jj i.;,- INTERVIEW .QUESTIONS       . /         ,.  '••


  1. Was  there  anything"positive or negative which struck you in  the-
    information  you:, were asked to read about the waste site?
 2. Specifically, anything positive  or  negative  about  the  "
    notice?        !••  •                         ,-...-
                                                           water .advisory
  You may  want  to go back and /glance at it
                                            again.
3. Specifically, anything  positive  or  negative iabout ,the
   s c en e r i o ?
                                                          management
  Ypu  may want to. go. back and glance at it
                                            again,
              «n
            Agency,  (EPA)  provide the public about such
                                                         sites.?
Was there more i

you would preferred
                information
      ,  or a different kiod of information, that
in reading about the hypothetical waste site?

-------
9. Do you belong to any community groups such as Che P.T.A1. , Kivanis,
   Elks, or others?    YES  .       N0.^        .                     .

     If YES:                    '  ;    '   '    ."           -
     Which groups are you a member of?          	"'•  •   . •'	.
     Do you attend meetings regularly?	

     Do you hold any offices in these groups?^
10. Which of the following most accurately describes your  frequency of
    attendance at religious services?
     I DON'T ATTEND AT ALL.,	1
     I ATTEND SEVERAL TIMES A YEAR.....2
     I ATTEND AT LEAST ONCE A MONTH....3            •
     I ATTEND EVERY WEEK.......	.....4      :       '        ..   = .'
     I ATTEND MORE THAN ONCE A WEEK....5

11. What was your approximate family  income  frora  all sources^ before
    taxes, in 1987?                             .
     LESS THAN $9,999... .'...1
     $10,000 TO $19,999	2                             ;
     $20,000 TO $29,999.....3
    • $30,000 TO $39,999	4
     $40,000 TO $49,999.....5
     $50,000 TO $59,999	6
     $60,000 OR MORE	.7                          -

12. As far as you know, are there  any toxic  waste problems  in your
    immediate area?  YES 	 NO 	                     .>
     If YES, what is the nature of these  problems?
   Have you been personally  involved  in  this  issue?   YES  ;   NO	
      How? 	 .            	,	              .

13. Are there any controversial  public policy iasues, not  related to
    toxic wastes, in your  community?  YES	    NO	
    If YES, what is the nature  of  this  issue?
         you been personally  involved  in  this  issue?   YES^	   NO.

    If TES, how?	    '

14. Ar« you aware of  any  toxic waste site problems  that have been
    discussed in newspapers,  magazines, and T.V.  shows? (Love Canal,
    for example).  Specifies     	         , '• \    •  •
    Any  in your State?

-------
 something about Che;
                            PERSONAL INFORMATION
 Finally, we need some  backround information  on'you so Chat we know
kinds of people  who participated in this  research.
 1.  Your present age:
       years,
 2.  Sex:   MALE	1         *'    .  .-     ' -     I   .
      -, - "FEMALE	J2              '.     "   '' .;   vi'.    '  /   ,    .'•-..

,3.  Are you presently married? :                  [
       YES......1   ••'    '   ••  '   .   '      "...':    '•  '   ./    '-. .  '
   ;    NO	2    •               '   .          ;''.,'..."
                   . -'l •            '       , , '    •     J    _     ,     „
4.  Do you have children?
      YES	..1 -   '.   ' .^    '--''.    '       ";•.:•.   .'..'•-
      NO.......2 '' i  •    '     -.•'•'....].    .     "•   ...'••
      (If  "YES," how many?                       j-

5.  Do you own your own home?                    :
      YES....1    -  >,.  '    •            .-.'•'•'     '•  "   ' •'- •     •;
      NO.....2     ••['';.•"        '.           ;    :'  '  • " '  "    '   .'.- '
         :        •  • li    •"••" '•• "•        '  •      "•    .-I' ; -• • • '' -   ' -    • '..•- :'
6.  Are you presently:                          •!"•'''
      EMPLOYED (FULL TIME)....!                  J
      EMPLOYED (PART TIME)	2
      UNEMPLOYED	......	3
      RETIRED...	'...........4                  ;
      FULL-TIME  HOME>1AKER......5              '  .  '.;      ".'-,.

7.  If employed: Briefly describe the kind  of work you do.
           . •.        i.r   '        .   •',''"''"     ;"    '  •' •             '

8.  What is the  highest level of education  that :you have completed?
    (Circle .Number) !'                            «
      NO FORMAL  EDUCATION	.....	......... j.. .1
      SOME  GRADE SCHOOL	....2
      COMPLETED  GRADE  SCHOOL..'.	 3
      SOME  HIGH  SCHOOL	,.	....4
      COMPLETED  HIGH SCHOOL........;..	'......;...5
      TECHNICAL  OR  afHER POST^-HIGH  SCHOOL
           EDUCATION  (NOT COLLEGE).		.....6
      SOW  COLLEGE...........		...7
. •  •   • .    HAJOR?;	I'      .. '..-.     .     '      .-•   ;-  ..     ;     ;
      COMPLETED  COLLEGE	 . ."g
           MAJOR?    i'.   ' •  ;..' . •          	•'  '-' •     •  ' '  '     •'
      SOM1E  GRADUATE WORK		."9         .
       •'    MAJOR? .-I'-',                     '           ,   .. •  ..
      A GRADUATE DEGREE	 10
       '                    "              ''    '         '

-------
 6. From Che standpoint of environmental  pollution,  is  the United States
    Coda7 nore or less.safe than it was 30 years ago?
      MUCH SAFER	1
      SOMEWHAT SAFER	2                                           :
      ABOUT THE SAME	3
      SOMEWHAT LESS SAFE...4
      MUCH LESS SAFE	5

 7. Are federal government agencies, such as EPA,  more or less honest
    today than similar government agencies 30 years ago?
      MUCH MORE HONEST	1
      SOMEWHAT MORE HONEST	2
      ABOUT THE SAME..".	3
      SOMEWHAT LESS HONEST	4             .                         '  •  '
      MUCH LESS HONEST........5
                                     i                   '       ^
 8.  Are state government  agencies,  such as the. Department of
    Environmental  Resources (DER),  more or less honest today than
    similar government agencies 30  years ago?
      MUCH  MORE HONEST	1
      SOMEWHAT  MORE  HONEST....2
      ABOUT THE SAME	;...'•. 3'
      SOMEWHAT .LESS  HONEST....4
      MUCH  LESS HONEST........5

9.  In your estimation, do scientists know enough about  the impact  of
    chemicals on our health to adequately  protect us?
      YES,  THEY DEFINITELY KNOW ENOUGH	1
      I HAVE  SOME  DOUBT THAT THEY KNOW  ENOUGH.....2
      I HAVE  SERIOUS DOUBTS THAT THEY KNOW ENOUGH.!
      NO, THEY DEFINITELY  DO NOT KNOW ENOUGH	4             ,
                                                       /

10. Do you think  that  industries which use toxic chemicals  are  seriously
    interested in protecting  public health?
     YES,  IT IS ONE OF THEIR  PRIMARY CONCERNS		   1
     THEY HAVE SOME INTEREST  IN IT.....	.....	.........2
     THEY HAVE LITTLE INTEREST  IN IT	 i......         "3
     THEY HAVE NO INTEREST IN  IT AT ALL....	............!!! .'J4

11. Do you think that the government agencies  (federal/state) which  are
    supposed to regulate  the  chemical  industry  are seriously interested
    in protecting public health?
     YM. IT IS THEIR MAJOR CONCERN...	.........	,	.....1
     T1ET HAVE SOME INTEREST  IN IT	'......	 " * 2
     TH1T HAVE LITTLE INTEREST  IN IT		...!.'!	3
     THEY HAVE NO INTEREST IN IT AT ALL.	 ]'.] ""4

-------
11, Thinking back lover  Chat  list of activities, you  just
                                                            w«t
           PTTHTTP «««e,c   SUPPORT ** IN EVERYTHING,  EVEN MARCHING IN
           PUBLIC PROTESTS  OR BRINGING,A LAW SUIT.......'..	..i


                                                        ,.Y. .,.,..|? .'..;,. .2  •

        THEY PROBABLY WOULDN'T CARE WHETHER I GOT INVOLVED  QR  NOT...3  \

        THEY PROBABL'ir WOULDN'T SUPPORT ME AT ALLL	... '..  ... '.     4  ''


                            select  that answer which most  closely  reflects


 1. In.your estimation, how much  can  you trust local officials to
    you the truth in situations like  this?
      YOU CANNOT TRUST THEM AT ALL.	1      V
      YOU CANNOT TRUST THEM VERY MUCH....2
      YOU CAN  TRUST THEM SOMEWHAT........3     ' :'
    ;  YOU CAN  TRUST'THEM A LOT.',.,..., i... 4   ;   :           .
,:-     •           :   -if- •  •  "•''•.'...-..,.     . '• ..   .jr,'   •  .    •     • ,  .  .•_.;;'

 2' De Irtmen8timati^' H°W ****•'<**'7** trust  State officials,  like the

      YOU CANNOT TRUST  THEM AT ALL.......1      '
',"     YOU CANNOT TRUST  THEM, VERY MUCH....2      ,:
      YOU CAN TRUST THEM SOMEWHAT,	...3
      YOU CAN TRUST THEM A  LOT...........4      |

3. In your estimatibn,  how much can you trust Federal officials,  like


      YOU CANNOT TRUST  THEM AT ALL	   1      I
      YOU CANNOT TRUST  THEM VERY MUCH... '.2
      YOU CAN TRUST THEM SOMEWHAT........3 "   .!  '
      YOU CAN TRUST :rHEM A LOT....	..4      i
             .'.;  . il.-'-   ....'    . ••  •  .  7       * f  •  • • .     ,       - • •

4' ?h^r.Vthaf r*'; ^ " reflly  P°8iible  to  control a  problem  like
      l^WQ 'ftWTWTf B*T V           •   '        '      i    ^w™ l»W miC, COuuDUnXCV •
      • •w 9 •*&* " ••••••••••*      '1:-             '
.      FiOBABLY,  AT LEAST I THINK So!!!.*2         i
      P10BABLY,  NOT....................3         1
      no, DEFINITELY]NOT..	......'.u         ';'.'
                     i          •'   ','-."      f    '"•$•'">    '•',••.-  „.'•.",   '•'•,.'-'
5. In situations like  this do you think  the Federal  Agencies use the
   Ivailabl!CClVe ClM^p »«thod available, the :ch«apest  cleanup method
   available,  or something that is in between?
      THE MOST  EFFECTIVE METHOD, REGARDLESS OF COST....         l
      AN  EFFECTIVE METHOD WHICH MAY HAVE QUESTIONABLE ELEMENTS'
         BUT PROVIDES; PROTECTION AT MODERATE COST......         2
     THE CHEAPEST METHOD,  WHETHER EFFECTIVE Q^ NOT.	........ .'..' 3


1 f   -'  '   '   .     '   \"\ ..-..,-,     ...-.".     '    -     ;t',-. ;.., .  '.- ; '•...,'-  .  '•-''.:*'.  ..-',-' ,':
                                                 8 :.
                                                 r •

-------
10. Here is a. list of  things  that  people can do in situations like -this.
   For each activity,  select  that  option that reflects how -likely it is
   that you would engage  in  that activity.          ,     '         .

     OPTIONS                                .                            >
        I DEFINITELY WOULD DO THIS...	1    '            ,    ,
        I PROBABLY WOULD  DO THIS. .. 1.-.-,. .2'.
        I PROBABLY WOULDN'T DO THIS......3,      ,  - '   .
        I DEFINITELY WOULD NOT DO  THIS...4                    ,

   LIST          "           "-.•••.- •;•    .•""    .   •  ;:     •     '  ,

     TRY TO GET MORE INFORMATION ABOUT,THE  NATURE OF  THE PROBLEM  '

   '  TALK ABOUT THE PROBLEM WITH FAMILY AND "FRIENDS _____

     TRY-TO CONVINCE FRIENDS AND RELATIVES  TO USE BOTTLED WATER

     SPEAK DP AT A PUBLIC MEETING  WHICH INCLUDES EPA  OFFICIALS .    /

     MARCH IN A PROTEST PICKET LINE,  EVEN IF IT MEANT BEING ARRESTED

     TRY TO CONVINCE FRIENDS AND RELATIVES  TO MOVE OUT OF THE AREA    '

     SERVE ON A COMMITTEE OF CONCERNED  CITIZENS _	

     GO DOOR TO DOOR .TO CONVINCE YOUR NEIGHBORS TO GET MORE INVOLVED
           IN THE PROBLEM _____                              '
                                         '        \ •   '.    -        "•
     WRITE OR'CALL MY CONGRESSMAN  TO  GET EPA TO REMOVE ALL THE  WASTE
           IMMEDIATELY __/                    -

     ORGANIZE A MEETING TO PROTEST THE  WAY  THE  PROBLEM IS BEING HANDLED

 . '  TESTIFY AS A CONCERNED CITIZEN AT  A LEGAL  HEARING   '       ;     ~

     BUY BOTTLED WATER    -

     HAVE A WATER-TREATMENT SYSTEM INSTALLED AT MY OWN EXPENSE  _____

     MOVE OUT OF THE AREA AS SOON  AS  POSSIBLE

     SUE         -..-",    '             .'•'-"-.'    ' •'"       •  '

-------
 9. Now here  is  th&t  same list of activities that may be  a  threat  to
 health and safety,1   Again',  .tell me how much oi: a threat  you think
 drinking.and  bathing in the local water is conpared to each activity.
 HOWEVER. THIS TIME,  MAKE YOUR JUDGMENT ON THE BASIS OF HOW HUGH A THREV
 YOU THINK THE. WATER  WAS AFTER THE EPA CLEANUPj'	
 IS DRINKING THE LOJCAL
 WATER MORE OR LESS •OF
 A THREAT THAN-SMOKING
 TWO PACKS A DAY FOR
 ONE. YEAR?         ;;
   (After Cleanup) i
   ^^«_^AHV i f
                   I
 IS DRINKING THE LO^AL
 WATER MORE OR LESS] OF
 A THREAT THAN LIVING
 NEAR A NUCLEAR POWER
 PLANT FOR ONE YEAR?
   (After Cleanup) j:

 IS DRINKING' THE LOCAL
 WATER MORE OR LESS I • OF
 A THREAT THAN DRIVING
 A CAR FOR TWO HOURS A
 DAY FOR ONE YEAR?
   (After Cleanup) "!''..
'IS  DRINKING THE LOCAL
 WATER MORE OR LESSJ;OF
 A THREAT THAN RIDING A
 BICYCLE  EVERY DAY FOR
 ONE YEAR?  ,       [
  (After Cleanup)
                   I.
                   11
 IS  DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN HAVING
 ONE CHEST X-RAY IN !bNE
 YEAR?  .         i   |; \
  (Aftar Cleanup) j;
         '. ,•    "'  '.' 1L
 IS  DRINKING THE LCKIAL
 WATER M01E OR LESS OF
 A THREAT THAN EATING
 FRUIT  EVERY DAY FOR: ONE
 YEAR WHICH WAS SPRAYED
 WITH PESTICIDES WHILE
 GROWING?           j
  (After  Cleanup) i
                          MUCH .' ..,   SOMEWHAT    ;     SOMEWHAT
                          MORE OF   MORE OF   TOE    LESS OF
                          A THREAT   A "THREAT  SAME   A THREAT'
MUCH  •
LESS .OF
A THREAT

-------
    Below is a list of activities which tnay also  involve a-threat, to.
    a person's" health-and safety.  Tell me how much of a threat you think"
    drinking and bathing in the  local water for one year is compared•to
    each activity.  MAKE YOUR JUDGMENT ON THE BASIS OF HOW ;MUCH OF A
    THREAT YOU THINK THE WATER WAS BEFORE THE EPA CLEANUP.  For example,
    if you think that drinking and bathing in the local water for one
    year is about the same degree of threat as riding a bicycle every day
    for one year, you would circle number, 3 (THE SAME).    :
                         MUCH
                         MORE OF
                         A THREAT
SOMEWHAT
MORE OF   THE
A THREAT  SAME
SOMEWHAT  MUCH
LESS OF   LESS OF
A THREAT  A THREAT
 IS DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN SMOKING
 TWO PACKS A DAY FOR
 ONE YEAR?
   (Before Cleanup)

 IS DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN LIVING
 NEAR A NUCLEAR POWER
 PLANT FOR ONE YEAR?
   (Before Cleanup)

 IS DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN DRIVING
 A CAR FOR TWO HOURS A
 DAY FOR ONE YEAR?
   (Before Cleanup)

 IS DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN RIDING A
 BICYCLE EVERY DAY  FOR
 ONE YEAR?
   (Before Cleanup)

'IS DRINKING THE LOCAL
 WATER MORE OR LESS OF
 A THREAT THAN HAVING
 ONE CHEST X-RAY IN ONE
 YEAR? (B«fore Cleanup)

 IS DRINKING THE LOCAL
 WATER MORE OR LESS OF   :
 A THREAT THAN EATING
 FRUIT EVERY DAY FOR ONE
 YEAR WHICH WAS SPRAYED
 WITH PESTICIDES WHILE
 GROWING? (Before Cleanup)

-------
                        •  I - POST-QUESTIONNAIRE i  '••'".
                     i   ' .-,       .      -          -,k ., -    .....  ..  .  ...

 1.  How concerned would you be about  living within one mile "of this site
    prior to the cleanup attempt?                •    -
     •EXTREMELY CONCERNED..	. .	....
      SOMEWHAT CONCERNED!.	................21
      NOT TOO CONCERNED.......	3'
      NOT CONCERNED AT ALL		.......	4 'i-              '   '  ' " •  '

 2.  How concerned" wou^d, you be about  living within  one mile  of  this  site
    after EPA had completed the proposed cleanup  plan?
      E3CTREMELY CONCERNED..	..."	........ 1                -'.'..
      SOMEWHAT CONGESTED.	.".. 2 !
      NOT TOO CONCERNED.....	'.............'....3;
      NOT CONCERNED AT ALL.....................4 ;   .   .  .  ,

 3.  How probable do you think it is  that you would  Buffer  health  problems
    as  a result of living near this  site prior to the cleanup attempt?
      EXTREMELY PROBABLE.*	.1                j        ,
      SOMEWHAT PROBABLE.........2                !              ^
      NOT TOO PROBABLE....	3                '
   .   NOT PROBABLE At ALL...... .4        '
                   • " t: '     '  :'•''.
 4.  How probable do you think it is  that you would  suffer  health  problems
    as  a result of living near this site after EPA  had completed  the
    proposed  cleanup?           .,        s        i
      EXTREMELY PROBABLE........1       '         f.
      SOMEWHAT PROBABLE...	2       "        ;
      NOT TOO PROBABLE.:	....3                T
     NOT PROBABLE AT;,ALL.......4           '     j;

 5.  How probable do you think it is that your family members would suffer
    health problems as  a result of living near this site prior to  the
    cleanup attempt? j
      EXTREMELY PROBABLE	.1                :"''-.        -
      SOMEWHAT PROBABLE....	2                I
     NOT TOO PROBABLE	....3                ;                   :
     NOT PROBABLE A* ALL	.4              .  !.                      '
     ..          'p      • '     ••''..". ' \ •.,   '   f'.. ' '       . '   ' •
 6.  How probable do you think it is that your family members would suffer
    health problems as  a result of living near this site after EPA had
    completed the proposed cleanup?              :
      EXTREMELY PROBABLE..... 1                   1
      SO«WHAT PROBABLE......2                   1
     HOT TOO PROBABLE.......3                 '  ['
     NOT PROBABLE AT: ALL....4    .  -  •- '        J          .
                                                 F
 7.  In  cases  like this,  do you think that the  company responsible.for
    the contaminationiproblem knew that their  waste handling policies
    would eventually threaten public health?     I
      I THINK THEY KNEW FULL WELL WHAT WOULD HAPPEN............. 1
      I THINK THEY HAD  SOME IDEA THAT IT MIGHT CAUSE PROBLEMS... 2
•'•'    I THINK THEY WERE NOT SURE WHAT MIGHT HAPPEN.	"--..3,
      I THINK THEY REALLY DID  NOT KNOW WHAT WOULD HAPPEN..	4

-------
 6.-Do you chink that you have suffered health problems due' to exposure
 to  hazardous -chemicals in the water, soil, or air?               ,
    YES,  I DEFINITELY SUFFERED HEALTH PROBLEMS FROM CHEMICALS........... 1
    I  SUSPECT I  HAVE SUFFERED HEALTH PROBLEMS BUT I'M .NOT '.CERTAIN, .... J.. 2
    I  DOUBT THAT I'VE SUFFERED HEALTH PROBLEMS FROM CHEMICALS	3
    NO,  I  DEFINITELY HAVE NOT SUFFERED HEALTH PROBLEMS FROM- CHEMICALS,..4

 (If you answered "1" or "2" to the above question please answer question
 number  7.   If you  answered  "3" or "4" go on to question number 8.)

 7.  If you think you have suffered health problems due to hazardous
 chemicals  what  is/was  the nature of those problems?
   What were  the  chemicals  and where did they come from?
8'. Have you ever heard of  the  chemical  benzene?    YES...1   NO...2
       (If "yes" continue,  if  "no"  go  to  question #9)

   8a. Do you know where benzene  comes  from or what it  is used for?


   8b. Is benzene a health  hazard?  YES...1   NO...2  If "yes," what
       kind?_	       '   '  	             '.  -

9. Have you ever heard of the  chemical  trichloroethylene OP TCE?
       YES...1    NO... 2                             '  '  • ";  ^
      (If "yes" continue, if "no" go to question  flO)

   9a. Do you know where Crichlorethylene  comes from or  what is is used
      • for?
   9b. Is trichloroethylene a health hazard?   YES...1   NO...2   (If
       "yes-," what kind of hazard?	
10. Where .do you get your information on the  issue  of  toxic  chemicals?
     (Circle as many numbers as are appropriate and respond  to  the
      questions to the beat of your ability.)

    . 1. I DON'T READ OR HEAR MUCH ABOUT THE SUBJECT

     2. TELEVISION?  Do you remember what program(s) your  saw on  this
          subject?                .

                           . .   '   •   ',      . . '        .     ,\  ..
     3. MAGAZINES?  Specific magazines? .-'   • '.    ''  .  '    -',-:'

     4. NEWSPAPERS? Specifics?                            ';'.''.

     5. RADIO?  Specifics?_	             ' '  ,  -. -      - :'

     6. OTHER SOURCES?

-------
                                                t1 •-. •'
                                                t:..'
 3.  Here are Chose .same health  problems  listed above.
                                     Out of each !00
    new c,ases- diagnosed each year,  how  many  would you estimate are,
    caused by toxic [.chemicals  from', waste  sites?,  For example,, if you
    think that about 5 out of  every 100 new  cancers  diagnosed in 1987
    were caused by coxic-chemicals  from waste  sites  that would be your
    best guess.  I'fj.you think  it was  60 out  of one hundred th,en 60 would
    be your best guess:.  If you think that toxiii  chemicals from waste
    sites do not cause any cancers  you  would pick 0  as  your answer.   I
    realize that you may not be an  expert  in this area,  but give me  your
    best estimate.  iREMEMBER: OUT OF  EVERY 100 NEW GASES OF EACH HEALTH
    PROBLEM WHICH -WERE DIAGNOSED LAST YEAR, HOW MANY DO  YOU THINK WERE
    CAUSED BY.TOXICiCHEMICALS FROM WASTE SITES?I
     .ADULT CANCERS
      CHILDHOOD CANCERS
      (How many new cases  out:  of  100 were  caused  by
      'toxic chemicals  from waste  sites?)

     __ (How many, new  cases out  of 100 were  caused
          toxic chemicals  from waste sites?)
                                                                        by
      LIVER,  KIDNEYv  & BLADDER PROBLEMS
                         (How many new cases  out  of
                          100 were caused by  toxic
                          chemicals  from waste  sites?)
     LUNG  PROBLEMS  (NOT CANCER)
                  (How many new cases'out  of  100 were
                   caused by ;toxic chemicals  from
                   wastes sites?)   ,
     BIRTH DEFECTS
     MISCARRIAGES
     LEUKEMIA
     SKIN PROBLEMS
i;	(How many new cases out oif 100 were caused by
;       toxic chemicals from waste sites?)

1	 (How many new cases out of 100 were caused by
v     toxic chemicals from waste sites?

|(How many new cases out of 100 were caused by toxic
1 chemicals from waste sites?)
••|i •• ,       ;   .      •    .   '. •    •  •  . ,      • -,•,.'"
'h	^_ (How many new cases out of 100 were caused by
i;  ,    toxic chemicals from waste sites?)
     OTHER SERIOUS CHILDHOOD  DISEASES
                         (How many new cases out of
                          100 were caused by toxic
                          chemicals from waste sites?)
4. Ar« th«re any health  problem we  haven't mentioned that you think
might b« strongly related  to  toxic  chemicals in the environment?
                                                u •.
                                                -r,
5. When it comes to!:the  issue  of  toxic  chemicals
youself; (Circle th^ apropriate number)
     WELL INFORMED.L.......... 1
     SOMEWHAT INFORhJED,	.2     '
     NOT VERY INFORl|JED..... ...3
     NOT INFORMED AT ALL	.4
                               do you consider

-------
                             '  INTRODUCTION     '            ;  •   -..•" .

 We are researchers affiliated.with Penn State University.  As you
 probably know, there is an ongoing Federal program designed  to-  cleanup
 some of the more serious hazardous chemical waste sites across  the
 United States.  H'owever, attempts to clean up these sites sometimes run
 into problems and take longer than they should.  We want to  explore,
 with you,  some issues related to communicating risks related to
 hazardous  waste sites.  We will ask you to read .some material and answer
 some questions related to1that material.  We-will also ask you  other
'questions  concerning your opinions about dealing with toxic  chemical
 wastes. This will take about one/half hour.   All we are'interested in
 is your beliefs and feelings about this important problem.   Can you give
 us approximately a half hour of your time?

 Thank you  very much.

 First,  let me make it very clear that the issues we are going to ask you
 to respond to are taken from toxic waste sites across the country and
 are not .something that is a local threat.  This research has nothing to
 do with a  real toxic waste site in your area.  Do you have any questions
 about that?

 Now,  before I ask you to read some material on our make-believe site let
 me ask  you some preliminary questions.                    ,
                         I  FIRST QUESTIONNAIRE I                   '

    How concerned would you say you are with this issue of toxic chemical
    wastes  sites? (Circle the appropriate number)
      VERY  CONCERNED		....!'
      SOMEWHAT  CONCERNED....	......2
      NOT TOO CONCERNED....... *	...3 '" •                       .
      NOT CONCERNED AT ALL..	...... .4
 2.  Here  is  a. list  of health problems that may be connected with
 exposure to chemicals.   Would you indicate how concerned you are
 that  you, or someone close to you,  may suffer this problem because of
 exposure to chemicals in the'air, soil, or water?

          Tell me if you are:
            GREATLY CONCERNED	(1)
            SOMEWHAT CONCERNED... (2)        .,
            NOT TOO CONCERNED....(3)
            NOT CONCERNED AT ALL.(4)
         (Circle the appropriate number)          ,.  >  .  '  ,
                              .•'•.'         .' .    .-     ' \ .    --•
      ADULT  CANCERS
      CHILDHOOD CANCERS
      LIVER,  KIDNEY,  & BLADDER PROBLEMS
      LUNG PROBLEMS (Not Cancer)
      BIRTH  DEFECTS
      MISCARRIAGES
      LEUKEMIA                  ,
      SKIN PROBLEMS
      OTHER  SERIOUS CHILDHOOD DISEASES  '
1
1
1
1
1
1
1 .
1
1
2
2
'2
2
2
. 2
2
2
2
3
3
3
3
3
3
3
3
3
4
4
4
4
4-
4
4
4
4

-------
 '         '   '   APPENDIX B       ;




Pre-iand Post-Manipulation Questionnaires

-------
EPA Holds  Public  Hearing;  Cleanup Commences
AC Che  public "hearing-Che  citizen'advisory commie tee-made  it  known  chat
there had  been  some  disagreement  between  them and  the  EPA  over  the  issue
of incinerating some of  the  soil-.   However,  EPA  experts  argued  that
incineration was  quite safe  and far less  expensive than  digging up  the
contaminated soil and trucking it  to  an approved hazardous: waste  site.
The-citizen advisory committee eventually agreed with  the  EPA experts.
Four months later, heavy equipment  moved  in  as EPA began to carry out
its plan.

-------
                    j         .                 -.-'.'.'
                    i_;,- .  (Augmented Citizen Participation)         '• -   .

             .  -  ' '  ,ji"'-;' '-1  , MANAGEMENT : SCENERIO j, •      ",-,_','   _'

 Eighteen months  later you learned  that  the site had been added to the
 National Prioritie,!* List of hazardous waste sices which are eligible for
 cleanup funding  under the federal  Superfund program.. Six months after
 that, the EPA announced that  it  was  starting a remedial investigation
 and feasibility  sttidy to decide  what to do about the barrels,  liquids,
 and contaminated soils .that remain behind  the 'fences of the old plant.

 Citizens Group Forms'   ".'•;'•'            ;     . ,  ..
 At the same time that the site was being added-to the National       ,
 Priorities List  a group of homeowners in the two neighborhoods adjacent
 to the site formedJA, group called  "Citizens Against Hazardous  Waste in
 our County" (CAHWOG).  They expressed their 'dismay that the waste  had
 not been removed in the, 18, months  since the initial problem was noted
 and that EPA refused to promise  that the waste would be removed shortly.
 The group also was 'upset to learn  that  EPA would explore options that
 might leave some oil the contaminants at the site and might burn
 contaminated substances on site.               f

 Citizens Group Geta Technical Aaaistant Grant \. -
 Shortly after EPA,announced that it  was beginning to study the problem,
 EPA community relations people met with leaders  of CAHWOC and  other
 citizens who live riear the site, to  get some idea of what they felt
 should be done to clean, up the site.  CAHWOC applied for and received a '
 grant of $20,000 from EPA 10 CAHWOC.couls  hire its own, experts to help
 the community understand what EPA was doing.   !
                   .! i    '       '   '• '.        .9-   ;       •   ". •   •    .

 EPA Agrees To Work With Advisory' Committee    i
 At the same time that CAHWOC received the  grant,  EPA also agreed to work
 with a citizen advisory committee made  up  of two members of CAHWOC,  two
 other local people,;1 one local official,   a  chemistry professor  from  a
 nearby college,  and an expert from the  State Department of Environmental
 Resources.   This committee was given the right ;to review all EPA plans
 involving the site.'  In addition,' EPA agreed that, a majority of the
 cjmuittee had to approve any proposed solution before it could be put
 into effect.       ;;   '      ,           ,  "    .;.'•.';.       ''...,
                    '!.'."'    -    ,     '       ",*'''      '''••.'   '. '    '"
 EPA Issues Cleanup iPlan After Reaching  Agreement  With Advisory Committee
 Sine months later,  EPA published a fact  sheet  and reported in  the
 newspaper what it planned to do  to remedy  the  situation with the
 contaminated site.  ';EPAalso announced  that  it had placed  copies of  the
.full report in the local library.  Tn the months  before the plan was
 officiAlly publi»he|d the citizen ao.vj.sory committee  had reviewed EPA
 propoMla and discussed them with local citizens.  With the advice  of
 local people,  the advisory committee insisted on  several minor changes
 in the EPA plan  as  |origirially prbposed.   The  final  plan was a  result  of
 the combined effort of the EPA experts  and  the advisory committee.
             '  ' '    !!•' •''•''  -';   ,  ,      •      :\ ."  • '•  '- .  ••  : \   ''. ' .•   '
 EPA announced that  jit intended to use. extraction  wells  to  collect some
.of the contaminated! water,  air stripping to  remove  contaminants  through
 evaporation,  and incineration of some of the most  contaminated  soil  on  ,
 the site.   The most contaminated .soil was  that directly under  two
 Lagoons  and under the piles  of barrels.   Some of  the  water—was
 contaminated with volatile organic  compounds  that  EPA felt  c'ould be
 dealt  with most  saf|ely by leaving them  at  the site-in large'air-tight
 containers.

-------
 EPA.  announced  chat it intended Co use extraction wells to collect s'cne
 of the  contaminated >water,  air stripping ta remove contaminants through
 evaporation, and  incineration of some of the most contaminated soil on
 the  site.   The moat contaminated soil was that directly under two.  •
 lagoons  and under the piles of barrels.   Some of the water was
 contaminated with volatile  organic compounds that EPA felt could be
 dealt with  most safely by leaving them at the site in large air-tight
 containers.                                >                           ',

EPA Holds Public  Hearing; Cleanup Comneneea
At the public  hearing the experts employed by .CAHWOC complained that
 they had some  trouble getting answers  to technical' questions  from EPA
and that they  agreed  with some  of the  local  residents who questioned  the
safety of incinerating some  of  the most  contaminated soil.  EPA experts
argued that incineration was  quite safe  and  far  less expensive than
digging up  the  contaminated  soil  and  trucking it  to  an approved
hazardous waste site.  Four months later, heavy  equipment moved in  as
EPA began to carry  out its .plan.

-------
                          (Indemnification•
     - - ,-• '   ---_,' -..;  ' '  -' I  'MANAGEMENT  SCSNERIO I-  '••••.•"•••-/'•"'    •

 Eighteen months later you .learned'that  the  site  Had  been added to-the
 National Priorities List of hazardous waste sices  which are eligible for
 cleanup funding under the federal,  Superfund program.   Six months after
 that,  the-. EPA announced'that 'it was starting>a;remedial investigation  -
 and .feasibility study to decide what  to do  about  the  barrels,  liquids,
 and  contaminated,soils, that remain behind the fc'ences  of the old  plant.

 Citizens Group Forna        *               'V  '
 AC  the  same time that the. site was being added !to  the National
 Priorities  List a group of homeowners in the  two neighborhoods adjacent
 to  the  site formed a group called "Citizens Ag«inst Hazardous  Waste in
 Our  County" (CAHWOCp.  They expressed their disimay that the waste had
 not  been removed in the 18 months since the  initial problem was  noted
 and  that EPA refused to promise that the waste would  be removed  shortly.
 The  group  also was 
-------
                             (•Standard)
                         I  MANAGEMENT --SCENERIO !'     , .'.         '.,

 Eighteen months  later you learned, chat  the  sice  had  been  added  to  the
 National Priorities List of hazardous waste sites  which are  eligible  for
 cleanup funding  under the federal  Superfund program.   Six months after
 that, the EPA announced that  it was starting a remedial investigation
 and feasibility  study to .decide what to do  about the barrels, liquids,
 and contaminated soils that remain behind the fences of the  old plant.

 Citizens Group Forms                      '•-.--.         ;
 AC the same time that the site was being added to  the  National
 Priorities List a group of homeowners in the two neighborhoods adjacent
 to the site formed a group called  "Citizens Against Hazardous Waste in
 our County" (CAHWOC).  They expressed their dismay that the waste had
 not been removed in the 18 months  since the  initial problem  was noted
 and that EPA refused to promise that the waste would be removed shortly.
 The group also was upset to learn  that EPA would explore .options that
 might leave some of the contaminants at the  site and might burn
 contaminated substances on site.

 Citizens Group Gets Technical Assistance Grant
 Shortly after EPA announced that it waa beginning to study the problem,
 EPA community relations people met with leaders of CAHWOC, and other
 citizens who live near the site, to get some idea of what they felt
 should be done to clean up the site.   CAHWOC applied for and received a
 grant ot $20,000 from. EPA so CAHWOC could hire ita  own experts to help
 the community understand  what  EPA was  doing.

 EPA Issues  Cleanup Plan                                             ,
 Nine months  later EPA published a fact  sheet and reported in the
 newspaper  whatsit planned  to do to remedy the situation with  the
 contaminated site.   EPA also announced  that  it had  placed copies of the
 full report  in the local  library.   During the next  month comments  on the
 plan were  invited and a  public meeting  was  held.  At the meeting EPA
 recorded public  comments  that  later resulted in minor changes in the
 cleanup  plan.

 EPA  announced^that  it  intended to use extraction wells  to  collect  some
 of  the contaminated  water, air stripping to  remove  contaminants  through
 evaporation,  and  incineration  of some of the most contaminated soil on
 the  site.  .The most  contaminated soil was  that  directly under two
 lagoon*  and  under the  piles  of barrels.   Some of  the water was
 contaminated with volatile organic  .compounds that EPA felt could be
 dealt_with'most  safely by  leaving  them  at  the site  in large'air-tight
 containers.           .'      '   ...  ~        .'>  •   •    ,  "   ,,-'...'.  ••'

 SPA Holds Public  Hearing; Cleanup Commences
At the public  hearing  the experts employed by CAHWOC  complained  that
 they had some  trouble  getting  answers to  technical  questions*,from EPA
 and  that they  agreed with some of the local  residents who  questioned  the
 safety of incinerating some of the  most  contaminated  soil.  EPA  experts
argued that incineration was quite  safe  and  far  less  expensive than
digging up the contaminated soil and trucking it  to an  approved
hazardous waste.site.  Four months  later, heavy equipment  moved  in  as
EPA began to carry out its plan.                        ,    •   " v<

-------
                         WATER ADVISORY NOTICE [
Testing has determined that the local water  supply contains unacceptably
high levels of triehloroethylene  (TCE) and benisene.  these are chemicals
commonly used in industry and as  components  of [ common household  products
such as cleaning; agents, glues, paint. strippers, and as an anti-knock
additive in gasoline.                          |  ,                   .
  •  -       •   _ •              ,          -  ,      1
AC high levels these chemicals may cause cancer, neurological
impairment, liver -and kidney damage,  and possibly birth defects  and
misca'rriages.  The \ chart below illustrates theinumber of people  who can
be expected to die per year as a  result of being exposed to your water
for one year.  Some other health  hazards are included for comparison
purposes-.   .       • •         ' _ •     '      '     \

                   ~~     ANNUAL CHANCE OF DEATH .    '       [   '.  '.      ~~
              -     : (Assuming One Year of Exposure)
                  (Deaths per 100,000 People Exposed)
      Level of
      Danger
              Smoking Two Packs of Cigarettes i  Day  (200)
       200

       175

       150

       125

       100

        75

        50

        25

         0
                    Hang Gliding (125)
Scuba Diving (75)

      Skiing (50)

            Driving
                                    I    I    Drinking and Bathing in Your
                                                Car (25)
                                                 Water (2)

                                                  I Drinking One Diet
                                                   Drink Per, Day (1)
If you feel that your water presents a danger
health then we recommend that you consider
until remedial action can be taken.
                                               :o  you  or  your  family's
                                            purchasing bottled water

-------
                       • '. "icn  jis.cj.ai:ner;       :         '                  •

                            WATER ADVISORY NOTICE  ' • "  •

  Teating has determined'.chat the local water supply  contains  unacceptablv
  high  levels of .trichloroethylene (TCE)  and  benzene,.   The3e  are  chemicals
  commonly used_in  industry, and  as components of common household products
  such  as cleaning  agents,  glues,  paint  strippers,  and  as  an  anti-knock
  additive in gasoline.

  Ac high levels these chemicals may cause  cancer, neurological
  impairment, liver and kidney damage, and  possibly birth  defects  and
  miscarriages.   The chart below  illustrates  the  number of people  who can
  be expected to .die per year as a result of  being exposed to your water
  for one year.   Some other health hazards  are included for comparison
  purposes.
        Level of
        Danger
                           ANNUAL CHANCE OF DEATH
                      (Assuming One Year of Exposure)
                    (Deaths per 100,000 People Exposed)
                Smoking Two Packs of Cigarettes a Day (200)
         200

         175

         150

         125

         100

         75

         50

         25

          0
Hang Gliding (125)
      Scuba Diving (75)
      I- •_
            Skiing (50)
            I
                  Driving a Car (25)
                        Drinking and Bathing in Your
                        1     Water (2) '••         "
                                                    Drinking One Diet
                                                    Drinlc Per Day- ( 1)
 If  you fe.l  that  your water presents a danger to you or your family's
 health then  we  recommend that you consider purchasing bottled water
 until  remedial  action can. be taken.

 NOTE:  The numbers  presented above are our best estimates.   But it is
 important to realize  that  they are only estimates.   They are based
 primarily on research with laboratory animals,  usually rodents,  who are
 fxhihi^trm?X hi«*do>" of che Chemical being tested.   If the animal
 exhibits health problems,  we then extrapolate from  these high doses to
 the very low doses usually, received  by  people who may be exposed to the
 chemical.  The risk assessor chooses  the  assumption,  or best guess  that
 appears least likely  to  underestimate the risks.  An  attempt,is  made to
overestimate rather than underestimate  risk.   But it  is an educated
guess.

-------
 The DER issued an advisory Co all residents of Che area serviced by
 wacer drawn from wells chaC had been concaminaced by chemicals seeping
 through the soil.  Your home is approximately ;one mile from the source
 of Che chemicals and is .serviced by wacer dravn from Che concaminaced
 well.3.    '    '        •    '      '     '     '    •    '
 Wichin two days ,EPA" (Environmental Protection Agency) technicians
 appeared on the siite'of the old manufacturing plant.   After collecting
 fresh soil samples,, along with DER people,  the EPA decided to  build  a
 new drainage ditch j;to divert water from the area of Che barrels  and  the
 smelly lagoon into,a new lagoon.  This new lagoon was constructed  so
 that it had a thick clay bottom alongf with a strong plastic liner.   The
 EPA said that the new lagoon would keep more chemicals: from seeping  into'
 the water supply and would control the situation until a more  permanent
 solution could be devised.  The EPA also fenced in the area where  Che
 manufacturing plant: had stood and posted signs warning people  not  to
 trespass because of dangerous,  hazardous chemicals.
                       -v£3g3ffiS!F»r ^>
                       -  ZgSSSB&feSasaS'^.^--'
The EPA also released the following water advisory to the press and sent
copies to the homes! of all homeowners serviced by the affecced wacer
supply:     •   •,'••••   . •         •     •       j-       • '"

-------
 Approximately two years ago people began noticing char the drinking
 water had an unusual smell to it.   The local newspaper carried several
 letters from concerned citizens, .some .of whom complained '• of health
 problems which they thought might  be linked to the possible
 contamination of local water,supplies. .For example,  several members  of
 one family have been suffering for about six months  from ;an unexplained
 rash;  two cases of childhood leukemia have been diagnosed  in the  same
 neighborhood during the past year,  and a growing number  of older  people
 have had to seek professional help for arthritis.

 Local  officials  contacted  the Pennsylvania Department  of Environmental
 Resources  (DER)  and requested  testing of  local  water supplies.  The
 tests  indicated  that  the chemicals  TCS (trichloroethylene)  and benzene
 were present  in  the water  at  levels  above  those approved by  the
 Environmental  Protection Agency.  The source  of this contamination was
 traced  to  the  site  of a manufacturing plant which had  gone  out of
 business and had been torn down years ago.  The old plant  site now
 consists of trees,  brush,  timber, and a few marshy areas.  There are
 several  lagoons  in  which chemicals had been disposed of by  the old
manufacturing  plant along  with « number of decaying 50 gallon drums
 scattered about which are  spilling  their contents onto the soil.  The
photograph below was taken  on part of  the  site.

-------
INTRODUCTION TO  THE  MATERIAL ON THE HYPOTHETICAL  WASTE SITE
 •  ".  . .  .'•  :< $ .   '. •-'•.:'.-'•'  '..'•  ••.";:•      .yf..-"':"'  • •'•'  '•'••.'•;'. / '
Now I would like to  Have you read some material dealing with
an area which  has the  problem of hazardous chemical  pollution.
I want you  tojimagine  that you live in th« community  in which
this  site is located.   In fact, imagine  that you  live within
one mile of the  source of the contamination.   After  you have
read  the material you  will be asked to answer  some questions
concerning your  reactions to the hazardous chemical  site.

-------
        APPENDIX A




Hypothetical Site Scenario




   vRisk Communication




    Management Options

-------
Sandman, P. M.  1986  ''Explaining Environmental (Risk-.M' U .S .EPA,  Office  o
   Toxic Substances, ^Washington,  D.C.           V     .     .   '    ;

Scott, W.J.' 1988  "Competing Paradigms in the Assessment  of Latent  .•
   Disorders:  the Case  of Agent Orange." SOCIAL PROBLEMS, 35:145-161.

Sharlin, H. I.  1987.  '^Macro-Risks,  Micro-Risks, and the Media:  the, EDB
   'Case." In  B. Johnson & V,T.  Covello (eds.) op^ cit. 183-197.  -     .

Slovic, P.  1987 "Perception of  Risk," SCIENCE 236: 280-285.        ,
                   '']'.     "  '•  ' ''   ' •  '  •• ' •     '•[• ••  .   •  •: '
Slovic,'P.", B. Fischhoff, & S.  LichCenstein 1979'"Rating  the Risks.,"
   ENVIRONMENT, "21:14-20.   '       "        *  ,    |   '•''"'
                     ). •     •• • v '     : .'   .•      -   . • t •  '     "' .      '  '
Slovic, Pi, B. Fischhoff, & S.  Lichtenstein 1980; "Facts  and Fears:,
   Understanding  Perceived Risk." In R.C.  Schwing & W.A. Abers, Jr.
   (eds..) SOCIAL  RISK: ASSESSMENT:  HOW SAFE IS' SAKE ENOUGH?  New York:
   Plenum Press.  181--214. '    ..   .    .           ;!
           -    •      ":       •   '         '.'•••  L •'''•'"-•
                     I       .               ^       i
Tarr, J.A., and C. Jsicobson 1987  "Environmental Risk in  Perspective."
   In B. Johnson  & V.t.  Cbyello (eds.) op.  cit.,  1317-344.
               ....i1     '        .•••••.-.   ^.''  ''-'-   -•'•         •'
Travis, C.C.,  S.A. Richter, E.A.C. Crouch,  R.  Wilson, E.D. Kleraa 1987
   "Cancer Risk Management." ENVIRONMENT,  SCIENCE,''AND TECHNOLOGY'21:-•
   415-420.          !' '                      . '    I '     '  '    ' -   ,'
                .  :   .l"'      '•          '. .          !••'  ' , -.-     ',   .  .. •-
                     !'.           ;                 h1 :
Warr, M. 1985  "Fear|of Victimization: Why are Women and  the Elderly
   More Affraid.?" SOCIAL SCIENCE  QUARTERLY, pp681-702.

-------
  rowlkes,  M.-R.  & P,Y.  Miller 1987 ."Chemicals and'Community at Love
     Canal. '  In  B.  Johnson & V.T.  Covello 
-------
        '.  .            •;"   REFERENCE BIBLIOGRAPHY  i   .  '  .   '•
                  ' '   'k.  ..  ..--.'  •    ''..•••.-,'..'•• ;i  •   .   ""•,•'.••,
Armo.ur, A.  1.987 "Repo!rt of .the Siting Process  Tasjc Force on Low-Level
    Radioactive Waste Disposal:  Opting for  Coopera.tion." Energy,
    Mines  and Resources Canada.  •  •-      .     .    '
Aronson,  E. ,  J.  Turner, & J.M. Carlsmith  1963  "Communicator's  '.      '
   ^credibility and communication discrepancy." JOURNAL OF ABNORMAL AND
    SOCIAL PSYCHOIOGYey: 31-36,    .   '     '. , ; •  ['  ',   '     .-...-'
                       .              ..         . .              .-     ..
 Booth,  W.  1988 "Socia|l Engineers Confront AIDS. " SCIENCE,  242 : 1237-1238 .

 Bord,  R.  J.  1985 "Opinions of Pennsylvanians  on Policy Issues  Relted to
    Low-Level Radioactive Waste Disposal." Institute, for Research on Land
    and  Water Resources.  83pp.               ,     j

 Bord,  R.J.  1987 "Judgjments .of Policies Designed  do. Elicit  Local  '
    Cooperation on LLRW Disposal Siting: Comparing  the  Public  and
    Decision  Makers'." NUCLEAR AND CHEMICAL WASTE MANAGEMENT.  7: " 99^105 .
                      I-' .         "         •      -  '•" [          i
 Bronstein, J.M. 1987 if'The Political Symbolism of Occupational  Health
    Risks." In B. .Johnson & V.T. Covello (eds.) THE SOCIAL  &  CULTURAL'
    CONSTRUCTION OF RISK. D. Reidel Publishing Co. |. pp 55-78.^

 Covello, V.T. ,  P.M.  S'andman, & P. Slovic 1988 "Risk Communication,  Risk
    Statistics & Risk Comparisons: A Manual  for Plant Managers ." Chemical
    Manufacturers Association, Washington, D.C.   !   '
                   ''"',.:•'    • , V'    "' •   "'  ••- -I' '.-"• '. •: v"  ..
 Covello,' V.T. ,  P.  Sloyic, & D. von Winterfeldt 1987 "Risk  Communication:
    A  Review  of the Literature." Unpublished Manuscript.
                      i .        .        '    „  " .  '   ^ .       ...    .      ....
 Douglas, M.  & A. Wild;aysky 1982 RISK AND CULTURE:!  AN ESSAY ON  THE
    SELECTION OF TECHNOLOGICAL AND ENVIRONMENTAL DANGERS. Berkeley:
    University of Calif orni'a Press.-             ,;,".',
                      ! ..'--•'.        • • '•   .     ".["',•    '  • '.
 Fisher,  A. .1987 "Risk1 Communication: Getting  Out (the Message About
    Radon/" EPA .JOURNAL 13:727-28..  '.'   • •   '••.-.  !           •       - ....
                      " •              ..    =         i""'  v    .
 Fischhpff, B.  1981 "Hot  Air: The Psychology of .C0;2-Induced Climatic
    Change."  In J.H.  h,'arvey (ed.) COGNITION,, SOCIAL BEHAVIOR, AND THE
    ENVIRONMENT, Lawrence Erlbaum Associates ,  Hillsdale,  N.J.
 -   .  •         ,   -^  ••}'• "..-         -   •-   ;.   •-'-     J"  ••••-•;'•  - .:•"•'  - -.-  .',
..Fischhoff, B.,  P.  Slovic >S. Lichtenstein,   S. Read,  &  B. Combs  1978
    "How Safe is Safe EnoughV" POLICY SCIENCES, 9:;  127-452.          "   .
       '..','         i; .'    '.'.'•-.•   _.••••  .  ,   -[•-  \    '••'., '          .  '•
 Fitchen,  J.M. ,  J.S.  Heath, and J. Fessenden-Raden  1987 "Risk \Perception
    in  a Community Context: A Case Study." In  B. Johnson and V.T.
    Covello,  THE SOCIAL & CULTURAL CONSTRUCTION OF;; RISK,  D. Reidel
    Publishing Co.  pp 31-54                       i    .

-------
 The lack of trust,  especially in industry and local government but also
 in  science and people's ability to manage complex problems,, 'indicates.
 that  communicator credibility and local representation are crucial
 issues.   The data of this study indicate .that concern with man-made
 toxins  and the,,perception of imminent threat are indeed .ubiquitous  ,
 (O'Riordan,  1983).   Therefore,  anything ah agency can 'do to appear    •:
 forthcoming is likely to have more positive than negative outcomes.1
 Although the uncertainty manipulation did not have a strong impact, it
 did appear to lessen post-manipulation fears somewhat.  An admission",  by
 an  agency,  of limited knowledge,  along with an explanation of what it
 does  know and how it knows  it,  may gain support among those who
 understand the pjrocess of science to some degree.  Also, because local
 government often  inspires so little confidence in situations like these,
 it  is crucial to  communicate information directly to concerned- and
 affected citizens.   There is much anecdotal evidence indicating that
 small,  informal meetings run by trained EPA community -relations
 specialists  have  reduced citizen  fears at Superfund sites.   These
 meetings  are a form  of increased  community involvement.   Large,
 officious  public  meetings that  serve a cathartic function are of little
 value.  •                                                           _, . .  .

 Perhaps  the  strongest  message imbedded in this research  is  that waiting
 to  communicate scientific-technological information until a hazardous
waste response action  has occured  is simply, too late.  Education plays a
 significant  role  in  the  risk evaluations noted above.  Risk
 communication  has already taken place through the media,  the schools,
within interest groups,  and  between relatives and friends.   Although
much has been  made recently•about  the need to teach humanities  and
 classics  in  the schools  a strong  argument  can and should be made that  •
 people also  need  to  understand  the  technical  environment -in which they
 live.   A public information  program,  sponsored by governmentiand
 reaching down  through  the schools,  would help citizens better understand
 their world  and its  problems  and, hopefully,  help them to make  more
 informed decisions about  that world.,  A broader  public information
program is consistent  with the  purposes  and  legislative  history of both
CERCLA and SARA.

-------
                                  .90
  . •'' scenario'that ha's'^a  somewhat  consistent,  though small,  impact on.
      explained'variance is/the  indemnification.manipulation.  ,    ,    •>.  •
      Apparently' the1 inde'minification  information! is  one more  signal
      about how serious, this  problems  really  is.  [Respondents  seem
      to. be' reasoning : that . if i-t  were  not  extremely serious then *
      this option1 would not be  provided.     • .•    ;.''-••.   '      :  •'- '  •'-
  .   ;.  •      ••.     .. |;:,   r[. V--'-'  -.      *  :  f'.'" .      -:  •-.   •.:••'/'
Implications  for Risk!Communications at  Superfund|Sites        ...
                     .\,  ' '  •   '           >         4"j     ..'•."
First, the notion of the "signal  potential"  of  certain stimuli is
strongly supported in'[ this^ research.   The  pictures of the'leaking  50 .
gallon drums  and of tiie technicians dressed  in  protective clothing plus--.
the description of the' contaminated site totally  overwhelmed any" other
information.  The data, both  structured  and  unstructured, consistently
indicated that those items  of information  communicated the "-real"
seriousness of the problems.  Subsequent tables or, statements  were
either ignored or viewed as lies.   From  the  perspective  of the layperson
facing such a situation this makes-, a great deal of sense.  Why would1 the
government be' here, and why would their  people  feel compelled  to be so
careful, if this situation were  not a  serious health  threat.   To' then -
present data  indicating the d,egree  of  threat  is minimal  is to  present a
contradiction.  Government  agencies must realize  that their total
pattern of behavior constitutes  the risk communication.^ Serious  .
consideration must be!,given to. how  to  structure that  pattern in ways
that sends the kind of message  the  agency  is  trying to send.   In one
national meeting a noted risk communication  expert was asked the most
effective way to present risk information.  He  smiled and said, "Hire a
good P.R. person."  While  that  response  may. be  taken:as  ,a'cynical  remark
it may also reflect his understanding  that risk communication  is more
than official notices!or data structured in  Various ways.  Messages are
sent by behavioral styles,  costumes, communicators attitude, certain
kinds. ofr trucks rumbling down city  streets,  fences and signs,  leaking
barrels, odoriferous,and colored water^ supplies,  and  a host of other'
factors.  Perhaps  thei image,  the gestalt,  is  more; crucial than the
pieces of formal information  released  by the agency.    .

Second, the fact that! so .many people were bothered by the length of time
it took to execute th'ia cleanup  must be taken into account.  None of the
information provided about  the  site communicated  why the remedial-   (
process at Superfund sites  cannot quickly be completed.   From a risk
communication perspective,  it is . important to ensure" that, expectations
do not outrace realities.   The  social  movement  literature consistently   ..,
calls our attention  to the power of unrealistic expectations ,to generate
outrage.  The agency should consider  providing  additional information on
the  remedial  process and how  long most cleanups take.  This information
should be communicated early  in the remedial 'process.  Local Citizens
may  not be happy with/ the  projected length of the! project, but they will
not  have to experience dashed hopes and  the  frustration associated with
that.   '      •   '-   i   "'..-,•      .    -;   " "-:\ ,: ;. ••'•'  •   ' •   ••. .-;:

-------
                              89
 *The extent to which the risk- communication was  either  disbelieved
  or simply ignored is nothing- short of  remarkable.   If  the  risk.
  communication fail's to reflect the subjective risk-it  is
 .dismissed.  Even those who evaluated the  risk communication  in  a
  generally positive way did not use the information  it\contained
  in>making their judgments. •

 *People overestimate the the risks associated with a  toxic waste
  site in comparison with both Voluntary, e.g., smoking, and
  involuntary,  e.g.,  annual chest x-ray, risks.  Fear  of man-made
 . toxins appeara to involve well-formed attitudes and  beliefs.
                                               •./,'-'    '••   ,  •  r •
 *The four elite groups (environmentalists, civic leaders,
  elected officials,  and business, people) are more alike than
  different on  prior attitudes,  level of concern,  health risk and
  comparative health risk estimates, and trust."1  Elected officials
  are somewhat  more trusting,  but similarities .are still more
  striking than differences.   Perhaps everyone becomes an
  "environmentalist" in this  kind of situation.

 *Behavioral intentions  scale  from behaviors involving very little
  personal involvement and  commitment to thosse that mean putting
  one's  self on the line publicly.   Almost everyone would get more
  information,  about  66  percent  would buy bottled  water, about
  50  percent would  encourage  others to buy bottled water, and"  ,
  approximately Half  would  take  political.actions.  Again,
  environmentalists do not  differ from other elites on this issue.

 *The variation in  levels of  concern are disproportionately
  explained  by  trust  items.. In  fact,  the trust items  consistently.
  explain  a  significant  portion  of  the  variance regardless  of the
  dependent  variable  under  analysis.   In addition, when health
  related  items  form  the dependent  variable,  then  demographics,'
  specifically  gender  and level  of  education,  become important,
  contributors  to explained variance.   In general, women, are  more
  fearful  and more  likely to assume harm coming to'them or  their
  families as a  result of the"  contaminated  .site.   The  more  highly
  educated are'more likely  to  reflect  levels  of estimated health
  risks more consistent with scientific  estimates.  Their levels
  of both  general and  specific concern,  however, mirror those of
  the less educated respondents.                    "

*The other  cluster of variables  that explain ,a significant amount
  of variance in the dependent variables  is  the attitudes toward
 hazardous waste that people brought with  them to the research
  setting.  Prior attitudes and involvements  in hazardous waste
  issues explain most of the variation  in the behavioral  intention'
  items.         ,                                  ;     .

*The uncertainty manipulation,and  the variations  in management
 scenarios had extremely small effects., At  the biyari-ate  level
 there .is some indication that the uncertainty manipulation',may
 have increased credibility'to some extent,  At the multivariate
 level the effect virtually disappears.  The one  management   ,

-------
     . SECTION XIV: SUMMARY AND  IMPLICATIONS  FOR RISK COMMUNICATION''
  . '             -      . I         -                   i ;
 This.research was designed to  illuminate  some  ,of. the major factors
 affecting consistencyjbetween  objective and 'subjebtive  risks  in
 Superfund-like situations.  Perhaps  the most significant  finding was  the
 virtual absence of impact of the risk communication itself and  the
 management scenerios.| The lack of impact of, an  independent variable  can
 always be attributed to a weak manipulation.   In  this case the
 uncertainty manipulation can clearly be a candidate" for  that  accusation.
 Perhaps the information on uncertainty was  insignificant  given  the
 •plethora "of other types of information, some of which were much more
 dramatic.   However,  th;e "relative lack of impact oi:  the management     /
 scenerios  is less open to the charge of insignificance to  the respondent.
 A large body of research argues strongly for the  Importance of  citizen
 invovement in decision making in reducing fear and  building trust.  But,
 an  analysis of the pre' and' post manipulation data;  and the open  ended
 material at the end  of;the research,  demonstrates [unequivocally  that  the
 these  types of information were unable to compete [with the beliefs and
 attitudes  that the respondents" brought, with them  tio  the situation. . The
 following  conclusions |weaye the net produced by this research.
' -           '  -         ; _              . '          • ,   }  '     .,-._',,
     *Most people enter  a situation-involving  toxic waste''problems with
    •<  a^dramatic  lack |of consistency  between objective and subjective
       risk estimates.  Almost  everyone is  highly concerned and a
       sizeable minority  disproportionately attribute the incidence of a
       number  of serious  diseases  to man-made toxic> chemicals.   Risk
       estimates are  clearly  related to levels  of education.  That means
       that  it  is  certainly  possible,  through training, to bring the
       average  citizen's  risk  estimates more  in  lin;e with  scientific
       risk  estimates.  'However, education  has no relationship  to levels
       of concern.  The, implications .are  that even jif subjective  risk
       estimates are brought  closer  to scientific risk estimates  it may
      have  no  bearing  on  public intransigence and  Distrust of  government
      agencies  involved  in  the  cleanup.            ;   .

     *There. is  considerable skepticism about government,  science,  and
    .  the  possibility  of managing these  kinds of problems  safel'y.
      Skepticism  of this scope makes Communicator  credibility  highly
 '.     problematic.    T     '                    :   I                 '   V •
-••..-..:          ;;,. -•••    ..   ;    , •.:  •   ^  [.. ,; ,    ,  '.   ,.  .    .
     *Trusfin  local government scales with  trust  in industry, not  with
      trust in  state aiid federal governments.   In-addition,  trust  in
      local government is lower than  trust in other  levels  of  government.
      This result^is identical with results  found  in surveys toward
      radioactive wastes (Bord, 1987).  The  problem  for the EPA  is  that
      gaining the cooperation of local officials.will not  necessarily
      convince other citizens that  their health  and  safety  are being
      protected.                                   '"  '
     *Almost everyone
      a sizeable
        highly concerned with  the|situation,  and  for
minority the cleanup  fails  to  reduce; that .,concern.

-------
 attributed the time delays to general bureaucratic ineptitude but  others.
 attributed it to a lack -of knowledge about how to clean up such,a   •
 situation.  Apparently the seriousness of the problem is. indexed by  the
'fact that it is officially defined as one and subsequent informatiqn  is
 often'viewed as part of a "cooling out" process or just plain .lying.<
 Many respondents are reasoning 'that if the problem were not a serious .
 health threat, then there' would not be government agencies involved  and -
 testing programs in progress.  Any information that fails to affirm  this
 perceived seriousness is viewed as misinformation.              ••

 Two common assertions from the risk communication literature were
 affirmed by the open-ended comments: comparisons of voluntary .with
 involuntary risks are viewed by some respondents as_ silly or misleading;
 and, people want to know the long-term, chronic health effects  and not
 just an annual chance of death:

-------
       -A rull 40 percent of the ' espondents  critically, mentioned '   ''
        tne.length of time that- • : * cleanup  process  took.   No  othe-  single
       .topic, was mentioned more jften or with "such  consistent  negative
        evaluation.   11^; some cases the respondent  generalized  to
        all bureaucracies and their relative  inefficiency
       *0ver  20  percent!.of the respondents alluded to  their  distrust
               igencies;;  industry, and/or government.'Many  responses
-uver ^u percent:, or the respondents alluded |to
 of  the agencies^  industry, and/or .government.' Mi
 were of a general nature, such as the following;
           ".It  gives  a  feeling of dishonesty on the; part of all  the
            agencies  involved.  I don't know that i|t is stated,  but
            it  is  implied."              .    .
      ^Approximately. 15 .percent_of the sample-criticized the annual
       chance of death'format.   They wanted the long-term probability
       of getting any  of  a  number of,, specific health problems.

      *Almost 12 percent  criticized  the comparison!of voluntary with
       involuntary risks.   On  the other, hand,  a few of the respondents
       liked the comparisons.

      *Several people who.got  the management  scenario that included
       enhanced citizen;, participation  mentioned this in a positive vein
       However,  it did :not  appear to have  a significant impact on their
       judgments. •     •''              -             ;•     -
                      '\:  '      : '••-.'  "    '-•     '"'  |.  '""   ~      -
      *The more highly Educated respondent-was  likely to want  more
       information and more highly technical  information while the
       lesser educated :was more likely  to  assert  that there  already
       was too much.information,  "certainly more  than anyone would
       want to read."  L           .      "           .! '      .
       \         '"      '!'--'•            .    '       ~i.
      *Respondents  frequently mentioned' their  fears|  of  contaminated
       water.   The  point  made was that .water is a necessity and  cannot
       be  avoided.      r                            j-

      *Approximately  one-third of the 'respondents Consistently stated-,
       and implied,  that  even though the information  provi'ded was
       adequate  they  simply did not believe it.  They wanted to  know
       where^the  statistics  came from and why anyone  should believe that
       drinking  the  contaminated water was  similar |in level -of risk to
       drinking diet  soda.                          ;

      *Finally, approximately 6 percent of  the sample felt that  the EPA
       should_have acted more paternally right from: the start.   They felt
       that  citizens  should  be  ordered  not   to  'drink.! that' water* and'that
      bottled water  should  have been provided.    [

Summary     •    '  .    ,!>         •       '  ,          f  ' '  '

Twb themes  dominate  these open-ended  remarks:  government ag'en'cie* take
far too long to deal with ..problems  and agencies,  industry,  and
government  cannot be  trusted- in cases  like this.   Some respondents

-------
                                  '03
           SECTION XIII: RESPONSES TO OPEN,-ENDED QUESTIONS  ,

One problem with structured questionnaire  formats  is  that  they,  tend  to
define the situation for the respondent.   Although  the  post-manipulaiion
structured questions appeared to tap those dimensions of  respondents'
attitudes, beliefs and behavioral intentions  that  are central to
assessing consistency between; objective and subjective  risk, a  number of
open-ended questions were posed at the end so  the  respondent could .
define the" situation for himself or herself.   In one  sense,.these
open-ended questions are validity checks.  They help  us determine
whether there is correspondence between ou-r definition  of. the situation  •
and that "of the respondent and whether our interpretation  of the results
is reflected in the respondent's thinking.  Four composite  open-ended-
questions were posed at the end of the questionnaire.   They started  in
very general  ,   terms, encouraging the respondent  to structure his  or
her answer independently.   In the face-to-face interviews  these
questions were asked verbally.  The following  constitute  the open-ended
questions:                                         -.'•'<••..

     *was there anything positive or negative  which struck  you  in
      the information you were asked to read  about  the  waste,site;

     *specifically, anything positive or negative  about the "water
      advisory notice, (you may want to 'go back and glance  at it   ,
      again);

     *specifically, anything positive or negative  about the "management
      scenerio," (you may want to go back  and  glance  at it  again); and,

     *what kinds of information should agencies like  the  Environmental
      Protection Agency (EPA) provide the  public about  such sites,
     (was there more information, or "a different kind of  information,
     " that you would have preferred in reading about  the  hypothetical
      waste site)?

Results of the Analysis of the Open-Ended  Questions

Although not' every respondent answered every  question almost 500
responses were provided.  Many of the responses simply  a.ffiirmed that the
information was adequate and offered little of a critical  or
constructive nature.  For example:              ...

     *"there was a lot to read but the information,was  interesting
       enough to keep me reading;" or,

     *"the information presented a situation  that  was probably,a  '
       typical one - examples such as that shown should sensitize
       the concerns of persons exposed.          ,

However,  the data did reflect recurrent  themes and provided some
valuable  insights.  The following reflect  the most robust patterns in,
the responses.                                        -     . .-+..-'

-------
                                   34'
                      I •         "       •            ' •  • •    '   ' , *    '    ' ' - '
  •   *Believing  that • problems like this' are  difficult  to  control-'is
      positively  associated with.- a ,number of' activities  as" is  believing
      that  science  does, not know enough about' problems - like" these.. ':

   . . *The greater the-'religious involvement  the  less likely  people are
      to get  involved'..in .political, activist  activities,  and  the more -'
      likely  they are i.to get involved, in talking about  the  problem.,
                -"" .    |                • .     ' .   '   •.'_"''•
Sunmary of.Section  XII . '    '                       ;  :        ,
     .---"- ~ ~            .  -» .         ,         ...        t     ,           •„,•"'
         '.'•••   '    !   "  "  •"  • .     •'.'•'.•';   ;£'•'•'   ••;•'-'•;..>'••••.
The .following outline?!' the most dramatic findingsInoted above.      ;

     *The formal  risk^information  manipulation had  virtually no
      impact.  People'js  judgments  of relative risk  are determined  by
    ..  beliefs and attitudes they bring  with  them t:o the situation.
    '  '   •'.  •'   •. '. •  :'   • •''"'' '•"   '•'•.•" '•''   -   • i' •  •:•     i    ;  ••"••
     *The indemnification  managment scenerio. acted  as a "signal" that
      this  problem  is [substantial  and had an  impact on people who
      received it.  This result highlights the possible impact  of  all
      sorts  of "signals" such as  technicians,  in protective clothing,    '
      fences and  signs to  keep people out,  and other visible-clues.
            1  •   •     i-    ..      " '    •     •'  • '  \-- •••••...
            - .         .'-•         •  ; "    -   -'  • •    '  ' I- ' ' :, -  • " i  ' ' '      / ,
     ^Trust  in government,  especially local  government, is by far  the
      the most.important and  consistent  set  of variables  in accounting
      for risk estimates.   Thi.s  fact  raises  some interesting
    .  implications because  agencies must work with local  government  yet
      have  little control  over their  quality  or the degree of trust  they
      enjoy
                                                  •..fe-
     *Female gender is. consistently  related to overestimating risks.
     . This may help explain why  so many  local protest groups are
      organized and led by women.     '••'.'..•   . [         •   -  -

     ^Education is predictably related to  accuracy! of risk comparisons
      and the type of 'activity that  people will jOjin.  Those with   '
      higher levels of], education are  more  accurate! in their risk
     .estimates and more likely  to be involved in pyert, political
      activity.  The better educated, however,  hav^ only slightly
      lower levels of concern.,                     [         t
       "  ;'   .-.   '  :   ']:, '  •  •••    • •'.••   •    "   " ':  I.' '.:•-"".  "..•'.••••'•  •
     *The belief that toxic waste sites ,are able tb be controlled by
      existing managment techniques  is consistently and predictably
      related to concerns and behavioral intentions.   This is another
      belief brought to toxic waste  problems  that has an impact on
      peoples'  judgments.         .                  !,.          ,

-------
                                  •83
     Table XII-10: The Additional Explained  Variance  for  Ea;ch  of  the
                   Major Independent Variable  sets.' for  Each  Behavioral
                  , Intention.  ,            • •          .. .   •     .

                        Actions  Talk  Personal   Others-  Political  ••'
     PRIOR ATTITUDES      7%
     TRUST   , '            5%
     DEMOGRAPHICS  '       2%
     RISK COMM. &.MANAG.  2%
     SHARED VARIANCE      1%

     EXPLAINED VARIANCE  17%
12% •
. 2%
2%
1%
3%
2%
6%
• • 5% '
1%
3%
.20%
17%
12%
 5%
 0%
 It'
 4%

22%
,  7% •
~  2%.
  1% ,
  1%
  1%

12% .
Several findings emerge from these  tables   '           •

     *The overall amount of explained variation  in  behavioral  intentions
      is less than noted in previous tables.  This'  is  probably  a  result
      of overt behavior, varying in its demands  for public  boldness,
      being strongly related to personality variates and  the, possibility
      of sanctions from any number  of sources including employers,
      friends, relatives, and acquaintances.  These unmeasured  variables
      may be accounting for much of the unexplained variation.

     *Prior Attitudes has the. strongest impact on overall variation with
      one minor exception: demographics and.trust play a
      substantial role for "personal" actions.   Those with  less
      education and women are more  likely to .take personal Actions.

      *The indemnification management scenario appears as a  significant
      variable in all of these scales.  As noted previously,  this
      option cued those respondents who received it that  the  problem
      was of sufficient magnitude to encourage them to act.
                                  1        '          •               t . f
     *Those of lower education are more likely to take personal action
      and talk,  but not more likely to get politically involved or_to
      take part in public behaviors..  This is consistent with a corpus
      of studies demonstrating that higher levels of education are
      strongly associated with overt,  public participation.

     *Women are more likely to take personal actions and convince others
      to do the same, but are not likely to get  involved in  political or
      public activist behaviors.                    '      •   .

-------
      Table  XI1-9:  Order of Variables and•„Summary
       :.       ,    ' Five' Behavioral Intention Seals
          Action  • -. I

      PRIOR. CONCERN  * >•
      TRUST  LOCALS -r . '
     .INDEMNIFICATION*
      MARRIED*  •- '    .':-:•"-
      PRIOR  INVOLVEMENT*
      UNCERTAINTY DISCL.+
      RELIGIOUS  SERV.-:
      TRUST  GO.Vt!.- .   |:
                     „' •

      Mult..  R=. 40     :•
                Others
  .THE  SCALES,
•  .  Talk

PRIOR CONCERN*
MARRIED *
INDEMNIFICATION*
POSSIBLE TO CONTROL-
"PRIOR INVOLVEMENT*
AGE*
TRUST LOCALS-
RELIGIOUS SERV.*

Mult. R= .44    •
R2= .20
                      Coefficients  for the
                      s.        .   '.
 Personal

TRUST'LOCALS-.
AGE*' •
FEMALE GENDER*    '
PRIOR CONCERN*
POSSIBLE TO CONTROL^
PRIOR.SICKNESS*
INDEMNIFICATION*.'
SCIENCE KNOWS-

Mult.,R=*.40      . '
R2= .17
             Political
          PRIOR CONCERN*
          TRUST LOCALS-
          INDEMNIFICATION*
          POSSIBLE TOj CONTROL-
          PRIOR INVOLVEMENT*
          SCIENCE KNOWS-
          FEMALE GENDER*
          Mult. R= .47
          R2= ,22".-  !
            PRIOR CONCERN*
            INDEMNIFICATION*
            PRIOR INVOLVEMENTn
            TRUST GOVERNMENT-
            KNOW CHEMICALS*
            RELIGIOUS: SERV.-
            TRUST LOCALS-
            SCIENCE KNOWS-
                  •  • •..['•• • '
            Mult. R= ;. 34
            R2= .12
Table XII-10 is also a summary  table presenting  the  additional  explained
variance of each of the major independent variables  for  the .five
"action" dependent variables.         .-..'•

-------
                                  0 -
     Table XII-8: Regression. Analysis for "Life-Risk" with  the Four
                  Major Independent Variable. Sets.          ;

       (NOTE: Explained variance with all variables included  = 20")

                                              '  .          MULTIPLE ,R

         ALL VARIABLES  .             •              .   •   '  -'< - ..45
PRIOR, ATTITUDES AND INVOLVEMENT ' ; /-
TRUST • ' ' ' .,
DEMOGRAPHICS , ' ,''"..
*RISK COMMUNICATIONS AND MAN ACME NT OPTIONS
ALL, MINUS PRIOR ATTITUDES AND INVOLVEMENT
ALL, MINUS TRUST , ' •-..',.
ALLJ 'MINUS DEMOGRAPHICS \
ALL, MINUS RISK "COMMUNICATIONS & MANAGEMENT OPTIONS
.18'
.25
.37
._10
.43
.40
.31
.-44.
     Additional Explained Variance Accounted for by Variable .Set
         PRIOR ATTITUDES AND INVOLVEMENT
         TRUST                  '
         DEMOGRAPHICS
         RISK COMMUNICATIONS AND MANAGMENT OPTIONS
         SHARED VARIANCE
 2%
'4%
10%
 1%
 3%
     *"others" is a three variable scale  including  convincing  others
      to buy bottled water, convincing others  to  leave, and  going
      door to door to get others involved; and,

     *"political" is a five variable'scale including  speaking  up
      at public meetings, marching, writing members of Congress,
     •organizing a protest meeting, and testifying  at a public hearing.

Table XII-9 -presents:the order of variables, the  direction of  the
relationship, and the summary coefficents for  the five scales.  This
summary is .presented because of the significant overlap in each scale.

-------
                                     so
  and life-risks..  Once,; again,  the', pa'ttern  of  results for the- two tvpe= • -
 ^risk are quite, consistent.  .Demographics  account ' for . the brunt-of.-
       Table XII-7: Regression Analysis  for'"Voluntary~Risks"'with the
                    Four-Major Independent Variable ; Sets. •

        •  (NOTE-:  Explained variance with' all variables  included = 32").
                        1 *         ' .    -      •         f "  '    '   • • '
                        i'               , "            >-••'-,•
                        I'         •     .           ,   j      Multiple R

          ALL VARIABLES!!     '   ,   -'•"•-         ""•  ''T  '  '  '• /  ,7 "  "•  •
         •:  ,  •           r-    . •            •     • . •    t    .'•.., j / '.•
\ '.''•'' ' . ''':,',
PRIOR ATTITUDES AND INVOLVEMENT
TRUST / - ]; .''-."'' '.-,•'• . . ' :
DEMOGRAPHICS |; ' " /
RISK COMMUNICATIONS ' & MANAGEMENT .OPTIONS
ALL,
'ALL,
' ALL ,
MINUS
MINUS
MINUS
PRI'OR ATTITUDES AND INVOLVEMENT
.TRUST , •
DEMOGRAPHICS . '.
,-'• .•-•.•' -• " - /-' ••
,: . :• • ...• .3.6 	
-••••:• :". -.35 . :
'•'• '• • -.46 ' ••
•': .10 •
• . 54
••• ' . •' - V51 - , '
. .40
         ALL, MINUS  RISK COMMUNICATIONS & MANAGEMENT OPTIONS  .56
        •..--.'    ; .ii  .-.•-•".•     ...'.-.   •   • f--:'.   •-,   :;.;  : ••
      Additional  Explained  Variance Accounted for by Variable  Set '
         PRIOR ATTITUDES  AND  INVOLVMENT
         TRUST     '    • j!   .   . .. ••   '.  .  . .  -   .'.'•_• -j
        •DEMOGRAPHICS  1;    ;             ,         . j
         RISK COMMUNICATIONS  AND MANAGEMENT  OPTIONS'
         SHARED VARIANCE        '        .    <•.*''[
 6%
16%
 1%
 variance.  In particular,  the  level  of education produces the largest
 effects^ The lesser educated  are more likely to overestimate the danger
 of the site, even aftez; variance for trust  and prior attitudes and.
 irivolvement is taken into  consideration.  However, | trust and prior
 attitudes and involvement -still account .for significant variation.  The
: consistency .of the impact  of trust is  impressive.
 Accounting for the Behavioral Intentions
           '

                         •          -   .    .      -.        .               .
 As  noted previously, behavioral intentions  form  one  overall scale termed
 "Action" and four subscales:
              : '         '."*      - ' •      '            ";-"''•'"     •
      *"talk"^is a thre« variable  scale including the  respondent's
       intention to get jmore information, to  talk with  others \about
       the situation, arid to serve on :a commitee  if [needed;
                     '!;'"•• •-' .''•  • >'  -  ""•-•  :  .    1:.^; :  ••  •'-'•'  -'.•      '
      *''personal" is a 4hree variable scale  including  the  respondent's
       intention to buy bottled water, to Install  a (water". treat ement
       system, and' to moy.e put of  the area;,    .   -I       •-•••••

-------
                                  79

levels of concern before the experimental manipulation.;- Note that the
experimental manipulations have almost no impact on-these comparative
risk judgments".  Getting the enhanced citizen participation'management
option slightly reduced overestimates of involuntary risks.  These
overestimates, nevertheless, are high.  This in spite of the fact that
some of these comparisons were graphically presented in^the information
provided respondents..  There could be no more dramatic  information
confirming the relative absence of impact- of the risk communication
itself.                         .            '                      .   ,  '
     Table XII-5: Regression Analysis for "Voluntary Risks.'"
                VOLUNTARY RISK-AS THE DEPENDENT VARIABLE

     Ind. Vars.         Multiple R   R2Change   Simple  R
     EDUCATION              - 398
     TRUST GOVERNMENT       .479
     PRIOR CONCERN          .508
     TRUST LOCALS        '   .525
     FEMALE GENDER          .533
     AGE      '           .•  .534
     MARRIED                .545

         Multiple R  =  .55
         R2 =  .30
                 .159
                 .070
                 .029
                 .018
                 .008
                 .006
                 .007
-.557
-.233
 .544
-.288.
 .364
-.146
-.555
Be.ta

-.364
-.190
  .137
-.109
  .084
'-.086
-.127
     Table XII-6:  Regression  Analysis  for "Life Risks."

                   LIFE  RISKS  AS  THE DEPENDENT VARIABLE

     Ind.Vars.   '.        Multiple R   R2Change   Simple R
      EDUCATION,     •       .300
      TRUST  LOCALS          .351
      MARRIED              .390
      INCOME             '  .399
      PRIOR  SICKNESS       .409
      PRIOR  CONCERN        .402
      POSSIBLE TO CONTROL  .427
      CITIZEN PARTICIPATION.433
       MANAGEMENT OPTION
                 .090
                 .033
                 .029
                 .008
                 .008
                 .009
                . .006
                 .005
-.300
-.201
-.115
-.060
-.053
  .100
-. 150
-.050
 Beta

 -.290
 -.149
,-.223
 -.101
'-.104
  .086
 -.101
 -.057
:  \
          Multiple R
          R2 = .19
.43
                                              ,'            •   .   *
 Tables XII-7 and XII-8 present the regression of the four major
 independent variables on both the health risk assessments of voluntary

-------
       Table XII-4: Regression  Analysis  with.Health! Risk Assessment
                    Dependent Variable -Regressed onj the Four Maior
                 .   Independent Variables.  ,      .  •   '-'''•''.''  7 '•
        ass tne
          (NOTE: .Explained variance with  all.' the .vatiabl.es included

                                                        • •  "' "Multiple
          ALL VARIABLES.!!
          PRIOR ATTITUDES AND INVOLVEMENT          "''••
         .TRUST '   •     i;   ;'        '-•••,      :r.
          DEMOGRAPHICS !    .  .  -• '    -"         •      'I.
          RISK;COMMUNICATIONS AND MANAGMENT OPTIONSt
   .3.6.
-   .42
' -' .'42
   .06
         ALL, MINUS  PRIOR ATTITUDES AND INVOLVMENT ''
         ALL, MINUS  TRUST .  •'  '       .      .   '•   1 '.-.".
         AL'L, MINUS  DEMOGRAPHICS            ""   -'[''.
         ALL, MINUS  RIS.K  COMMUNICATIONS & MANAGEMENT OPTIONS
                       , - •    -   - ,   ,      -  «        . L ' '  . , : • 'r  '  .

      Additional Explained Variance Accounted for .bV Variablp
         PRIOR ATTITUDE;!? AND  INVOLVEMENT    -
         TRUST  .       !' .   -•'.'.         '''••.   1
         DEMOGRAPHICS  :    - '    ' .         -      -  j
         RISK COMMUNICATION AND MANAGMENT OPTIONS   !
         SHARED VARIANCE (two or more  variable  sets1)
                       • ''•'   -' •   ..--' '• •   '.,  '•'    •    •  *^
                                                    I-
   ..'55
   .49
   .49
                                                               Set
   ' 9%
   ' .'9%'
    1%
   11%
 nsk assessments than they were for  "qoncern."   Gender plays the most
 important  role among the demographics.  Women overestimated  sickness •
 possibilities much more than did men.  Prior attitudes -and  involvement
 also contribute Substantially to the explained/variance.
Accounting  for Compara1:ive Risk Judgments
Recall  that  a  previous[discussion indicated that  tHe  comparative  risk
judgments, comparing the risk of drinking the water, with nine more
conventional risks.before and after cleanup, brokeidown into  two  scales-
one  "Delved with_voluntary risks such as smoking, W riding,  and  bike'
riding, and, one  involved with risks that are difficult to avoid  such as
getting an X-Ray  and eating fruit sprayed with toxics.  Tables  XII-5  and
•XI1-6 present  the regression analysis for these two scales.  v      '
The  results shown in these two tables are reasonably consistent.  Level
or education -accounts  for most .of the variance in the comparative risk
estimates..  This simply,;means that  those with higher education  are
making more accurate rijsk judgments.-  , On the other Wd,  unreasonably
high estimates of the  risk of drinking the water are associated with   ,
distrust in government !at  all levels,  being single,! and having  High

-------
     Table XII-3:  Regresssion Analysis for "Health."
                   "HEALTH" AS THE- DEPENDENT VARIABLE•
          Vars.
Multiple ,.R
                                      R2Cha
nge   .Simple R  • Beta
J. LIU « v a. J. J •
FEMALE GENDER
TRUST LOCALS
.EDUCATION
KNOW CHEMICALS
PRIOR SICKNESS
TRUST GOVERNMENT
UNCERTAINTY DISCL.
AGE
, Multiple R1=
R2 = .33
.357
'.451 , .
.493
.516 - •
.531 . .
.545
.555
.564
,57 -

.127 ..
.076 .
.040
.023'
' .016
.014
.012'
.010 .


.357
-.323
-.258
-.279
.221 •
-.303
.010
.024


.215
' -.206 •'
--. 179V. ;
-. 154 :
:•"•-. 128
; -. 122-.
1 114-
.100
.. :

induced illnesses.  The summary of the above  table. is as  follows:

     *women make less -accurate health risk  estimates  than do  men;-;  ,

     *not trusting local government  raises  estimates  that illness
      will result from living near the site;

     *higher education is associated with  lower  health  risk  estimates;

   •  *knowing the chemicals  is associated with  lower  health  risk
      estimates;

     *prior judgments characterized  by overestimating the health risks
      of toxic  chemicals are associated  with  overestimates from the
      hypothetical site;                     '.'-.-'-,

     *not trusting state and federal governments is associated with
      higher health, risk estimates;              .                 .

     *being older  is  associated with higher health,risk judgments.

Again,  the explained  variance (R2=.33)  is  substantial.   Table XII-4
presents  the results  of  regressing the  four major independent variable
sets,  prior involvement  and concern, trust, demographics, and the risk
communications  and management options,  on  health risk assessaraenfs.

As  before,  trust-makes  a significant contribution to the explained
variance.  However,  the  demographics are much more important for health-

-------
                 •  ,                 .•-">    n.        , __ .
*De»ographics is made up of  the  following  .six variables:
                                  Uvel  of
                                                                 feraa.V-
       *Risk communication and managraent  options .are  the four dummy
        variables Abased; on whether  the  respondent  received  the uncertainty
        manipulation, the indemnification management  scenario  the    •   '
                         partlcipatiori  °Pciotl.  and [the  standard management
  Table XII-2. presents Jfche. results of regressing  ttle  above  four  variable
  sets  on. concern together and in combinations.   In these -types  of  tables
  the first coefficient; is the strength of the relationship  of all  the
  TH^±T  Varia^es,with Particular dependent variable under  scrutiny.
  The next  four coefficient, are the strength of  the  relationship of  each
  set of  independent  variables,  by itself, with the' dependent 'variable.'
  The following four  coefficients are the, strength bf the relationship of
  all the independent variables  when one set has be'en removed from  the
  analysis.  This  gives: another  picture of the relative impact of the
  variable.  The  last five coefficients are the percentage of additional
           e                                                         °
 The above analysis highlights  the  importance of trust in accounting for
 its^r* M H exPla"s an  additional  12  percent | of the variance by
 itself   No other variable set  comes  close to trust  in importance.
 One.s level of concern is more  a function  of whether one trusts various
 governments and the management  process i than  of  other factors,?  If trusJ
 is absent,then^the manipulation of  information  will  have little impact
 "objectiveC"edlble C0mmunicator there can  be n° information  defined  as'
                       !.     '     -          .    •'•• "I ''•"..•,
 Accounting  for Health iRislc Assessments            i
                       rh'                - •         •. ^        •    .
 Another,  more specific1,  dimension of people's reactions  to toxic  waste
 sites  is  their fears  about  health problems.  This idependent  variable is
 a  four item scale  we  simply lable "health":  how  likely is it  that you
 h!^axa*      a?d y;°Ur ifamily. would  g^ sick  from laving near  the  site
 both before and after cleanup.   Table XII-3  presents the  regression
 analysis  for  this  dependent variable.             .[

 In ^e above  table female gender has the strongest' relationship with '
 health.    This is consistent with  many other studies demonstrating
 thar women  have more  concern  about  a,large  variety: of risks (,Warr,1985)
 variance  .iT   S m8fi-n.mfk«  * significant  contribution to explained
 variance,as does the  respondent's level  of  information about  the key
 chemicals.  Two previously  undetected  variables  emerge in this analysis-
 prior^sickness  and  the ^uncertainty  disclaimer.  Prior sicknesJ is a two'
 ranra^e-SCann  encoinPa!?sin8 che  respondent's  judgement of -how many adult
 cancers in 100  are  caused by  chemicals  from toxic waste sites afld
whether the respondent ;;feels  he  or  she has  experienced chemically
                                       ' • '. •  -.   .  "  I. '  •     '     . •  .   '•

-------
     *income;                     .                 .-..-.•              '

     *belief that it is possible to control such a.  problem;

     *prior concern with hazardous waste, and      .          '   •-.  -

     *education.                    .
                                                -•- •'         i     /
Another useful way of decomposing this data is  to regress  the, four major
sets of independent variables on the dependent'variable  to assess their
relative impact on the explained variance.  The four major sets..of
independent variables are: prior attitudes and  involvement;'  trust;
demographics; and the risk communication - managment option's.

     *Prior attitudes and^involvement  is measured by three^scales and  a
      single measure of levels  of concern.  The scales measure beliefs
      that waste sites are-causing much  illness, knowledge of specific
      chemicals, and prior involvement in hazardous waste  issues.

     *Trust is measured by two  scales  and two  single variables assessing
      trust in local government  and industry,  trust in  other government
      bodies, and trust in science  and the management  of, risk (because
      there is no overall scale, the  statement that the  relation between
      trusfand level of concern is  .43  means  that  the  multiple r
      between these trust'measures-and concern is  .43.
     Table XII-2:  Concern  as  the Dependent Variable           ,

         (NOTE:  Explained variance with all variables included
                - 26%)
                                                          MULTIPLE ,R
ALL VARIABLES
PRIOR ATTITUDES AND INVOLVEMENT ,
TRUST . , '
DEMOGRAPHICS . ....
RISK COMMUNICATIONS AND MANAGEMENT OPTIONS
." -51
. 27
.43 ,
.: , .29
' '.08
         ALL,  MINUS PRIOR ATTITUDES AND INVOLVEMENT           .48
         ALL,  MINUS TRUST                                     •38
         ALL,  MINUS DEMOGRAPHICS                      . .... .47
         ALL,  MINUS RISK COMMUNICATION AND MANAGEMENT OPTIONS .51

      Additional Explained Variance Accounted for By Variable Set
         PRIOR ATTITUDES AND INVOLVEMENT
         TRUST
         DEMOGRAPHICS                           .
         RISK COMMUNICATIONS AND MANAGEMENT OPTIONS
         SHARED VARIANCE (two or more variable sets)
 3%
12%
 4%
 0%
 7%

-------
       Table  XII-1:  The^Regression Analysis for "Concern"

                    "CONCERN" as the .Dependent Variable;

       Ind. V.ars.     '  Z  .Multiple'R  .  ^Change  'Dimple R
      TRUST LOCALS     ]i ' , ..' .388
      TRUST GOVERNMENT jr .    .416
      PRIOR INVOLVEMENT     .436
      KNOW CHEMICALS   I  '.    .455
      •MARRIED          ;      .468'
      INCOME   :        ::     .486
      POSSIBLE TO CONTROL    .496
      PRIOR CONCERN    j i    .503
      •EDUCATION        .-    • .509

         Multiple R= .51
         R2 = .'26   ' ,  <:
  .151
  .022
  .017
* -017
.  .012
  .017
  .010
  .008
  .005
-.388
-.303
 . 100
-.150
-.076
-.130
-.21,0
 .180
-.120
Beta

-.259
-.136
 .134
-.117
-.206
-.138
-.090
 .100
-. 071
                                    1   • •• •  •       x T  •
      (Note: Beta is a Istandardized slope  coefficient).
 In the above table the;; order of the variables  indicates  their relative
 importance in accounting for variance in  the dependent variable.   The
 Multiple R is the cumulative correlation  of the  independent  variable(s)
 with the dependent variable while the Simple R is |the zero-order
 correlation between the single independent and the1 dependent variable."
 Note that Che Multiplej'R and the Simple R is identical for the  first   '
 independent variable. -;;The R2Change is a  coefficient indicating how much
 additional explained  variance is accounted for by the independent
•'-c-i-.ble..  Beta is, a  standardized slope coefficient indicating  the
 relative change in Y,  the dependent variable,   for .each unit  change  in X
 an independent variable.                           [  '   •             .   •
                       \              .       '       5 '.'.''"       •  •--•.,
 In these tables. the. fiir-st few variables account  for most of  the
 explained variance.'  "Concern" is  primarily the  result of two trust
 scales:  trust  locals  and  trust government.  Trust locals  is  a scale with
 two  variables:  tru8t  in local officials and the belief that  industry  is
 seriously interested  in protecting public health. ;Trust  government is a
 scale  with six variables:  trust  DER to tell the truth;  trust EPA to tell
 the  truth;  the U.S. is i.safer now than  it  was  30 yesirs  ago;  the Federal
 Government is  more hon«st  now than 30  years ago;  the DER  is more honest
 now  than  30 years  ago; iand,  Federal  and  State  agencies  are  seriously
 interested in  protecting  public  health.   The  following  variables have
 a  lesser  impact:       : •      '                     1   ,
     *prior information!1 and  involvement;

     *levels of knowledge about  the  two  chemicals  cif interes-t;

     *marital status;  j  .'•''•'       . •      [
                    '   I...'-    •  ..'     '•''.-'      .!•' .-  :      .

-------

-------
                                    72  :
       SECTION XII: EXPLAINING THE RESULTS:'  MULTIV^RIATE ' ANALYSIS  . .'

 The  goal of this section is to.delineate  the.majot  correlates' of the
 primary dependent variables.  This will be  accomplished  by  using
 stepwise multiple regression techniques.-,' 'This-  technique  enters
 variables in the order of their importance  in accounting 'for  variation
 in  the particular dependent variable.  The  discussion  for each' dependent
 variable will follow the same format: ^first, the Analysis will be
 presented,  including all the independent variable*,  that '-individually
.account for at  least -one additional percent of-explained variance  in  the
 step-wise-regression;  'then,  a regression analysis!will be' presented-
 that  isolates the four sets  of dependent variables.  Figure 1  summarizes'
     FIGURE  1:  A  SUMMARY OF THE VARIABLES USED INiTHE-MULTI-
              ,  VARIATE ANALYSIS .           -. .  '  -•>'',.  •   ;
       DEPENDENT VARIABLES
       ' I              Ii i,   -  '
      LEVEL OF CONCERN     • .  '.'
      (,a two-item  scale)
           :        • '  '; i  • ' •

     • HEALTH RISK  ASSESSMENTS
      (a four-it em sca!J.e)     ,•

      COMPARATIVE -RISK: JUDGMENTS
       , A.  Voluntary F.isks
           (a six-item; scale)
      .  B.  Life Risks |;   •
           (a four  item scale)

      BEHAVIORAL INTENTIONS'
        A.  Actions    r
           (A Scale composed of
            the following four
            scales)   ;
     •   B.  Talk  _    >   •'   .' ' '
           (a three-item scale)
        C.  Personal   /
           (a three-item'scale),
        D.  Others    ••!;
           (a three-item scale)
        E.  Political j:
           (a five-item scale)
  .  .INDEPENDENT VARIABLES  '
     ' ••' '   ' \   '  '  -•   •'  ,••-••  '•
  PRIOR ATTITUDES &  INVOLVEMENT
  A. PriorjConcern  (one item)
  Bi Prior[Sickness (two item
           i   . .       scale)
  C. Know Chemicals (two item
  -•   -   '  ' !  .        'scale)
  D. Prior [involvement
       (a  tWjo-item scale)
 TRUST
  C.
  D.
Trust [Locals
 (a tWo-item scale)
Trust povernment
 (a si^x-item sea 1 e)
Trust [Science (one item)
Possible to Control,  '
 (one [item)
 DEMOGRAPHIES
  A. Female,
  B. Education  .
 .'C. Age "  :[ '.     "  .   '.
  D. Income!          •      •
  E. Attends Religious Service
 .F. Married   .          •
           1       •'       " -; •
 RISK COMMUNICATION ANC
.MANAGEMENT;OPTIONS
  A. Uncertainty Disclaimer
  B. Enhanced Citizen
   -  Participation
  C. Indemnification"';
  D. Standard Procedures*'
                                                  ••I '••

-------
Summary    '                 •          '                             -. •   '

This  section has  dealt  with  the  following results:  :

      *levels of concern are  primarily a function,of a person's
      'assessment  of  the likelihood  of health, problems developing
      as  a results of  living  near  the'site;

      *levels of .conc.ern also  relate significantly  to  trust in       \
      government,  industry,  and  technology;

      *the variations  in messages  given the respondents had no impact
      on  levels of. concern nor on  the health risk  assessments;       ,

      *more accurate  risk assessments and comparative, risk assessments
      are strongly related to trust in institutions;

      *levels of concern,  health  risk assessments,  and comparative: health
      risk assessments  each  is able to account fo.r substantial variance
      in  intentions  to  take  actions to deal  with the  hazardous waste
      threat;  and            '   .                 .          ;

      *the uncertainty disclaimer  is slightly,  but  significantly, related
      to  the degree  that  living  near, the site is viewed as less
      threatening than  other  risky  activities.

This  section has  described the relationships among the key variables in
this  study.  Trust in  institutions  is strongly related to estimates of
health effects and estimates  of  health effects are strongly related to
concern.  However, this section  has not included attitudes citizens-held
before they learned  about the hypothetical.site, nor  has it included the
demographic variables  used in earlier sections.   The  next section
employs multivariate techniques  that permit  a sorting out of the
relative  importance  of  each  set  of  factors,  including prior attitudes
and demographic variables,, in explaining consistency  between objective
.and subjective risk  assessments.

-------
       Table XI-6: Correlations  of  Behavioral  Intentions' with Selected
                  .Measures of Levels  of  Concern,  Health 'Risks,
                   Comparative Health  Risks, and  Trust,
                1
                2.
                3
                4
Levjel of  concern, -after  the  cleanup  '   .
Probability of  illness,  after  the cleanup
Threat from living near  the  site vs driving a car
Trust federal officials  to tell! the ' truth        :

•' !
GET MORE INFORMATION.
TALK WITH FAMILY 'AND FRIENDS

CONVINCE -OTHERS TO' USE ' •
BOTTLED WATER
SPEAK UP AT- A 'PUBLIC MEETING
.WITH EPA i 'OFFICIALS
MARCH IN A PROTEST, EVEN IF
IT MEANT BEING ARRESTED
CONVINCE OTHERS TO LEAVE
SERVE ON A COMMITTEE
GO DOOR TO DOOR TO CONVINCE
NEIGHBORS TO GET' INVOLVED
WRITE OR CALL MY CONGRESSMAN
TO GET EPA TO REMOVE WASTE
ORGANIZE A PROTEST MEETING

TESTIFY AT A LEGAL HEARING

BUY BOTTLED WATER' '
1: ' , . . .
INSTALL WATER TREATMENT
. . SYSTEM
MOVE OUT OF AREA J, . '
SUE . i;

n t- /-»«*• -i MM «•• «-u . i_j 	 L • f • . ^ • .
      ^   __  _—  _. ,.„_.  ,^,l.«,sul=llu upuionsj tne inaemnitication option is
positively  related  to  serving on a committee (0.16).  One explanatibn
for this result  comes  from one respondent who said |  "If they are willine

the nrnbT  ^ ^^'S >*• v,''  "" What they Said itl the °^^l notice
the problem must be horrible."  For some people, the willingness to
provide  indemnification did not provide comfort, but communicated a
message  that the danger of living near the site wa«  indeed'-high.

-------
 completion  of  the  cleanup involve bel.ieving that the federal officials
 are  telling the  truth and thinking that it is possible to manage'a
.situation like this  one with little or no threat to public - health.  ' This
 is  logical;  people who think that government is lying to them and  that
 the  problem cannot be managed without threatening public health have
 little  reason  to conclude that the cleanup reduces their risk of health
 problems  from  living near the site.

 As with levels of- concern and health risk assessments, the managraent
'options fail to  account for  any variability in comparative health  risk
 assessments.'  The  uncertainty communications variable, however, does
 correlate slightly with most of the comparative risk assessments.        :
 Having  the  uncertainty disclaimer seems to reduce the degree to which
 living  near the  site is seen as more threatening than the other  '  ,
 activities.  The strongest correlation coefficients are wi'th living near
 the  site  compared  with smoking cigarettes: -0.18 (before the cleanup)
 and  -0.17 (after the cleanup),- both significant>at the .01 level.   The
 three  other significant coefficients are -0.12 with driving a car,  after
 cleanup;  -0.14' with  having a .x-ray, before the cleanup; and -0.13  with
 eating  sprayed fruit,  before the cleanup.  This suggests that providing
 the  disclaimer has a slight  impact on improving consistency between
 expert  and  popular opinion in comparative risk assessments.   Section
 XII, the multivariate analysis, provides a more detailed assessment of
 the  impact  on  this factor.                                          ,

 Behavioral  Intentions-                                            '

 Different behavioral intentions correlate with many individual measures
 of  Levels of concern,  health risks, comparative health,risks, and  trust.
 Table XI-6  shows the single  measure of levels of concern, health risks,
 and  comparative  health risks, respectively, that in general correlates
 most highly with the behavioral intentions.  In other words,: rather than
 present the correlations of  the behavioral intentions measures with all
 ten  trust measures,  the table includes only one trust variable, how much
 can  federal officials be trusted to tell the truth.  This variable wa-s
 chosen  because it  is the trust variable that correlates most'highly in
 most cases  with  the  behavioral intentions.measures.

 Simple  bivariate measures of levels of concern, health risk assessments,'
 and  comparative  health risk assessments account for substantial
 variation in respondent intentions to take action ,to deal with the-
 problem.  The  relationships  are strongest for. behavioral intentions
 involving private  reactions  to the problem (e.g.., using bottled water or
 convincing  others  to use bottled water) and slightly less strong for
 political responses  (e.g., contacting Congress).  Trusting federal
 officials is related to behavioral intentions although the correlation
 coefficients are much lower.

 Receiving the  uncertainty disclaimer seems to slightly decrease the
 likelihood  of  certain political reactions as'evidenced by correlations
 of  -0.17  with  protest marching; -0.1.6 with contacting Congr-e.s-s; and
 -0.15 with  litigating.  Some behavioral intentions are also slightly

-------
                                     •63

                       smoking is  the lo.e.r
                                                     oio.c
       Table XI-5:
Correlations of Change  in  Comparative' Risk   ''<
Assessments with Trust, After  Cleanup Completion

                             va  smoking
                             vs  near  nuclear  plant
                             vr,  driving  a  car
                             va.  riding a bicycle
                 1 - Threat of living near  site
                 2 - Jreat.of .living near  site
                 J -• Threat of living near  site
                 4 - Threat of living near  site
                      ,              —          —  .-.  — ..__»» Oa>fc/i^J''— •*•"   '
                 5 = Threat of living near site W  annual chest x-ray"
                 6 - Threat of living near site vs|  eating sprayed  fruit
      LOCAL  OFFICIALS!'TELL TRUTH
      STATE  OFFICIALS': TELL TRUTH
      FEDERAL OFFICIALS.TELL  TRUTH '
      POSSIBLE TO CONTROL  PROBLEM
      ENVRNMNT SAFER  'THAN  30  YRS AGO
      FEDS MORE HONEST THAN 30  YRS AGO
      STATE MORE HONEST  THAN  30 YRS AGO
      SCIENTISTS KNOW!HOW.TO  PROTECT US
      INDUSTRIES CONCERNED WITH HEALTH
      STATE & FEDS CONCERNED W/  HEALTH
(1)

.23*

.24*
.24*|
.20*1
(2)

".15*
.16*
.21*
.17*
.14*
                                     .13*
                                                                   (3)
                                               .20*
                                               .18*
      LOCAL OFFICIALS flELL TRUTH
      STATE OFFICIALS JTELL TRUTH
      FEDERAL OFFICIALS TELL TRUTH
      POSSIBLE TO CONTROL THE PROBLEM
      ENVRNMNT SAFER-THAN 30 YRS AGO
      FEDS  MORE HONEST  THAN 30 YRS AGO
      STATE MORE HONEST THAN 30 YRS AGO
      SCIENTISTS KNOW HOW TO PROTECT US
      INDUSTRIES CONCERNED WITH HEALTH
      STATE &  FEDS  CONCERNED W/ HEALTH
The strongest relatioiships  between  trust  and  changing opinions" with the
                                               69
                                             (1) ,
                          -15*
                                     141
                                     (2)
                                             119
                                    .20**
                                    .16*
                                    .14*
                                    145
                    10.9

                    (3)

                    .15*

                    .1.6*--
                    .22**
                    .16*


                    .15*.
                                                                  135

-------
                                   •D,
      Table  XI-4:  Correlations of Comparative Risk Assessments with
                  Trust,  After Cleanup            '       • , '  -
         1
         2
         3
         4
         5
         6
Threat
Threat
Threat
Threat
of
•of
of
of
Threat of
Threat of
living near site vs
living near site vs
living near site vs
living near site vs
living near site vs
living near site vs
smoking  •     ,
near nuclear plant
driving a car
riding a bicycle:
annual chest' x-ray
eating sprayed fruit
     LOCAL  OFFICIALS  TELL  TRUTH
     STATE  OFFICIALS  TELL  TRUTH
     FEDERAL  OFFICIALS  TELL TRUTH
     POSSIBLE TO  CONTROL THE PROBLEM
     ENVRNMNT SAFER'THAN 30 YRS AGO
     FEDS MORE  HONEST THAN. 30 YRS AGO
     STATE  MORE HONEST  THAN 30 YRS AGO
     SCIENTISTS KNOW  HOW TO PROTECT US
     INDUSTRIES CONCERNED  WITH HEALTH
     STATE  &  FEDS  CONCERNED W/ HEALTH
     LOCAL OFFICIALS TELL  TRUTH
     STATE OFFICIALS TELL  TRUTH
     FEDERAL OFFICILS TELL TRUTH
     POSSIBLE TO CONTROL THE  PROBLEM ::
     ENVRNMNT SAFER THAN 30 YRS AGO
     FEDS MORE HONEST THAN 30 YRS  AGO
     STATE MORE HONEST THAN 30  YRS AGO
     SCIENTISTS KNOW HOW TO PROTECT  US
     INDUSTRIES CONCERNED  WITH  HEALTH
     STATE & FEDS CONCERNED W/  HEALTH
                                 (1)

                                -.28****
                                -.26****
                                -.20**
                                -..'11*
                                •.20***
                                -.12*  '
                                ... 19**

                                (4)   ,

                                ..21***
                                .. 25****
                                ..25****
                                -.14*
                                •. 14*
                                -.21***
                                •'.16**
                                    (2) ,

                                   •.2 5****
                                   -.21***
                                   -.19**
                                   -.12*
                                     11*
                                     16**
                                           (3)

                                           .23***
                                           . 28****
                                           . 27****
                                           . 14*
                                           .14*
                                           .11*   -
                                           .15*,"

                                           .18**
                                           .19**
                                    24****
                                    14*
                                    18**
                                    17**
                                    11*
                                    1.9**
                                    11*
                                          -.17**
                                          -.17**
                                          -.17**

                                          -.12*'
                                          -.15*
The strongest relationships  are  between  the  comparative risk assessment
from living hear the site and  trusting government  officials  to tell the
truth.  In comparing living  near the  site  with  the other activities,
people who trust the government"  to  tell  them Che  truth perceive the risk
from living near the sl^.2 to be  less  dangerous  than do the people who do
not trust government.  This  holds across all six  comparisons,  whether
voluntary activities such as smoking  or  relatively involuntary ones such
as eating fruit sprayed with pesticides.                 ,    ^

How effective one views the  cleanup in reducing the'risk of ."living near
the site in comparison with  other risks  is also a  function of  trust in
government.  Table XI-5 on the. next page presents  the  correlation
coefficients between changes in  risk  assessments  and the trust variables.
Change is categorized as either  none  or  some improvement in  the

-------
                                    ••66
  adequately  protect.u-s;  every trust /item ,significantly correlates wich'
  health risk assessments.. .People who trust institutions and think ^  is '
  possible to manage  the  site  without  risking the community's ''health'have-
  risk .assessments  similar  to  those ,of the  experts;.   Individuals with low
  trust _ considerably  overestimate., the.  risks from living near  the site  '•
  Individuals with  higher trust  are also more, likely, to substantially
  reduce their health  risk  assessments with the Completion of the cleanup-
  individuals with 'lower  trust'are less willing tojbelieve that the
  cleanup reduced;their risk from  living near the site.'
                  '•  -: !•'   .---.--.-.', :- ••„•';,'..   \,  •'.":'.-,'   •   ••''"' •.
  Table XI-3 includes  only  the results for  the  question of•the respondent '
  suffering health  problems, not his or her-familyj   The results for the
  family question are  almost identical so they  arefnot  reported"      '   '
       Table XI-3:  Correlations of Health Risk Assessments  with .Trust-
                     Prjpbability of illness before  cleanup
                     Probability of illness after .cleanup         '
                     Decrease in probability of illness with  the  cleanup
                                                        (2)
(3)
LOCAL OFFICIALS TELL TRUTH
STATE OFFICIALS TELL TRUTH
FEDERAL OFFICIL.Sl'TELL TRUTH
POSSIBLE TO CONTROL, THE PROBLEM
ENVIRNMNT .SAFER .'THAN 30 YRS AGO
FED'S MORE HONEST;, THAN 30 YRS AGO
STATE MORE HONEST' THAN 30, YRS AGO
SCIENTISTS KNOW HOW TO PROTECT US
INDUSTRIES CONCERNED WITH HEALTH
STATE & FEDS CONCERNED WITH HEALTH
' , 1 ! ' '
i ' , •
r ...
-.20** . -.30****
-.27**** -.38****
-.19**
-.32****
-. 17* r -.29****
-.23*** -.25****
-.15* -.16*
-.19** 1 - -.20**

-.14*


-.20**
-.21**
.16**
.21***
.21***
.21***
. . 14*


• -
• 1.3*.
..21*'*".;
          ,         ' .   I    •   ' , "        ' .    '      *    '-'•','
 The uncertainty-risk  communication  and' the  managment option variables
.have no direct impact ion health  risk  assessments.]  Regardless of whether
 the.respondent received the uncertainty disclaimer  or which of the three
 management options were presented,  there  is no impact on health risk
 .assessments.                                                   ,
                       ['    :         '       •   . ,   • _ 'E1  .-,   • •    ,  _ \ \       • '.

 Comparative Risk Assessments                      !
                ~~~—~fT"~	~   -•    '•        -•• f •/-"•••'.•  - ..-   - ,•  '  • , •

 Trust not only relates to assessments'that  one will  become  sick  from
 liv.ing^near, the site,  ;but also, with the comparative  risk, assessments.
 Table XI-4  presents  the correlation coefficients cif  the com'parative risk
 assessments with trust after the cleanup was completed.   Coefficients
 tor before  the cleanup are similar, but slightly lower  due  to the
 reduced variance in  thle comparative risk assessments,which  was explained
 earlier.           •    ,; '• .       ,   :              . i                  , .

-------
      Table XI-2:  Correlations of Levels'of 'Concern'with/Trust'
                     = Level of concern- before cleanup
                     = Level of concern after cleanup    •  ,
                     ^Decrease in concern .with - the cleanup
                                         (1)
(2)
(3)
LOCAL OFFICIALS 'TELL TRUTH
STATE OFFICIALS TELL TRUTH
FEDERAL OFFICIALS TELL TRUTH
POSSIBLE TO CONTROL THE PROBLEM — .
ENVIRNMENT SAFER THAN 30 YRS AGO "-.
FEDS MORE HONEST THAN 30 YRS AGO
STATE MORE HONEST THAN 30 YRS AGO
SCIENTISTS KNOW HOW TO PROTECT US
INDUSTRIES CONCERNED WITH HEALTH -.
STATE & FEDS CONCERNED WITH HEALTH
16*
17*,
16*
17*
23***



24***

-. 29****
.•- — . 34****
-.32****
-. 25****
— .36**** •
. -.14*
-.20**

- . 20**
-.19**
. 28****
. 3 3****
•-.31 ****'
.18** .
.31****
.17**
. 20**

.13* . ...
.23*** .
The dimensions  of  trust  discussed  in Section VI  include the question of
whether scientists know  enough  about the impact,  of chemicals on our
health to adequately, protect  us.   This  variable  stands out in Table XI-2
by its failure  to  correlate with level  of concern.   The general question
of what scientists know  simply  is  unrelated 'to how concerned citizens
would be living within one mile of the  site.   On the other hand,  level
of concern is related to judgments of whether it is possible to manage  a
hazardous waste site  safely.  These "two findings,  alou^ with the strong
relationships between trust in  institutions and  lev.el of concern,
suggest that the respondents' evaluations are driven r je by concern
with institutional capabilities than by concerns with the limitations of
science.     .                                     •

Finally, reg'arding bivariate .relations  with level  of concern,  the impact
of the uncertainty communications,  (uncertainty disclaimer or not)  and
the management  options (standard,  enhanced participation,  or
indemnification) is nonexistent,   there is no relationship between which
uncertainty communication was given,  nor which management option was
chosen, and level  of  concern  (See  Appendix A for the text of these
messages).  Instead,  consistency between objective  and subjective risk,
the lower levels of concern,  relates  strongly to lower health  risk
estimates, to lower comparative risk estimates,  and to trust in
institutions and in the  belief that  it  is possible  to clean up the site
with low risk to the health and /safety  of the community.

Health.Risk Assessments         .....'         -'.'•-.•        ;

Table XI-3 demonstrates  that  how people assess the  likelihood  of'their.
becoming sick from living near the site is -strongly related-"to the trust
measures.  With the exception of the  item that measures  the"belief that
scientists know enough about  the impact of chemicals'on  our health to

-------
      Table XI-1: Correlations' of Levels/ of Concern with Risk
                  Assessments
                     "Risk Assessments  Before Cleanup
      PROBABILITY OF ['HAVING
      PROBABILITY'OF!HAVING
      THREAT -OF
      THREAT OF'
      THREAT OF
      THREAT OF
      THREAT OF
      THREAT OF-
LIVING' NEAR.
LIVING'' NEAR
LIVING NEAR
LIVING NEAR
LIVING NEAR
LIVING NEAR
PERSONAL HEALTH PROBLEMS
FAMILY HEALTH PROBLEMS   >'
                     t

SITE VS SMOKING     ]
SITE VS NEAR NUCLEAR - PLANT •
SITE VS DRIVING A CAR,     ;
SITE VS RIDING A BICYCLE
SITE VS ANNUAL CHEST jX-RAY
SITE VS EATING BRAYED FRUIT
                      Risk Assessments After Cleanup
      PROBABILITY OF iHAVING PERSONAL HEALTH PROBLEMS
      PROBABILITY OF [HAVING FAMILY HEALTH PROBLEMS
                     5              >       '      ' '-I • • '        '  •
                                 VS SMOKING     ;
                                 VS NEAR NUCLEAR I PLANT
                                 VS DRIVING A CABl
                                 VS RIDING A BICYCLE
                                 VS ANNUAL CHEST |x-RAY
                                 VS EATING SPRAYE.D FRUIT

                             significant at the .J05 level
                             significant at the .pi level
                             significant at the .001 level
                             significant at the .0001 level

      (These  significance  levels  will  be used  in Jthe remaining
       tables  in  this.!, section.,  . The coefficients are tail's.)
THREAT'
THREAT
THREAT
THREAT
THREAT
THREAT



OF
OF
OF
OF
OF
OF



LIVING
LIVING
LIVING
LIVING
LIVING
LIVING
•{••;.
i"

NEAR
NEAR
NEAR_
NEAR
NEAR
NEAR
* ss
** 3
*** S
SITE
SITE
SITE
SITE
SITE
SITE
sigi
sigi
siei
! **** = Q 1 y.
. 4 6****
. 50****

. 29****
• 15*
. 31****
.2 7 ****
,27****
:19**   .
                                               .58****
                                               .59****

                                               .39****
                                               .29****
                                               .38****
                                               .37****
                                               .33****
                                               .22***
Although the. correlation  coefficients  are not aslhigh,  level of concern
is also a function  of  the level  of trust, respondents have in government,
industry, and the possibility  to control a problem like this one so that
it presents  little  or  no  health  threat to the community.   Table XI-2
shows these  relationships,  and reveals that  the deduction in the level
of concern brought  about  by the  cleanup is also ii function of trust.
Those ^.ndividuals whp  trust the  government,  industry,  and the -  "
possibility  of managing a hazardous waste site without  threatening the
health of the community have a greater reduction!in  their level of .
concern with living near  the site  at' the completion  of  the cleanup than
do less trusting individuals.

-------
      SECTION XI: EXPLAINING THE RESULTS - BIVARIATEtRELATIONSHIPS  ,

 The firsc ten sections of the report have explained the research,
 presented the measures, and reported the results in terras of
 frequencies, variations among,the six subgroupings,  and correlations
 with demographic varaibles.  This section is an examination of the
 bivariate relationships that may account for variance in four sets of
 variables:                              •             ..'-•'

      *levels of concern,  correlations with health risk assessments,
       comparative risk assessments,  trust, andAthe uncertainty
       communications and  management options variables;        ;

      *health risk assessments, correlations with trust, and the
       uncertainty communication and management option variables;

      Comparative risk assessments,  correlations with trust, and
       uncertainty communication and management options variables; and,

      *behavioral intentions, correlated with all the above listed     •
       factors'.                           '          "          ' ;   '   '

 The next section is a multivariate analysis of these factors along with
 prior attitudes and demographic attributes.  In this section the intent
 is  to examine simple relationships among factors measured after
 respondents  read the hypothetical waste site scenario.

 Levels of Concern                .         '

 Table XI-1,  on  the next page,  shows  that levels of concern are strongly
 related to  risk assessments.  Levels of concern are most strongly a
 function of  a person's assessment of the likelihood of health problems
 developing  as a result of living near the site.  Levels of concern are
 not just a  function of an assessment.of the likelihood of sickness, but
 also are a  function of how dangerous the respondent views living near
 the site in  comparison with other possible threats to a person's health
 and safety.                ...

 The strong -relationships  between risk assessments and levels of concern
 support the  arguments of  those who stress the importance of risk
 communications  to effective implementation of environmental statutes.
 When there  is a substantial gap between objective and subjective risk
.estimates,  the  result is  a level of concern inappropriate to the given
 risk.

 In  all instances, the correlation regarding judgments of the situation
 after the cleanup is stronger than the correlation regarding judgments
 before the  cleanup.  This results from an attenuation of correlation
 from the lack of great variance in levels of concern before the.cleanup;
 with most respondents reporting themselves extremely concerned,
 correlations involving levels of concern cannot attain high levels.
 With more variance in opinions after the cleanup, higher correlations
 are possible.       •                ,

-------
                                   5-2
                     •I !               '     '       . 'f            '
This initial description of the beh-avioral intentions  produces  several
conclusions:         ':•   '                  '•   .-   .- •.   " :.  •.

     *most respondent!?  say that they definitely, would  engage  in the
      traditional',  low-commitment activities of getting more
      information,  talking about the problem, and' serving  on  a
     ' committee;     >p                  .          I     •  .   ••

     *most respondents  say that they would definitely'use  bottled
      wate-r, and many',; say they would take other s^teps .to protect
      their families :such as installing a 'water t'reatrae'nt  system
      or even leaving1the area;             •     '
                     I!     -  '  .  ••   .; "•  -'"' '.''.' '   '['.:','-.' '  .'•'• ' ,   '
     *most people wou|ld try to convince others to get  involved  and
      to use bottled water;
                     I!-'.."

     *most would take!'political actions, with the1 most popular
     "option being  contacting their representative in  Congress  to
      put pressure  on;EPA;
     ^behavioral  intentions  fall into four dimensions  that  scale  in
                      (           •      •           - -/
      likelihood  from talking about the problem, to taking  personal
      steps to protect  the family's health, to-convincing others  .to
      confront the  problem,  and finally to taking political  actions '.
      designed to pud pressure on  the government;-

     *members -of  environmental groups are not morle likely than
      anyone else to  take action - when faced with a toxic.waste
      problem near  one's  home, almost everyone becomes greatly
      concerned;      |:        . •       '        •   ••!•., .  ' '.     .  '   • •  .  .

     *the less well educated and women are the most likely to drink
      bottled water and take other steps to prdtejct their health.,
      but they are.no.t  more  likely to engage in most other activities;

     *the political behavioral intentions rarely 'correlate with any o.f
      the demographic; variables, suggesting that other factors
      influence personal  decisions to participate1.
   -   -  '  '  '     -     !!.-••••...'•' '•-  :    .   'P  '  ••'.••  .'.     - ..   -
The next section  of this  report looks at how these "other factors" -
including uncertainty!'communications'and management options - account
for variance in key citizen  attitude's and intentions.

-------
Women and Che less well-educated  are  more  likely  to  react  by  ad ope Lug  •
these strategies designed  to  protect  the  family's'health.   Encouraging,
     Table X-3: Correlations  of  Behavioral  Intentions  with
                Demographic Variables.      .    ;

                         Age  Male   Relg  Marital   Home   Incm.  Educ
     GET MORE INFO          ''

     GET OTHERS TO USE        -.12*
          BOTTLED WATER
     MARCH IN A PROTEST,

     GET OTHERS TO MOVE       -.23**

     SERVE ON COMMITTEE

     GO DOOR TO DOOR TO       -.13*
        INVOLVE NEIGHBORS
     CONTACT CONGRESS TO
         PRESSURE EPA
     BUY BOTTLED WATER        -.17*

     INSTALL WATER
      TREATMENT SYSTEM
                           .13*
           -.11*
                                       -. 14*
                  .16*
               .16*

               .13*
     LEAVE THE AREA
-.13*
-.12*
-.11*  -.13*

       -.18**

       -.17**


       -.15*
                      * =  significant  at  .05
                     ** -  significant  at  .01
others not to drink the water  and to  get  involved  is  primarily
correlated with gfender; women  are more  likely  to  say  that  they  would
encourage others to drink bottled water>  to  get  involved,  and  even to
get others to move. .•              '        '-              '     '

The five political items all involve  a  public  commitment  to 'influence"
events at the site.  The'findings regarding  the  political  items are
different from those relating  to the  other activities  in  that  propensity
to engage in the political activities is  rarely  related at all  to  any  of
the demographic variables.  The three activities  related  to Meetings -
speaking up, organizing, and testifying - do not  correlate with any of
the demographic variables.  Contacting Congress  correlates .weakly  and
negatively with income and education; the lower  the  income and
education, the more the  intention to  contact Congress.  Marching in a
protest only correlates with one variable; negatively  with religiosity.
Attendance at church services  seems weakly to  depress  the  intention to
march.                      -  ,.  •                           "''-•;•

-------
among  the.  groups.   In other words-,  the similarities among  the  grouos,  far
outweigh the  differences  reported in Table X-2. | These results must _be
viewed with caution i.since the number of respondents in some  of the
groups, i.s  quite  small.    .'-"   .       -    .  '.  •[  • .
                    Jr •' '•-.'".  - ' t ••;'•.     .•''''  ' , '   '' 4  •' " .',','   " ' '    •'-,.'.
The decision  to  take  specific actions is -a function of one's  level of
concern with  the problem,  assessment of the. likelihood that  the  action
will be effective,  eind., the extent.one feels comfortable .with 'the
activity.  Some" individuals,, for .example, said tihat they would never
walk in a-  protest  msirch,  regardless of their level of concern.

The scenario  of  this' study -creates  a situation which, -as reported  in
Section VII,  would greatly concern  the vast majority'of•the  respondents.
In this situation  almo;st  everyone- becomes an emrironmentalist  concerned
with protecting  the |health and  safety of themselves and their families..
The great  variations  among activities appear to'be. a function  of  their
perceived  appropriateness -and effectiveness,  rat'her.; than previous
exposure to environmental literature or, general [level of. concern.     :'
Although,  as  reports^.in previous sections, the lless well-educated
express  the most  concern about the situation and the greatest
inconsistency with  expert risk assessments, they are not the most- likely
to engage  in  every  activity.   The activities that require working with
others in  an  organizj.ed manner (committee service; organizing meetings)
have a lower  level  of participation among  the leiss well-educated  than
among members of  the;-four groups who are presumably more used  to    -
attending  meetings  (environmentalists, elected officials, civic leaders,
and business  persons').   The respondents in the siample with graduate
degrees  also  are  less willing to engage in these, activities. 'In  light  •
of the fact that  the members  of the. first/four groups chose to join
organizations, perhaps it is  not surprising that: they look to  an-
organizational approach in dealing with the wastie problem.

The less well-educatied are the most willing to tiake individual'actions
to protect themselves and to  try to convince others to do so.  They are
the most, willing  tojbuy bottled water and  to flee the areai
Section VI  reports  that elected officials are dispropbrtionately
trusting  of EPA representatives. .This does not [discourage these  elected
officials,  however, tfrom contacting Congress to^pressure EPA  to remove
all  the waste  immediately.   Majorities of every[group, except  the
business  community  and the  well-educated, claim]that'they definitely
would  contact  their'local representative.  The less well-educated may
have doubts about engaging  in committee work, ..but. they have few-qualms
about  contacting Cbiigress.                      '[            .
                  - • - i      , '  •     .         -   .  ,t.  • •   •       \    ...
Different  dimensions!  of activity have different •demographic  correlates.
The  scale  that  involves getting more information, and serving on a
committee  correlates!, with home ownership.  People who own their own
homes  are  more  likely to state that they intendito engage in this  type
of  traditional  citizen activity.  Changing one1!! personal-behavior —
using  bottled water,;' installing a water treatment 'system, ' and  gven
leaving  the  area --^correlates negatively with .education and male  gender.

-------
 on Che POLITICAL scale, that' person very likely scored .'nighlv on  all' the
 other scales.   Conversely, if an individual did hot score/, highly  on  the-
 TALK scale,  it is 'unlikely that he or she would have claimed  intentions
 to change water uses, to convince others to make changes, or  to 'become
 highly active  politically.  These scales- will be used in  thev
 multivariate analysis of section XII.              .       •
      Table X-2:  Behavioral Intentions-Among Groups.

                 %• who would definitely .engage in the activity


GET MORE INFORMATION
TALK WITH FRIENDS
CONVINCE OTHERS TO USE •
BOTTLED WATER
SPEAK UP AT A PUBLIC
MEETING WITH EPA
MARCH IN PROTEST EVEN
IF ARRESTED
CONVINCE OTHERS TO MOVE
SERVE ON A COMMITTEE ,
GO DOOR TO DOOR TO GET
NEIGHBORS INVOLVED
CONTACT CONGRESS TO
PRESSURE EPA
ORGANIZE A MEETING
TESTIFY AT A HEARING
BUY BOTTLED WATER
INSTALL WATER TREATMENT
SYSTEM
MOVE OUT OF AREA
SUE
' N =
Envmt

91% '
83
44

39

9

13
52
23

52

17
26 ,
61
13

18
5
23
Elect

100%
95
•' 42

53

1.1

,0
58
26

63

32
. 63.
• 74
37

5
5
19
Civic

97%
77
39

52

13

16
58
27

61

29
42 '•'
65
26

19 '
10
31
Busns

' '91%
86
36

41

9'

9
46
10

. 4.6

27
41. •
59
29

9
0
22
' Lo-Ed
!
. 89%
85
54

39

• 7

11 .
37.
17 -

61
'
11
37 •
80
39

. -2$ :
13 .
47
• \
Grads

90S--
83
'38

33

12 -

•' 7 ' -
' 39.;
19

39

; 19
32
: 53 .
19

11 '
4. • .
.59
Members of environmental groups might.be  expected  to  engage  more  heavily
in activities to deal with the problem  than  others.   After, all, by
joining a local environmental group  people assert  their  concern with
environmental protection and their willingness  to  participate  in  group
activities.  These expectation's are  not borne out  b.y  the data  bfTable
X-2.  Variations are greater among the  items (e.g., intend to  sue)  than

-------
 concerned with, the situation.  Almost everyon'e  says  Chat  he  or  she would.
 talk with- others and ;.try to get more inf o-rmationj  A  majority  claim they
•would definitely 'change their personal habits byjd-rinking bottled water
 or installing a water treatment system..  Forty-five  percent  probably or
 definitely would leave the area.  Not only do these  citizens claim they
 would take personal action, most say they would  encourage others  to take
 action;  majorities would .urge their friends  to  use bottled water  and '
 would go door to door themselves to convince  their neighbors to get
 involved with the problem-.'  Forty percent say they probably  or
•definitely would try;:to convince their friends  to  leave  the,
 area.  However, these numbers probably overstate^ what, people actually
 would do since the costs of indicating a behavior  on a questionnaire,are
 considerably less than the costs, of actually'engaging in that  behavior.
 Besides talking about; the problem,  changing , their personal behavior ,  and
 urging others to act ,  ma jorities  claim  they definitely or probably would
 take political action t'p change the situation.  'Majorities checked that
 they definitely or probably would  contact 'Congrebs to put pressure on
 EPA, speak, up at a public meeting,  and  testify  ap a legal hearing.
 Forty-three percent Claimed they would  organize 'k meeting to protest  the
 handling of the situation and  thirty-four  percent said they would march-
 in a protest picket line, even if  it  meant being! arrested.
                     '!•     ,          ' •  -•          r '    •'-    ••'••.
 The only item listed! that failed  to generate  substantial support i.s  ,
 litigation.  Several ^respondents  noted  that  the problem with this option
 is that the scenario!rdid not clarify  who  could  be a successful target of
 litigation.  They said they were  not  opposed  to 'supporting litigation tp
 force faster action, !;but did not  see  how  this would be possible.,
                    ."'.-•            .              J.    - •     '•-.'.
 All of the items froijn' Table X-l were  submitted  tja a Guttman scaling
 program.  What emerged was four scales  that  covered the different
 dimensions of citizen actions:  •    '             [  ,       ' • .            ,
                     1 •      , ' •   '    ; •     ...     %  •....,'•
                     r        •.''•'        '''*'•' •>           ' . •
      *a three-variable scale (labelled  TALK) , that includes getting
       more information, talking about the  problem, and serving on •
       a committee;- these acts  do not require a public commitment or
       great, changes ;ir  _••'-•       '
| . • •'
ERS),
            .          ~ -- -       -    •         .    .  .     ..•..    •     .
      *anothef  three-variable  scale (labelled OTHERS), that, involves
       convincing  others  t'o: use bottled water,  get involved in the
       problems, and  leave  the area;              [
                      |. ' •              "    . • •    - '  •- -t   •        -•' '.
      *&  third  three-variable  scale (labelled PERSONAL) ,  that includes
       using bottled  water,  buying .a  water treatment, system, and moving
       out  of the  area; and                       I
     '  "      ••   '    £  -.  •    -   '    • •     • '  '   1    - -  ••• '.  \   -.  ' •
      *a  five-variable  scale (labelled POLITICAL):, that  involves  ,
       contacting  Corigress,  speaking  up at a publjic meeting, marching,
       organizing  a protest  meeting,  and testifying.            . ' •  .

 The scale  scores .from  these scales combine into |an overall-s-cale of
 behavioral intention^.   The order of the scales [is TALK, PERSONAL, ,
 OTHERS,  AND ..POLITICAL.-  This  means that if an individual scored highly
                    .  r   .•-•'.-              4.  :  -  ••".
                                                 '4?--

-------
          SECTION X:  THE HYPOTHETICAL SITE  -  BEHAVIORAL INTENTION'S

 When  faced  with a threat to one's neighborhood,  an individual may ace
 to remove  the  threat,- flee the area, or react  with some combination o)
 the two.  Table X-l presents the frequencies of  these behavioral   '•
 intentions.
      Table X-l:  Responses to the Question, "Here  is  a  list of things
                  that., people can do in a -situation like this.   For
                  each  activity, .select that option, that reflects how
                  likely it is that you would, engage  in  that activity.*

                    1=1  DEFINITELY WOULD DO THIS
                    2=1  PROBABLY WOULD DO THIS
                    3=1  PROBABLY WOULD NOT DO THIS
                    4=1  DEFINITELY WOULD NOT DO THIS

GET MORE INFORMATION
TALK WITH FAMILY AND FRIENDS
CONVINCE OTHERS TO USE BOTTLED
WATER
SPEAK UP AT A PUBLIC MEETING
WITH EPA OFFICIALS
MARCH IN PROTEST, EVEN ' IF IT
MEANT BEING ARRESTED
COVINCE OTHERS TO LEAVE AREA
SERVE ON A COMMITTEE
GO DOOR TO DOOR TO CONVINCE
NEIGHBORS TO GET INVOLVED
WRITE OR CALL MY CONGRESSMAN
TO GET EPA TO REMOVE WASTE
ORGANIZE A MEETING TO PROTEST
THE HANDLING OF THE PROBLEM
TESTIFY AT A LEGAL HEARING
BUY BOTTLED WATER
INSTALL WATER TREATMENT SYSTEM
MOVE OUT OF .THE AREA
SUE
*See the appendix for
(1)
. 87%
79
46

34

10

11
38 -
19

49

16

34
67
" 27
17 .
7
the exact
(2)-
11%
19
41

41

24

29
45
42

30

27

38
25
34
28
17
item
(3) •
2%
1
12

20

40

44
16
32

17

46

23
8
31
39 \
52
wording.
(4)
1%
r
i

3

27

14
2
8

5

11

5
1
9
13
25

N
173
173
172

,171 .

170

. 170*
,173
171

173

173

170
172
,170
168
165 '

            —        w  — -  —	— — — *• «•.—** w >**>_. v/fc. A.Lii_^i4o^.i_jr  O t  COtlCSlTTlj
these numbers suggest  that  most respondents think they would »be  quite

-------

-------
variance in the comparable risk assessments.  Although women  are
somewhat more likely to overestimate  the  threat  from drinking  the  water
vis a vis the other risks that are voluntary  and  the young  somewhat ,
underestimate the danger of.riding in  automobiles,  education  level
accounts for significant variance in  comparative  risk assessments.  As
education level rises, respondents are  less worried about drinking  the
water in comparison with the other threats.                 '.       .

This initial analysis of comparative  risk assessments produces  stv=r-il
conclusions:

     *there is a different underlying  attitude .toward risks |that      •  •
      are voluntary versus those  that  are involuntary;.
      (this simply confirms much  of  the work  done by Slovic  and
       Fischhoff)            ,                    -,-•-'•

     Respondents overestimate the risks,  of drinking the  water  from1.  .
      the site, especially before the  completion  of the cleanup;

     *many  respondents became less fearful of drinking  the  water
      after completion of the cleanup,  although  substantial.minorities
      still viewed drinking the water  as  more dangerous than  other
      activities;                                         , -'

     *environmental group members are  somewhat  less likely  to
      overestimate the comparative risk from  drinking the water than  are
      the rest of the sample; and             /              '

     Comparative risk estimates  are  strongly related to  education as
      the well educated are consistently  more likely to make  judgments
      consistent with those of the experts.        ~        :

-------
  impossible  to  avoid, _ The f irsc ',caie,: ,nc LudesUomoarxsoas' of.   <
  che water with  smoking,' riding  in a  car,-and riding a b-v- •=" -  b
  berore and  after  the  cleanup,  The- s,econd  scalfe  inclader = om^risons  -
  drinking the .water.with.having an annual  che-st! x-rair and' -a-'ine  -ui^
  .sprayed, with pesticides - again, .both before .and  after the" cleanup.  Yte
  variable that compared  drinking the, water  wittv living near a nil'"
•  power plant failed .to  fit,intp either scale,  ] ,  ' .          ™
       Table IX-3:, Correlations of' Comparative  Risk  Measures with
        •  '     :.   Demographic  Variables.    ... ].       .'.-'..      "•':
      -Before the.Cleanup
       Drinking the .water
       compared with ?!''
                   '  i    *
     '.  SMOKING CIGARETTES  '
                     I'.    :,
       LIVING NEAR NUC PLANT .

       RIDINGXA CAR DAILY  '
                     I -     '
                     "},] -, ' -  .-, , ,4 V ,

     .RIDING A BICYCLE DAILY

      ANNUAL X-RAY -  !

      EATING FRUIT '  j.  ' ''<-

      After  the .Cleanup   .-—
              .  '     - |  . •• , :  ' .  :
      SMOKING  CIGARETTES
               Male ...

               -.22**
Age
              "-.15*  '-.,10*
                  "v, '       - "

              -.15*  ' "'-••

              -.15*
( -        i . • , -  -   I       ,
jMarital'  Incm  • Educ

;t .». ,-     i ;•;'.-; 16** -.'35****




        ,  ,     •   -.29****

                 '...'-. 34****

                  -.15**   '

               .  • -.27****
      LIVING NEAR NUC  PLANT

      RIDING A CAR DAILY

      RIDING A . BICYCIE DAILY

      ANNUAL XrRAY

      EATING FRUIT
the coefficients  are
Religiosity  and home
correlated signifi
              -.16*
                                                            -.. 14*  -. 32'****.- "
                      -.12*
              -. 14*
                                .•12*'
                        '-. 31****

                        -.34****

                        -.22***

                        -.29*,***
                      * * significant at .05
                     ** = sifnificant. at .01 ,
                        = significant at .001.
                   ****=• significant at .0001,
  ^tau.'s,  the  ordinal analog o|f Pearson',s  r.
  ^ownership are  not  included [above because  they  never
cantly with any comparisons.  !   	
Table IX-3 illustrates, the -ability'of ..education [level-.to', account' for
                     f •*'".'-. ^  '...       •    •  * -','. („ ."  '.---,.''• , . . ' . i , ,

-------
     [able  IX-2 (continued)  Comparative Threat Perceptions    ,

                •   Envmt •  Elect   Civic   Busns   ' Low Ed  ., 'Grads
            Drinking the Water vs Having One X-Ray Per .Year
                    22%
                     8
MUCH MORE
SOMEWHAT MORE
SAME           22
SOMEWHAT LESS   9
MUCH LESS '      0

          N =  23
         MD =
32%
16'
26
16 '
11 '

19
- 52%
 26-
 10
 10
  3
V
 31
14%
46
23
 9'
 •9

22
 •53%
 23
, 11
  4 r
  9,

 47
 31%
 33
 '22'
 10
;" 3

 58
  1
       Drinking the Water vs Eating Fruit  Sprayed with  Pesticides
MUCH MORE
SOMEWHAT MORE
SAME
SOMEWHAT LESS
MUCH LESS
N =
MD =

5%
23
50
14
9
22
1

21% .
,11
v42 -
16
11
19


16%
19 ,
45
.7 V
.13,
' 31


9%
'27 '•_•
46
9
.9
22

\.
21%
"43 ' -
30
4 "
2
47


7%' , -
.19
46 •
18 •
1 1
57 ' •
-' : 2'' '

The less-well educated consistently  overestimate,the threat from
drinking the water at the  site.   They  stand  out  in the degree to which
they find drinking the water more threatening than anything,,  even  •
smoking two packs of cigarettes  each day.  On the other hand, ;the risk
assessments of-those with  graduate-degrees are close to the risk ./
assessments of experts who see  little  danger in  drinking the  water.     ;

Among the environmentalists, elected officials,  civic activists, and the
business community, one might have expected  .the  members_of environmental
organizations"to be .particularly- concerned with  the environmental
problem and, therefore, overestimate the risks posed by the^'site.  This
is not the  case, however,  as the'environmentalists do not differ greatly
from the others.  When  they do/differ,  as with the comparison of
drinking the water with' riding  a car,  the environmental group members
are mora consistent with  expert opinion than are other Centre Region
resident*.        •                     '             .      ..'',-..

Among the other  three  groups,  the civic activists are most likely, to
overestimate the danger of drinking the water.  This group would stand
out even more  if civic  activists who are also environmentalists were
removed from the analysis.                                   ,            ,

When Guttman scaling  techniques are used to  examine these...comparative
risk variables  no  single  comparative threat  dimension  is  found*
Instead,  two scales  emerge: one composed of  voluntary  risks  and a second
composed  of potential  environmental' risks that are difficult or

-------
Table IX-2: Comparative Thr-at  P»r-on^-'  '             •
 ;           Che Cleanup.        ^epcions  Among  Groans
                                                 Before
               .Elect,,  Civic    Bu^L    LOW -Ed   -Grads',


Drinking the Water vs Smoking Two. Uk.  a  Day'
' MUCH; MORE . ].,.Q%
SOMEWHAT MORE i 13
.. ' SAME .. . j 9
SOMEWHAT 'LESS ': 35-
MUCH LESS " L44
i ' '
• - N = 23
MD = r

..«' • 16% . oj[--..- 28%
J 16 • 9 , ,o
' 5 " 3' ' ' 5 .21
« ' 11 - ^ •'"
Jp : 59 .' 13

.. 19


31


22
1
-•

47 ..-

-3S
- - . 7
••' • 12 .
•29 !
48

58
1
Drinking the <;Water vs Living Near a Nucl '
MUCH MORE 126%
SOMEWHAT MORE 'J35
SAME : <26
SOMEWHAT LESS • ti 9
MUCH LESS , 'i'4
' i -
N = 23
• MD = r
' |M
Drinking the Water
:' MUCH MORE in
SOMEWHAT MORE 2fc
SAME :4
SOMEWHAT • LESS 39
rMUCH LESS . |:3
'• 'N - 2,13
Drinking, the
MUCH MORE - 302
SOMEWHAT .MORE '2''
SAME ,"
SOMEWHAT LESS 26
MUCH LESS 17.

N = 23 '
MD a
42%
21
16
16
5
19
29%
29
26
16
0
31
41%
14
36
5
,
22
40%
32
• • 17
9-
,
' ' . , 2 v
- '; 4?'
• . , ^ . .
vs Dri<
21%
26
• 11
16
26
19
Water
22%,
17-
n'
28
17.

is
i
; (continued



i . • '



; , ''
^ing a Cai
39%
19
.
19
• .19
' 31
vs 'Riding
40%
20
3
13
23

30
. 1 ...
on next



-
r for Tw
.
23%
23
9 •
18 :
, 27
22
•
o Hours E
47%
23
- 13
.11
.-'•••' O
47
idli U
21%
38
21
V 2~
58
•_'-.!

- 17% .
24
5
31'
24
59.
a Bicycle Daily
32%
14
18
18
18.

22
page)

,
- •
(
57%^
30 '
4
• '' 7
' ' 2

46
|

' )' •


. * '
20%
17
12
31 .
20

59



- . - '


-------
What 'is startling in Table IX- 1  is  the  enormous  overestimate  or
threat from drinking the . water  in comparison with  the  voluntary ^
activities of smoking, riding  in an automobile,  and  bicycling.   ,ne
information provided in  the water advisory  notice  iSee Appendix A) .seems
not to be reflected in citizen  responses  to' the  comparative  threat
questions.  Onlv one-third of  the sample .checked chat -drinking  the  water
is much less dangerous than  smoking two packs  of cigarettes  .each day.
Although  the table  showed  that  the  threat of  death from riding  in a  car
is ovef twelve  times  that  from drinking, the water '.bef ore the cleanup   • .,
overLlf the sample  reported  that  driving a  car is  less dangerous  than
drinking  the water.          :  .   •'••.'•              ,

Before the  cleanup,  two-thirds viewed drinking the water as  more
 threatening than living  near a nuclear power plant.. After the cleanup,
 opinion Same  equally divided with most repondents finding-one on y
 "somewhat more  of  a threat"  than the other and 26 percent -ewing t hem
 as  equally  threatening.   These moderate views suggests little dread of
 either drinking' the water after the cleanup or, living near a nuclear
 power plant.                                     ,
 drinking the water is viewed as quite  threatening,
 that many respondents perceive eating  fruit  treated  with  pesticides  as
 quite dangers.  From'some comments made  during  the oral -terviews  it
 is apparent that not all  citizens  are  aware  that  most  fresh   fruit
 available in the stores has been  treated both .with f
 cost-harvest preservatives.  Many  people may not  make a
 oecween  toxic  chemicals  from waste sites and pesticides.

 The  comparative  threat  perceptions of  drinking the water  .changed/for

                  ^
  smokin  as more threatening than drinking the water even before the
  cleanup, there was less room for improvement for post-cleanup  .
  perceptions than there was for the other less ^eat*m
  Sost of the changes are slight improvements in the "^
  perception of drinking the water, i.e., from "much more of^a threat
  "somewhat more of a threat" or from more of a threat to  the same.

-------
                                                           fro.-
                                                          °ne
   ,      ,    ---:-'. w<=  =11 so  asKed
    a health and:;safetv  from'driving the ia
-------
-he
    "'our health risk estimates (that you and your tamily wou.i surrer .....;
       problems before and" after ..the cleanup) form a single uu.-.san
        This constitues a single measure of  Che consistency Between
 xoert and oooular opinion; the higher the. scale score, . thj = greater ,,e
Is-i-nated likelihood of sickness, and the. greater the -gap  Deepen
popular and expert opinion.  This scale will be part, or che.-ulEw.riate
analysis of section XII.            .

Health risk assessments do not correlate  at  all with any cbmmon
                    lf«
                                               The second
                                                                     th«
                                                              -
 Before examiniag the impact of variations of ^j^/^j6^.^08
 Secpion XI,  we can make some conclusions regarding health risk
 assessments:                          .                        ,
      *over one-third of the entire sample, and        1
       of the less educated, report an extreme probability  that
       health problems would result .from living near  the  site
       before the cleanup; •

      *with completion of the cleanup, half the, sample either  reduced
       their health risk assessments  or were  not  too  concerned even
       before the cleanup;                        ,               .

      *environmental  group members  make similar assessments to those
       of  the rest of the population;

      *Che business  community has  the lowest .health problem estimates
       so  it  is  closest  to  matching expert .opinions;  and

      *women  and the less  educated are, substantially more likely to,
       overestimate  the probability of health problems.

-------
      •Table, VIII-2:,Responses, to  che  Questions ,]' "How arpbable'do vou"
                    ••think.it as  that  you  would .[suffer health problems
                    ,;:as a  result  or  living near [this site, crior to the
            '  -  .   ' .! cleanup attempt  (after. EPA [had completeTThe '
                   •-I/cleanup)?   _               ^ •.   .     •
       EXTREMELY
       SOMEWHAT
       NOT  TOO
       NOT'AT  ALL
;Envmt'

;.' 22%
 .65
•'  13
                              '• '  Prior.:to  the  ClJeanun     :
                             Elect '  Ci-v,jc •  .Busns    Low Ed
.  37%
  42
  21
   0
 100%
                N  =
      EXTREMELY
      SOMEWHAT
      NOT TOO
      NOT AT ALL
              v N
   23
 1 *-
 1
" I •
 ;J
 i1 : '•••
•in?



160"%

 ,23
 39%
 52
 10
,  0 .
100%

 31
.  18%
'.»|
  27j
   0[
100?
                                              22 L  .-' ,, 47
                                 After the Cleanup
          5%
         58
         37
        	0
        100%

         19
          16%
         .58
          19
        	7
        100%

         •31
          5%
       •JSl
        1001
                                              22 '[•
          26%"
          64
          10
           0
         100%

          47
 10%
 46-'
 39
	5
100%

 59
 th« tJT«       ?!1"   *rS' Wh° "" >e  "SUme^ C° b? bett"    orne
 F^tor %r "ge C1^zen'"are somewhat below  the  average  on the -dread
 factor,  the "extremely probable" response, before the cleanup.  .After
 the  cleanup,  they -closely resemble the sample as  a whole.   Although  they
 overestimate^ risk pf health problems, they  are 'not  alarmist In     Y'
 comparison with the i general .population.         I         larmisuin

 The  group  whose subjective assessments are most] consistent  with expert
 opinion  is the  business community.. One-quarter \ of  thes.e  private  sector
 managers,  administrators,  and owners thought thkt  the chances of  health

                                         the clenup-  Af ter
of !^"P "ntr"JKwi;^ ^e b^iness community iJ the less educated group,
?L«S S  :"       ~8h 36 ?erCent °f th.'M«pU!o£ less educated people
lowered _ their assessment  of the probability .of Health problems with
completion  of the. cleanup,  these probability assessments of the '
  c-°nsf^er_ the cleanup are higher than the probability : assessments
       rh   Juslness;; "^^^y  and  the highly educated for the situation
         ^'JT^'J   '^  ^  cleanuP>; 90 Percent of the..less educated
         that health-problems would be "extremeljy" or "somewhat"
probable.   .        ;|:          : _       , •..:'.•' ,L  .,....,   -.- i  • •   •.  .

-------
      SECTION  VIII:  THE HYPOTHETICAL SITE -'HEALTH- RISK'ASSESSMENTS  ;  ,

After reading the information ab'ou.c the hypothetical Superfund' sire,
respondents  immediately recorded their levels of concern and. their,"
assessments  of  the  probabilities that they or/and their families would
suffer  health problems as  a result of living near the site both before
and after  the cleanup.          ...                   -
     Table VIII-1
Responses to the Questions, "How probable do-you
think it is that you (your family members') .would,
suffer health-problems as a result of living hear.
this site prior to .the cleanup (after EPA had
completed the proposed cleanup)?         >. -
EXTREMELY
SOMEWHAT
NOT TOO
NOT AT ALL
N =
You
Prior
40%
46
14
0
100
174
You
.. After,
14%
55 - •
28
, 4
100 ,
174
Family
Prior
44%'
46
•• 10
0
100
'174
; Family
After
' 16%
; 54 '
' - 26 "
5-
100 ••-•'.
174
Completion of  the cleanup  has  a  significant impact .on reducing
assessments that health  problems  would  result  from living near the site.
Forty-one percent of  the sample  reduced their  assessments that they
would suffer health problems,  and 44'percent reduced their assessments
that their families would  suffer.   Most of  this  improvement comes from
people moving  from the "extremely probable", response instead of from the
"somewhat probable" response  to  a-lower probability. ..If the,"extremely
probable" response, is viewed  as  respresenting  a  dread factor',  completing
the cleanup reduces that factor 'from almost half of the sample to, at
most, one of six persons.  While  an impressive improvement,  these
changes do not connote-consistency between  objective and subjective risk
assessments.   Scientists would evaluate the probability of health
effects from living near the  site after the cleanup, as "not probable at
all," and certainly not  higher than "not  too probable."  Two-thirds of
the sample disagree,  placing  the  probabilities higher. ,,!•''

Table VIII-2 presents results  from the  six  sub-samples of Centre  Region
citizens .regarding the individual's assessment of  the probability that
he or she would suffer health  problems.   Because results from the
assessment for the family  are  similar,  we report only the results for
the respondent's own health.                               ,.-.,'-

-------
•uch.aY3it,e warrants extreme  concern!
                 • i "       '      /
   .	—~ n	-—^-	:	

    Table VII-2:  Responses of  the Groups
                                                              i5
                                                ?na« ..»*«  living  aear
                                            to trie  Question   "
                                           abouc -"-
       EXTREMELY
       SOMEWHAT
       NOT TOO
       NOT AT ALL
  measurng this concern  '
 Th6 only demographic [variable  that  correlates with 1^ i" e  •
 living near the siteibefore  eieamin  i.       -^ P  level  of Concern  with
 level of concern ,and jageL  sH           :   ^[ relationshiP Between
'The  initUl analysis ^ .lwil, o,

     'there is . broaji consensus 0£ great  concerj before' the cleanup;

                                                        af Cer
     •==.2:
      aphic  variables Suggests that



-------
        SECTION VII: THE HYPOTHETICAL SITE - LEVELS OF CONCHA  ••

Twe Hrsc question asked after respondents read  about  the-hypothetical
Suoerfund site was, "How concerned would you be,  about  living  within  one
"H of this site orl£I 'P'the cleknup  attempt?"_  Over three-quarters  o:
the'sample checked  the "extremely concerned" option    Not  -.
person said that he or she would "not be concerned at  all.
Question asked for  Che level  of concern after  the  .cleanup  ^ been   ,
completed.  As..Table VII-1 shows, almost hair  the  sample reported ctiac
they still would be-extremely concerned about  living  near  the site.


     Table VII-1-  Responses  to the  Question,  "How concerned would you
                   be  about  living within  one mile of  this  site prior
                   to  the  cleanup  attempt  (and, after  the EPA had
                   completed  the  proposed  cleanup)?
            EXTREMELY CONCERNED
            SOMEWHAT CONCERNED
            NOT TOO CONCERNED
            NOT CONCERNED AT ALL
Prior

 78%
 19
  3
	0
100%

174
After

 44%
 46
  9
  2
                                                    100%
                                                    174
 More respondents retained the  same  level  of  concern  than  "ported  less
 concern resulting from  the  cleanup.  However,  almost 40 percent  of the
 respondents did indicate  that  the  cleanup would  reduce  their  level of
 concern.  The results fall  into  five categories:

      *forty-four percent  remained  extremely  concerned;

      *thirteen  percent-remained  somewhat  concerned;

      *three  percent were  not too concerned from the start;

      *thirty-four  percent initially were extremely concerned but
        became less  concerned after the  cleanup; and

      *aix percent  were  only somewhat concerned before the cleanup
        and were even less  concerned after its completion.
  As noted in the previous section with the analysis of
                            lf        -
                                                                       °f

-------
      Tmost res
       least some
ponder ts
                  interest
      *most people have
       protect us from
       believe that it
       control toxic was
       little.or no health
      *trust .does not
       in  local"
     :  the  ability to
officials
     *statistically
      are  frequent,
      trust.
      significant'demographic  correlates with opinions
     but do not account  for most, bf the', variance in
Having outlined a
describing their
Superfund site.
   skeptical
  levels
   believe that industry anf government ^
   <"•  in, protecting public isealth;   '•  ' '
         some, doubts that, scientists  know .enough  to'
       .the impact of chemicals, 4nd  over one-third
        vs  definitely or .probably |noi:  pos.sible'to
       istes at a Superfund site trust
als, State and Federal'agencies,  science,  and  ~
manage a Superfund  site sa£ely;
       community,  we turn ill the next section  to
   of  concern with living near a hypothetical

                                                 V
                                                ' ~

-------
what scientists know.
     Table VI-7: Correlations of Trust Measures  with  Demographic
                 Variables. (The coefficients  are  tau's,  ah' ordinal
                 •analog to Pearson's  r).         '           '
                     Male   Age   Rejig.  Marital  Home   Incm  Educ
     SAFER NOW

     TRUST LOCALS

     TRUST STATE

     TRUST FEDS
                .16**  ,

                .14*   .15**  .18**
                      .17**   .15**

                      .14*

                             .12*    .11*

     GOVT. CARES      .19**

     SCIENTISTS KNOW         .13*    .14*
STATE MORE
  HONEST NOW
FEDS MORE
  HONEST NOW
INDUSTRY CARES  ..13*
     POSSIBLE TO
       MANAGE SAFELY
                        .11*
              *  =  significant  at  .05
             **  =  significant  at  .01'
 ..13*
. 18**
                                                            .15*
                                              . .12*
                                       .13*
-. 18*

 .12*
                                             -.21**
       -.16**

        .11*
While  five  of  the  trust  questions  exhibit only one statistically
significant correlation  with any of'the'demographic variables,  trust in
local  officials  correlates  with all except,home ownership.    ,

In  conclusion,  the citizenry appears  skeptical with neither trusted
heroes nor  clear villans:

     *a .large  majority believes that  the country is less safe
       environmentally than  it was  30  years ago;

     *most  people  respond  that they can trust government officials.
       at  least somewhat, with cynicism greatest toward local
       officials, and support weakest  among environmentalists, and
       strongest among elected officials themselves;

     *few citizens judge State and Federal agencies to be less  ./honest
       today than they were  30 years ago;

-------
                                             nces  are   erious 1-
                                                         '
                       prpceccirig , public  health .
            probl5» U:
-------
~3 chemicals and the possibility • of'cleaning-up. a1 Superfund  sire  wi'.:h.our
threatening the community's health.  Fifty-seven percent  of       •
environmentalists either Seriously doubt  that  scientists  know  enough  or-'-.
definitely know that scientists .lack adequate  knowledge.   Only •
forty-eight percent believe that the,cleanup  either  definitely, or
probably can be'done without a health  threat.   Although  the
environmentalists are somewhat more 'skeptical  and the  differences
between them and everyone else are statistically significant at  the,-.05
level, these differences should 'not be overemphasized.  'After  all,  the-
environmentalists are almost evenly divided  among themselves on  these
questions.  .The difference is 'that they are  almost evenly divided,  while'
two-thirds of the other respondents have  a positive  reaction Co  what  is
known and what is possible.                       .                   '

The different dimensions ,of trust  are  clarified when Guttman scaling  ',
procedures are applied to the trust variables.   The  Guttman'  scaling
technique (Guttman, 1944) is used  to determine whether a series  of
attitude questions  tap the same  underlying dimension.   If the  items
reflect the same attitudinal dimension they  will produce a,scale  that is
both unidimensional (assessing the same dimension) and cumulative
(respondents who reply positively  to a difficult item  willrreply
positively to less  difficult items).   Actually a variant of  factor   ',
analysis, Guttman scaling allows an assessment of the  number of
dimensions involved in respondent  treatment  of a concept. As  used  in
this report, every  scale realizes  the  Guttman criteria of a  Coefficient
of Reproducibility  of at least 0.90 and a Coefficient  of Scalability  of
at least 0.60.            .                     ,   ,      ,

When Guttman scaling  is applied  to the trust items discussed  in  this
section, four separate measures  emerge: two  scales and two single
variables.  One  two-variable scale combines  trusting local officials  to
tell  the truth and  believing that  industry  is seriously interested  in
r-rrtecting public health,  this  means  that  the individuals: who are  most
trusting of local officials are  also most trusting of  industry.   They
are not necessarily the same people-are most trusting of State and
Federal agencies.         '    '  •   .  " •     •            ,  '

A _second scale  includes-six variables:               .

      *trusting DER  to tell  the  truth;              >

      *trusting  EPA  to tell  the  truth;     .       •           ,    .

      *judging  federal agencies  to  be  more honest, now than,
       they  were  30  years  ago;

      *judging  state agencies  to be more, honest now than
       they  were  30  years  ago;                   ;

      * judging  the country  safer now than 30 years ago from-.-.
       the  standpoint  of environmental  pollution;  and

-------
       Table VI-6:  Ress'ponses  to  two Questions.

                                                .1
             Question j;l:  In  your  estimation, do scientists know
                     •i:..   enough about the impact of chemicals  on
                     j.   our health  to adequately'-protect us?
            YES,  THEY (DEFINITELY KNOW ENOUGH     '(•
           • I  HAVE  SOME' QOQBT THAT'THEY, KNOW 'ENOUGH
            I  HAVE  SERIOUS 'DOUBTS THAT THEY KNOW ENOUGH
            NO, THEY DEFINITELY DO NOT KNOW ENOUGH   - '
  15%
  47
  23
  15
                                                            100%
                                                       '•N
                                                       MD
'173
'   1
           Questioner Do you think it is  really [possible  to control
                       a problem like this  so  that  it  presents little
                       or no health threat  to  the1 community?
                       •.•••;•            •      -: £'-  • '• '    • <
                   'YES;,  DEFINITELY
                   PROBABLY,  AT. LEAST  I THINK SO
                   PROBABLY NOT
                   NO,:; DEFINITELY. NOT        r  •
 health    This  belief does, not flow from great trust  in  science; elected
 officials  actually fall below the mean on the question  of  scientists
 knowing  enough to protect public health.  What may be happening is  hat
  n'sta fanl Sf'^S-'? ^ "? C°ntrolied <4  -fleet their trus
 «rli«%£J    ^    Pf?"lals t0 8lve them the truth.  We reported
 of  laell  nff' n°f •u5Pr"ln«1y-  elected officials! are much' more trusting
 officiS.fi!      "e °ther Clti2ens-   ^ fact,: the elected      *
 official.-are much more trusting of both State and Federal
 administrators than is the  general public.   No elected official checked

 t^st^h^'v^rmuc;11"111 f  aU"  3nd °nly ^ 1Q ^ Ch6Cked "7« H
 trust them very much.    As  one respondent  stated,["I've had a lot of
, • , .      ,            T	   ~"^j  7e never givien me any reason not  t'n
believe them »hen the^  sav they  can do something.!'
Differences among theipther five groupings arsnof great.~  8of
botrthrKoiuS^c^tiitrLvr'rthrnLitrirftctrSorjjfsures8
    •     •  •"      -   '•:,./ .-,-:.      .'   4'..  :•.--••       -   •'"

-------
     Table VI-5: -Responses to  the Questions,  "Do  you  chink  that  -..'•
                 industries which use  toxic  chemicals  (the;goyt.
                 agencies which are  supposed  to  regulate  the, chemical
         - ,       industry) are seriously  interested  in  protecting
                 public health?         .

               "YES,: IT IS ONE OF THEIR PRIMARY' CONCERNS  . i  '  '• .
                THEY..HAVE SOME INTEREST IN  IT                 .
                'THEY HAVE LITTLE  INTEREST  IN  IT-  ••
                THEY HAVE NO INTEREST  IN  IT
                     OVERALL
ENVIRONMTLT'S.
ELECTED OFFS.
INDSTY
YES •
SOME INT.
LITTLE INT.
NO INTEREST

N =
MD =
6%
57
31
6
100%
173
1
GOVT
.26%
64
9
1
100%
174
. 0 • .
INDSTY
"4%
44
. 44
9
100%
23
0
GOVT ,
13%
74
13
0
100%
23
0 .
. INDSTY
22%
72'
6
0
. 100%
. 19.
0
GOVT '
42%
58
0
0
100%.
19
0
Trust in environmental  issues  involves  an assessment of the capabilities
of scientists and  science  itself  as.well  as  views about the honesty and
integrity of institutions.   A  person may  believe that officials are
honest and committed  to doing  the best  they  can, yet still' have little
confidence because of a belief that  basic scientific and technical
knowledge is lacking.   Table VI-6 indicates  that there is considerable
doubt both that  scientists  know enough  about the impact of chemicals to
adequately protect public  health  and that it is possible to manage a
Superfund site so  that  there is  little  or no health threat to the
community.  Only"one  in six.respondents checked that scientists
definitely know  enough  and that it is definitely possible to control the
problem at a Superfund  site so that  there is little or no health threat.
On the other hand, only one in six checked that scientists; definitely do
not know 'enough  and an  even smaller  number are .equally certain that it
is impo«»ible to control the problem at the  waste site.  Most
respondents are  unsure  about what scientists know and about whether it
is possible to adequately  address health problems emanating from a
Superfund site.  This- ?-...>cs-s that  many citizens will greet any
communications involving health and  safety assurances, in the context of
toxic waste management, with a good  ,bit of skepticism.

-------
                                                  1  ...          V
      "",'."'''    I   '    ' '    '     "   '     •  '.*•.,•"        .     .    ,
 simply  believe  that government  has  never  demonstrated'high levels of
 integrity.  We ..did. ndt :tap '.that 'possibility.     .[.'"  /    '     '     •
                 ' . '  •       .       . ..  .'    .     ..i-  '-''..'....

 There.- is some skepticism .toward officials',  yet .f;ew respondents  say they
 cannot  trust officials at all and  few  perceive that agencies  have- become
 less  honest. '       !"  "  •        '.   '    .   '     t''
      Table VI-4.- Responses to the .Question,  "Are! Federal govt..
                 .agericies, such as  the EPA,  ( State- govt. -'agencies ,
             '  .   such as the Department  of Environmental Resources.)
                 .more or less honest  today than.similar government
                  agencies 30 years  ago?'         !•
                                           EPA
DER
HUGH MORE
SOMEWHAT
ABOUT ! THE
; SOMEWHAT"
MUCH LESS

-


HONEST
MORE HONEST
SAME
LESS HONEST
HONEST . ..

. N«'
. MD=

5%
30
42
18 -
5 .
100% .
' . ;;
171
3-' "..-

. , 7%
28 ..--.-
' • • -47 , .
15
4
, Too*
;"- ' ' • '.'
t . ,
; .! 170
A
1 • . • ..-. . > •
Similarly,  few  respondents report that either th« industries  that manage
toxic chemicals  or  government agencies that regulate those chemicals
have no  interest  at all  in'protecting public, health,,  On the  other hand,
only .one-quarter  of respondents report that protecting public health  is
the "major  concern" of  the regulatory agencies.  iThese results are
steady among  civic  activists, the business community', and both those
with graduate degrees; and  those without a college? diploma.  The
noticeable .differences are among the elected official's, who are quite
impressed with  the  public  health concerns of froth; industry and the
regulatory  agencies,  i Also different .are the environmentalists "who," not
surprisingly, are generally less trusting.  Table; VI-5 presents the
results  for the entire'sample,  the environmentalists, and the elected
officials.  -Although  members  of local environmental groups are decidedly
less convinced  that government  regulatory agencies are strongly
committed to  protecting  public  health,  none of them in the sample 'argues
that the regulatory agencies  have no interest at [all in protecting
public health.  Only  13  percent credit  the government with "little
interest."  On this and .other measures  the environmentalists are
somewhat different  from  other groups., but the differences are not stark..
                                                 .4-
                                                   •

-------
 is composed of elected officials .themselves.  The municipal i tie's  of  the'-
 Centre Region have reform-style governments with professional  managers.
 iitcle patronage, and.no scandals within recent memory:.  However, -this
 relative Lack of trust in local officials  is consistent with .research.. en
 other environmental -risks such as low-level radioactive waste  (3ord
 1985).                                '-''.."                 ;

 Table VI-3 breaks .down the trust-for-local-officials  for different
 segments of the community.   Not surprisingly, local' officials  are cruise
 Crusting of themselves-.  Still, the overall sample divides evenly    '
 between the "can trust" and "cannot trust" sides.  Only "'among  members of
 local environment a 1^ groups  and among the less-we^l-educated.respondents
 do distrusters substantially outnumber respondents who trust local
 officials  somewhat or a lot.   Of course, the relatively small  size of
 each  group indicates caution in interpreting these results'.  The
 results,  however, are consistent with research on p_ther environmental
 hazards (Bord, 1987).                                .
      Table  VI-3:  Responses to the Question, "In your estimation, how
                  much can you trust local officials to give you-the
                  truth in situations like this? '                ,

                    ENVNT   ELECT   CIVIC   BUSNS   LOW ED   GRADS
NOT AT ALL
NOT VERY MUCH
SOMEWHAT
A LOT

N=
MD=
22%
39
39
0
100%
23
0
0%
5
' 47
47,
100%
19
0
19%
16
42
23
100%
31 "
0
14%
14
50
23
100%
22
0
15% ,
41
39
4
100%
46
-' ' l
10% :
31
48
12
100%
59
0
In, the  interviews  several .respondents explained their reasons for their
relative  low rating  of  local  officials.  .They said that local officials'
would be  reluctant 'to provide the full truth because of,, two concerns:
that citizens might  respond with unreasonable demands and that- the
reputation  of the  community as  a safe, good place to live might be
harmed.   State  or Federal  officials  might  be less sensitive to
protecting  the  reputation  of  the community, and therefore, more
forthcoming with honest  information.   •-.

Another way to  examine honesty  is to put  it in the context of whether
institutions are more honest 'today than  they were in the past.   Table
VI-4 illustrates that most  respondents believe that State and Federal
agencies  with responsibilities  in the environmental arena are at least
as honest today as they  were  30  years ago.   The fact that in 1988 the
executive branches of government were headed by a Republican in
Washington and a Democrat  in  Harrisburg  seems to indicate that  ,£he
perceptions of honesty are  not  tied  to affection for either political
party.  On the other hand,  perhaps Americans are generally cyncical  and

-------
  conditions safer tha'n" chose' of  the  lat-e' 195.0 ' s /
       Table VI-T: Responses  to  the  Question,  "Fribm.the standpoint .or *
                   environmental  pollution,  is  the-AJnited: States today
'more or less safe. .than
•„ (N ;= 173; MD = 1)* ".
'. .. '' ' [MUCH SAFER ''•
. '



*MD=Missing
ISOMEWHAT SAFER
"!ABOUT .THE SAME
^SOMEWHAT LESS SAFE
•MUCH .LESS SAFE
1 . . . .
Data •'.'•'
it .was !
1 ' ..
'6%
. 20
- -wi.*.d.i.ww. wu o..u co u Jvl a. v
30 years ago?
,

8 • '<• •' ; •' '' ' "
.' 32' r. ' ' • ;' ;. '
35

. 100% i •-. •
\ • ' •• -'
 More directly  related to questions of risk communications  are/ questions '
 that ask  if  officials "can be trusted ,to tell the!  truth'in  situations"
 involving  to*ic  waste problems.  Table VI-2 reports,that,  although  few
 citizens  are totallyi.trusting of government of fiicils,' mosf respondents '
 find officials at. least somewhat trustworthy.   '•'• '  '  '           '
      Table VI-2:  Responses  to the Questions, "iJ your estimation,  how
                   much  can you .trust local, officials, State officials
                   (such as the Department of Envi|ronmental Resources)",
                   Fed«?ral  officials (such as the Environmental
               •    Protection Agency) to give you the truth in situations
                 '•'  like  this?               '     , i      '     '  :     ; '
         YOU CANNOT TRUST THEM AT  ALL
         YOU CANNOT TRUST 1HEM VERY  MUCH
         YOU CAN TRUST; THEM .SOMEWHAT
         YOU CAN TRUST THEM A ,LOT
                                              Local   State   Federal
 11%
 33
-•48
	8
lOOX
                                              MD
k ;,, - /.
E "-. 8%'
28
. '56
' 7
; 100%
: N-I72 .
: ;'MD» 2
•11%
,28
51
11
. 1.0.0%
/ ' N-I72
MD= 2
.Although these figures'are not a resounding  expression  of  confidence
 that  public officials; can be trusted to provide  t!he  truth  in a Superfund
 situation,  neither do', they report a view of  publiic officials as entirely
 dissembling.   For each level of government,  more [respondents are on the
  can  trust   side than; on the "cannot trust"  side.j  Although  the people
 may be  skeptical, of official pronouncements, few. ifind public officials
 totally  untrustworthy!. ,                   "        [

 Perhaps  surprising in these data,is the lack of gireater! trust  for local
 officials,  especially! in light of the fact that ill percent of  the sample
          '   '     •   '            '               ''"     '

-------
                      -  .  SECTION 71: -TRUST '    . '.                -

 In  Che  attitude changg Literature, communicator credibility has been '
 equated wich expertise' and trustworthiness (Aronson, 'e't.'ai., 1963.;
 McGuire,  1969,;).   Of these dimensions trustworthiness must certainlv    ;
 rank  as the most  crucial.   All of the other-factors ;depend on
 trustworthiness for their  effectiveness..  Expertise cannot be credible
 without a belief_ in the knowledge base of the expert and faith that ;the
 expert  is acting  with integrity.   Attractiveness can actually be a
 detriment-to effective communication if the audience 'thinks.that the
 communicator is .us.ing his  or her  attractiveness in a dishonest or
 manipulative manner.   If communicators are distrusted,  nothing they can
 say will  bring  ab'oiit greater consistency between objective and    -
 subjective risks.  ^Besides involving assessments' of the honesty .and
 integrity of insitutions,  an individual's trust calculus may include
 judgments of the  ability of scientists and engineers to control problems
 such  as the one presented  in", this study..  Before assessing levels of
 concern arid 'the relative accuracy of'risk assessment it is essential to
 describe  the environment of trust in which the communications took place.
 The focus is on a  multidimensional view of trust,  including hoy citize'ns
 evaluate  whether  scientists know  enough to protect  us from harmful ,
 chemicals.   Within this broad context of trust several  dimensions are
 explored:   _                   ,       .            •

      *the perceived honesty of federal,  state, and  local government;

      *the degree  to which  industries' that use toxic wastes, and the,
      governments  that regulate those industries,  are seriously
      interested "in protecting public health;  and

      *the ability  of  scientists and  others involved in  toxic  waste
      managment to adequately protect public "health.           •  "  '

 First,  before the  trust findings,  we report the responses  to  a question
 that 'asks  if the United States is  safer  today than  it was  30  years ago
 regarding environmental 'pollution.   These data are  presented  here in
 part  to provide a  context  for interpreting the trust  question responses.

 As noted  in  Table  VI-1,  two-thirds of the sample believe that the
 country is  less safe  today than it was  30 years ago.  Despite the rise
 of the  environmental  movement,  the formation  of the U.S.;Environmental
 Protection Agency,  the passage of landmark Federal  acts, and  a
 substantial  increase  in efforts to control environmental pollution by
 state and local governments,  few  respondents  judge  the  situation  as
 safer in  terms  of  environmental 'pollution.  This finding is consistent
 with  other  research and speculation  and  suggests that risk
 communications  are likely  to  occur in a  context of  widespread
 apprehension.

Majorities  of the  environmentalists,  civic activists, business  leaders,
 and the general sample judge  the  United  States to be  environmentally
 less  safe now than 30  years  ago.  Elected  officials  are  unique  in,
 holding a majority (58 perceiit) who  find  existing environmental
                                     V

-------
In general,
                    emale., ' the more
                            childC8n
These  items -are  also




Summary of Section  V '
The following  is a




     *the levels of




     *while levels
                                               '-.--!!;
                                               *.
                                in  subsequent



                                '     •:   I '.
                                         4 .
             summary of the more importsJt -..results :




             I concern/ measured by any•standard, are high;
      c.nctr.aod birth defects trigger  the moit"iV..ri


                                                   '

                   20 percent report actual
                                                      ebout





                                                         toxic
                                                         fn  this  lssue;

-------
 the results of a.general question about level of.'information. -While a
 majority of these  respondents view themselves as' somewhat' or well
 informed it is clear that the vast majority view themselves as    •
 marginally informed on this issue.             .

 More  specific  assessments of knowledge tapped respondents'  knowledge
 about benzene  and  trichloroethylene,  the two chemicals' included in the
 hypothetical  site  description.   Table V-7 presents the results of the
.analysis.   Th'ese results are not surprising. -Benzene is a  somewhat" •
 common chemical  and enjoys wide name  recognition.   Besides  the above,
 respondents were asked open-ended questions, about  the .source or use of
 these two  chemicals and the specific  health threat they posed.
 Approximately  15 percent of the sample accurately  identified the
 source(s)  of benzene while less than  10 percent so identified the
 source(s)  of trichldroethylene.   Similarly,  while  22 percent labeled
 benzene  as  a carcinogen only 12 percent so labeled trichloroethylene.

The correlates of  information accuracy are predictable:  the more highly
 educated,  males, and somewhat older people know more about  these
 chemicals.  These  items are later used in a knowledge scale and related
 to various  dependent variables.              '

When  asked  the source of their  knowledge 16  percent  say  they have not
 read  or  heard  much  about his subject.   The remaining 84  percent get
 their  information  from,  respectively:, television,  newspapers,  magazines,
and radio.  Television news  programs  and documentaries such as 60
Minutes  or  20/20 are disproportionately picked  as 'sources as are~
Newsweek and Time Magazines  and  local  newspapers.  Prestige and
educational sources  of  information, such as  the  New  York Times,  Harpers,
 professional journals,  and  .educational television,   are  severely
underchosen.

To assess respondents'  levels of  involvement in  the  toxic waste  issue
they were asked  how  aware  they were of toxic waste problems,in their
immediate area,  the  state,  the nation,  and  in media  coverage.   They were
also asked  specifics  about•awareness and  their  level  of  actual
involvement..  The following,  summarizes  the  results of  those  questions:

     *approximately  50  percent of  the  sample reports  awareness  of
      local, state,  and  national  toxic  waste problems  and can  provide
      specific .examples.;

     *bnly 20 percent contend that they  are  personally involved  in
      this  issue locally;

     *involvement tends  to mean attending meetings, writing  letters
      to .officials, being members or officials of concerned
      organizations, or  consultants; and,

   •  *Love Canal  tends  to be  the example most often recalled.  x

-------
                    ••-. .    •-•   ••  ,- • •.  ..  • ,-, .   ,,,  -     .-•.-. •.,,
rrom waste  sites  may .us.eem .anomalous at  first; glance.   With further   '  '
reflection,.there is !'no inconsistency  here  because one's level of  '
concern may  represent:  a-summary judgment involving the . likelihood of
problems  emanating from, chemicals in general.   The second measure, cases
per 100 resulting from waste sites Specifically,]asks  for a judgment of-
comparative  causes.  'The better educated are not[more  concerned than the
less well educated,  but the  former are much  less|likely  to. attribute    -
health problems ..to wa;ste sites specifically.

Prior Knowledge and  Lssue Involvement    -:

Prior knowledge was  measured in a number of 'ways -  Table V-6 presents'
     table V-6:  Self ..Estimates of How Well Inf oraed--Resporidents are
                 on  tHe  Toxic Waste Issue (in percentages):  N=174.

       Question: When it  comes to the issue of  toxic  chemicals
                 do you consider yourself:       '•

              WELL  INFORMED.......'	03% ,-"''''
              SOMEWHAT  INFORMED-.	.. . 30%
              NOT VERY  INFORMED.	43%
              NOT INFORMED  AT ALL	.04%
                     I '.'-",         100%
              • '      •!•'    '    "   ;"  -.•     ''
     "     ~~:   ":•  : ," -'.  •  :   ~     '  ••    ,
     	I1  "  '  •   • •   •' .  •   '   .   .  • 1	
                 ••'):.'.'•''.    ,-   :      •     . t .  " ' ~ ,
     Table V-7:  Levels  of Specific Knowledge.    [
    ..    ,   •  •      -JL    .-".'.'      .  ..  ;.  •! :':•.:.-.•  •'
       Question: Have you heard  of the chemical 'benzene?

              YES	[I'.'	75%   •                V' ' '"'•:
              NO	ji	..25%'  /  "•,'  .          ! "

       Question: Is benzene a.health  hazard?
                     i:'    '     ,    ''  •       -  .
              YES.... !i". .....54%
             • NO	|i.......-,46%

       Question: Have; you heard  of-the chemical Crichloroethylene•or
         .   -     : TOE?};    -   '    '  •   ,    -•    ;.-  -['.

     .    •     YES....CV. ....53%  '"        .        t '
              NO..-...J...	.47%;.  '           '.-'   '[••• •'..

       Question: Is trichlproethylene a health haizard?.
              ' '      ii-  .'•'..  '     ••   ••'•-.
              YES	ii	39%  '
              'NO..'.....I",:	61%  "-'     '     "

-------
 instructive.  This pattern of  results  indicates' that,  for  a  many' people.
 man-made toxins are 'a major evil  in mo.dern  society.  This  supports  the
 notion of a culture of  fear and  the idea  that  fear  of  man-.made  toxins
 is, at least for a portion of  this sample,  a'well-formed attitude/  •

 While we know of no study that attempts to  estimate  the 'number  of
 cancers,  or other health problems, caused by toxic waste sites,  there
 are two reasons.to believe .that  these  respondents are -overestimating trie
 implication of toxic chemicals from waste sites  in.the.etiology  of
 cancers and other-health problems: it  is highly  unlikely that toxic
 chemicals are causal factors in most or all cancers; and, man-made
 toxins cause fewer cancers than natural toxins  found in peanuts  and
 other common foods (Travis,  et.al., .1987).          '

 In attempting to determine who,is more concerned, two demographic
 characteristics demonstrate somewhat consistent  patterns of correlation:
 gender and  education.   In general, women'are somewhat more concerned
 than men and overestimate more than men the number of diseases caused by
 chemicals in the environment.   However, the most instructive
 correlations are those of education with the level of concern and with
 estimates of various  diseases  caused by man-made chemicals in the
 environment.  Table V-5  presents  these correlations.      • •          '
         Table V-5:  Pearson Correlations Between Levels of Education
                    and Concern and Incidents of Various Health Problems
Health Problems

ADULT CANCERS
CHILDHOOD CANCERS
LIVER, KIDNEY, BLADDER
T.UNG
BIRTH DEFECTS
MISCARRIAGES
LEUKEMIA
SKIN PROBLEMS
CHILDHOOD DISEASES

       *P-.02
                              Level of Concern

                                 N.S.
                                 N.S.
                                 N.S.
                                 N.S.
                                 N.S.
                                 N.S.
                                 N.S.
                New Cases Per 100
                                 r
                                 r
          ***P-.001
            N.S.- NOT'SIGNIFICANT
.'is*.
. 19**
r =
r =
r =
r =
r =
r ='
r =
r =
r • =
-. 32***
-.32***
-.26***
-.20**
-. 26***
- . 24***
-.20**
-.31***
-.28***
The above table dramatically  illustrates  that  accuracy  of  the
respondents' risk estimates is  strongly affected  by  levels  of  education
while the concern.levels are  less  affected.  For  every  health  problem,
levels of education decrease  the risk  estimates for  each of the  problems
that we examined.                                  .

The finding that the better educated are  somewhat more  likely,to express
a high level of concern with  skin  problems and childhood diseases  and
much less likely to overestimate the proportion of .new  cases arising

-------
   sites."  At  least  one]; in three, .respondents believye .that- 1,1  percent br'  •
   more of these diseases  'are caused by toxic chemicals  from waste sites.

   In addition  to  the  above respondents were also  asjke'd  the following
1   open-ended .ques t ion:  <•;     •'      . '      '      •• '  f '  •"   •'••'''."•.   ••  „

        Are there  any  health problem we haven't mentioned  that 'you   :  '
        think may  be  strbngly related-to', toxic chemicals  in the,    '.    -.  " '.
     ;  "environment'?-   • '»•    .'     .  - • ••   ''.  .,    '   -^ __   .  ;••-,    •_       ,   .^

.   Nine pe.rcent of. the respondents indicated various neurological problems -
   while 8 percent  namedji"mental  health problems."  [In addition,.  2'percent
   mentioned Parkinsons  Disease.  .There appears to b|e,a  significant  portion
   of • the population who]:holds  toxic chemicals, accountable  for some  of our. .
.   mo~st dreaded neurological" problems.      ..  ' '' ..   !      .--.""'    '       '  •  /•
     '  -  .       -    .     I , '                • •   -       4  .   • •• .;.'  ' .'•:  -     •  ,
                    '     j \     .....   , .     . •      ,ifc. ,ri, .   j , ' -  .
   The  scope and'depth of  people's beliefs, about the! implication  of
   man-made toxins  in  some  of modern-^society' s most [dreaded health problems
   bode ill for those  attempting  to construct objective  risk communication.
 >  What data,  and  what sort  of  message  construction !or delivery,  could
"'•  allay the fears  of  those  who believe that 50 percjent :or  more of adult:
   cancers ar,e caused  byj'toxic  chemicals from waste [sites?   Furthermore,  it
   is reasonable to assume, that those who^hold thesej. beliefs /are'  the 'ones
   who'would be most vociferous in protesting an agehcy's handling of a
   toxic waste problems, j: This  possibility  will be explored in later
   chapters. '  ',        . .!'  ,        "  . ''  '
   Finally,  a question was designed  to determine to what degree  people  felt
   that the,ir own health:!had  been  affected by. exposure £o  toxic  chemicals.
   Table V-4 presents the percentage distribution'for that question..
           Table V-4: Estimates  of.Having Suffered'Health'Problems
                      (iii percentages):  N=l74.
           Question': Do you think  that  you have suffered health; problems  .
                     .due!to exposure  to hazardous chbmicals in  the water,
                     soil,- or  air?/     .    ,   -     ,[-.-:,    •       -

              YES, DEFINITELY.		i. .6%                  '  '
              I SUSPECT i HAVE. .'. ..	 ...121   '      :   .'   ,
              I DOUBT .THAT I HAVE....		...161
              NO,  I DEFINITELY  HAVE NOT.............[±2  '
            .•  '     -    .1:.-':'.    :  •   •••.•.  ••••••=.••'' 1-:'1   'lOQ  '
  The  results of this table parallel  the results no^ed in Table V-3.
  More than one in four^respondents feels that he or she has suffered
  health problems as a  result  of  toxic  chemicals.
   ••        -             -i, • •        ..-.".••'   -••-..! •   --•  ^ ::'   * ,
   The  extent to which a|significant  portion of- this|sample is willing  to
   attribute cause for.numerous health  problems to man-made toxins  is-

-------
 experience health  problems  as  a  result  of 'exposure to toxic  che-icais.
 What this measure  does not  tell  us  is' how  probable they  think that   ; ;
 health risk is.  Table V-3  presents  the results  of a question de-slgned
 to assess public beliefs  about how many new cases  of a  specific- class of
 health problems are a result of  exposure 'to toxic  chemicals.  .Ln' other
 words,  this question attempts  to determine  the extent to  which  people
 view toxic chemicals as a primary cause of  certain classes:of health- .
 problems.  While there may  not be good  scientific  answers  to  these   :   ••
 questions, which makes it difficult  to  use-"these answers  as  benchmarks   -.
 of the accuracy of .public judgments,  they can serve as a  sort of
 Rorschach test of  public  fears.  In  other words, when people  interpret
 somewhat ambiguous stimuli  their predispositions and decision biases
 become apparent.                         '   ,               .           ;
      Table V-3:  Estimates of.How Many New Cases Per  100 of'Nine
                 Different Health Problems are Caused by^Toxic
                 Chemicals from Waste Sites (in -percentages): N=174-.*
      ADULT CANCERS

      CHILDHOOD
            CANCERS

      LIVER,  KIDNEY,
       BLADDER PROS.

      LUNG  PROBLEMS
       (Not Cancer)

      BIRTH DEFECTS

      MISCARRIAGES:

      LEUKEMIA

      SKIN  PROBLEMS

      OTHER SERIOUS
       CHILDHOOD
       DISEASES
Zero
Cases
. 1%
1
3
One ..
Case
11%
' 12
10
2-5
Cases
29%
28
35
6-10
Cases
19%\
17
14
11-25
Cases
•" 16%
18
17 '-.
26-49
Cases
10%
8
8 :
-50+.
Cases
9%
• 11
1 •• 7 •
8
3
5
6 .'.
5.
14,
12
14
14
14
22
28
27
33
, -20
18
'.17 . •
18
.1.0
.ir ._.
15
• 18
14
12
19 '
6
10
9.
- 12
10
                             9

                             6

                             .7

                             6
31
-.17'
                                                  14
       *The percentage  remaining  in .each  case  reflects  the fact that
        a few respondents  chose not  to  answer  these  questions.
The results in Table V-3 dramatically  illustrate  the  beliefs  underlying
many people's fears of  toxic  chemicals.  .Almost one  in  five  respondents
believe that 26 percent or more .of  adult  cancers,  childhood  cancers,
leukemia,  and skin problems are caused by  "toxic  chemicals  from waste

-------
 to each of those health  oro'blems.. •
        Table ,V-2: Level of Concern  for  Specific Problems (in      •  .'
                   percentages):-N=l74.      '  •   I                '
          N            ' '                  '          !          .--"       '.'..'
        Question: Here! is a list  of  health'probleras  that may be connected
                 -.with; exposure  to chemicals.   Wodld you indicate how
                  concerned you  ar,e  that  you,  or someone close,, to you,
          • '    :   may suffer  this problem because of exposure to
                  chemicals in  the air,  soil,  or Water.                  :
                      •          •   .  •••--_  |-
                      ;, 1 = Grea.tly Concerned   .
                      ;; 2 = Somewhat  Concerned
                     . ,, 3 = Not Too .Concerned
                       4 = Not Concerned  At All
                      - 5 = No Response
• i ;'"'•'. t ' '


ADULT CANCERS ':: .
CHILDHOOD CANCERS
LIVER, KIDNEY; & BLADDER PROBLEMS
LUNG PROBLEMS (Not Cancer)
BIRTH DEFECTS' ; ' .
MISCARRIAGES .'. . . '.,•'.'
LEUKEMIA . /

SKIN PROBLEMS - ;. / ,. ;-
.... - ^ •,
OTHER SERIOUS , 'CHILDHOOD DISEASES
(!) .(


57% 2
57 2
40 3
44 3
61 2
50 2
49 3

39 3

48 2
2) .(3).
-•--
'/ , • •
8% 8% .
, '
4 . ' . 10
8 16
4 15
0 10
> ;
7;, 13
2 , 9
; v ,
'4 "(: 22
••-.'•
, '
8 ... 14
' •
, (4)


/ 7%
. 8
5.
.. i.
, 6
8
; 9
,'s

5

8
(5)

i
. P%
1
1
• ' ;1
1
I- ,
. 2",

0

2
Cancers *nd  birth defects predictably generate the highest levels of
'concern.  What  is noteworthy about Table V-2, .however,- is the small
variation in concern expressed across all classes,!.of health problems.
Seventy-three percent of the sample expresses somie level of conern even
for  skin  problems,  which presumably are not life threatening.  These
results reinforce the .assumption that fear of man-made toxins may
exhibit characteristics  of well-formed attitudes and beliefs.  Only 7 to
10 percent of this  sample consistently expresses jlit.tle or no concern
that chemicals  may  cause health problems for: the respondent or someone
close  to  them,        :
Levels of concern  i
most of  these
    illustrate one dimension of people's  fears,  that .is,
respondents believe .that they or someone  close  to  them may

-------
     SECTION  V:  PRIOR ATTITUDES,  ISSUE INVOLVEMENT..  AND INFORMATION"

 Douglas  and  Wildav'sky (1982)  argue that- something" akin to a
 fear has  developed in Che Uni-ted States with respect to many man-made
 technological  and  environmental.risks.   If that argument is valid then a
 large proportion  of the population carries a predisposition' to react
 strongly  and negatively to the imposition of'certain man-made risks.
 While it  is  true  that not all communities react the same way to
 analagous  risks,  it raa'y be that  fear responses have taken1 on che
 characteristics of -well-formed attitudes.  Well-formed attitudes are
 easily triggered,  usually simply structured, and invoke considerable
 levels of  behavioral commitment.   They  are extremely difficult to change
 with ordinary  information-education campaigns  (for  a discussion of this
 issue as  it  pertains.to AIDS  education,  see Booth,  1988).   Well-formed
 attitudes  act  as  conceptual  filters for incoming information and, to
 some extent, determine how that  information is processed.   This chapter
 examines  the attitude-belief  structure "concerning toxic chemicals that.
 people bring with  them to" a  chemical waste situation.   It -explores the
 conceptual filter  that exists prior to  exposure to  the hypothetical
 waste site,  the respondent's  level of both general  and specific
 knowledge  about toxic chemicals,  and the source of  their information
 about  this issue.
      t                        ,                   •

 General and  Specific Levels'of Concern        ~
 (See Appendix  B for a Complete. Copy of  the Questionniare)

 Two  questions  attempted to tap respondents'  depth of concern about the
 toxic  chemical issue.   A"general  question was  asked first and  then a
 more specific  question assessing  level  of concern by nine specific
 classes of health  problems:  adult  cancers,;  childhood cancers;  liver,
 kidney, and  bladder problems;  lung problems  (not cancer); birth defects;
 miscarriages;  leukemia;  skin  problems;  and,  "other  serious  childhood
 diseases."   Table  V-l  presents the results  of  the general question.
        Table V-l: Level of Concern  (in  percentages):  N=174.

          ' VERY CONCERNED	V.		54%     '
           SOMEWHAT CONCERNED.	..'...43
           NOT TOO CONCERNED	 3
           NOT CONCERNED AT ALL		..0
                                         100%
A majority of these respondents  report  the highest  level  of  concern  and
97% express being somewhat or very  concerned.  The  fact that  virtually
the entire sample expresses some  level  of concern gives credence  to
Douglas and Wildawsky's culture .of  fear thesis.  However,  a  general
expression of concern can mean many  things.  In an  effort  to  ferret  out
some specifics respondents were  asked their  levels  of  concern for nine
specific classes of health problems.  Table  V-2 presents  their* responses

-------
general examination .cI-
hazardous waste  cleanup
highly involved  in any
risk assessment  issues  inlthe context of a
,  among a sample  of  people; not unlike thos-e  found
important community"'issuej

 '"'-•'.'..'.•       1           .      •   ..'  '
          •  - •     '.   '  •  i'   • ' ' ..':: •'....   "      •

-------
 As noted in the  first  paragraph  of  this  section.,. the  sample  is  not
 designed to be representative  of  any  particular  population.   There  is,
 for instance, a  strong  overrepresentation  of  citizens  who  are active  in
 community affairs; after all,  elected  officials  comprise  11  percent  of
 Che sample but under one-tenth of one  percent  of  the  population of  the
 Centre Region.  The purpose of the  research was.  not  to explore  the
 attitudes of a representative  population but  to  .examine the,-consistency
• between objective and  subjective  risk  estimates  among  different types of
 people.  The sample provides the means  to -achieve  that objective.

 Conducting the Interviews                     :'   ,      - ..'

 Each interview was conducted in one of  four ways:

      *orally and individually.;

      *in writing in a group setting with a principal  investigator   •  ••
       present;                   '   .              :

      *in writing and individually;  or,                            ,

      *by mail.     •                 •

 When working individually with a respondent, we asked  for  the
 individual's preference between doing the survey in interview form or in
 writing.   Most chose to write the responses, although  a majority of the
 elected officials preferred the interview format.  In  the  group
 sessions,  we asked the respondents  to fill out the questionnaires/in
 writing before discussing reactions as a group. .'In some cases,
 particularly with elected officials and members of environmental groups,
 they mailed completed questionnaires to us.  Of the individuals who
 requested  questionnaires at environmental group meetings,  only  two
 failed to  return them.   A substantial majority of the  questionnaires
 were filled out  in group settings with one of the pfinicpal
 investigators  present,  approximately 15 percent were filled.out
 individually and mailed to us,. approximately 10 percent were done in an
 individual  interview session,  and several were sent through the mail and
 returned to us.

 Concern Chat the data collection format might influence the results
 proved unfounded.  Testing for significance by analyzing all the crucial
 data by the type of data collection fo'rmat failed to show  any
 significant relationships.   However, the small numbers  in  some of the
 formats suggest  caution in assuming no differences.

 Summary                  .               . .  '   ' ,

 The data are 174  completed interviews/questionnaires with  residents of
 the-Centre  Region who were asked  to react to a hypothetical, hazardous
 waste  site.   Inferences made from the sample should be generalised to
 other  populations with  great caution.   The sample overrepresents
 community  elites.  Instead,  the data are intended to permit a more

-------
  among  ,ne  four groupings .   Although i'c wpu Id have been ' poss ibL/
  compare group  dir ter^nc-es. with multiple-member individuals*  elud
  tnere_is no  tneoretically. sound. reason to do ,6.j.  :in. al? communi-ies' •
  certain. individuals  perform multiple roles and nU cros.-cu^?^    '  '
  identifications.  Three of „ the groupings, public officials      * •'   • ''
  environmentalists, and civic  activists, ; have in common a willingness to
  Participate  in voluntary,  public  activities  in the community.  !he
  fourth, group-, member a "of the  business  .community, I .have interests and  "
  resources that commonly lead  to' public involvement.
               .       " ! '  '.             •         •'(.'•""     '    ••'•'.,"
  Identifying the Interviewees  ..             .
         1              H'  •     >,    •' • '    '.','•*:

  Interviewees were initially identified through  slveralJmeans :


       *elected officials .received,  phone 'calls -from  the senior
        researcher on. the project, Bob O'Connor;  -J


    '   *1"ders-of: local '.environmental groups were I contacted 'and    '
        asked  if  the. principal investigators could '{.address, a meeting  to
        ask  for cooperation;                        >'     '           &
   •   :.           .  .   ii •  : •..-.•  ••     •  : .-...;,  --|r   ,  ••...••-,-.• ;.. '. ;.

       *instTuctors  of  .Continuing Education courses'  whose  students .
        are generally older,  long-term residents, were  asked to   .:
       devote  one  class  period  to the survey in re|turn for a
       discussion  of social  science and hazardous [waste by the       ;
   ,  .  . principal1, investigators;       .-.".•''   [      -    '
           '        '        '"             '       '    '      '   '
                               .                     .

              °J'univef-8i^  secretaries  were asked to attend a lunch-hour
       session to compl-ete the questionnaire and discuss the topic in
      *leade;rs of church, groups were  asked  if  the principal
       investigators could attend, meetings  either to seek volunteers
      •or to conduct -the survey and a discussion  session.

 These methods produced: over 100  interviews^  AfteJ  reviewing the
 demographics of those -interviewed, a research assistant  went  -    •
     nei^borh1^361^'^ nei8hborhoods -to -conduct [additional interviews
     neighborhoods chosen were those whose  resident's  were
                   am°"8,the interviews previously [completed.  We were
                                                      of interview, with ,
Of the elected  officials  called,  the' only one not "[interviewed was a
supervisor who  asked  to respond by mail and who neVer returned the
                                   n0t ln the State! at.the time -of the
                                        -     '      '
  nlnd«H       ,    interVleWS  " the Sprin8 and Sunffler °£,. 1-988, we
concluded th.t. adequate  numbers  of interviews in, all key groupies had
been collected to permit the  comparative analysis called for in^he
research design.       |.            '              ••   ,

-------
Table IV-2: SAMPLE DEMOGRAPHICS
EDUCATION :
HIGH 'SCHOOL
TECHNICAL- SCHOOL
SOME COLLEGE
"COLLEGE GRADUATE .
SOME GRADUATE WORK
GRADUATE DEGREE

FAMILY INCOME
UNDER $10,000
$10 - 20,000
20 - 30,00'0
30 - 40,000
40 - 50,000
50 - 60,000
60 f . •''..';

HOME OWNERSHIP
OWN HOME
RENT •• ' '


OCCUPATION
STUDENT •
PROFESSIONAL
BUSINESS
NOT IN LABOR .FORCE
SKILLED - TECHNICAL '
''LABORER .
"
ATTEND . RELIGIOUS SERVICES,
NOT AT .ALL
SEVERAL TIMES YEARLY
MONTHLY
WEEKLY
MORE FREQUENTLY

*MD - Missing Data
(In Percentages)* .' : '

6% '• ' ' '.
5 ' • • '
26 . ' -.• ' .' "
''17' '. - • . "
13 > • • "
34 • , '.-''.
100% ' N .= 174 • ' - . '-•
' ' •
',.-., ..- .
10Z-- •'...' • -
' •- 15-: . '.
•-15
.. 24 .."•••
' 11
•'. 9 ' •••.- . -,' , • . '.
16 . .
- 100% N = 160; MD - 14

55%
45 . . . • •' .••.-.'
100% .
. • ..' , ,. N = 174 . ;
'" •
. 19% (Undergrad and grad)
14 (e.g., professor , .lawyer)
14 (e.g., 'manager , sales)
18 (e.g., retired, homemaker)
17 (e.g., nurse, technician)
20 (e.g., clerical, factory)
.. 100% N = 154; MD = 20 '
j ' • ~ -
- . - ' --
26%- ,
22 '
20 •
25 , .-•-'.; • ..
7 •
100% N - 171; MD =» 3
"'.'-' "
These four categories are not mutually  exclusive...  In  practice,  no
individual falls into ,all four categories,  although a .couple  fit. into
three categories and several fit  the  criteria  of  two categories.  For
example, one person runs a property management  firm, is  active^n
several civic organizations, and  is a borough  councilman.  These
cross-cutting memberships reduce  differences in attitudes  and opinions

-------
  parameters is obvious:
 In. selecting our. sample, we- deliberately  chose
                                                            residents
  accounted, for observed variance .in consistency .between object vand
  subjective  risk assessments.   The characteristics that di ferentiate our
  sample  from the general population rarely account for significant
  differences.   Table  IV-2  demonstrates that the .ample achieves    •
  substantial variation. on  all  of the demographicjcharacteristics:   '

  In selecting  the  sample,  particular attention w^ given to including
  civic activists,  elected  officials,  environmental! ft.,  and the business
  community.  The sample  has  31  civic  activists,  19 elect* officials   23
  environmentalists, and. 22 members  of  the .busis, community
   thLionc      *T h*aVily  inV°1Ved  " volary organizations such
   the Lions Club.  The intention  was  to  select  -active organization
members, not simply nominal  joiners.,  To  be identified as  a civic
a!CKV:LSt^a resP°ndent eich"  is an  officer in an  organization or  a
member of two organist ion with  regular attendance at meetings.

                      are menibers  of th«  boroughj council °f State
                    <:5mmissione" ' the- aup.rvi.ot.  in the  townships  in
                           C3X collect°« ^ those  townships with
                         The CTt7 C0nmissi°"" -d  the tax collector
                         cpunc1-1 Ambers  and supervisors serve  for  a
 the    nrr
 job.  «.
The eavironmentali.ts;! are 'identified neither by (heir holding any
                th         .an  actlve  identif icatijon with environmental
                the community .                   |
   '
ch«i
chemists,  small
                                                          as individuals
               private  sector  in  a managerial,'professional,  or
                         In  this .study  they  include,  for example,
                        owners, purchasing agenejs,  and  the  owner'of  a

-------
      significant social upheaval..   Even  during ,our  interviews "with
      members of environmental  groups  no  one  mentioned  the  local
      Superfund site or evidenced any  major  concern  with  the other  local
      problems.

      Third, areas of the county outside  the  Centre  Region  range,, from
      Amish settlements to mining towns.  In  the  northwestern, section of
      the county-,  approximately one  hour  away, from the  Centre  Region,
      concern -over toxic wastes allegedly dumped  in  abandoned  strip
      mines received wide publicity  during the period of  interviewing for
      this study.   Efforts to locate a solid  waste facility in  that  area •
      also generated heated'debate.  Interpretation  of  our  .data from
      areas outside the Centre Region would have  been overly complicated
      by ongoing hazardous waste controversies in some  communities.   Bv
      restricting  our focus to the Centre Region,  we were better able
      to guarantee that, the hypothetical case was the central stimulus.
      for.the respondent.   While this-enhances the study design, it
      limits the extent to which these results would be relevant in  an
      area characterized by high levels of concern and  controversy about
      hazardous  waste.

      Table IV-1:  Sample and  Population Demographics (in percentages)
                  N
174
         AGE
             18-24
             25-34
             35-44
             45-54
             55-64
             65+
        GENDER
            MALE
            FEMALE
        MARITAL  STATUS
            SINGLE
          •  MARRIED
            46%
            54
           100%
            41%
            59
                           POPULATION*
                               56%
                               15
                               10 ' *
                                7
                                6
                              	6    :
                              100%
 52%
 48
100%
 66%
 34
                                 100%
                              100%
        *Source:  State  College  Area  School  District
                  1985 Census
The Sample                          '

Table IV-1 illustrates some .of  the  differences  between  sample  and
population characteristics.  The  impact  of  students  on  the--population

-------
              SECTION:
                       IV':  SAMPLING  AND  SURVEY IMPLEMENT -T^O

 desire to *ake the sapple »ore     ,r ble
                "     i
  n
                                          "«Plin   techniques were  used
K
had 6
the sample
In  this  section  we'describe the location, the sample, how the
interviewees  were  identified,  and how the interviews were Conducted.

The Centre  Region    •'.        :    •  '   .


The "Centre Region"  of'Centre  County/is  the  official name of an area in
the center  of Pennsylvania.  -In 1980,  the Bureau [of the Census
designated  Centre County as  the State  College Metropolitan Statistical
Area.   The nucleus of,this  new  MSA  is  the Centre Region which is
composed of the Borough of. State  College  and  the Townships "f Allege

            f ?u0n;,Ja™«S'  a"d Patton'   Thes« municipalities in 1980'
           of the 112,000 residents  in the county!.   We  chose to limit
           to the Centre Region for  three reasons:

     First,  the Centre Region provides a  substantial diversity  of
     people    The area includes .the main 'campus  6JTthe  Pennsylvania
     State Unxversity, severaL advanced technology  industries,
    •traditional  manufacturing industries in  the glass  and cuy  sector
    3 growing ssrvic(j ssctor oarti'cnTaT"ivT«v»^^ki« -.— j         • •    '*
          f         '           f*** w^wt^^a^ I.JT j.u  llOtfiLS aHQ  ITSCrfiStLOn   3TlH
    some farmers.   Ttrere is diversity in demographics, but  also an'
    essential homoegeneity of exposure to situations regarding
    hazardous waste  in  the .community.   Differences in attitudes are
      ikely,to arise  from factors>asured by'the questionnaire  not
    from experiences related to drastically different community, events.
                        Re?ion,does h^« ^ hazardus waste site on the
                         List (Centre County Kepotie), but that site
        generated  little publicity in recent yea rs.  However, there
    h«« been  some  recent, .mild,  concern generated by rumors of

    -lfu^f   Wat6r in a  nearby Village and a Sfish kill caused by a
    rbfckdronTg SeWaSe.treatment ^cility.   Whiile the8e events form
    a backdrop for community concern,  there has been no significant
    is iiobS f«y.m any of  these  cases.   In othe|r, words,  Centre Region
    hLpf   K'31m  ar/° "f^.0"1"  areas in the United  States which
    have some historyof toxic pollutants,  some citizen' concern," but no

-------
at these factors both  individually  (the bivariate measures in section
XI ^ and in raiiltivariate  analysis  (Section .XII).   Before turning to-a
description of  the  results,  a  brief explanation of the 'sample an'd the
data collection methods  is  necessary.   -.                 '  ;

-------
  category, behavioral intentions,  may serve as  an indication of -he
  intensity of concerns;;  ..Overreacfeipns  to  the situation  su-h-as" aa
  Cne area or'.buying bottled water, may  indica.te . a failure to" achieve
  consistency between objective and subjective assessments of the
  31tuation.      .      :,  . .   ...    - . '   .'    ..                 '  •

  The Independent Variables     '-

  Variations  in, consistency between'objective  and  subjective  risk*
  assessments,  and other dependent variables/  may  bk  a  function  of
  variation in four sets-of independent variables:  I      .  '

       *.risk .uncertainty communications and management  options  'the
        quasi-experimental design of this study described  above;'
      *attitudes  citizens bring- to a hazardous waste situation  prior
        to  any  EPA invoylement;              /  '--.'   : |'    '_  _. ,	;; '

   :  • *trust in government,  science,  and industry; [and   '      ' .••

      ^demographic  factors  such as education,  age, (gender, religiosity'
        and marital  status,, that predispose citizens!  to form attitudes
      ..consistent or  inconsistent  with expert  opinijon.'

 The, first set of independent variables are the risk uncertainty
 communications and management  options available to1  the Agency.   Other
 farmr.- i™,,-  „.. Account  for significant variation in achieving

                       •1.'.. ''••''  ,  ',;' ."•', -    /•' -..-- v|;•'-..'•  '>;-:'v'~- *  '••  '••.  '.
 Attitudes/citizens havg  prior  to  learning about tht  hypothetical
 Superfund situation may  influence whether they develop  risk assessments
 !hnS^^n^W:Lt:hueX?ert:;:0pinioa'   Before respondent.b  read  the material
 about  the hypothetical lease they  answered question^  that  explored  their
 levels  of concern with!toxic waste sites,  beliefs Lhat  they personally
 may  have health problems due to exposure  to hazardbus chemicals
 estimates o| health problems: emanating from.: toxic ^aste  sites,  and  their
 knowledge about toxic chemicals.           •               :    '

 Regardless of  how EPA communicates at a site  or manages a  cleanup
 hve^^HClt!:uen8  reachifonsistency with expert  opinion may'be influenced
 by whether they trust the government, science,  and industry.  We
 therefor* examine the levels of trust in government (local;  state   and
 national),  industry,  and science - both whether scientists know enough
 to protect us  and whether it is possible to control a problem like  the
 one presented  in  this study.                       j     .

Finally, demographic  factors may stand as  surrogates for life
experiences  that  predispose  people toward  consistency or inconsistency
with expert  opinion.  We  examine whether difference's, ur gender,  age
income,  education, marital  status, and religiosity help in" Any way  to-
explain variations  in the dependent variables.     .             * .    '.

In-accounting  for variations in consistency with exlpert  opinion, we -look

-------
     *the granting  of  a  technical' assistance  grant  to'the Local
      citizens  group;                 .   .                    -'.'-.

     •*Che issuance  of  the  remedial  investigation and feasibility.
      study;                              .         '

     *a public  hearing organized  by  EPA; ,and      . —       . '  •        -

     •"commencement  of  the  cleanup.  .

This option  is  characterized  by a common  cleanup scenerio in which EPA.
accepts comments as  required  by law,  but  makes  cleanup  decisions
without directly involving  local  citizens  in  evaluating cleanup options.

The  citizen  participation  option, which  is  actually  an  expanded
"standard" option,  includes the elements  in the "standard" option, with
an important addition: EPA  agrees  to  work with  a citizen advisory
committee which is  given the  right  to review  all EPA plans and  to
approve any  proposed solution before  it  is  put  into  effect.   The  final
plan is described as a joint  effort  of. the  advisory  committee and  EPA
experts.                           ,                      : - .,-;•-'.   '. "

The  indemnification  option  also includes  the  elements noted  in  the
"standard" scenario, but adds 'a section describing an indemnification
program by the  State.  The  State  would, under a special experimental
program, purchase the home  of anyone  within one mile of the  site         .
providing the owner  had first tried  to sell the property in  the-private
marketplace  and had  been unable to get fair market" value-for it or could
not  sell it  at  all  pver a  12 month  period.

The Dependent Variables                          ,-                 ,

The  purpose  of  this  study  is  to examine the impact of variations in
uncertainty  communications  and raanagment options on  achieving
consistency between  objective and subjective  risk estimates.  We measure1
this consistently through.four sets  of dependent variables:

     *the level of :concern with-living near the site, both; before
      and after the  cleanup;

     *expectations  of suffering health problems from living  near the
      site, both before and after the cleanup;                          ,
                                                         '    •'
     *risk assessments of  the danger  of living  near  the site compared
      with other situations,  both before and  after the  cleanup;  and,

     *behavioral intentions when  faced with a Superfund site in the^
      vicinity  of one's home.                         ,   ,              '

For the first, three  sets of variables, the  achievement  of  consistency
between each option  and citizen attitudes,is  indexed, respectively, by
low levels of concern, low  expectations of  health problems;,  and viewing
living near  the site as safer than smoking  or driving a car.  The  fourth

-------
        received  fay people who may^e exposed to ' Che" chemical .  -he  -isk
               /i00383  ^ aSsumPc^> or best gdess, that aooears'
              Likely  toh underestimate the risks.  .4n attempt  is'made  to
              tim                                                     '  '
                                                  .
       overestimate rather than underestimate- risk!/
       educated guess. -|:; .            •               -f
                                                      But ir  is
                                                          "
                                                                an
                       .                                     .
  The risk communication. is  set  in the context oft "Wafer Advisorv   .   '
  Notice  put out. by  the EPA .(See Appendix A) , •- Tht advisory- describes the
  chemicals causing the problem  and the threats  to|,he.lth they  ose  %n
   •annual chance  of death" comparison chart is. presented which   *  ~
  graphicallly compares the  contaminated water risk with common risks such
  as> smoking,  hang gliding,  scuba diving,  skiing,  driving a car, and
  drinking diet. soda.  .Only  voluntary risks were chosen because .research
  indicates that  forced comparisons  between voluntary and involuntarv   "
  risks  are resented _  (Oovello , et.al.,  1987),  Current conventidnaL wisdom
  in risk communication- indicates, that:.  '     ,    [          '

      Comparisons ar^ useful in  informing people Lab'out risks,  and
                      • I: - - -    .    '    ' •  .     •.•.'.';[•••.•,    .
       *the comparisons should be those comparable  to  the  focal  risk
       ..in degree of voluntariness.                •    -
                       !'     -  •"   :         '"''     .  f..'.    '   .
  In  other words,  the rvisk format- chosen here may generate  some  hostility
  because  of  the  nature of the comparisons made.  We will be able  to
  assess  that  possibility in the -comments solicited! about the nature  of
  the  communication.    i                           '•(
                       ir-- •.   •',--..-•...• ' -•  •  .." f  '.  '."-'" '    -  .•'•   .
                      •',.   The Management Options  I              .
.'             •      ,  ' I:' '   '.           ,       •- '  f'   .      ,'..•' '
 The  literature  on  environmental conflict resplutiion points' to  two
 classes  of options  that  can  be offered a concerned community
 incentive-compensation  options and power-sharing options  (-Bord, 1987)   '
 -:;fn-l^:coinPen"*ioti  options tend to be material rewards or payments
 designed  to restore  equity or  to  push the reward-bost /ratio 'toward the
 or«vfrf T^ °f the.conl:inuum-   Power-sharing  options  are designed  'to
 kev J«iiiC*  c^-l"n*--^«ro«P»  with 3°^  ^iliby  to have L impact on
 key decisions concerning site  operation  and  monitbring.   Each  set of
 options has been suggested to  moderate local concerns: one set appeals
.to pecuniary motives while the  other  appeals to  control motives.   One
 study suggests that  the most vocal  opposition to  Locally  unwanted land
 uses, those who^hold unqualified "NIMBY"  (Not  In iky Backyard)  attitudes
  "  ™°r"      abl7 PredisP°sed  tow«d  power  sharing options  (Bord,      ?
       •        •  •      ].,..'•   .         :  '       i .-...--       -      • • •
                  options
Three management
"standard" scenerio, a
indemnification
                 scenerxo
The  "standard" scenerio includes the following:
     *the  formation of
      added  to  the
  are systematically variect in  this  research:  a
participation in decision-making scenerio,  anH an
• (See Appendix A).
                       a concerned  citizens'  group!' when the sife was
                   National Priorities List;       i  ;

-------
 The scanerio was constructed  so  as  Co  include ' some .of  Che  routine
 features of a Superrund  situation.   Attention  Co  a  oossible. problem was
 created by locals noticing  that  the  drinking water.', had an'"unusual"
"smell and  the  local newspaper  reporting  letters  from  concerned
 citizens who linked possible  health  problems with the  smelly water/  The
 list of health  problems  included  one family experiencing an  unexplained
 rash, two new cases of childhood  lukemia  in the  same.neighborhood,  and a
 growing number  of old people  reporting  problems with arthritis-.-   Note
 r.hat the relative ambiguity inherent in these  stimuli  parallels  the
 ambiguity usually present is  real situations.  The water had an
 "unusual," not  "bad," smell and  the  reported health  problems are  well
 within the range of routine health problems that  could be  found
 anywhere.                                                        , .

 Repondents are  then told that local  officials  contacted the  StaCe
 Department of Environmental Resources (DER) which proceeded  to conduct
 tests revealing benzene and trichlorethylene levels in drinking water
 above those recommended by the Environmental Protection Agency (EPA).
 The source of the contamination  is identified as  an abandoned
 manufacturing plant which has several waste lagoons and decaying  metal
 drums on the property (a picture  is -included).  The EPA is called in,
 technicians take samples, construct a temporary holding lagoon, and
 fence off the area and post signs (a picture is included).   This
 sequence of -events is similar to  those  characterizing  the beginning
 stages  of many Superfund sites.

 Constructing the Exogenous Variables

                          The Risk Communication

 Although the uncertainty characterizing risk assessment is frequently
 alluded to there has been little systematic research on variations  in
 information about uncertainty and reactions, to risk communication.  On
 the one hand,  cognitive psychologists tell us that people desire  .
 certainty and .that information indicating less-than-perfect  knowledge is
 upsetting to those who do not fully understand the nature of science.
 On the  other hand,  the admission of limits to knowledge, by  tho^se
 experts representing government agencies.,  could have the effect "of
 increasing their credibility.   Perhaps   the strategy of  projecting the
 image of the  omniscient scientist raises totally  unrealistic
 expectations  that must be'eventually dashed.  Research  on social  "
 movement* indicates  that raising expectations  can be an important step
 in the  erection of a rebellious public.   An admission £f uncertainty
 versus  the absence of such an admission constitutes the risk
 communication variate.   One-half of the questionnaires   includes the
 following paragraph  at the end of the advisory:

      NOTE:  The  numbers presented above  are our best estimates.  But
      it is  important to realize that they  are  only estimates.  They
      are based  primarily on  research with  laboratory animals, usually
      rodents, who are given  extremely high doses  of the chemical being
      tested.   If the animal  exhibits health problems,  we then  *
      extrapolate from these  high doses  to  the  very low doses usually

-------
 The 'Quasi-Experimental /Design"
 The following diagram  illustrates  the, quasi-experimental design: .

          A Hypothetical  Superfund  Situation With.Subj.ec.es,   '   • '•  .
      Instructed  to  Imagine  Living  Within One Mi lie 'of the Site
      Risk Communication.I
      A Comparative Risk       '   .
      Communication Emphasizing   .-
      the 'Uncertainty Inherent
      in Risk Estimates

      The Management Scenerios
       Management Scenario'I:
         Standard

       Management Scerierio II:
        Enhanced Citizen Participation

       Management Scenerio III:
        Indemnification
 Risk Communication II
 A Comparative .Risk    '
 Communication Not..Emphasizing
 the  Uncertainty Inherent '"
-in Risk  EJstimates         . .
 The Management  Scenerios
  Management-Scenario' I:
    Standard        '  •
       • T''"'".' .     '
  Management  Scerierio II:
   Enhanced Citizen  Participation
   .   '•   \  •••"-.••:.•   •
  Management  Scenerio III:;
   Indemnification
 As presented above, l:;he design is a, simple  two-way  analysis  of  variance
 with two categories in one variable and three .categories  in  the other.
 However, as noted in :the research proposal, this  study  is  an attempt  to
 ferret out the more important factors making up peoples' decisions  vis  a
.vis risk communication at Superfund sites.  As such,  it has  an
 exploratory nature as well as the quasi-experimental  nature  diagrammed
 above.  A number of attitudinal factors are included  as well as
 demographic variates.:  This implies that the analysis will incorporate
 correlational techniques as well as analysis of variance techniques'.
 While a multitude of .uncontrolled variables subtracts from the  decree of
 control, there is enough known about possible impacts to make reasonable
 inferences,  especially if the results are strong[and/or consistent.  The
 next section describes the design ..in .greater, detail.
                 ' '    | }~          '       :         'I .'."'•''
 The Hypothetical Situation        ".  '

 Two elements comprised qhe attempt  to create a realistic Superfund
 situation:    -       ';.'''   ,   ,„   •' ' .   .    ...-["     •".,••.     •
                      i-     -'      '        -••*•.' V  • •'  ;•
      *iB»tructions to' the respondent that,  in effect,  asked  him or
       her to role-play wiu p«.rt of  someone living; within one mile'
       of the source pf a toxic waste problem (Se€i Appendix A); and

      *a  realistic  scenerio that included pictures of decaying 50
       gallon drums and people "in  protective dotting collecting
       samples  (See Appendix A).


-------
                               SECTION  III
                       Research Design  and  Measures
Introduction
Risk communication at actual  Superfund  sites  is  a  complex  process.
There are multiple sources, multiple messages, and audiences.   These  may
change at any time given new  information  or.input  from ad  hoc  groups
that are successful in getting-attention.   Research in actual  Superfund
settings is valuable but has  problems of  interpretation due  to the  lack
of researcher control over the stream of  events.   On the other hand,
carefully controlled research  on  risk perception and communication  tends
to totally eliminate the very  features  that  define an actual risk
communication situation.  The  study reported  here  is an attempt to
straddle those two research 'extremes.   A  quasi-experimental  design  was
selected so that somewhat complex stimuli  could  be presented in a
somewhat controlled fashion.  A hypothetical  situation was presented  in
a questionnaire/interview format  along  with  pretest and postest
questions.

Several considerations guided  the design  of  this research.  Fi'rst,  the
goal was to put research subjects in as realistic  a situation  as
possible.  .Second, given the  research hypotheses,  the exogenous
(•independent) variables had to include-differences in risk message
characteristics and differences- in compensation  and involvement-in-
decision-making options offered to those  facing  the hypothetical risk.
Third, measures had to be constucted for  beliefs and cultural
understandings that past research has suggested  are important.   Fourth,
there had to be a way to measure  crucial  individual characteristics
that could cause different responses to risk  messages.   Finally,
respondent reactions to the matrix of information  provided had. to be
assessed to give guidance on  the  effectiveness of  the manipulation  and
insights into subtle, perhaps unmeasured  factors influencing their
responses.

-------
     •^individual  variations in beliefs acid sociarinstitutiorial
     _ attachments' wi
      Superfund-like ^situations.
While, this research
implications, may. -be
chapter outlines.the
operational  processes
is -at best .a. preliminary .first  step',  its
revolutionary for 'agencies, llike EPA.   The next
 translation of, the above  abstraction's into
 I.,have the greatest  irapac'
on reactions to

-------
 the threat^ of property devaluation, community stigmat iz-ation,  possible
 decreases in community economic, potential', . possible • out-migration.
 chronic stress,  and other social-definitional problems that do not -lend
 themselves readily to quantification.   These are the.kinds of  factors
 that the typical risk communication' does not attempt  to,.'deal with.  An
 impressive body  of research points to,four major sources' of'concern for -
 residents facing situations such as Superfund Sites (Armour',  1.987):

      *perceived  risk as shaped by official and unofficial' information .on
       the characteristics of of the toxic substances  (it is •important' to
       realize that the mere presence of the EPA specialists may be
       sufficient "data" to trigger high concern);

      *perceived  inequities in the distribution of  costs and benefits
       (anger in  being singled out for  such nefarious  treatment);

      *feelings of  loss of control over forces affecting the quality of
       one's  life and community;

      *and, a general lack of trust in  big business  and bvg government.

 Ideally,  perceived risk should vary somewhat  with  the  amount  and  quality
 of  risk information reaching local citizens.   The  problems  of perceived
 inequities couTd be approached with some sort of compensation package,
 this  option  is presently  not available at Superfund sites.  Feelings of
 loss  of control  and an absence of trust  may be ameliorated  by some sort
 of  intensive public participation program:  much depends on  the  depth of
 the  lack  of  trust.   The central  problem addressed by this research can
 now be  stated:                       :                   .    ;

      *what is the  relative impact  of risk communication variates,  equity
      enhancing  variates,  and' local control and  trust  variat'es  in        .
      shaping people's  judgements  about,reactions to a Superfund  type
      situation.                                     '

The complexity of  the.above  problem is magnified by the possibility that
all these factors  vary  by  some poorly  understood personality  and
demographic-personality characteristics.

The following hypotheses,  based  on  the above  analysis,  our  own  research
on fear^of radioactive waste  (Bord,  1985) and  the experience  of others
 faced with the problems of siting  risky  facilities  (Armour, 1987),  guide
 the research to  be  subsequently  discussed:

     *risk communication per  se  has  relatively little  to  do with
      structuring  reactions  to Superfund-like  situations;

     Compensation  designed  to redress perceived inequities.will have a
      greater impact on reactions  than risk information but that
      impact is  also small;         •                      _.. ..

     *control-granting options will have a greater  impact than  either.
      risk communication factors or compensation;

-------
  tenuous.   Studies  do'rie'in somewhat- sterile''Ubo.ra'dory''environments ' us in?
  low or noninvolving  ,s t imul i,. in noiisocial . si tuatiloris. cell use  little.
 .about the  structuring, of  public reactions  in redl Life situations'"
 Furthermore,  the anharisins generated by such research are of -such-'
 generality., even being  contradictory,  that applying them in'actual  risk
 communication  situations  is little more than the[ application  of common
 s.ense judgmen-t.      |  .      ,-..-.     ,   ,  ,| ...
             .        'i. .  ' '••     :    "   .      •• •  .1':' .   •  •   /
 Studies of how human, decision  making departs 'frcjm statistical
 .rationality also do  little 'to  inform' the risk'communicator.  To know
 that people do not -approach low probability,., potentially deadly .outcome
 events the way an  expert  would says nothing  abouit how to lead .them  to'
 think like a statistician.   Since  it is  extremely, difficult, to get  '   ,
 .students to think  statistically even after several statistics courses
 it.seems unlikely  that, the 'general public  can be! so trained within  the.
 temporal parameters  of  a  real  risk event.
 Social and cultural approaches  to  risk  communicajtion sensitize us to the
 possibility that publ'ic reactions  may have- relatively little to do with
 actual risk messages ;per but a  great deal  to  do Vith issues of public
 trust and social-definitional dynamics.  In actuklity,  risk
 communicators in the; field are, pointing  to  basically the' same phenomena.

 Implications for thisS Research        -           1
     •^••>V^B««BB««^^B««Bn^B_Ba ^^^••M.^^^^M^™M^MMI.           ,        I '                    1 '
 It  is  understandble that EPA seeks a better  fit  between "objective" risk
 estimates and public j reaction.  It is also logical  that the area of risk
 communication be targeted as central in pursuing}that  end.-  However,  it
 must be realized that  some very fundamental  issues  in  risk communication
 have either been, under-researched or not researched  at  all.   Most
 importantly,  little has been done 'on the actual,  market ."for risk
 communication information.  .We know, for example; that  even educated
 people  are generally poorly informed on most issues .and that many
 factors compete for people's time and attention.I The  current  policy
 trend  that emphasizes  education and information  as  a solution  to many
 social  problems is based less on any evidence ofjthe efficacy  of such
 approaches than on the. political reality that modern American  society,
 which  stres.ses  individual autonomy and decision making,  leaves' no  other
 functional possibility.   Also,  existing risk Communication  research does
 little  to inform us about the relative impact of [risk  communication, per
 se, on  public reactions.-  The research reported here is  designed to take
 a first step  in answering those questions.        '
Public Response  and  Risk Communication

Public response  is  a  function-of multiple factor's!, of which risk
messages may play a relatively minor part (Armour^, 1987; Bord, 1987;
Krimsky & Plough, 1988).    Certainly risk information plays some role in
shaping perceived threats  to  health and safety.  JHovrever, this
information can  come  from  official and unofficial sources, be unevenly
distributed in a given  community,  and vary in terms of its'"'•"•'
technical/scientific  accuracy.   A number of other! factors may have an
equal or greater impact on,public reaction: the restriction of .land use,

-------
 cotamunicacing  technical  information will result in more "appropriate"
 public  reaccions  to risk,  especially .the chronic,  low probability kind
 defined  earlier  in this  chapter.   The uncertainty inherent in estimating
 risks of this  kind provides  the ideal context for social-defiriit'ional
 processes  to dominate  (Scott,, 1988; Tarr,  1987;•Sharlin, 1987;(Mazur,
 1931).   The  social construction of risk provides  a symbolic      '
 representation of  culture  and interest based conflicts.  This
 perspective  helps  us understand 'why some Love Canal residents downplayed
 the  riskiness .of  the uncontrolled  chemicals  and others saw it as an  " •
 unmitigated  disaster.  Their reactions had little to do with underlying
 cognitive heuristics,  the  nature  of the scientif ic/techni.cal . information
 provided, communicator,credibility, or the strangeness and dread
 engendered by  knowledge  of the substances.   Their reactions: reflected
 the  major interests encompassed by their key roles and stage in
 life-cycle.  Those with  young children and unpaid mortages despaired and
 fled while those nearing retirement and hoping to sell their property
 saw  little need to panic (Fowlkes,  1987).   Similarly,  if an affected
 public strongly desires  the  use of land officially defined as
 "contaminated" they are  unlikely  to exhibit  much  fear in the face of
 technical risk messages  (Gale,  1987).   The socio-cultural perspective
 argues that  the symbolic-political  aspects  of risk communication are
 more central to the nature of public reactions than,the scientific
 credibility of the messages  or simple message construction variations.

 The  Experience of  Practitioners

 A related theme characterizes the  recorded  experience of many ,of those  •
 who  have spent time on environmental risk  "firing lines."  In personal
 interviews with several  EPA  community-relations people,  including some
 toxicologists, the following points were reiterated:

     *there is very little public  demand for general  scientific-
      technical information;  and,               ,

     *information  demands  and public reactions vary significantly from
     * case, to  case.  •                 .•-...
These practioners had a  readily
to bear on the issue of risk communication'is  structured in ways that
make generalizations to real risk  communication settings extremely

-------
        .   .      .    '.•       •       .     •       ••...   • •    ...••
 'lictie public;interest.  On  the  other  hand,'  if doroorate malfeasant -a'u
.' be linked to- the was;te.. problem,  or  if  there  appears/to be an unusual
  local incidence, of cancer, or  if  a  "save  the children" group
  arises,these constitute "news."   While  Sandman C1986)  and othe-s have
 ,made'suggestions. for;; improved.communication  betwieen  spbnsoring'agencies-
 .and  newspeople. there'is little reason  tp  s-uspect  that  scUntific-
  technical debates over chronic risks, will  soon dominate, newspaoer and
  r*aiPVTQirm r r\\i <=» T- ^ r+ A I  ?*.». ;  £  -... _ i_  j_c  ..         .  .  .  _   '  ^ - -  .
 television.coverage.]  Even if such debates were.
                                                   included  .in the major
    ,.   .         .- —  -!-         • 	 — — .— «»» v» ••»  "*-•"• *s. *. t*^i.LiUCVl  1. LI • L Cl 6
 media  it  is highly doubtful that many people'woujld' pay  them  any
 attention.   More  important', -given-the scientific* uncertainty  surrounding
 risk estimates and estimates of cleanup efficacy:,  experts with
 credentials can always be found to provide  a' countering view.   The  role'
 of  the media in risk (.communication is not likely? to  change significantly
 in  the. near future,  i:            •        •           -.=»&.
                     ir   ,     •              •  '

 The Soci-o-Cultural Tradition

 The public's understanding and reaction to risk also has been  approached
 from a socio-cultural  perspective.   Douglas,and tfildavsky. (1982) contend
 that Americans  view  the  world  as  frightening and[increasingly risky  in
 spite of significant ireal  increases in health arid longevity.   They  link
 this pessimistic view: to an interaction between cultural notions of
 purity and  progress and  the realities  of conflicting interests  generated
by^multiple  lifestyles and organizational  affiliations.   Others would
point to the  growing ;" evidence" that modern  technology,  generally
-promoted by  big business,  developed by  corporate [science,  and regulated
by big government, is  beyond the  scope  of  ordinarjy  understanding and
ordinary control and management (Perrow,1987).   BJronstein  (1987:223)
presents  this perspective  succinctly:
      "...the verified'existence of a  risk  is  not
      the danger to be:; publicly recognized...   The
                                                  sufficient for the
                                                   public  understanding
                      ,--     -     W   	--—.•*.  fr**. v J. * W  HIAUC^ 9 UCL11U, .L L
      of danger bears only a tangential  relationship to the objective
      evaluation of the riskiness of the substance], but a  close
      relationship to the political impact  of  identifying  the  risk."
   -            ' •  • -   ir-. •     ••:.'••    ..'•-.-'  ^ .•      ,    '     ' .   •
 In  other words, risk identification and public  reaction to risk may be
 viewed  as a challenger'to existing institutions  rather than an imperfect
 digestion of information on risk.  This perspective helps  us  to
 understand why public.reactions to risk situation^ tend to take on  all
 the  elements of social movement development (GerLach, 1987).  Risk
 estimation and communication can be viewed as driven by the demands  of a
 modern  protest movement.   From this perspective there is little-
 likelihood of ever being able to satisfy the most[vocal elements of  the
 contentious 'public.    Every response from agencies to demands  for more
 and  better information can be challenged as inadequate and result in a
 new  set  of demands.   Furthermore,  the increasingly litigious  nature  of
modern America almost insures  that  a  significant subset-of environmental
 risk situations will  wind  up in protracted 'court deliberations
 (Jasanoff,  1987).       '
From a socio-cuttural
modifications in risk
                      perspective,  there is little reason to expect that
                      estimation  procedures and improvements in
                          ; ,   '.    ,   ..-,.•     -.  4. '  _  ...-,.

-------
       ^people respond co problems as they see them noc as :hev ar-=
        (Fischhoff,  1981);    '  •      •    '              ',•'."

       *new information will be assimilated to fit existing'attitudes.
       (Slovic,  ec.al., 1979);  and,                          '

       *risks  distributed unequitably tend to generate more outrage
       (Slovic,  et.al,., 1980).       '  :        •

While  the  above noted body  of  literature has certainly increased our
knowledge  of  judgmental biases in certain kinds  of risk'decisions the
implications  for risk communication are  unclear.1  These same  authors go
on  to  note that the  perception and  acceptance of risk is  rooted in
social and cultural  factors (Slovic,. 1987).   Because it is difficult to
manipulate these social and cultural factors or  the imagiriab'ility.
memorability,  likelihood of fatality,  catastrophic potential,  signal
potential,  or voluntariness of a given risk  in a risk communication
program, we have limited scientific information  to guide  the  risk  •
communicator.   In fact,  each of  these  factors is open to    :    • -   "'
social-definitional  processes:  these qualities are not inherent.in a  ;
stimulus or configuration of stimuli,  they are created in the  process  of
communication itself.   Risk communicators  can, and should, provide
information, that helps the  community-understand  the scientific backround
for agency decisions  and the cleanup process itself,  but- many  of the
critical variables that influence individual decision making are beyond
the control of  the risk communicator.  This  is one reason that equity
issues  demand attention.  Addressing equity  concerns  has  the potential
to  temper  opposition  based'  on  unreasonable fears.

The. Impact  of the Mass  Media -.'•-.

Research dealing with  the impact  of the  mass  communication media  on the
perception  of risk produces  consistent generalizations:

     *balanced media  coverage  of  man-made  risk situations  is unlikely
       since the  criteria for newsworthiness  tends  to  be conflict, drama,
      human interest,  negativism,, photographability,  exclusivity, and
      newness (Peltu,1987).

In a recent analysis  of  network  evening  news, Greenberg,  et.al.(1988)
basically  reiterate the  above  noted  quote:

     "Risk  as calculated by  scientists .had little  to  do with the  amount
     of coverage provided by the  three networks'  evening  news
     broadcasts.  Instead,  the networks  appear to  be  using the
     traditional journalistic determinants of news  (timeliness,
     proximity,  prominence,  consequence, and human  interest) plus
     the broadcast criterion of visual impact  to determine the  degree  ,
     of coverage of risk issues."(p. 28)

Because "news," by definition,  requires  acute incidences  to promote
public attention, the  long-term nature of  Superfund problems and  the
nature of the information needed  to  understand the  problem generate

-------
 improved risk cotnmunication. .-• However.  knowing  i'hat 'people tend so fUu's.
 on, Che outcome of a^given risk, and  ad just  insuf If icient ly provides some"
 guidance -oh.message;construction  if  it  is' known| which,"outcome people are
 likely, to use as anianchdr. 'in, the  "tangled .webs"  of'risk communication
 at  'actual Superfund'sites., the anchors,  or benchmarks,  that people use
 in  risk estimation may "float" to, some  degree a^  mixed  messages'are"
 received from the•media, the experts, arid. Interpersonal communication
 networks.   Further research is needed "to  bridge!the gap''between an  '    '
 understanding of how; people make  decisions in .highly  controlled,-
 laboratory  situations and how they utilize multiple channels  and  '•
 messages in complex,'; shifting, situations.    ',• •,  .'        •  ,      ",-'  ,'

 While-the .literature on decision making biases  and judgmental heuristics
 is .stimulating and interesting,  it provides- limited-.useful- information
 about  effective riskii communication. (Perrqw, 1984).               '

 The  "Rating the Risks" Tradition .   '  :    '•''..

 A number of psychologists from the decision heuristics  school  have
 applied  that-perspective specifically to .individual biases  in  risk
 estimation.   This  isj; the "rating  the risks"'tradition .of research which  '
 applies  psychometric, methods,  using paper and pencil'  tests, to  very
 limited  samples  of- people.  .Psychometric methods!'ask  subjects  to'compare
 objects  or^concepts  using their  own,  individual,! anchors or benchmarks
 for  comparison.  These researchers also  have generated'a set  of
-principles ^that  have:become  dogma  in the risk perception literature.'
The  following  are  the. most  frequently repeated aiphorisms:   "
      *events that are highly  imaginable and menlorable, which have a
      high l-ikelihood -of being fatal , , and high catastrophic potential1
      are .-more likely to be  feared  (Slovic,  Fishhoff,  LichtensCein,  '
             ••      -   • •      '            -         -'     •-••'
        .   -               _...  ,...                -       .    -  .   .
      *things that are unobservable,''  unknown.,  new,  arid have delayed
      1 consequences ire more  likely to  be fe'aredj (Slovic,  1987);

      *the" limited power of peoples'  information! processing capabilities
      tends to result in overconf idence,  in errotieous  decisions, and a
      a desire for certainty  (Fischhoff,  1981);[

      *presenting relative risks in a comparatiyb format  enhances'
      understanding but may mobilize negative  reactions if the
      comparisons appear specious or manipulativ^ (Covello,  et.al,
       -               •                           -        • -    -  •       "
                  .•          ••         -       -     •-  ••••           •
      *events that have "high signal .potential"  (somewhat  insignificant
     events that are interpreted as indicating tlU potential  for  future
     catastrophe), .such as a small accident  in an unfamiliar  and  feared
     system,  will increase fears dramatically (Slovic,  1987)-        ,•  ' -
             ..   .. .:!..'.;.. ,; ..-. :  •-.-   .---.  ;..'-• ,.f:; ;,.v •.-,,.-,'.     :'
      *greater -risk i,,s tolerated, if that risk is I voluntary^ "immediate,  '
      known precisely, and controllable (Fischholf, -et.al., 1978);

-------
                 * and, the behavioral outcomes of the change are unknown. '      - .

            The aphorisms generated by this research deal with aspects of  the
            communicator, the communication, and -the audience.  This research
            tradition is tied to Carl Hovland and the Yale Communication Program
            which began in the.late 1940's (Hovland,  et.al, 1949).  Basically, they
            tell us that the. communicator should be trusted, interesting,  and'
            attractive; me.ssages should be at the appropriate level of complexity,'
           'clear,  use audience.arousing devices but make sure. they are not aroused
            inappropriately or at too high- a level;  if the audience is somewhat
            highly  educated and heterogeneous then-messages should be presented In-
            different ways and include multiple perspectives so that audience
            members perceive an attempt at'objectivity.            ,

            The shortcomings of this  body of literature,  especially as it  pertains
            to  the  risk communication problem,  should  be  obvious.   In contrast to
            the typical attitude change experiment,  communication at Superfund sites-
            tends to be in an  emotionally charged, atmosphere,  there are multiple
            communicators  and  message channels,  the  focal issue  has  a history and
            and this  translates  into  preexisting attitudes that  act as;filters • for
            incoming information,  the "official" communicator  represents  a'•
            government  agency  that  may have  less,than  total public confidence,  the
            topic is  complex and characterized  by  enough  uncertainty to generate
            mixed evaluations  even  from experts,  information transmission and
            interpretation  involves ongoing  social networks, and  the message   '
            recipients  have  the  power  to  choose  which  information  they want to
            expose  themselves  to.   In  this, communication  atmosphere  the above noted
            aphorisms are  little better  than basic common sense.

           The Cognitive Heuristics Tradition          "   ,        ,

           The work on decision-making and  problem solving  heuristics, at  least
            that part of it  applied to  risk  information^ processing,  has centered  on
            typical departures from statistical  "rationality."  The  leaders in  this
            tradition are Kahheman,. et, al,  (1982).  Generally, people  fail to  take
           sample^size, population characteristics, and.objective odds into  account
            in  their decision making.   Furthermore, in low  probability, high  risk
           situations, people tend to anchor on the negative outcome and adjust
           insufficiently downward for the low probability.  McClelland, et, al.,
           (1986) indicate  that there may be two modal responses  to low
           probability, high risk, situations: some people  focus on the low
           probability and dismiss the risk as not worth worrying about and others
           focus on the outcome and become overconcerned.  The determinants of
           these different decision making .styles are unknown.

           This body of research shares some of the problems noted in the  attitude
           change research-  It tends to be laboratory research characterized by
           the  following:  a single, low-involvement,  issue to be judged in an
           atmosphere free of social  interaction and pressure.  The principle
           generalization, that people do not make decisions the way an ideal
           statistical thinker'would, should come as little surprise and, at least
           to this  point,  has  not translated into solid  recommendations for
.

-------
 a judgment made  by  she' sponsoring'' government agjencv and- is at "-as-  *s
 .political in. nature  as  it . is  scientific      •  I•   '"         '  •
                  "•';••.-.„.':"   '      .;]''.  •  '  .    •..  .
 Risk Communication: .'What' Do We Know      '•     ["   -  •"'-.'

 The emerging  subdisedpline o.f risk communicatio.il  has' evolved" from the
 following:          ; •  •        ..,-.'.'        '               '
     -.* che attitude'change  research  from psychology,. 'social psvchologv
        and marketing;                •   %          .             - .     &. .

      * the work on decision-making and heuristics from-cognitive
       , psychology, and to,a  lesser extent,  micro-economics;

      * research on. risk estimation being done'primarily by       ..   -.
        psychologists;     -    ,                  -|

      *. research on the impacts of the  mass  coramimicatibn media;
      * case stu'dy research and theorizing  emphasizing  the  social-
        cultural context of risk; and
                                                              V
      *  the recorded accumulation of the experience  of  those  charged
        with risk communication responsibilities [in  actual  field
        situations.   •',                        ,   4      ,
    '  •          '     •!•     •     '.       •         'I    '•;•••..•• i-  ••'-
Most  literature reviews have somewhat uncritically  combined  insights
gleaned  from-these  diverse sources.  This is to ibe  expected, and  is a
valuable  contribution,  in the formative stages 6|f an .intellectual
endeavor,   However,  the combinatorial process teWds to neglect serious
questions  of suitability and actual substantive 'contribution.  While a
thorough-  critique Is well beyond the scope of thjis  chapter* there are '
some  relatively, obvious points that need to be made.
  •  •  •               . •             •               i                '   .
                                                 f- '.-.•...         ...
The Attitude Change  Tradition       .   .         I  -   -
                     I - -  '  ";       ,    ,       "     ,'p " '   •      ,      '  '
The attitude change "literature in social  psychology has gelled into a
number of  concise principles that can be  found inmost introductory'
textbooks.   This  research exhibits  the followingj-characteristics:

     * .it  is done in,laboratories' using college students as subjects;

     * it generally  explores the  impact of a  single independent
       variable on some measure  of  attitude change;
                     . i     '   -     •'            • i  •-.•'.•••'•..
     * it  involves topics  that engender little  emotional involvement
       or  the level  of  emotional  involvemient  goesi  unmeasured;

     * the research  situation  precludes social  influence;

     * the topic  seldom involves  risk  of any  kind-       '       "-
         •   .-.••   •   ;:   "•      •"                i   ••  -.'.-   .-.  •"    •••
     * the results tend to reflect  very small changes  in attitudes;

-------
                      SECTION II:. RISK COMMUNICATION              •..'..

    "Because of the complexity of risk communications .and  the  subtlety  of
    human response to 'them, it is extremely difficult, a priori,  to know
    whether a particular .message will adequately- inform its recipients"
    (Covello, et.al.,  1987: 31).                '                ,- '

      ...risk communications in.their social context resemble  tangled
    webs,  in contrast  to parallel series of sender/receiver interactions"
    (Krimsky & Plough, 1988: 298-299).

 Introduction                                                      '

 There presently exist several excellent reviews^ of both a theoretical
 and practical nature, of  the risk communication literature (Covello,
 et.al.,  1987;  Covello,  et.al, 1988).  These reviews guide the reader
 through  the..morass of materials  that bear,  in widely divergent ways and
 with differing, degrees  of applicability,  on the risk communication issue.
 It  is not  the intent  of this  chapter to simply reproduce those
 discussions.  Instead, the focus  will be on generic problems plaguing the
 very concept of risk  communication,  the difficulty in doing research to
 answer the kinds  of questions risk  communicators ask,, and how the
 research reported here  is designed  to address a  subset of these
 problems.

 First, the major  threads  characterizing research on risk communication
 will be discussed and their relative contributions highlighted.   Then,
 the basic  assumptions of  the  risk communications enterprise will  be
 reviewed and  their difficulties  noted.- Finally,  the implications for
 research on risk  communications  will be discussed  and  cast into  the
 framework  of  the  research to  be  reported  here.

 It  is  absolutely  essential  to keep  in mind  that  this report  is concerned
 with  a particular type, and context,  of risk:     •

      THOSE  SITUATIONS -IN  WHICH A SOMEWHAT HISTORICALLY AND SOCIALLY
      UNIQUE COMMUNITY FACES AN IMPOSED, MAN-MADE,  DREADED  SUBSTANCE
      THAT  IS GENERALLY  CHARACTERIZED AS SOMEWHAT ENVIRONMENTALLY
      UBIQUITOUS,  OFTEN  HAVING MULTIPLE  EXPOSURE  POSSIBILITIES, BEING,
      DIFFICULT  TO  ELIMINATE EASILY OR COMPLETELY,  AND HIGH SCIENTIFIC
      UNCERTAINTY  RELATIVE TO  EXPOSURE AND LONG-TERM IMPACT.

This  is the nature of most  Superfund  situations.   We specifically bound
 the  problem in  this way, to  make  it .clear  that  the  communications
difficulties in this  situation may be quite distinct from  those
 situations in which a communicator is attempting to  mobilize people in
 the  face of an  imminent and known risk  (hurricane),  a natural  risk
characterized by  some uncertainty in  terms of  long-term impact (radon),
or a  self-selected risk characterized by somewhat  complex  risk-benefit
outcomes  (a new drug).  Communication effectiveness  in the-latter cases
is usually defined in terms of number of 'people mobilized  to tatce
recommended actions while in  the Superfund case effectiveness  is  viewed
as the "appropriate"  intensity of community response.  "Appropriate"  is

-------
Their responses 'are rins tractive and add.further[weight  to' our  '
interpretations of  tihe  structured questionnaire^ results.,
Section XIV, "S-unrnarv' and  Implications for Risk [ Communications ,"  is  a
discussion of  the .niafjpr  findings :and what they say: about  communicating
risks iii Superfund-like  situations. •           j .    .•'.'  ','   .  ',.  -' .

-------
  section VII,  The Hypothetical  Site:  Levels  or  Concern,"  consti-uteV --e
t  results of questions asking  respondents  how concerned  they  would  be
  ^ving within one mile of the sice both  before  and  after  completion' of
  the cleanup.  This comparison provides one measure  of  the credibility  o-'
  the cleanup scenario.  Various  categories of  respondents,  are  explored
  ror variations in pre- and post-cleanup  levels  of concern..

  Section VIII,  "The Hypothetical Site: Health  Risk Assessments," also
  examines judgments both before and after the,cleanup.  This section
  however, focuses, on the issue of perceived health risks to  self and'
  family.   Again,  variations in health ris.k concerns are compared across-
  categories  of  respondents.                              .'...;'

  Section IX,  "The Hypothetical Site:  Comparative Risk Assessments,"
  presents the data related  to  the relative accuracy' of,the 'individual's
  risk  assessments.   Respondents make  judgments on how'risky living near
  the waste site is  in  comparison to a  number  of voluntary and involuntary
  risks:  smoking,  living  near a nuclear power  plant,  driving a car,  riding
  a bicycle,  chest  x-rays, and  eating  fruit sprayed with pesticides.
  Both  pre- and  post-cleanup judgments  are  made.  Besides providing a
  measure  of  relative accuracy,  this procedure  provides another  index of
  Che respondent's  faith  in  the cleanup scenario.   These judgments are
  related  to  the respondents' demographic characteristics.

  Section  X,  "The Hypothetical  Site: Behavioral  Intentions," describes  the
  distributions  of responses to  various  behavioral options  one might take  '
  in a situation such as  that described  in  the  hypothetical  situation.
 These options  range from those that demand very  little  personal
  sacrifice or involvement to those that demand  putting one1s^self  on  the
  line in a public situation.  Fifteen  of the behavioral  options
 effectively scale into  four subscales  that can be logically  interpreted'.'
 These scales and their  relationships,  to respondent characteristics are
  •-*, j :ii3sed.                                       ,

 Section XI,  "Explaining the Results:  Bivariate Relationships," examines
 the  factors  that relate to four sets.of variables designated; as
 dependent variables in this research:  levels of  concern, health risk
 assessments, comparative risk assessments, and behavioral intentions.
 This  section sets the  stage for understanding  the complex pattern of
 relationships that characterize the overall results.

 Section HI,  "Explaining the  Results-:  Multivariate Analysis," presents a
 series  of stepwise multiple regressions that  provide  estimates  of the
 relative contribution  of a  substantial number of independent  variables
 on the  dependent  variables  noted above.  The  independent variables that
 account  for  the most variance  in the  dependent variables change with the
 dependent variable being, analyzed.  The pattern of responses, although
 not what  was  originally  expected,  is  consistent enough to' generate a
 reasonable interpretation.

Section XIII, "Responses to Open  Ended Questions,"  is an overview of
what the  respondents wrote  in  answer to a  number  of open-ended  questions.
Respondents were asked about their reactions  to the hypothetical
situation, the  management sceneries, and the  risk communication itself.

-------
                "  "  "j:.  .- -.SECTION  I:  INTRODUCTION--'• • - '•' "  ' ". ; ' '  •  ••'
                    ' ijv  .   : •  •' /. '.•..;•   •   '    .  '*-- '   •.  ^    .-  ••-..  ' • -  .  •- •
  The, goal of the research reported here  is  to: provide'.information usefu,;
  ror 'designing a risk'communication  strategy" that  will', reduce        •' '
  disjunctions between perceived and  objective-risk levels.  'In addition
  the _ hope--.is that these, results will  further  the, EPA's  goal, of .helping'
  citizens make their;.own risk,management  decisions (Fisher.,- 1987).'   .-"

  This  report is comprised of a series  of  topics that  lead.the  reader
  the  study;  describe ;;the .study .itself, present  the results-  of  the	
  analysis,-  and discuss  the implications  for risk].communication.   '
into
 Section  II,  "Risk Communication," is an .overview .and  critique  of the
 risk  perception-risk communication literature..  This  section  raises
 questions  about  the/ability of past '.research to', answer  the  questions  '•• '
 that 'risk  communicators !ask.   It is generally assumed that  formal  risk
 communication;is  important, but, in fact, therejare weir-founded reasons
 to question  that  assumption.   Other factors may.]be more  important  in
 structuring  public  reactions  to Super-fund situations'. Hypotheses'
 .suggesting these  other  factors are .presented atjthe end  of'this . section.
 Section III,  "Research  Design and Measures," presents a discussion  of ,
 the research  design,!  the  key concepts,  and'"'the construction of-measures •
 to represent  the  concepts.   The quasi-experiment-al design and  its
 rationale are discussed in  detail.   "Before" and "after" attitude,    •
 belief, and behavioral  intention measures are outlined.           '  ,   "
 Section IV, "Sampling and  Implementing the Survey," describes the
 sampling procedure'and how the  survey, was implemented.  The actual
 sample is considerably larger  than .initially proposed and selected to  ,•
 reflect the,range of|; social  groupings  central tci community 'responses to
 toxic waste threats.]  The  sample  includes environmentalists, 'business
 persons',  civic activists,,  elected  officials, : and" working class people.
 An approximately even mix  of males and .females w'as sampled.

 Section V, ,"Prior Attitudes-, Issue Involvement,  and Information,"
 delineates  the results ..of  the pre-manipulation measures .of general and
 specific  levels of concern and  estimates  of  the [health, effects of
 exposure  to man-made;t6xins.  These  items are eslsential in assessing the
 impact of the risk communication and-the  degree of correspondence
.between objective,and subjective risk.  In addition,  levels of,  and    ,-
 sources of,  knowledge about toxic  chemicals  are .tapped by multiple
 measure*.        •'!-.-'          -  . / •'   •• '     • *-.•'.-.'•'   .•••  •'..  . .  .'.

 Section VI,  "Trust,"! presents the  results of  the analysis^of the  various
 items assessing trust.   Trust is an essential  component of- the
 communicator credibility problem.  If  communicatbrs  are not trusted,
 then  the  quality or format of the  risk information is  meaningless.
 Distrust  means disbelief.  Measures, of. trus.t  in  local,  state,  and
 federal governments,;industry,  and science,are  included .in-this
 research i       "    , '   •., •    '   •     •',.-''"''"

-------
 Limitations.                            '•',"'      ,

 It  Is  possible_that  the  reason the' hypothetical site information had so
 Little^impact  is  that  the  differences between'the manipulations were not
 dramatic  enough.   Events  at  an actual site take place .over rime and are
 often  highly dramatized  by the media.  A'study of,.this type cannot
 duplicate  the  reality  of  an  ongoing site conflict.   Respondents 'provided
 feedback  indica-ting  that  once  they, had seen the pictures " with the      ,"
 rotting 50 gallon  drums  and  the technicians in protective clothing their
 attitudes were- formed.  'The  management scenerios and the, risk messages
 were processed  through a' lens  colored''darkly'by .the. initi,al 'information..

While  the sample size  is adequate  for .some analyses it is  too small for"
 robust  subgroup analysis.  Therefore, subgroup analyses must  be viewed
 as suggestive .and  judged in  the context  of. overall  patterns and other
 available research findings.     .

-------
      "•Che communication- network  tends  to  focus  too  heavily  on local
       officials and inotables , who  tend  to  engender  little  trust  in   •
...  •    situations like .these, and not -enough  on  local  residents.
      -          .    •];   .   ,- '.• .  .   - ''  .  • •'  •    -fc.":    •
                    V-'  .Implications  for  Policy!'" •

 A number of recommendations  follow from our  analysis
       -'       .-   •:''   •      •   ..-•'-•  -.' '.   i   •' •  .: ...   '.   '.   •  •'
      *The public neeids a better understanding  of  the  risk  cleanup
       process and constraints on the agency  so  ihat their  expectations
       more closely match reality.  Educating-about  waste  cleanup  may
    , •  be as important as educating about  the- characterisitcs  of  the'
       risk.     '   j                      ''I--'.1'-
                                                 focus only on local
*Educating  about  risk and  the cleanup process should begin as soon
 as  the agency .begins its  involvement in the community.  The agency
 can  prepare  some  generic  material,  in laypers.ons'  terms, about
 what.is  involved  in  a cleanup action and jirobable time frames.
 Locals should also be informed about the•sampling and testing
 process, the rationale for  the protective  clothing used by
 technicians, and;why determining risk takes time.

*Agency personnel  must be  made aware that Virtually every act
 and  pronouncement is viewed  as a statement!: about relative risk.
 Programs designed to elicit  good will may  actually be perceived as
 indicating high risk.

information and education programs  cannot
 officials and notables.   These people do not enjoy high
 credibility and tend not  to  be tied into Crucial information
 networks.  Information- education programs! are  probably more
 effective at the  neighborhood level.       [

*It appears that the  average  citizen is  not! being adequately
 prepared to make  risk  decisions  in  modern  [technological society.
 There appears to  be  a  need for a broad  publlic information program
 about environmental"  risks that can  help citizens make  their  own
 risk management judgments.   Given the increasing need  for citizens.
 to pass judgment  on  difficult  trade-off options, it  seems
 appropriate touring  such science,  technology,  and society
 training early •• to a  person's  educational experience.   Such a
 program of dispensing  objective  risk  information is  consistent
 with the purposes and  legislative history  'of CERCLA  and SARA.

               :  Limitations  to  this Research
Any individual piece of  research  has  significant! limitations.and
shortcomings.  The research  reported  here  is  no exception*   Two problems
require mention: problems with  the  stimulus materials  and sample size

-------
 Respondents seated that they would take actions including getting -.ore
 information,  using bottled-water, leaving the area, and contacting    -
 Congress to put pressure-on EPA.  Intent to become involved'at  the  level
 of actions requiring the least commitment (e.g., talking, serving on a
 committee) correlates strongly with great concern'with toxic .waste
 before learning of the particulars of the hypothetical case.  Personal
 actions, such as,using bottled water or leaving the area, -correlate:
 most  with distrust of local officials,  low education,  and feminine
 gender.   'Convincing- others to become involved or to use bot'tied water
 relates  strongly to high levels of .concern with toxic Waste issues
 recorded before the hypothetical case was presented.   Finally, political
 actions  (e.g.,  contacting Congress,  organizing a protest meeting) again
 correlates most strongly with prior concerns and issue involvements.
 For talking,  involving others,  and political actions,  exposure to the
 indemnification option has a slight,  yet statistically significant,
 impact on. encouraging those actions.                           .
                                           •"•-'- i
 Poor  risk estimates seem to be generated more by distrust than by
 anything else,  although less education  and  feminine gender also  have
 independent impacts.   The  intention to  take personal  action to protect  .
 one's  health  (e.g.,  buy bottled water)  also seems  generated by these
 same variables.  However,  the other actions (talking,  involving  others,
 and political)  come more out of a history of concern  with hazardous
 waste  and involvement in hazardous waste Issues  prior  to this  '
 hypothetical  case.                                        .   •

         Problems in Communicating Risks  to  a Distrusting Public

 The problem for EPA in  devising a strategy  to achieve  greater
 consistency between subjective  and objective risk  estimates  is. that  many
 key predictor variables  -  trust in local  officials, gender,  education,
 prior  attitudes and involvements  - are  factors  that are  difficult or
 impossible  to control by means  of a  simpl'e  risk  communication  program.
There  are  strategies, however,  that  appear  worthy  of more  attention^

Trust  is  very important.  The most cleverly  crafted risk messages     '
are unlikely  to overcome distrust.  Distrust  is  a:function  of  many
 factors,   some which  the  agency  has a  certain  degree of control
 over and  others which it can do little  about.  Our  respondents,  in
 open-ended  questions  and group  discussions, 'indicate that  three
 site-related  factors  promote distrust:

     *a general lack  of understanding of  the  cleanup process as well
      as  'problems in  risk estimation;

     *mixed messages - official communications tend to define  the risk
      as  manageable and minimal while the behavior and costumes of the
      agency personnel, and possibly  the sights  and smells
      characterizing  the site itself, indicate that it is.-very  ,      .-
      dangerous; and,                                          f

-------
                  Concern  is .High.and-Risks  Overestimated

  Levels of concern were uniformly high,  even  aftkr. the cleanup.   In the
  health risk estimates and the comparative  risk [estimates,  most   '•••  '
  respondents severely overestimated-the  risks  frjom living  near the  s< te
  Most said' chey. definitely would take  some, action  such as  using 'bottled
  water,, contacting Congress, or-even  leaving  the: area.            "
                    • ••           -                i,-        '•  •
  The four subsamples ' (members of environmental groups,  civic leaders,
  elected officials, and business people) are'quite  similar' on all•of'
  these measures.   Elected officials are more  trusting, of everyone,  not
  just themselves,-but the similarities are'more [striking than the'   "
  differences.   When faced;with a response action, almost everyone seems
  to  become an  "environmentalist."
        .   "          ','.          •              . '  |-  '••'••-•'•-'.•'
  Essentially the  risk communication and management  options  have only a "
  slight . impact  on  consistency between subjective and  objective .risk
  estimates.  Providing the uncertainty disclaimer has a .slight tendency
  to  increase message^credibility,  and' therefore trust, for  some
 • respondents.-  ,The  availability of  indemnification, however,  does ,not
  re'duce  concerns and health risk  estimates,., but actually increases them
  somewhat.   Providing;  indemnification signals  to [.some respondents that
  the  situation  is  so iserious  that  their'health is in .immediate danger.

      Prior Attitudes,'  Trust,  and  Respondent Characteristics Important

 Other factors  account  for much of  the variance  £n  how people view a
 hazardous waste situation.   Three  sets of  variables account for risk
 •estimates: prior attitudes about,  and involvement  with,  hazardous waste;
 trust; and demographics  such as gender and education.
 Level of concern correlates  strongly  and  negatively,with two scales:
 t:vasting local officials and  industries;  and  trusting state and national
 agencies.  State and, federal  officials  are  trusted  more, than local
 officials.  The key 'finding is  that many  peopleiviev local  government
 officials as working closely  with, local hazardous waste generators,  and
 see other TeveIs of government  on- a different  dimension.  People who do
.not trust local.officials to  tell  them  the  truth are overly concerned
 with living near the site.  Because a majority o'f respondents  would  not
 trust local officials in this matter, there are-high levels of concern
 with living near the site,.                      '
 Health, risk assessments also vary with the trust! variables,  but  gender
 and education are also important determinants.   Women and  the  less well
 educated are much more likely to overestimate their chances  of suffering
 illness from the site.                     •   . - .  •                   '

 In  comparing the risk of living near the site with voluntary risks such
 as  smoking,  women,  those with less education, and those who express  less
 trust  judge  the site as more risky.   For the comparison withi involuntary
 risks  such as eating :fruit treated with chemicals, those with 1-fess
 education and those who express less trust judge!the site  as more risky.

-------
                              EXECUTIVE  SUMMARY
  The purpose of  this  research  is  to  learn  what  risk communi-ation '
  stracegies can  achieve greater consistency  between the  sub jecciv- ' r< s<
  estimates or citizens and  the objective risk estimates  derived  bv
                                                                   •
The Research Process

                        '    a countv that- had
                                   "
  n>h        K.                                  a  countv  that-  had
  neither a highly publicized Superfund site nor- other dramatiic" hazardous
  waste controversies.  The sample was stratified to include members"?
  environmental groups,  business persons,  civic activists, and  elected
  officials as  well as a segment from the general population. .    '

  Each  interview began with questions about the respondent's concerns with
  hazardous waste,  risk estimates of health problems caused by hazardous
  waste sites,  pn.or  involvements in waste issues,  and information about
  chemical  waste matters.   Then,  respondents were asked to imagine
  themselves  living within  one mile of a  Superfund site.   They read a   •
  typical Superfund scenario  including a  response action,  an EPA water
  advisory  notice that small  amounts  of some contaminant  had gotten
  into  the  water  supply, the  formation of  a  concerned  citizens eroun and

  e
                                                              Ntional
    k     n    ,       descriPcion  of. <*e  site  included  pictures  of
 leaking 50 gallon drums and  technicians - in  protective  clothing taking
 samples .                         ,                             °     ° -

 The scenario ^ provided to half of the  respondents -included  an  uncertainty
 disclaimer with the water advisory notice.  The disclaimer is  a '        '
 paragraph explaining that the risk estimates in the water  advisory
 stud?LareThery =af f°US *«i«te. based on extrapolations  from animal
   ?h   ;h     I     " madS  Chat scientiscs Attempt to  overestimate
 rather than underestimate the risk.  The other half of  the  sample did
 not receive this disclaimer.  '                               ^pj-e

 One-third of  the respondents  read a management option that  included
 enhanced  citizen participation through EPA agreeing to work closely with
 a  citizen advisory  committee  at  all stages of the remedial action.
 Another third  of the respondents received  an indemnification management
 option th.t involved. the State agreeing to purchase the home of anyone
 within one mile  of  the  site at the, fair market  value prior to
 identification  of the hazardous  waste problem.   The final third of the
 sample received  the  standard  EPA management  procedures  including giving
 hearings    aS313tan«  S"nt  to  a citizens'  group  and  holding* public


After  reading the material, the  respondents  were asked  to record  their'
 level  of concern from living  near  the  site,  health risk  estimates
comparisons of the risk  from  living near the site  with  other voluntary'
(e.g., smoking) and  involuntary  (e..g.,  eating fruit  sprayed -with
insecticides) risks, and actions  they  would  take if  faced with  fhis type
or situation.       '                                              .   '    .

-------
                            TABLE OF CONTENTS'
• XECUTIVE .-SUMMARY;	!<.,.... .'.. . .' .	:-. . .'.


   SECTION I: INTRODUCTION....... ,1 ....... ...'..'..


   SECTION II:  -RISK  COMMUNICATION.'... . .'.	


  . SECTION 'III: RESEARCH DESIGN  AND MEASURES'. .'.


   SECTION: IV :  SAMPLING AND  SURVEY  IMPLEMENTATION^.. . . . v .'....'....'	21


   SECTION V:'PRIOR  ATTITUDES,  ISSUE INVOLVEMENT],AND  INFORMATION..28


   SECTION vi:  TRUST.'...•....'............	,'!.;....;.	.35
                   14
   SECTION. VII:  THE HYPOTHETICAL  SITE - LEVELS OF CONCERN	-:46


   SECTION VIII: THE ^HYPOTHETICAL SITE - HEALTH 'RISK ASSESSMENTS...48
   SECTION IX:  THE'HYPOTHETICAL -SITE -
                                 COMPARATIVE RISK
  SECTION X:  THE HYPOTHETICAL  SITE  - BEHAVIORAL
ASSESSMENTS.......51


INTENTIONS........57
  SECTION  XI:  EXPLAINING THE  RESULTS  - BIVARIATE .RELATIONSHIPS. . ... 63

         .   •-'   • :. '"  i;- '.,-  : /••.. ':•' '  .•: '•  ••:  .  =,  ••'{*••'.' '  •"-•.••'••  ''* •   ••• '•
  SECTION  XII:  EXPLAINING THE  RESULTS  -  MULTIVARIATE  ANALYSIS.'. .. .72-


  SECTION  XIII: RESPONSES TO  OPEN-ENDED  'QUESTIONS.„.......	....85


  SECTION  XIVV:  .SUMMARY AND IMPLICATIONS  FOR RISK COMMUNICATION. .. .88^
         •   -•    ...  ;•  .    . -.;   -    ..,.    •     .,^..-..  ....   .- .  ;-.

  REFERENCE BIBLIOGRAPHY..	........ . . .\ .. .,,,./..,;,..;... .'..'....:  92


  APPENDIX A:~  HYPOTHETICAL SITE'SCENARIO        I

              •RISK COMMUNICATION    '       -      I    . '                '

      .     .    MANAGEMENT OPTIONS. .........	[.-..,.	......'..,95
  APPKHDIX  B :  PRE- AND POST-MANIPULATION QUESTIQNNIARES. ..	.... .106

-------
        ACHIEVING  GREATER, CONSISTENCY
    BETWEEN SUBJECTIVE AND OBJECTIVE RISKS
               Richard J. Bord •       •
       Associate Professor of Sociology
               . Donald J. Epp
     Professor of Agricultural Economics
              Robert -E.  O'Connor
   Associate Professor of Political Science
               Prepared for the
United States Environmental Protection Agency

   Cooperative  Agreement No.  CR-814592-01-1
  Environmental Resources Research  Institute
   •   The Pennsylvania State University
     University Park,  Pennsylvania 16802 /
                 April 1989

-------