ML20133C733
| ML20133C733 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 11/21/1980 |
| From: | Ham K, Miller R, Tokuhata G PENNSYLVANIA STATE UNIV., UNIVERSITY PARK, PA, PENNSYLVANIA, COMMONWEALTH OF |
| To: | |
| Shared Package | |
| ML20132B167 | List: |
| References | |
| FOIA-85-285 NUDOCS 8507200558 | |
| Download: ML20133C733 (43) | |
Text
- . ?, '.
HEALTH RELATED BEHAVIORAL IMPACT OF THE THREE MILE ISLA h'~
PART II
.\
'Ir Report submitted to the TMI Advisory Panel on Health Related Studies '
of The Pennsylvania Department of Health November 21, 1980
- Peter S. Houts, Ph.D., Principal Investigator with
- Robert W. Miller, Ph.D.
- George K. Tokuhats, Dr. 2.H., Ph.D.
- Kum Shik Han, Ph.D.
~
~ [..
s
- The Pennsylvania State University College of Medicine and N
- The Pennsylvania Department of Health g 8507200558 850524 PDR FOIA LODGE 85-285 PDR
. g-Contents Chapter 6: Characteristics of persons distJessed during and af ter the incident at Three Mile Island Chapter 7: Summary of findings with relevance to public policy 9
9 l
s f
I l.
i
- -I_
Introduction Findings to be presented here are Part Two of a report to the TMI Advisory .
i Panel on Health Research Studies. The acknowledgements of assistance for Part One also apply to Part Two with two additions.
We would like to acknowledge the help of Ms. Jane Todd Cooper in organizing material used in sections ofs thi report.
We also appreciate the secretarial assistance of Mrs. Robin Long.
Part One contained a description of methodology utilized, sample selection ,
description of reactions during the crisis period and persistence of distress over a nine month period from the original crisis.
Part Two contains two chapters: Chapters (, and 7.
Chapter 6 includes two topics: 1) characteristics of persons who were distressed during the initial crisis in March and April ,
1979, and 2) characteristics of persons who increased or decreased distress over the nine month period from April 1979 to January 1980.
Chapter 7 is a summary of findings from both parts of this study which have relevance for public policy .
We feel, in view of the videspread public interest in events during and after the incident and the possibility that these experiences will influence future public policy, that there should be a su= mary of findings with qualifications and conditions necessary for drawing implications for public policy.
- In the introduction to Part One of this report it was stated that Part Two would include a discussion of the impact on the Three Mile Island Incident on the Health Delivery System.
This material will be presented by Dr. Teh
. Hu in his report on the economic impact of the Three Mile Island Incident .
O
Chnpter 6 CHARACTERISTICS OF PERSONS DISTRESSED DURING AND AFTER THE THREE MILE ISLAND INCIDENT
Background
Stress resulting from the Three Mile Island crisis was largely psychological .
The crisis did not cause physical damage and the disruption due to evacuation was limited because people returned to unchanged homes and normal routines.
Furthermore, no immediate health effects from radiation were observed. Although the immediate crisis situation was mitigated by mid-April, the future of the facility and how it was to be " cleaned up" remained unresolved. Publicity about problems at the plant and possible future dangers continued for many months.
Public reference to the incident was continuous from March 28, when the crisis at the nuclear power plant was first publicized, until the end of }by.
There were daily references to the situation at Three Mile Island in the
- Harrisburg Patriot and the Harrisburg Evening News and frequent references in the daily papers of York and Lancaster, two other major population centers within 15 miles of the p,ower plant.
From June until the January survey, references were less frequent but at least weekly with the exception of November. When the Kemeny report was issued in November, daily references to TMI occurred again for approximately I week.
Local weekly newspapers also printed many articles on TMI in April and May with another group of articles in November.
In addition the Readers Guide to Periodical Literature lists over 100 articles about TMI printed in magazines with national circulations between April and December 1979. -
C As a result of this continuous publicity, situation at Three Mile Island changed from a short term crisis to a chronic, unresolved problem for many people living near the facility. While the percent of the population who were 5
very distressed dropped sharply by July, there remained a substantial number
of prop 12 who were otill vary cencerntd about Three Mile Island nine months after the original crisis (see Chapter 5 of this report for substantiating data.)
The purpose of this chapter is to examine what characterized persons who a) were distressed during the crisis period and b) who changed their levels of distress from April, 1979 to January, 1980.
Population Two surveys studied individual differences in response to the crisis among the population living within 0-5 miles of TMI. As explained in Chapter '
Three, 692 persons who lived within 5 miles of TMI were interviewed in July about their experiences and feelings during the 2 weeks in April immediately following the crisis at TMI and for the 2 weeks in July immediately preceding the interview. At the end of the interview, each respondent was asked if he/she would be willing to be re-interviewed at a later time. Five hundred eighty two people, or 81%, agreed, and their names and numbers were recorded.
No identifying information was taken from those who refused. In January, 1980,
.g.
467 of these 582 persons who agreed to be re-interviewed were contacted and 403 were interviewed a second time (58.2% of the original sample of 692).
Nine of those contacted refused re-interviews, 46 had disconnected phones, while nine did not answer on four calls. ~
l.
l A comparison of demographic characteristics of the January sub-sample with l the original sample of 692 is shown in Table 1. The demographic characteristics l
of the July sample and the January sub-sample are remarkably similar. '
The l
! January sample includes slightly more married homeowners who evacuated during the crisis. Other differences are quite small, with the percent of males exactly 52.1% in both samples. .
l l
l l
- 3- .
COMPARISON OF DEMOGRAPHIC CHARACTERISTICS D
OF JANUARY SAMPLE AND TOTAL SAMPLE TOTAL JANUARY (Sample N = 692) (Sample N = 403)
Respondent's mean age 42.4 41.6
% male 52.1 52.1
% widowed, separated or divorced 15.5 15.0
% single 10.9 7.0
% married .
73.6 78.0 Mean family size 3.1 3.2 Mean years of schooling 11.9 12.0
% with income over $10,000 80.0 82.1
% owning homes 77.8 83.1
% of respondents who evacuated du' ring crisis '
50.1 55.1
% Pennsylvania Dutch
- 42.0 41.4
- By Pennsylvania Dutch is meant descendents of German and Swiss immigrants who settled in the area in the 18th and early 19th centuries.
-S- -
Mt5murse of Strema Three measures indicating personal distress because of TMI were collected in both July and January. g
- 1. In the July survey, respondents were asked, "Please tell me whether (you) were upset during the Three Mile Island crisis? Were you extremely, quite, somewhat, or a little upset?" In January, the ;,
i question was rephrased,"How upset do you feel now about the sit-uation at Three Mile Islandi' e i
- 2. In the July survey, respondents were asked, "On Friday, March 30, how worried were you about the safety of other people in your family! Were you extremely worried, very worried, somewhat worried, a little worried, not at all worried?" In January the question was rephrased to ask, "How serious a threat do you feel the Three Mile Island Nuclear Station is now for you and your family's safety? Was it a very serious threat, a serious threat, somewhat of a threat or no threat at all?"
- 3. In the July survey, respondents were asked if they had experienced any of 15 stress-related symptoms during the 2 weeks immediately foli che Three Mile Island incident. If they responded "Yes" to any symptom, they were asked if they thought this was a result of the Three Mile Island incident. The number of symptoms attributed to Three Mile Island can be interpreted as an indication of distress felt about Three Mile Island. If a respondent felt he/she was experiencing physical symptoms because of Three Mile Island, this was
~
taken as an indication that he/che was upset about Three Mile Island. -
. In the January survey, respondents were asked if they had experienced the same 15 stress-related symptoms during the past 2 weeks. They were also asked if they thought their symptoms could be due to TMI.
-e-Nota that these ara reported symptoms, not symptoms actu:lly experienced. As explained in Part One of this report, many factors, in addition to actually i having a symptom, can affect symptom reporting. For our purposes, though, attributing synptoms to Three Mile Island can be taken as an indication of being upset about TMI without assuming that the symptoms actually occurred.
The distress levels during the crisis period and in January 1980 for each of the three distress measures are reported in Tables 2, 3 and 4. These tables include cross tabulations showing how many distress scores increased ,
decreased or remained the same for each measure.They show that, for all measures, the tendency was for distress to decrease over time though the change for attributed symptoms was not statistically significant .
I e
0 a
m 4
,- - . _ - . _ _ .- -- , - _ , - .- - . - .. y
TABLE 2
.CROSSTABUIATION OF RESPONSES TO QUESTIONS ABOUT 110W UPSET THE RESPONDENT WAS DURING Tile TliREE MILE ISTAND CRISIS AND IN JANUARY,1980.
UPSET JANUARY, 1980 4 3 2 1 0 extremely quite somewhat little not upset don't know Totals 4 extremely 16 25 20 . 18 14 1 94 3 quite 2 18 19 14 6 0 59 UPSET 2 somewhat 8 6 12 14 22 DURING 1 63 i TML g4 1 little 1 7 16 19 40 CRISIS 0 83
- 0 not upset 2 3 7 17 73 1 103 don't know 0 0 0 0 1 0 1 Totals 29 59 74 82 156 3 403 Item scores (extremely - 4, quite = 3, somewhat = 2 little = 1, not = 0)
TMI crisis mean = 1.90 January 1980 mean = 1.31 t test of difference between means - 6.46, 399 d. f. , p < .01 Correlation between the two upset scores = .48, p< 01,
TABLE 3 CROSSTABULATION OF RESPONSES TO QUESTIONS ABOUT RESPONDENT'S CONCERN FOR SAFETY DURING THE TMI CRISIS AND IN JANUARY,1980 (January, 1980)
Il0W SERIOUS A THREAT DO YOU FEEL TMI IS FOR YOU AND YOUR FAMILY'S SAFETY very somewhat no threat don't serious serious of a threat at all know Totals extremely 42 53 35 23 2 155 very 10 19 b IIOW WORRIED 27 13 1 70 8 WERE YOU ABOUT somewhat 3 12 34 TIIE SAFETY OF 23 0 72 OTilER PEOPLE little 3 6 23 20 IN YOUR FAMILY 4 56 ON FRIDAY, not at all 5 2 10 MARCil 30, 1979 29 0 46 don't know 0 0 2 1 1 4 Totals 63 92 131 109 8 403 Correlation between concern about safety during TMI crisis and in January 1980 = .40, p< .01 (Heans are not comparable for the two time periods since a 5 point scale was used in the July survey and a 4 point scale was used in the January survey.)
4 TABLE 4 CROSSTABULATIONS OF NtHBER OF SYMPTOMS ATTRIBUTED TO Tile SITUATION AT THREE MILE ISLAND DURING Tile CRISIS IN MARCH-APRIL AND IN JANUARY, 1980 NUMBER OF SYMPTOMS ATTRIBUTED TO TMI JANUARY, 1980 0
or more 5 4 3 2 1 0 Totals 6 or more 5 1 0 0 0 0 5 11 NUMBER OF SYMPTOMS 0 0 0 4 1 1 0 10 12 ATTRIBUTED TO TMI 3 1 1 0 1 1 1 8 13 DURING CRISIS 2 1 0 2 0 *1 0 10 14 1 0 2 1 3 0 1 21 28 O 6 4 1 3 4 1 301 '320 Totals 14 9 5 8 7 3 357 403 Mean number of attributed symptoms during TMI crisis = .598 Mean number of attributed symptoms in January, 1980 = .566 t test of difference betteeen means = .34, 402 d.f., not statistically significant.
Cortclation between number of symptoms attributed during THI crisis and in January, 1980 = .37, p <.01.
.'t e e m .
Demographic characteristics 1 Sex, income, education and marital status have all been shown to relate to reported emotional distress (Kessler, 1979).
The most common pattern has been that females, low income, low educational and widowed, separated or divorced persons are especially likely to report physical or psychological distress.
It should be pointed out, however, that there is disagreement among researchers as to the universality of these relationships or what causes them.
For example, there is considerable debate about why women usually report more physical and psychological symptoms than men.
Some argue that women's role is society is more stressful than is men's.
Some argue that women are more willing to acknowledge sickness publically and some argue that there is a biological reason (Mechanic, 1978). We will not b-e to resolve these issues in the findings to be reported here, but it is arcant to recognize that there are many interpretations possible for findings in this area.
The following demographic characteristics were available for respondents in this study:
age, sex, education, income and marital status.
Coping A number of researchers have proposed that coping plays a role in how people react under stress. .
Cohen and Lazarus, (1979) suggest that how one views stress influences how one reacts to it. If the situation is viewed as a challenge from which gain, growth or mastery may occur, coping will be directed at gaining control and feelings about the situation will be positive. However, if the situation is viewed as one where mastery is not possible , coping efforts are likely to be defensive and feelings about the situation are likely to be negative.
This suggests that the degree and severity of distress experienced as a result of the Three Mile Island Incident may be affected by whether people feel thay can affect the situation or whether it is a problem over which they have no control.
O
Questions about twenty coping strategies were included in the July and January surveys. In the July survey respondents were read a list of C
15 coping strategies and, sf ter each one, were asked if they had made use of it. These strategies are listed in Chapter 4, Table 4.6. The strategies ,
were divided into two groups: " avoidance" strategies (Type I in Table 4.6) ,-
and " social" strategies (Type II.in Table 4.6) on the basis of a factor analysis. E Each respondent was given an " avoidance" and a " social" score consisting of the.
number of strategies he/she used in each group.
The avoidance and social strategies were primarily concerned with reducing anxiety. One other question also concerned how people dealt with anxiety: "During the two vecks following the Three Mile Island incident, did you take any tranquilizers or nerve medicine to help you feel less tense?"
In addition, four other questions were included in the surveys which asked about strategies to deal with perceived danger from TMI.
- 1. "Has the use of cow's milk for your family changed since the TMI incident?"
- 2. "Did you do any of the following things to protect the health of household members? Kept family inside, sent them out of the area,
. had tests for radiation, changed their diets."
I
- 3. "Did you leave the area during the 2 weeks following the Three Mile i
l Island incident?"
! 4. "Have you, personally, been active in any organization or gone to any meetings to influence what happens at TMI?"
Social Support Cobb (1976) has reviewed a number of studies indicating that social support may serve as a buffer to moderate the effects of a stressful situation on an
- individual. By social support is usually meant having persons available to give A
encourager.ent and support. This usually occurs through friendship and family relations.
t
- ~.. _ - . . . _ _ . . - - ~ ~ _ - . _ _ - _ . ~'
Anothsr kind of cocial support, which va will call " community roots" can come from sharing values with others in the community and having commitments to community institutions.
Community roots implies a sense of involvement with the community as a whole rather than with specific individuals. Several questions were included in these surveys which are indicators of community roots including being " Pennsylvania Dutch" (the predominant cultural group in the community) coming from a family which has lived in the area for several generations, attending church and owning a home. Following are the questions included under these definitions of social support and community roots.
Social Support 1.
"How many friends or relatives do you have that you feel close to, that is, people that you feel at ease with, can talk to about private matters, and can call upon for help?"
- 2.
The following four questions concerning availability of help were summed into one score.
a.
If you were so upset in the middle of the night that you needed someone to talk to immediately, is there someone immediately available.within your household whom you could '
talk to?
b.
If you were so upset in the middle of the night that you needed someone to talk to immediately is there someone immediately available outside of your household whom you could call on?
- c.
Do you have any difficulty finding people with whom you can discuss upsetting personal problems?
d.
Do you have any problems you can't discuss with any friend or relative? (A "no" answer scored as positive. )
e Community Roots 1.
"As far back as you can remember, how long has your family lived ~
in this area?"
- 2. "Do you own or rent your home?"
- 3. "How often do members of your family attend a church at the -
present time?"
- 4. "Many people think of themselves as coming from a certain ethnic background. Is your family background primarily Pennsylvania Dutch?" (Pennsylvania Dutch is the predominant ethnic group in the area.)
Sensitivity to Stress A number of studies (for example, Canter, Imboden and Cluff,1966) have shown that persons with a history of emotional problems are especially sensitive to and react strongly to stressful situations. Furthermore, studies on illness behavior (Mechanic,1978) have shown that symptom reporting is frequently associated with mental distress and with a tendency to attend to one's bodily state. In addition, there are unique features of the Three Mile Island incident which might make certain persons especially sensitive -
durin'g and following the crisis: 1) seeing the cooling towers from one's home or place of work and, 2) having young children or children in utero at the time of the crisis.
Information was obtained on the following individual characteristics which are possible indicators of sensitivity to stress during and af ter the TMI crisis:
1.
Whether a respondent had a history of any of 13 chronic physical illnesses including arthritis, heart disease, diabetes, skin disease, etc.
2.
Whether a respondent had a history of any of 4 mental problems:
insomnia, severe depression, severe anxiety or nervous breakd on w requiring hospitalization.
3.
Whether anyone in the household was pregnant at the time of the Three Mile Island incident.
4.
Whether there was a person below six years of age in the o househ ld
-5.
The frequency with which the respondent consumed alcoholic beverages.
(The assumption here is that high consumption of alcoholic beverages may indicate sensitivity to stress which leads to the consumption of alcohol (Chotlos and Deitler,1959).
6.
Whether the respondent had experienced a life crisis other than ree Th Mile Island within the past year.
The crises were taken from Holmes and Rahe's life events scale and included death of a family member or close friend, serious illness or injury, serious trouble with the law, loss of a job, divorce or separation, serious financia l troubles or any other serious problem that was upsetting to the respondent (Holmes and Rahe, 1967).
7 .
Whether the respondent can, on a typical day, as he/she goes about his/her daily business at home or at work, see the Three Mile Island plant.
8.
Whether the respondent describes him/herself as having any of the following personal traits: 3
(
- a. is sensitive and introspective. ~
(This measure has been reported t
to be associated with both psychological and somatic symptom reporting (Mechanic,1979 and 1980)).
I L
l -
- b. has a tendency to complain about symptoms.
- c. rates his/her general level of health as poor rather than good.
C
- d. says he/she is a person who gets upset often. y I
Method and Results Multiple regression analysis was utilized to analyze the relationships L.
between 1) distress (upset, concern about safety and attributing symptoms to TMI) and 2) demographic variables, coping, sensitivity to stress and social $
support. Analyses were carried out in four steps. First the relationships between demographic variables (age, sex, education, income and marital status) and the three distress measures were studied. Second, relationships between the seven coping strategies and the three distress measures were assessed holding constant (i.e. controlling for) demographic variables. Third, relationships between the six social support and social roots measures and the three distress measures were assessed holding demographic variables constant.
Fourth, relationships between the thirteen sensitivity measures and the three distress measures were assessed again holding constant demographic variables.
In each of the above analyses, relationships to distress in July were studied as well as relationships to change in distress from July to January. .
Results Relationships between the demographic variables and three dependent variables in April, 1979 are reported in Table 5. Results show that younger persons, women and more highly educated persons were most upset during the crisis in March and April of 1979. Ibrrfed persons were more distressed ,
and single persons were less distressed than widows, separated or divorced persons. Thus the rank ordering of distress for the three marital status groups was: married persons most distressed; widowed, separated or divorced intermediate; and single persons least distressed. It is interesting to note that women were more likely than nen to report symptoms which they
r
-uw
,e
- attributed to TMI. H; wever, since females generally report higher symptom rates than males in surveys of this type, this finding may not be indicative of feelings about Three Mile Island.
Several of the results in Table 5 are different from those commonly reported, namely that younger, more educated, married persons experienced the most stress. Most studies of symptom reporting have found the opposite. This is likely due to some unique features of the Three Mile Island crisis:
the technical nature of the problems which would have had more meaning for more educated persons and the fact that persons in the child bearing stage are most likely to experience the effects of radiation to either themselves or their offspring.
Table 6 shows that change in distress from April to .Tanuary is related to only one demographic variable: age.
Since this is one out of 18 relationships studied, it is likely that this finding is due to chance.
Our conclusion is that, .while a number of demographic variables are related to distress during the crisis, they are not related to change in distress over time.
6 0
0
,- i.
TABLE 5 RELATIONSHIPS BETWEEN DEMOGRAPHIC VARIABLES AND DISTRESS IN APRIL, 1979 *
(Only statistically significant relationships shown.
Direction of relationships shown by sign.)
I upset worry attribute g about symptoms safety to TMI age - -
sex * - - -
education + +
married ** + +
single ** .
income y.
- Male is higher score.
- Compared to widowed, separated or divorced. Married and single are higher Scores.
A plus sign indicates that a higher score on the demographic variable is as,sociated with being more upset during the crisis in April, 1979. -
The above findings can be summarized as follows:
- 1) Younger people tended to be more upset and more worried about safety than older people.
- 2) Women tended to be more upset, more worried ebeut safety and core likely to attribute synpcoms to TMI than men.
- 3) Higher educated people tended to be more upset and more worried about safety than lower educated peopler ,
- 4) Married people tended to be more upset and more worried about safety than widowed, separated, or divorced persons.
- 5) Single persons tended to be less worried about safety than widowed, separated, or divorced persons.
s
_p TABLE 6-RELATIONSHIPS BE'1VEEN DEMOGRAPHIC VARIABLES AND CHANGE IN DISTRESS FROM APRIL 1979 TO JANUARY 1980 (Only statistically significant relationships shown. Direction of relationship shown by sign.)
change in change in change in upset worry attribute about symptoms safety to TMI age
+
sex
- education married **
single **
income
- Male is higher scored. -
- Compared to widowed, separated or divorced A plus sign indicates that a high score on a demographic variable is associated with either maintaining high levels of distress and/or with increasing distress over the 9 month period from April 1979 to January 1980 Blanks indicate that there was no statistically significant relationship.
Relationships between coping uttategies and distress during the crisis (with demographic variables held constant) are reported in Table 7. There is a strikingly consistent pattern to these findings: persons who were most g I
distressed during the crisis used more coping strategies. Furthermore, Table 8 shows that, for'll of the 21 rela'tionships studied, persons who stayed upset through the nine month period from April 1979 to 0
January 1980 were also more likely to be active copers.
C This finding that most people who used coping strategies remained upset could be due to the unresolved nature of the Three Mile Island Crisis. As explained above, problems at the facility were continually in the news during the nine month period from April to January and groups for or against re-opening the facility attracted a good deal of public attention. Under these circumstances coping did not change the situation and, as a result, people who tried to cope may have felt particularly frustrated. This would be especially true for persons who use political actions as a coping device.
9 l
l
~
I l
l I
TABLE 7 RELATIONSHIPS BETWEEN COPING STRATEGIES AND DISTRESS IN APRIL (Only statistically significant relationships shown.
Direction of relationships shown by sign.*) i upset worry attribute about symptoms safety to TMI avoidance coping strategies + + +
social coping strategies + + +
change use of milk + + +
kept family inside + + +
evacuated during crisis + + +
-r:
participated in political activity re: TMI + + +
took tranquilizers during crisis; + + +
1
- Plus sign indicates that persons who used that coping strategy were more distressed during the crisis in April 1979.
Age, sex, education, i
these analyses. income and marital status have been controlled for in e
i i
i i
r
_ . . , -. . , . . . . _ _ _ _ . . _ = _ _ - , _ . - , _ . _ . , . . .
C l
TABLE 8 RELATIONSHIPS BETWEEN COPING STRACEGIES AND CHANGE ,
f IN DISTRESS FROM APRIL TO JANUARY c
(Only statistically significant relationships shown.
Direction of relationships shown by sign.*)
change in change in change in upset worry attribute about symptoms safety to TMI i'
avoidance coping strategies + +
social coping strategies +
change in use of milk -" + +
kept family inside 1
-evacuated during crisis + +
participated in political activity re: TMI + + -
tooktrankuilizersduringcrisis + +
- A plus sign indicates that persons who used that coping strategy tended to remain distressed and/or to increase distress over the 9 month period from April 1979 to January 1980. A blank indicates that there was no statistically significant relationship.
Age, ser., education, income and marital status hive been controlled for in these analyses.
Relationships between social support and distress measures as well as between community roots and distress measures, with demographic variables held constant are reported in tables 9 and 10. These results are less consistent than those found for coping behaviors. Availability of help is associated with fewer attributed symptoms during the crisis and having more friends is associated with decreasing upset and worry about safety from the July t to January surveys. These findings are consistent with the hypothesis that social support helps to buffer the effects of stressful experiences.
.However, only three out of twelve relationships were statistically significant.
Therefore, .the overall pattern of findings provides only weak support for the social support hypothesis. Findings for " community roots" variables are contradictory. Home ownership is associated with higher upset at the time of the cris3s as well as remaining upset through January. Church attendance, on the other hand, is associated with lower upset scores during the crisis.
Overall, the findings in . tables 9 and 10 do not give strong support for either social support or community roots having a major effect in moderating stress during and after the TMI crisis.
e y
C
TABLE 9
~
RELATIONSHIPS BETWEEN SOCIAL SUPPORT / COMMUNITY ROOTS AND DISTRESS IN APRIL g (Only statistically significant relationships shown. I Direction of relationships shown by sign.*) l upset worry attribute .
about symptoms O safety to TMI Social Support number of' friends availability of help -
Community Roots i . length of residence own home or rent (own home high score) +
r; frequency of church attendance -
Pennsylvania Dutch ,
J Relationships control for age, sex, education, income and marital status.
, *Minus sign indicates that more social support is associated with less distress!
The above findings can be summarized as follows:
l 1) Those with help available attributed symptoms to TMI less often than those whithout help available.
e
- 2) Home-owners tended to be more upset than renters.
i'
- 3) Church-goers tended to be less upset than non-church-goers.
[
e
.e- .
TABLE 10 RELATIONSHIPS BETWEEN SOCIAL SUPPORT / COMMUNITY RCOTS AND CHANGE IN DISTRESS FROM APRIL TO JANUARY (Only statistically significant relationships shown.
Direction of relationships shown by sign.*)
change in change in change in upset worry attribute about symptoms safety to TMI Social Support I I
i number of friends -
availability of help l 6
I
-_ Community Roots i
t l length of residence .
own home or rent (own home high score) +
frequency of church attendance Pennsylvania Dutch Relationships control for age, sex, education, income and marital status.
- A stress minus from sign April indicates that persons with social support tended to decrease to January. er th i The above findings can be summarized as follows:
1)
People with more friends tended to decrease how upset they were and how much they worried about their safety more than those with fewer friends.
2)
Home-owners tended to stay upset more than did renters.
I
~-
25 The relationships between sensitivity variables and distress in April 1979 are reported in Table 11. A consistent pattern is apparent here with distress in July related to four sensitivity measures: 1) chronic illness (both physical and mental), 2) being frequently upset, 3) being in poor health, '4) being sensitive and introspective. One interpretation of these findings is that poor physical and mental health may predispose an individual to emotional distress during the crisis. However, an alternate explanation is suggested by the parallel finding that being sensitive cnd introspective relates to distress.
Mechanic (1979 and 1980) has reported that persons who score high
- on the introspection scale also tend to report high levels of psychological and somatic symptoms.- He has suggested that this may be because sensitive, intro-spective people focus attention on bodily sensations which, in turn, helps them to remember their symptoms.
These sensitive, introspective personc may report more health problems before the crisis because of their tendency to focus attention on bodily sensations. They may report distress af ter the crisis for the same reasons. If this were so, the relationship between poor hedith before the crisis and distress af ter the crisis could be due to being. sensitive and introspective rather than because health problems make one sensitive to stress.'
Unfortunately, correlational data, of the sort reported here, do not allow us to select between these two explanations. We can only say that the correlations are consistent with both explanations.
This issue will be discussed further in
- Chcp te r 7.
l
' Table 12 shows that three sensitivity va,riables relate to change in dificress from April 1979 to January 1980: chronic physical problems, poor general heatlh and introspection. While there are not as many significant findings as in Table 11, all relationsips are in the predicted direction: greater sensitivity is -
associated with continuing distress.
l l
L - - _ - - - - . - - - -
-it.
There are two surprising findings in tables 11 and 12; 1), lack of relationships between distress and having family members below 6 years and 2) the lack of relationships between distress and having person a pregnant in the household.
These findings are somewhat surprising since these families were advised to evacuate during the crisis and considerable attenti on was given by' the media to possible radiation effects on children and To .
fetuses understand these findings it should first be noted that respondents with children under 6 or a pregnant person in the householdungwere and mostly y married.
We know from table 5 that young married people wereseamong most tho upset by the crisis.
Results reported in tables 11 and 12 (where age and sex are controlled) indicate that having a child under 6 or pregnant son in per the household did not add additional distress beyond bei ng young and married.
It is likely that young, married persons without children e view d th emselves as potential childbearers and therefore were as upset as were young married persons with children.
i e
O
s.
TABLE 11 RELATIONSHIPS BETWEEN SENSITIVITY TO STRESS AND DISTRESS IN APR (Only statistically significant relationships shown.
Direction of relationships shown by sign.*)
upset worry attribute about symptoms safety to TMI Chronic physical problems + +
Chronic mental problems + + +
Frequency of alcohol consumption prior to TMI Gets upset often
+ + +
Life stress problems in past year Poor general health
+ + +
Express feelings openly +
Tendency to complain Sensitive, introspective person + + +
See towers from home ;
+ "
See towers from work
+
Family members below 6 years old Persons in household pregnant during crisis 4
- A distressed.
plus sign indicates that persons who had high vulnerability scores tended to be Relationships have controlled for age, sex, education, income and marital status.
r TABLE 12 RELATIONSHIPS BENEEN SENSITIVITY TO STP.ESS AND CHANGE .
(Only statistically significant relationships shown.-
Direction of relationships shown by sign.*)
change in change in change in upset worry symptoms about attributed safety to TMI
' Chronic physical problems
+
Chronic mental problems-Frequency of alcohol consumption prior to TMI Gets upset of ten
. Life stress problems in past year r
-Poor general health
+ + +
Express feelings openly Sensitive, introspective person +
See TMI towers from home See TMI towers from work .
Family members below 6 years old Person in household pregnant during crisis
- Adistressedplus sign from indicates April that persons who had high sensitivity scores tended to remain to January.
Relationships have controlled for age, sex, education, income and marital status.
t I
l l
l .
l
. _ . . . , , . m.,_.-._ . , _ . . _ . _ . , _ _ _ ,__._,-.,_ _...._, ., . . , , . . . . , _ , - . . - _ . . - _ _ _ _ _ _ _ _ - _ . _ - _ _ ,
s.
1 Summary There is evidence that persons who are younger, more educated, p
married and female were especially distressed during the crisis. However, these demographic variables did not relate to change in distress over ti=c. .
j /
People who actively coped had high distress during the crisis and tended l ,/
{ to maintain that distress over time, possibly because the situation at T:!I i
remained unresolved and their attempts at coping resulted in frustration.
There were only a few indicators that social support may have helped to moderate distress during and after the crisis. Finally, persons with poor i
- mental or physical health or a tendency toward self-attention and evaluation tended to have high distress scores and to maintain their distress over the 2
nine month period from April 1979 to January,1980. One possible explanation i
for this finding is that people with poor physical or mental health were more vulnerable to the stress of Three Mile Island. However, there are
} alternate explanations for, . .
these findings and it is not possible, with the 5 data available in this study, to deternine which explanation is correct.
1 t
o
-i l
I t
I s
~
(
l
- 1 i
i I
i 4 *
-, _1__._, -. ,.. _,. _ , , _ . . . , , . _ , _ _ . _ _ . . _ _ _ . , . _ , , , _ _ _ _ . . ,,,m__. _,_,,,_____,_,.r__.._.,.-.-__-,.- , - . . - . ~ . -
Chapter 7
SUMMARY
OF FINDINGS RELEVANT TO PUBLIC POLICY The purpose of this chapter is to summarize the main findings reported in earlier chapters with special attention to those which have implications for public policy. Also included are qualifications which should be considered in any practical application of these findings. We recognize that public policy decision-making involves a broad range of issues, one of which may be social science research findings.
It is not our intention, therefore, to make recommendations for specific actions by public officials but, rather ,
to report clearly and carefully findings which may be of use to them in the complex process of developing public policy.
Before reviewing specific findings, one general qualification should be made which applies to most findings reported here.
Subjective, introspective
?*
reports, which constitute the bulk of data analyzed here, are vulnerable to selective memory and conscious as well as unconscious distortions. Usually these distortions are in the direction of making respondents appear more positive to themselves and others or of making past experiences consistent with present feelings. .
Certain descriptive facts, i.e. age, sex, education, etc. are most likely correct, but recollections where feelings are strong and where memory is long are especially vulnerable to distortion. For this reason it is best to look for patterns which are repeated across measures and across studies.
This has been the approach used in presenting findings of this study and will be the basis for the summary statements made in this chapter.
FINDINGS CONCERNINC EVACUATION Data from questionnaires and in-depth interviews indleste that, in most decisions to evacuate, the prinary reasons were because the situation seemed
d:ng:rcus cnd/or that information about the situation was confusing. Primary reasons for not evacuating were that respondents either saw no danger or because they perceived some danger but felt constrained to stay because of such things as being unable to leave the job, too sick to travel, etc.,
Interview data indicated that the decision whether or not to leave wa on an evaluation of the information available from the media and that mass panic (in the sense of people responding to other people's emotional states rather than to information about the situation) did not occur.
Generalizations relevant for public policy included the following:
1)
Information conveyed by the media was tLt main basis for decision to evacuate.
Two aspects of this information were especially important in the decision to evacuate:
a) information about possible danger b) inconsistencies and/or contradictions in information.
2)
There was considerable variability among residents in how information was interpreted resulting in:
a) evacuation and returning being spread over a considerable period of time which contributed to the efficiency with
. which the large number cf people evacuated the area .
b) a substantial percentage of the population eleccing not to evacuate at all It should be pointed out that the above conclusions are based on retrospective questionnaire and' interview data. Therefore, bias may
- I i
s
be introduced through respondents trying to aopear more rational than they were during the crisis or due to' distorting memories so as to be consistent with feelings at the time of the interview.
FINDINGS CONCERNING DISTRESS CLOSE TO TMI COMPARED TO FARTHER Ak Four measures of distress were studied: 1) how upset respondents felt about Three Mile Island, 2) how worried they were about the safety of themselves and their families, 3) whether they reported any of 15 stress-4 related symptoma in a two week period, and 4) responses to the Langner scale of psychological distress.
Results from the questions about being upset and worried about safety show that
- 1) from April to January 1980 there was a 50% drop in number of people quite distressed about TM1 and 2) in both April 1979 and January 1980 persons close to Three Mile Island were more distressed than were persons farther away.
Findings for the 15 stress-related symptoms showed a similar pattern to the questions about being upset and worried. Results showed that symptom reporting was significantly higher close to Three Mile Island farther away for the two weeks immediately following the accident in }hrch 1979. the first two weeks in July 1979 and two weeks in January,1980. For all three time periods, reporting rates were raised between 10 and 20% close to TMI compared to 40 to 55 miles away.
Results with the Langner scale of psychological distress, which was administered only in January, showed no signi,ficant differences between those close and far from Three Mile Island.
The above findings should be carefully qualified in considering their relevance for public policy. First, all of this information was obtained in the context of an interview about Three Mile Island which may have
affected respondents' expectations of what they should say. Second, symptom reporting, as explained in detail in Chapter 5, are susceptible to many factors in addition to actually having symptoms. Therefore, we cannot be certain that people close to Three Mile Island experienced more stress-related symptoms than did people farther away. It could be for example, s'
that people close to TMI paid special attention to their symptoms ( wondering if they were due to THI) which could help them to remember those symptoms.
This interpretation is consistent with the finding that a higher percent of persons close to TMI attributed their symptoms to TMI than did persons farther away. Unfortunately, the self-report, retrospective data available do not allow us to choose between these alternate explanations.
The fact that, in January, the Langner scale scores did not change with distance has implications for understanding the type and severity of distress experienced. The Langner scale was originally designed to differentiate persons in mental health treatment from the general population.
This suggests that the distress characteristic of mental patients was not higher close to THI in January. However, there is an important qualification to this statement. The Langner scale questions deal primarily with depressive symptoms and psychosomatic symptoms. Other areas of mental health are .
either not represented or only peripherally represented.
A detailed analysis of items in the Langner and in the stress-related symptom list was helpful in understanding the type and severity of distress close to n!I in January, 1980. A comparison of items in the two scales shows that the Langner scale asks about more-frequent symptoms than ,
does the stress related symptom list. For example, the Langner scale asks whether respondents had headaches often and the stress-related symptom list asks only if respondents had a headache in the past two weeks.
d .
Another difference between the two scales, which may have contributed to their different results, is that the Langner scale includes questi ons about psychological mood not included in the stress-related symptom list.
The above findings regarding stress close to TMI can be summarized as
% follows:
The number of persons with severe distress dropped shortly after the crisis, buc ~ between 10 and 20% of respondents close remainedto TMI sufficiently distressed nine months after the crisis to say that they were still quite upset about the situation, that they were still equit worried about the safety of themselves and their families and to report higher frequencies of stress-related symptoms.
On the other hand, it should be noted that the type and frequency _ of these symptomsrom were differen the type and frequency of symptoms assessed by the Langner a e. sc l There are two important qualifications which should be made e to the ab conclusion.
First, all these data are subjective and retrospective and therefore vulnerable to conscious and unconscious distortion. In view of the fact that some persons near Three Mile Island are suing for damages because of mental distress, the possibility of conscious distortion should be conside red.
Second, comparisons between persons close to TMI with persons r away farthe only shows differences between the two groups and does not include dis .
due 'to TM1 felt by groups faw away.
Therefore, these estimates of distress near Three Mile Island may be conservative.
The above findings are consistent with reports by two other investi gatory groups.
The Task Force on Behavioral Effects of the Presidentia n on
, Three Mile Island reported that scores on the " Demoralization Scal e," were
' raised immediately following the crisis period but had dropped to the p where they were not different from a control group over 100aymiles by aw July.
The demoralization scale and the Langner scale are highly correlated and even include some of the same questions.
Therefore, the findings with
, th2 L:ngnst cesle in J:nusry, 1980 coplicata cnd extund ths findings with the Demoralization scale in July,1979. The Presidential Commission task force also reported that distrust of authority remained elevated in July,1979. Dr.
Evelyn Bromet (Bromet, 1980) has also reported elevated scores on a 90 item symptom checklist for mothers of young children near TMI in December 1979 /
~
compared to mothers of young children near another nuclear power station several /
hundred miles away. The severity of this heightened distress is not discussed in sufficient detail in Dr. Bromet's preliminary report to allow comparisons to the two scales used in our studies, but her finding does indicate some persistence of distress into December 1979.
FINDINGS REGARDING LEVELS OF ALCOHOL, TOBACCO, SLEEPING PILLS AND TRAN0UILIZERS USE.
Responses to the July survey suggests that persons close to TMI used more alcohol, tobacco, sleeping pills and tranquilizers during the two week period immediately following the crisis than they did in July, three months later.
When many of these respondents were re-interviewed in January,1980, the use of these substances was essentially the same as that in July. Respondents who used tobacco and alcohol were asked if their use of these substances had increased, decreased or remained the same in the nine months since the Three '
Mile Island crisis. Approximately equal numbers reported increases as decreases with most respondents reporting no change. These findings suggest that the use
- of alcohol, tobacco, sleeping pills and tranquilizers increased only during l
the crisis period and did not persist beyond that time.
These conclusions also should be qualified since siceping pill and ,
tranquilizer users were not asked in January if their use had changed since TMI. Furthermore, all data are subjective and retrospective and therefore subject to conscious and unconscious distortion.
6 a' .-- - - -
n----. . . , , , - , - - , , - - _ - - , . --_,---,-- - -, .,-------.---,wm._--r-n -,,-,,.-,,- - - - , - - - - ~ . r- --- - - - - - ~ - - - - - - - - - - - . - - -
-7 FINDINGS CONCERNING HO DEMOGRAPHIC CHARACTERISTICS R INITIAL DISTRESS AND CHANCE IN DISTRESS OVER TIME.
Persons who were younger, married, higher educated and fe= ale reported more distress during the crisis period than did older, single, less educated
\ males.
The greater responsiveness of younger, married persons was possibly
\
due to their concerne about the effects of radiation on their present and future children and, since radiation effacts often have a long latency , concerns about their own future health.
More. educated persons were more upset possibly because of greater knowledge of the effects of radiation and the technical problems at the Three Mile Island facility.
The male-female difference is consistent with many other studies in which females have reported higher distress and more symptoms than men.
The reasons for this sex difference may include greater willingness to acknowledge feelings on the part of women ,a greater sensitivity to stressful situations on the part of women, or in the case of the Three Mile Island crisis, the fact that pregnant women and mothers of young children were told to evacuate.
The fact that being pregnant or having young children did not increase distress beyond the effects of being young and married was possibly due to the fact that most young ,
married persons considered themselves potential child-bearers. Again, the '
subjective, retrospective nature of the distress measures should be considered in generalizing from these findings.
TIUDINCS CONCERNING H0t/ COPINC REI.ATES TO INITIAL
_ DISTRESS AND CHANGE. IN DISTRESS OVER TIME All measures of coping were associated with at least one distress measure during TMI and all but one measure is associated with change in distress over time.
The consistency of these results are striking, indicating that people
who did things to actively cope with their feelings or the external situation were more upset and -remained upset over the 9 months following the crisis.
It is likely that, in this situation, coping was more a response to stress than a way to change stress. This may be due to the persistence of problems at Three Mile Island and feelings of frustration in trying to control one's
/,
own feelings or the objective situation.
,/
A qualification to these findings is that coping may have helped people remember distress. In other words, people who tried to control distress may be more likely to remember being distressed than people who gave up and didn't try to do anything about their distress. This explanation implies that selective memory may also account for those findings. This and other retrospective distortions could have been involved in the relationships that were found.
FINDINGS CONCERNING HOW SOCIAL SUPPORT RELATED TO INITIAL DISTRESS AND CHANGE IN DISTRESS There were a small number of significant findings in analyses of relationships between social support / social roots and distress. The number of statistically significant relationships was not large and direction of relationships was not always consistent. The overall pattern of these -
findings provides little indication that social support or community roots helped to buffer the stress experiences during and after the crisis.
FINDINGS CONCEIDTING HOU VLT.NERABILITY RELATES i jITIAL DISTRESS AND CHANCE 11 DISTRESS A consistent pattern emerged from these analyses indicating that persons who report a history of physical or mental health problems are more likely
o than others to report higher distress during the crisis period and to maintain that distress over the 9 months from April 1979 to January,1980.
This suggests that persons with serious nealth problems may have been especially
, responsive to the stress of the Three Mile Island crisis. However, there is an alternate explanation suggested by the parallel finding that " sensitive
'\ introspective" people tended to have and maintain high distress.
Since the
" sensitive and introspective" scale has been shown in other studies (Mechanic 1979 and 1980) to be related to both mental and somatic symptom reporting, it is likely it plays a role in the associations between health ratings and distress.
If so, it may be that people who attend to and focus on health problems (i.e. sensitive, introspective people) are more likely to remember symptoms before as well as after the crisis. If this were the case, the association between health ratings before TMI, and distress during the l crisis would be because people who paid attention to bodily states before the crisis did so during the crisis as well and not because sickly people are responsive to stress.
Competing explanations, similar to the ones above, are possible for j
most of the findings presented in this report. This is partly the result of having to rely on subjective, retrospective data and partly it is '
because of having incomplete data about a complex situation. tie feel that it is important that the public and the makers of public policy understand not only the findings of this study, but also the many interpretations that are possible.
Only with this full understanding can these findings be meaningfully related to public policy. ,
.k
_RECOetENDATIONS FOR RESEARCH CN FUTURE CRISES As explained above, we do not feel it is not appropriate for us, as researchers, to make policy recommendations concerning the situation at Three Mile Island. However, we do feel qualified to make recommendations in the area of research. Our experience in designing, and carrying out .
research on Three Mile Island leads us to make the following recommendations j
for carrying out similar studies in the future.
- 1) In this study we had to rely on retrospective reports of how people felt and behaved during the crisis. As explained earlier, retrospective data is vulnerable to many types of distortion which could have been avoided with data collected during the crisis period. We therefore recommend that governmental and private groups concerned with the effects of disasters make funds available which could be used immediately so as to begin data collection while crises are still in progress.
r
- 2) When this study wa's being planned, in April and May 1979, other investigators were planning similar studies. Several people commented, at that time, that there was " overlap" and " duplication" in these studies implying that this was undesirable. On the contrary, we have found, in the year and a half since this work was begun, that it has been of great value to have several independent investigators studying the effects of the Three Mile Isisnd crisis. Each study has had limitations, e.g. populations sampled, time of sampling, questions asked, etc. but, by noting consistencies and inconsistencies across studies is has been possible to identify patterns which support or do '
not support specific findings in any one study. It has also been possible to identify patterns across studies which have gone beyond what could be learned from any one study e.g. the relationships between
-g-
,s.
chsrt tcra stress demonstrated in the studies by the Presidential Commission and long term stress reported here and in Dr. Bromet's research. Our recomendation, therefore, is that, in studying future crises, that several independent investigators be involved.
\
l I
O e
1 1
I i .
a BIBLIOGRAPHY Bromet, E. Psychological, Behavioral and Social Aspects of the Three Mile Island Nuclear Incident, Pittsburgh, PA., May, 1980.
Canter, A., Imboden, J.B., Cluff, L.E.
The frequency of physical illness as a function of prior psychological vulnerability and contemporary stress.
Psychosomatic Medicine, 1966, 28: 344-350.
Choclos, J. and Deiter, J. Psychological considerations in the etiology of alcoholism. In D.J. Pittman (Ed.) Alcoholism An Interdisciplinary Approach.
Springfield, Ill. : Thomas, 1959. f In Cohen, F. Personality, stress and the development of physical illness.
Stone, C. , Cohen, F. and Adler (Eds.) Health Psychology San Francisco:
.Jossey-Bass Publishers, 1979.
In Stone, C.,
Cohen, F., Lazarus, R.S. Coping with the stresses of illness.
Cohen, F. and Adler (Eds.) Health Psychology San Francisco: Jossey-Bass Publishers, 1979.
Cobb, S. Social support as a moderator of life stress. Psychosomatic Medicine, 1976, 5: 300-314.
Dohrenwend, B. , Dohrenwend, B., Kasl, S. and Warheit, G. Final draft of the task force on behavioral effects of the President's Commission on the Accident at Three Mile Island, unpublished manuscript, October 15, 1979.
Flynn, C.B., Chalmers, J.A. The Social and Economic Effects of the Accident at Three Mile Island. Washington, D.C.: U.S. Nuclear Regulatory Commission, January, 1980.
Holmes, T.H., Rahe, R.H. The social readjustment scale. Journal of Psychosomatic Research, 1967, 11: 213-218.
Houts, P., et. al. Health-Related Behavioral Impact of the Three Mile Island Incident, Part I, Harrisburg, PA., April, 1980. .
Kemeny, J., et. al. The Report of the President's Commission on the Accident at Three Mile Island, Washington, D.C., October, 1979.
Kessler, R. Stress, social status, and psychological distress. Journal of Health and Social Behavior, 1979, 20: 259-272.
Langner, T. A twenty-two item screening score of psychiatric symptoms 269-276.
indicating impairment. Journal of Health and Human Behavior, 1967, 3:
Mechanic, D. Medical Sociology. New York: The _
Free Press, 1978.
Mechanic, D. Correlates of psychological distress among young adults: a theoretical hypothesis and results from a 16-year follow-up study. Archives of General Psychiatry, 1979, 36: 1233-39.
Mechanic, D. The experience and reporting of common physical complaints.
Journal of Health and Social Behavior, 1980, 21: 146-155.
m 4
"This project is funded, in part, under a contract with the Pennsylvania Department of Health. Basic data for use in this study were supplied by the Pennsylvania Department of Health, Harrisburg, Pennsylvania. Any analysis, interpretation, or conclusion based on these data is solely that of the authors, and the Pennsylvania Department of Health specifically disclaims responsibility for such analysis, interpretation, or conclusion."
" LEGAL NOTICE" '
"This report was prepared by the organization named below as an account of work sponsored by the Electrical Power Research
( .
Institute, Inc. (EPRI). Neither EPRI, member of EPRI, the organization named below, nor any person acting on their behalf (a) makes any warranty or representation, express or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or (b) assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method, or process disclosed in this report."
This report was prepared by the Department of Behavioral Science of The Pennsylvania State University College of Medicine, The M.S. Hershey Medical Center, Hershey, PA. 17033.
.