01
02
NOT READ LIST.)
PUBLIC 5
—*• Continue
—•* Skip to Question 48, part (4
heard on the radio. (PROBE
'ERViCE ANNOUNCEMENT.
SPECIFY 1
*~ ._
SPECIFY 2 _
SPECIFY 3
SPECIFY 4
' *~' - • ._
/
(4) Was that on TV?
a. YES.... •
b. NO .
Please tell ma th« „„
- ' - ' .
— . . 01
~ ' — — — — • ..'. 02
•— ^ —
' ' • -. . . 04
1 : • •
-•••••:. •••••-....:.... 01
02 •
lit
Public Service
Announcement
Was this a 01
Public Service
Announce- 01
ment? If YES,
circle 01 .01
01
t
-;'. • / • • "- • ""' ":
r-*- Continue
"—*• Skip to Question 4C
on page 4
NOT READ LIST.)
PUBLIC SERVICE AD WITH JOHN GLENN'
LITTLE GIRL WITH GAS MASK
'
PUBI-IC SERVICE
on TV. (PROBE FOR
ANNOUNCEMENT.
01
02
-•«• —— • •" "M n M rMMILY • ' '
!™~^ -DO, w^H;;:;;; I
06
3 •
Was this a
Public Service
Announce-
ment? If YES,
circle 01
Public Service
Announcement
01
n
01
01
01
01
01
-------
nave you seen a po&ier, reaa a utility- DIM insert, or heard a presentation/slide ;show about radon
in the past few months?
a- YES '••• • 01 —* Continue..
b- NO .'....,........:..... 02 —*» Skip to Question 40
Was that a poster, utility bill insert, or a presentation? (CIRCLE ALL THAT, APPLY.)
(D POSTER ;..' 01
Please briefly describe the poster or posters you saw. (CIRCLE ALL THAT APPLY. DO NOT
READLIST.)
BABY WITH CIGARETTES 01
HOUSE WITH ARROWS COMING FROM GROUND ....02 '
HEALTH AD WITH FOOD, EXERCISE, ETC. 03 , ,
ZIP CODES OF AREAS WITH HIGH RADON LEVELS 04 ...'-.
OTHER (SPECIFY) ... 05
(2) UTILITY BILL INSERT ..02
(3) PRESENTATION/SLIDE SHOW 03
What organization was responsible for this presentation/slide show? '
4D. Did you see any other printed information about radon? "- ; ,, • "
a- YES : 01 —> Continue
b- NO , 02 —> Skip to Question 4E
Where did you get this printed information? (CIRCLE ALL THAT APPLY. DON J READ LIST.)
THE STATE OF MARYLAND 01
THE 'US. ENVIRONMENTAL PROTECTION AGENCY : 02 , Who published this
WJLA-TV CHANNEL 7/RADON WATCH 03 information?
AMERICAN LUNG ASSOCIATION/AMERICAN CANCER SOCIETY . 04
AT A LOCAL STORE OR MERCHANT 05 ' ^ > '
ATTHE LIBRARY 06 -
A REAL ESTATE AGENT7TRANSACTION 07 -
TESTING COMPANY ,........... 08 - . .
OTHER (SPECIFY) ... 09 —-Sf-jL^ . . ' . ' •.
DONT KNOW '. 94
4E. Have you talked about radon with a friend, relative, or coworker in the past few months?
•a. YES 01 ^-*Continue
b. NO , '.......;. 02 —*• Skip to Question 4F
, Was that a friend, relative, or coworker? (CIRCLE ALL THAT APPLY.) '-"...
a. FRIEND 01 c. COWORKER ...03
b. RELATIVE 02
4F. Have you called the State of Maryland toll-free number for radon information? •'
a. YES ... '..... 01
b, NO, ,.:. 02 • . •
4Q. In the past few months have you learned anything about radon in some other way?
a, YES . • 01 -—*> How was that? '• . . : ; • •_
t> NC 02 ' •• • •''.'.
-------
wnicn government agency would you contact? {DUtNOTREAD LIST; CIRCLE THE AGENCY.)
a, MARYLAND DEPARTMENT OF THE ENVIRONMENT
b. MARYLAND DEPARTMENT OF HEALTH AND MENTAL HYGIENE
C. LOCAL HEALTH DEPARTMENT ',.....
d. U.S. ENVIRONMENTAL PROTECTION AGENCY
, e. Of HER '
f. DON'T KNOW (DON'T READ).
6.
If you .wanted to learn about radon-related health problems,
you trust the most to give you that information? (READ LIST;
which of the following sources would
CIRCLE ONE NUMBER.)
a. MARYLAND DEPARTMENT OF THE ENVIRONMENT
b. MARYLAND DEPARTMENT OF HEALTH AND MENTAL HYGIENE
C. LOCAL HEALTH DEPARTMENT
d. U.S. ENVIRONMENTAL PROTECTION AGENCY
6. FAMILY DOCTOR
f. SOME OTHER SOURCE (SPECIFY) ' '
g. NO ONE (DON'T READ) . . ,
h. 'DON'T KNOW (DON'T READ) ...:...,........ .,..>.....
7. People have different opinions about radon. How much do;you
following statements are your opinion? (READ ANSWER CHO/l
. ' -
a. IT IS IMPORTANT TO TEST MY HOME
TO FIND OUT IF 1 HAVE A
RADON PROBLEM
b. IF 1 HAD A RADON PROBLEM, IT
WOULD' BE COSTLY TO FIX
c. EVEN IF A RADON PROBLEM WAS
FIXED, MY HOME WOULD STILL BE
. WORTH A LOT LESS
Strongly
agree
. ... 01
... 01
. ... 01
Agree
i
. . . 02
...02 ..
... 02
Disagree i
I
03
. . 03
>
.... 03
Strongly ; j Don't ;
disagree i \ know ;
f\A • t\A
F\A ~ Q*
...04 94
....: 02
....03
....04
. ! . 05
.... '94
01
02
03
04
05
06
07
94
agree or disagree that the
JES AFTER FIRST STATEMENT.)
-------
UH. nave you liau youi Home tested tor radon? (CIRCLE ONE NUMBER.)
a, YES .'...-....
b. NO
01
, 02)
c. DON T KNOW (DON7READ) 94 f
Continue
Skip to Question 9
on page 7
8B. When did you get your results?
(If "don't know", enter "94/94") MONTH/YEAR
8C. Were the results over 4 picocuries per liter?
a- YES 01
b- NO ,,..., 02
C. DONT KNOW/DOW? fl£4Dj 94
Continue
Skip to Question 12
on page 7
8D. Some people may have done various things to reduce their radon exposure. For each of the
, changes I read, please tell me whether you are not considering, considering, planning.or have
done it.
Not Considering
Considering
Planning JDonej Don't!
I Know!
03
03
04
04
94
94
a. SPEND LESS TIMu IN CERTAIN
AREAS OF YOUR HOME, FOR
EXAMPLE, IN THE1 BASEMENT 01 02
b. 'STOP USING A BASEMENT
BEDROOM OR FAMILY ROOM ......... 01 02 .."'
c. OPEN WINDOWS AND TURN ON
FANS TO INCREASE AIR FLOW ' * i • " .
INTO AND THROUGH THE HOME
• • • • • • 01 02 03 • 04 94
d. CONTACT A CONTRACTOR 01 ,. 02 ..' 03 ..... 04 .... 94
e. STOP OR CUT DOWN ON | .
SMOKING INSIDE THE HOME .......... 01 02 03 '..'.. .04 .... 94
I INSTALL VENTILATION SYSTEM ......... 01 02 .......... 03 ..... ;04 .... 94
g. OTHER (SPECIFY) 01 .02 03 ...... 04 ... 94
V ' .
8E. Did you do any folldwup testing? :
a< YES •••••• -••• ;...-oi —*• Continue
b' N0 ; •'• 02 —+ Skip to Question 11
on page 7
8F. Were the results over 4 picocuries per liter?
a. YES . .-. ,; 01 \ !
b NO . ........ 02 > Skip to Question 11
c, DONT KNOW - ..:... 94 7 Ofl P*9e 7
-------
.... 01
a NEVER THOUGHT ABOUT IT -.,...-..... j,.;.. ^ ... :,r ..-.•• : •••
; 02
b. ' DIDNT KNOW IT WAS POSSIBLE , .--..•••: j. ••-.-• •.„• • • - v' '' y, ' • . . •
e. DONT THINK I HAVE A PROBLEM IN MY NOME {, ......••.
d. DIDNT KNOW HOW TO TEST .... I .......,.,..••••.••••••-•••'•: ."';' " " •
e. THOUGHT TESTING WAS TOO EXPENSIVE j ........ • • • ...-..••••.•:: •; • •;, ^
f. DONT THINK TESTS ARE RELIABLE /.. • • • • j-.v. •••••••'•• -y • • • ;'•::'''' '
g. 'NOT INTERESTED : : ..;....:..... • -: ••;•••' _•.
h. DIDNT KNOW IT WAS A PROBLEM IN THIS AREA |- • • • ''" 09'
i. WOULD RATHER NOT KNOW IF THERE IS A PROBLEM - -. -/T • • • • • • • • • • • • ;••'• • — ''' ^
j. NOTHING CAN BE DONE ABOUT RADON ANYWAY - - • • • -j : '"' ^
k. FIXING A PROBLEM IS TOO EXPENSIVE • ..,,:..,.. • • • • ^
I. CONCERNED ABOUT CONFIDENTIALITY ..., • • • • • • • • j • -• — ••••• " " " ' ' 13
m. JUST HAVENT GOTTEN AROUND TO IT - ..,,-•• j ?••••- '' ^
n. OTHER (SPECIFY) . : ' " . • h"~ ~~ " ' ' 94
,0. DONT KNOW (DON7 READ) • • • • ; |
a.
b. NO :
c. DONT KNOW (DONT READ)
•
your fuel bills by less than $100 per year. Would you make such a change?
(CIRCLE ONE NUMBER.), i • ^
a. YES ; /• -, p ; • •••;••• • • ^
- b- N0 ••<••" •'•••• '• ••-•••'••> ''";"-"l;';':;-'"":':""'':.^.::./. 94'"
; c. DONTKNOW.. ......,..•: :j::;-• • • .. _.- . . .
would you buy? (CIRCLE ONE NUMBER.) . |
a HOUSE 1 (Cheaper House with Radon) j ••.•<•••,
• , 02
b. HOUSE 2 (More Expensive House without Radon) ..-.. • • • • -,| :-.-•:••. •,-••••.-'• •• .-••••••••••• ^
c. ' DONT KNOW (DONT READ) .' :,..'.:..•...;•••.•••••••••••[•••••.••••••;.••:•"." " "
-------
Record
Response
13. Where does most radon in homes come from?
a. INDUSTRIAL POLLUTION
.......... •••••, ........................ ...... . ........... • ; . 01
b. NATURAL URANIUM IN SOIL .... .........
C. OR HOME APPLIANCES . . " " ' ....... ••.-......
............. ..................................... . ..... . . 03
d. DONTKNOWO>ONTfl£>iDj...... ................. 0 .
'•*•••••• • • ......... , ......... , ........ .... 94
14. -Which of the following best describes radon? Radon has: :
a. A SLIGHT ODOR .......... .............. '
................. ......
b. A STRONG ODOR ... . '. ............. ' ........... ""
' ........ ..... ••• ...... •"-•• ..................... . ---- '.;. 02
C. OR NO ODOR AT ALL..... . '
..... , ..... ••• ..... ......... ........ .......... ............. 03
d. DON'T KNOW (DON7 READ) . . . .-••'
.............. . ...... ....... '••••• ...... ........... ....... 94
15. When radon is measured in a home, which of the following will affect the level the most?
a. THE TIME OF YEAR IT'S MEASURED ......... .........
b. THE AMOUNT OF INDUSTRIAL POLLUTION AROUND THE HOME ......;....... .'-." ..... ' 02
c. OR THE NUMBER OF APPLIANCES IN THE HOME... ...... " ..... ' ....... :",-A-
*"*'**"**"•••••• ..... « i ........ «..,, Ww
d. DONTKNOWfDONTflEAD) '"
........ : ......................... ; ............... ........ 94
16. How can you test your home for radon? .
'.a. YOU CAN DO IT WITH A HOME TEST ............. . . .. .
b. ONLYTRAINED PERSONNEL CAN TEST ....................... .... ..... ••••••••••....
c. OR YOU CANNOT TEST FOR RADON •• ' ....... ......
. ........ ...... : • • • ........ ................ . ...... 03
d. DON'T KNOW (DON'T READ) ..... .......................
17. When do health problems from being exposed to radon usually occur?
a. WITHIN A FEW WEEKS ................ ...
b. INAFEWYEARS ' ' '• ............... " ..... •"" ' °1 ,
................ : ......... • ................ '• ...... ' ....... .... 02
c. OR NOT FOR 10 TO 30 YEARS
........... : ................................... • .......... . 03
d. DON'T KNOW (DON'T READ) ..... ...............
18. What kind of health problems are high levels of radon exposure likely to cause? '
'a, MINOR SKIN PROBLEMS ......... . ............... ..
b. EYE IRRITATIONS ' ..... ' ..... ;" ". .......... " °1
.................... ..... • ............... ............. ' ---- ....... 02
c. OR LUNG CANCER ...
......................... • ................................. 03
d. DON'T KNOW (DON'T READ) "'""'
........ ..... . ......... "• ................................ 94
19. What can homeowners do to reduce high radon levels in their homes? .
a. REMOVE THE APPLIANCES CAUSING THE PROBLEM/: ...... ...... ....... • ', Q1
b. HIRE A CONTRACTOR TO FIX THE PROBLEM. . -...:...- ....... ':.''•'.'."• ..... '•
c OR THERE IS NO WAYTO FIX THE PROBLEM.. :....v;. .; .............. . ,, . \ ...... •'. 03
d, DO.NT KNOW (DON'T READ) • ....... ...... '
..... •" ' ' ' ' ---- •"•: ...... ' ...... .' ' ^ ' ' ', ....... .. /94 ..
8 • . •'. '•' •'".';•..
-------
their homes. Please answer the following questions with your LASTlr
20A What was your last major purchase? (READ LIST. CIRCLE ONLY ONE.)
.WASHER OR DRYER 05
WATER HEATER ., • __ 06
DISHWASHER ;. ...!..,.....; 07
OTHER (SPECIFY) ______^___ ... 08'
a. FURNACE ...../... 01
b. WALL UNIT AIR CONDITIONER ......... 02
c. CENTRAL AIR CONDITIONER . • 03
d. REFRIGERATOR ' ." 04
e.
f.
g.
h.
20B. What was the approximate cost? (READ LIST. CIRCLE ONLY 6/VEJ
LESS THAN $500 ....01 '•
a.
b.
c.
d.
$501 TO$1,000 02
$1001 TO $2,000 03
MORE THAN $2,000 04
e. DONT KNOW (DON7 READ) 94
20C. What was the amount of time you spent on this decision? (READ LIST. CIRCLE ONLY ONE.)
, a. ON THE SPOT ..' ; 01 d '
b. ONE DAY 02 e'
C. SEVERAL DAYS 03 f.
SEVERAL WEEKS ...'..... ^
SEVERAL MONTHS ...............,..'...'. 05
DONT KNOW; (DONT flEADj 94
20D. Where did you obtain information.about this product before purchasing it? too LIST. CIRCLE
ALL THAT APPLY.)
a. SALESPERSON 01
b. .' MANUFACTURER. ....'. 02
c. IMPARTIAL CONSUMER REPORT ....... 03
d. NEIGHBOR OR FRIENDS .:....'... '..... 04
e. ADVERTISEMENT.,... ; •_._ 05
t, SHOPPING AFOJND ;_ 05
g. DONT KNOW |00WT READ) 94
a.
y°U made this P"3567 (READ LIST' CIRCLE ALL
MANUFACTURER REBATES . , /. ......; 01
b. LOW COST FINANCING FROM DEALER. 02
c. UTILITY REBATE OR FINANCING PROGRAM . ^...... 03
d. THE PRODUCT WAS ON SALE . .04
e. NO.SPECIAL INCENTIVES WERE OFFERED ........ 05
f. DONT KNOW (DON7READ) '.-'.:.-'......'.'.94
a. LOWEST LIFETIME COST 01 d.
b. LOWEST PURCHASE COST ........... 02 e.'
c. LOWEST OPERATING COST '.....'.! 03 f.'
FAMILIAR BRAND ...
, • i * • * •
BEST WARRANTY .."..'.
DONT KNOW fl'?ON7 READ)
04
05
94
Thank you very much for your cooperation.
Your answers will be most helpful in this study.
-------
-------
APPENDIX G
RISK COMMUNICATION MATERIALS
This appendix contains some of the risk
developed as part of this study.
corrmunication materials that were
RTI G-1
-------
13
C
LJ
(/)
Z
>•
I]
H
•5 s s s ,' sz_^zr~_
S . = •»• =
•"
^ Jllill
.£ i. •
'•j
-------
. w 11 tvtf-\ ^
t M L. D
• ,
'
'i
i
JJ
^
EZ
2
^
^
2.
u.
u.
1
~ V
"•" " " ~? o i
-• ""* "~ > ir -
• -= 5 •' £ — S =
~ — c •: 5-
s JE i "5 i ?
I* ?. 5 K I? •-
iUlJi
= u i< — ~ i!
'=i^ *
S2JiJsS
> £ - 5 Z S
IHtlt
:S- = r g'= 1
™* *™ — X <_T
™ ^ ^5 «* v ,ri
2 c. j ? *5 -5
~ ^ r _ ^
r - <••• - ^ ^
2 - r •*" 's x ~
f ~
— ; — —
IL iHJ
Z "*~ - /" -^ ' £• V
— . ^ . — •» "•
~* — — ~ n ? f
*f "* -i ~* t i ~
= z •' .5 "^ — '-
** £ 1 "• ~ ** *~ -
£ £ -" ~ - ^ _£ -
- ^ -^ -: ' — X — 5-
- . ^ -, •/:•-:•_
-z*. , ,
J~
w
.
s
z
^
2
•!•
C
i
i
i
j; , - ^
t; „; £-= -r
^1- • p" ** "^
i" C ~ 1C ~
= r -^ £ =
** « •• ~ r: ji
^- -' X ^ — 5
s ~ " t> ^ *-
Ti^ E-^ =
•• » ~ S > ta
5 ;• = Z .- >-
= "='^ 0 ^
H^ i ^-^^ -f =
S , a = r 5 2 •=
i "' s •
•— ,_•_ ••
» • ^ W *• •
w c. x 5 <
. - ^" "2 >. (
r P.' S i/ ]
2 * - - = •
tin •
2=2 S'... j
iS4 ^ • .i
s§2^ :
S u 3 5 u
i >r >- - f
$**: C
EZ «= J
1x§S - ' :
T — ^ S ~ <•
*<£-=r L
- a: tf = Ji H
0
/^
V7
•s
y.
•^ •" . r
C , 'x
*£ -. '' ^ ^
"= S £ c-| | 1
? •• 's »« **- C
Z i -'~ c 1 5
~ ^ ' -^ "•" ^ ^* • ™
J, ! ,
i • •
|
i
w •
•— | ^T"'.1'^*>w ~
* 1 r;:QHjii
*J J | JfwNf^^j
.c ' ' L>?7^ li^r^
! ^~rj |t^r'1
.-f i ?Q=I
asj . I J^
§ 1 kQliF^
5!
1 . -*
1 3
m jj S
N C J5
2 JS 'ii
5 6 2 E
- £ £ >
< - ' "3 uJ 2
- W '*•(/• j!
2 2 "c-j^f
^ — — ' ;
i d Sj(J
r--s. «•£&
u . -' li-sS
^ Z -r c a
5 S g^<
? 5 = -S I
5 2 Ell
M -J II 1
H
. — •
^^
«• c^ •
.^j ^ ^_
i f^ *•
lr •». — •*"«
C C "^ W
-* "^ § 5
| i= I
^™ H -^ ^
•^ c c ?
< ^ r ^
f c
5 3
S E-8
"* w £
C~^fe
i.- 'i"n"»iiiiii^CWi
C§ ^- t-i j| |.Sj
*ti ^^-^*' ff ?
U "• ||
P.: 4 J
f ' •(_ Ir;' \|
|jf j
a
•% ' .
*. ' \
•. ' ' 5
* • ' ' s •
a • • • .§ ' - -
•a • • S ' ' • -•".'."
* ' • • Mi f • f . ^ '%
8 5 ii - "n ' '' . -^^T
iii Aii (•^r:.
|j|| K |||I|i .^Ir^-'
If III .1 PllJr
•' ,. ii's
' 1 f 1 1 1
E J S -j -i
2 r <= -5 s
§!=¥l|
"= J ir •£ i
•* M ta* V
5 S "5 > 7
^* •*• " v ^
^1 c .« 3 c
-* " " ? ^ ^"
5 ',|"i"|j:J=
- ~ — !- "
~ 5 r £. ~ "
• ,~ , ~ :» 'J i
^ ' •} ™ ^'
ji ~ = "
— -3 „ ,. ,
11 f!
? C C *f
— • " ."^ '•
C ^ h. -i .;
ta u ;;. j;
.. >i™ £j s
-"7 t "•=
.2 5 x *,i 5
5:^ « J! c .
•j s -~i :; s
liHl-,
ETECTOR?
>ON D
*
LB
UJ
^
••
z
^
>
^
^ IMF'-
lilllli
i^Ilfll'
1 v si t • f o '
5 £ C •£ ** eg tr
•^ y, _. ™ -^ b> ".
- ^ < C "5 — V
5 £ — . £• c ** -5
rl^^ £ 2c " .
-* 2 - ^ c- — *t
.2 ^™ -^T M *^ *•" ^™"
E;ili=l!5
U
1
^
t
•jj
W
A
X
1
\
— >* ** « *• •"•
t? •— ^ — f^ -^
3 y — i ^ ^ *1
'E^z 2«|?
^ Jj 3 O ^ y 5
s ~ ^^ "?" Sf *•
"5 2- «^ *^ "i .« 3
^ '£ C r~c.v >•
y 3 "" ?> - " <"
-. ^ "E. v y •£ 3
3 2 a 5 g -*J §
' sl^'l-sla'
K5^ = =J Sr=
s 2^-55-a i
,' . - '
RTI G-3
-------
CO
QC
111
o
a,
z
o
Q
<
QC
T3
O
3 O
o-o
o
.5
i
CO
^ *n?
•^ -f.
£• 2
' . I
- o
H **
53 u
2 2
"u .2
>- i_
b
oc
I
^
o
o
T>
ro to
' 4M ^^^
QC Q
jfl (D
•U CO
(5 CM
re
-
0 00
CO
o
(0
d
I
•o cp
n ,'N
f «
| 6
f O
O CO
£ -
i
3
RTI G-4
-------
0)
-Q
•o
CO
cJ
tmm
3
o
d)
E
o
N
(D
0)
fi
"co
O
IT)
S3 «
N
V)
rJ\
O
'J-.
O
U
r;;
- U
T3
r"
2
to
o
v<
a)
L; *-'
§ Pi
S.-8-
DD I.Q
o
TD
05
w
o
•&
•o
(0
I
o
o
•IB
I
d,
1
o
ro
-------
RADIO PUBLIC SERVICE ANNOUNCEMENT SCRIPTS
Responsible homeowner (:3n) ' ' .
(no music, man's voice) .'-•._
1 , w
You're a pretty responsible homeowner, right?. You try,to protect
your investment and your family. Or do you? Have you tested your
house for radon? Radon is a radioactive gas that seeps into houses'
from the ground. High levels can cause lung cancer.. But there are
things you can -do. A simple test .can tell you if you have a
problem, if you do, it can be fixed. To- find out more, call the
I J ' : ' - ' '
State of Maryland radon hotline. 1-800-872-3666. That's
1-800-872-3666. ' : ... '.",•.-.'•..,
Responsible homeowner (:20) .... .. :.....
(no music, man's voice) '
You're a pretty responsible homeowner, right? .You try to protect
your investment and your family. Or do you? Have you tested your
house for radon? Radon is, a radioactive gas that seeps into houses'
from the ground. -High levels can cause lung cancer. ' To find out
more, call the State of Maryland radon hotline. 1^800-872-3665.
*:. . • .
Responsible homeowner (:in) ,
(no music,, man's voice) •
Have you tested your house for radon? "Radon-is a radioactive gas
that seeps into houses "from the ground. High- levels can cause'-'lung
cancer. Call the Maryland radon .hotline,, 1-800-872-3666..
RTI G-6
-------
radioactive
If someone told you (:3Q)
(music, woman's voice)' •• : •
What would you do if someone told you there
•home could, contain high levels of a
First you'd want some answers. You'd want
where it comes from, how you'could test your
what you could do about it if you did find :
these and.other radon questions, call the
at 1-800-872-3666. That's 1-800-872-3666.
If someone told you (:2Q)
(no music', woman's\voice)
was a cliance your
gas called radon?"
o know what radon' is,
home for radon, and
:it. For answers to
Maryland radon hotline
'What would you do if someone told you there
home-could contain high'.levels of a radioact
You'd.want to know.what radon is, where it
could test, your home for radon, and what you
you, die find it. Call,the Maryland radon
1-800-872-3666. . . .
If'someone told you (:"io) • "
(no -music ,'woman's .voice)
Some homes in Western Maryland contain high
radioactive gas called radon. For more info
radon hotline'today. 1-800-872-3666.
was a chance your •
ive, gas .called radon?
comes from, how you
could do about it if
hotline at ' ' •
levels of a
rmation,'call the
RTI G-7
-------
Heard of it (;30) . . _ . -. _
(•no- music, man's voice)
Radon, ever heard of it? A lot of people haven't, even though
it's been found in many homes in Western-Maryland.. Radon is a
dangerous radioactive gas that can cause lung cancer, it seeps
into homes from the ground. The Maryland Department of
Environment urges you to test your home for radon this winter.
To find out how, call the Maryland radon hotline. The itoll-free
number is 1-800-872-3666. That's 1-800-872-3666.
RTI G-8
-------
Neighbors/easy (;3Q) • •
Announcer: A public service message from the State of
Maryland. '"•-.'.
.(outdoor noises, mailbox)
Steve
Hey Bill, I just got .my radon test
Looks like we;'re ok. Have you tested
Bill:
Nah, with all those new health
up with, it's'not-worth the'effort
hazards ..they keep -coming
Steve
Sorry friend, that's where you're
problem that you can do something
test kit through.the mail,'put.it
couple of,days, then send it back
analysis. In a few weeks, they send
Bill: Sounds pretty easy.
Steve: it is. And now we kn.ow our home is safe.
Announcer:
For more information, call the
hotline at 1-800-872-3666. That
results in', the mail.
your home yet? . •
wrong. This is one
about. YOU order a
In your .house for a
to. the company for '
. N
you your results.
Maryland radon
's 1-800-87^2-3666
RTI G-9
-------
^Ml_0 I
Neighbors/cancer (;30)
Announcer: A public service message from the state/of
Maryland. . ' ' , , ' , ' ,
(outdoor noises, mailbox)
Steve: Hey Bill, I just got my radon test results in the mail.
Looks like we're ok. Have you tested your home yet?
Bill: Nah, with all those new health, hazards 'they, keep coming
up with, it's not.worth the effort.
Steve:
Sorry friend, that's where you're wrong. Did you know
.that' radon causes lung cancer? And it's been found in"
homes in Western Maryland? Listen, even if you're not
worried about yourself,, at least do -a test for your '
kids' sake,
I
Bill: You're really serious about this, aren't you?
' • ' : . • •
Steve: You bet I am. And now we know, our home is safe.
Announcer: TO find out how you can test your home, call the
Maryland radon hotline at 1-800-872-3666. .That's
1-800-872-3666. '
RTI G-10
------- |