Content-Type: text/html

               "Personal-societal" risk comparison
 
 
            Mass media, interpersonal communication and real-life factors
            in risk perception at personal and societal levels
 
 
 
 
 
            Eunkyung Park, Clifford W. Scherer and Carroll J. Glynn
 
 
 
 
 
 
 
            Department of Communication
            Cornell University
            Ithaca, NY 14853
 
            (607) 255-0354
 
 
 
 
 
 
 
            Paper accepted for a presentation at the annual conference of
Association for Education in Journalism and Mass Communication,
            Anaheim, CA, August 10-13, 1996.
 
 
 
 
            All correspondence should be addressed to the first author.
            e-mail: <a href="/cgi-bin/wa?LOGON=A3%3Dind9612D%26L%3DAEJMC%26E%3D0%26P%3D1349829%26B%3D--%26T%3Dtext%252Fplain" >[log in to unmask]</a>
 
 
 
 
 
              Mass media, interpersonal communication and real-life factors
              in risk perception at personal and societal levels
 
 
 
 
              Abstract
 
 
                The purpose of this study investigates factors influencing
personal and societal risk judgments about four health issuesDwater
contaminants, radon, AIDS, and heart disease.  This study replicated and
expanded the impersonal impact studies by incorporating a number of relevant
factors, such as community involvement, personal experience and self-protective
behaviors, as well as communication variables (both interpersonal and mass
mediated communication).  The data were collected through mail survey from a
random sample of 750 upstate New York residents.
                Major findings include that personal level and societal level
risk judgments are distinct; Interpersonal communication primarily influences
personal judgments; Mass media influences societal judgments for some topics but
not others; Community involvement is an important predictor for the
discrepancies between the personal and societal level risk judgments.
                This study also found that the functions of factors vary across
the four health issues.  Different issues may have different degrees of
desirability, perceived probability, personal experience, perceived
controllability and stereotype salience.  These issue characteristics should be
further explored in future "personal-societal" risk perception studies.
 
 
            "Personal-societal" risk comparison
 
 
 
            "Personal-societal" risk comparison
 
 
            "Personal-societal" risk comparison
              Mass media, interpersonal communication and real-life factors
              in risk perception at personal and societal levels
 
              Introduction
                "I never thought it could happen to me."  People often say this
when they get involved with an accident or various other negative life events.
Contrary to the conventional wisdom that "the grass is always greener on the
other side of the fence," many researchers have found that people often feel
optimistic about their future compared to others.  Studies concerning automobile
accidents (Robertson, 1977), disease (Harris & Gutten, 1979; Kirscht, Haefner,
Kegeles & Rosenstock, 1966), environmental risk (Weinstein, Klotz  & Sandman,
1988),  and other negative future events (Weinstein, 1980; Weinstein & Klein,
1995) suggest that people tend to be unrealistically optimistic about their
vulnerability.  Even for purely chance events, such as flipping a coin, people
sometimes show optimistic biases (Langer & Roth, 1975).
                The purpose of this study is to investigate factors influencing
risk perceptions.  Distinguishing the lay-person's risk judgments, which rely on
intuition, from sophisticated analysts' risk assessments, Slovic (1987) called
the former a risk perception and defined it as "the judgments people make when
they are asked to characterize and evaluate hazardous activities and
technologies" (p. 280).
                Slovic (1987) states:
                _ there is wisdom as well as error in public attitudes and
                   perceptions.  Lay people   sometimes lack certain information
about hazards.
                   However, their basic conceptualization of risk is much richer
than that of the
                   experts and reflects legitimate concerns that are typically
omitted from expert
                   risk assessments. (p.285)
 
                In a similar line of reasoning, Dunwoody and Neuwirth (1991)
argued for the multivariate nature of risk perception. They emphasized the
distinction between cognitive and affective dimensions of risk judgment.
Another approach to the multidimensional concept of risk perception comes from
Tyler and Cook (1984), who differentiated between individual and societal level
risk judgments.  This paper uses Tyler and Cook's referential distinction in the
risk perception to investigate factors affecting people's individual and
societal levels of risk perception.
                Several recent lines of research investigate optimistic biases as
they pertain to vulnerability, or risk perception.  Research findings in these
areas suggest that individuals see themselves as being somehow different from
others in terms of the probability of good or bad things happening to them.
That is, individuals believe that negative events are less likely to happen to
them than to others in society, while positive events are more likely to happen
to themselves than to others across a wide range of topic areas (Glynn, Ostman
and McDonald, 1995; Culbertson and Stempel, 1985).
                The tendency to see others as less fortunate than oneself has
been called "unrealistic optimism" (Weinstein, 1980; Weinstein, Klotz  &
Sandman, 1988) or "illusion of invulnerability" (LeJeune & Alex, 1973).
                Weinstein (1980) states that:
 
                   According to popular belief, people tend to think they are
                   invulnerable.  They expect others to be victims of
misfortune, not themselves.
                   Such ideas imply not merely a hopeful outlook on life, but an
error in judgment
                   that can be labeled unrealistic  optimism. (p.806)
 
                In a similar line of reasoning, Davison (1983) suggested that in
general, individuals tend to think that the media will have greater impact on
others than on themselves, a phenomenon he labeled the "third person effect."
Even though the unrealistic optimism and third person effect research originated
in different fields (social-psychology and communication respectively), both
share some common principles.
                First, both unrealistic optimism and the third person effect
research has found that individuals tend to have the illusion of
invulnerability.  Research on the third person effect finds that individuals
expect they are immune from influence by persuasive mass communications.
Second, both lines of research focus on negative, undesirable events or results.
Gunther and Mundy (1993), for example, argue that the self-other discrepancy is
smaller in messages that promise to benefit the individual or which advocate
socially desirable outcomes.
                Third, studies in both areas suggest that the discrepancy between
the self and 'others' judgments increases as the hypothetical others are defined
in more broad and global terms (Perloff, 1993; Perloff & Fetzer, 1986).
Finally, both areas have some implications for individual and social actions.
In the case of unrealistic optimism, the behavioral implication is that it may
discourage people from taking adequate precautions to avoid victimization.  In
the case of the third person effect, the result may be "paternalistic" attempts
to regulate media content to protect the "vulnerable others."
                Several studies have investigated factors influencing
differential risk judgments.  In one line of research, a group of researchers
focused on the possible connection between media coverage of crime and public
attitudes toward the risk of being a crime victim.  Some have found that media
presentations influence personal judgments about the likelihood of criminal
victimization (Gerbner, Gross, Morgan & Signorielli, 1980), whereas others
report that there is no such relationships (Tyler, 1978; Doob & MacDonald, 1979;
Hirsch, 1980).  One explanation for these inconsistent findings is dealt with in
the impersonal impact hypothesis (Comstock, 1982; Hawkins & Pingree, 1982; Tyler
& Cook, 1984).
                The major assumption of the impersonal impact hypothesis is that
people can and do distinguish between two possible levels of judgmentDsocietal
and personal (Tyler & Cook, 1984).  Individuals' beliefs about the larger
community and conditions of community residents in relation to some social
phenomenon form a "societal level" of judgment; individual's beliefs about their
own condition and risks form a "personal level" of judgment (Glynn, Ostman &
McDonald, 1995).   Furstenberg (1971) first suggested the separation of personal
and societal level concerns about crime, and several studies since then have
shown this distinction empirically (Tyler, 1978; Doob & MacDonald, 1979; Tyler,
1980, 1984; Tyler & Cook, 1984; Coleman, 1993).
                Another component of the impersonal hypothesis is that mass
mediated messages affect people's perceptions of the prevalence of certain
problems or risks within a society, but do not affect their perceptions of
personal risks (Tyler & Cook, 1984).  They later revised this component of the
impersonal impact hypothesis to develop the "differential impact" hypothesis
(Glynn, Ostman & McDonald, 1995).  The impersonal impact hypothesis, thus,
develops the unrealistic optimism thesis further by incorporating communication
factors, especially their differential impact on risk judgments.
                Culbertson and Stempel (1985) found that individuals tend to be
optimistic about their personal health care (personal optimism) and pessimistic
about health care in  society as a whole (societal pessimism).  They tried to
connect these findings to the media malaise hypothesis, and argued that people
rely heavily on media coverage for a picture of society as a whole, but draw
primarily on personal experience when considering their own lives.
                These studies suggest that personal and societal level judgments
are distinct and largely separate, and people do not necessarily draw personal
implications from their general views about society.  Findings that the mass
media influence only, or primarily, the societal level judgments about health
problems have important implications for health education campaigns through the
mass media.  One technique often used in mass media health campaigns is to try
to increase individual perceptions of vulnerability and thus encourage
initiation of self protective behaviors.
                Studies concerned with societal and personal level judgments
either have been experimental (i.e. not in naturalistic settings) or have
excluded some of the factors which would appear to be closely linked to risk
perception (Tyler & Cook, 1984; Culbertson & Stempel, 1985).  An exception can
be found in a study by Coleman (1993).   Using an earlier version of the data
set that is used in current study, Coleman analyzed various factors influencing
personal and social level risk judgments.  In her regression analyses, Coleman
used the aggregated risk perceptions across various health issues.  This
clustering, as she acknowledged, made it difficult to determine factors relevant
to specific health issues.
                Taking those limitations of prior studies into consideration, the
current study uses survey data which investigate four health related
risks--radon, chemical contaminants in drinking water, AIDS, and heart
diseaseDto examine the functions of various factors, not only mass media and
interpersonal communication but also personal experience and community
involvement, in people's personal level and societal level risk judgments.
                The first major research question of this study is whether there
is a distinction between personal level and societal level judgments in terms of
the four health and environmental risks.  Dunwoody and Neuwirth (1991) addressed
two of the components of risk judgment: cognitive and affective.  The cognitive
dimension refers to how individuals assess their own likelihood of coming to
harm whereas the affective dimension refers to the concern, worry, or dread that
people feel about risks.  Based on this distinction, the major dependent
variables in the current study are risk judgments on the cognitive dimension
examined on two referential levels: personal and societal levels.
                The second major research question concerns which factors
influence the personal and societal level risk judgments differentially.  The
variables investigated in this study are mass media exposure, interpersonal
communication, and real life factors.   Mass mediataed communication
                As reviewed above, the impersonal impact hypothesis suggests that
mass mediated communication affect only the societal level risk judgment.
Scholars like Gerbner, Morgan and Signorielli (1982) noted that a major factor
underlying Americans' attitudes about health is denial and unwillingness to
believe that catastrophic illness could attack one's own immediate family.  They
found that those who choose television programs as a main source of health
information are significantly more likely to be complacent, rather than
concerned, about health.
              Interpersonal communication
                The impersonal impact hypothesis distinguishes informal social
communications from communications through mass media.  Studies have found that
informal social communications differ from mass media influences on risk
judgments (Tyler, 1984; Tyler & Cook, 1984).  Thus, we could expect that
interpersonal communication affects the personal level of concern whereas the
mass media exposure influences the societal level judgments.
              Real-life factors
                Past personal experience is another important factor in risk
perception.  Tversky and Kahneman (1973, 1974) argued that personal experience
makes it easier to imagine situations in which the event could occur, leading to
greater perceived probability through the mechanism of "availability."
                Social context is crucial to understanding how people perceive
risks.  If people participate more in various community organizations, they
might become less individualistic and develop a feeling of community with those
around them.  This might, in turn, decrease the discrepancy between the personal
and societal level judgments as the individual feels more a member of the
community rather than different from and separate from the community.
                Stereotype salience also is a factor that might influence the
process of impersonal impact.  For many events, including various health
problems, people may have a stereotyped conception of the kind of person to whom
this event happens.  If people do not see themselves as fitting the stereotype,
the representativeness heuristic suggests that people will conclude that the
event will not happen to them, overlooking the possibility that few of the
people who experience the event may actually fit the stereotype (Tversky &
Kahneman, 1977; Weinstein, 1980).  If stereotypes of the victim carry some kind
of ethical stigma as in the case of AIDS, expressing personal invulnerability
tends to serve an ego-defensive function, and the optimistic biases will be
exaggerated.  We would expect this stereotype to be exaggerated by media
coverage of AIDS, thus leading to a greater discrepancy between one's personal
assessment of risk and a societal  level perception.
                In summary, this study tests the following hypotheses:
              H.1  Personal level risk judgment is distinct from societal level
risk judgment: Across the four issues, people will perceive higher risk at the
societal level than at the personal level, and also functions of various factors
will be different across the two levels of judgments.
 
              H.2  Mass media exposure affects people's societal level judgment
more than their personal level judgment.
 
              H.3  Interpersonal communication affects personal level judgment
more than the societal level judgment.
 
              H.4a  The discrepancy between personal and societal level
judgments will be greater in the case of AIDS compared to the other health
problems due to stereotype salience.
 
              H.4b  In the same line of reasoning, the impersonal impact of mass
media will be greater in the case of AIDS compared to the other three topics.
 
              H.5  If people are more involved in the community, the
discrepancies between personal and societal level risk judgments will be less.
 
              Methods
                Data
                Data for this study came from a mail survey of a random sample of
1,500 households in the Northeast.  The sample was constructed from a composite
non-redundant telephone list and drivers license list.   The study used a
12-page booklet questionnaire with a 3-wave mailing and telephone follow-up of
non-respondents.  Of the original sample, 351 were either bad addresses or
otherwise unqualified respondents.  Individuals from 750 households responded
for a 65% response rate.
                Fifty-one percent of the respondents were male, 37% had graduated
from college and 11% had not finished high school.  Age of the respondents
ranged from 19 to 94 with a mean of 50 and a median of 48;  males were only
slightly older than females as the mean values were 51and 49 respectively.
Twenty-three percent of the respondents had income less than $20,000 and 25% had
incomes above $60,000.  These demographic figures compare favorably with census
figures as shown in Table 1.
 
              Table 1: Study sample compared with census data
 
 
 
 
              Census (%)
 
              Study Sample (%)
              Gender
              Male
              52
              51
 
              Female
              48
              49
 
              Married, living with spouse
 
 
              63
 
              69
 
              Education
 
              Elementary
 
              12
 
              4
 
              Attended HS
              11
              7
 
              Graduated HS
              37
              32
 
              Some College
              17
              19
 
              Graduated
              23
              38
 
              Income
 
              Less that $10,000
 
              11
 
              8
 
              $10-29,999
              28
              32
 
              $28-49,999
              30
              26
 
              $50,000 or more
              34
              31
 
 
 
 
 
                Measurement[1]
                Issues investigated were chemical contaminants in drinking water,
radon, AIDS and heart disease.  Risk judgmentsDthe major dependent variables of
this study--were measured across the four health issues by asking "How likely do
you think it is that [a specific issue] will cause you health problems?" at the
personal level and "How likely do you think it is that [a specific issue] will
be a very serious problem for our country in the future?" at the societal level.
The answers were coded in a 5-point scaleD1 for "Not at all likely" through 6
for "Has already caused problems."
                There are five major groups of independent variablesDmass media
exposure, interpersonal communication, community involvement, personal
experience, and demographics.  Three mass media exposure variables were included
in the analysisDtelevision, newspaper and magazine exposure.  Television
exposure was measured by four questions about national network news, local news,
news documentaries and science programs, which are more likely to cover health
issues (Chronbach's alpha was.70).  Newspaper and magazine measures focused on
exposure to specific content: For newspaper exposure, three questions about
environment, health/medical and food/nutrition were used (Chronbach's alpha was
.87), whereas for magazine reading, two questions about health/medical and
science magazines were used (Chronbach's alpha was .38).  The reliability was
low for the magazine measures, but is judged acceptable because there are only
two items, and the Pearson correlation coefficient for the two items is .24
(p<.001).
                Interpersonal communication was measured by two questions for
each health problem: (1)  talking with neighbors and the others, and (2) talking
with knowledgeable experts about the specific topic.  For chemical contaminants
in drinking water, AIDS, and heart disease, the two questions were combined and
recoded as a 3-point scale from 0 through 2.  The reliabilities measured by
Chronbach's alpha are .48  for water contaminants, .56 for AIDS, and .52 for
heart disease.  These reliability measures are relatively low, but are judged
acceptable because there are only two items and the Pearson correlation
coefficients for the two items across the four health issues are .32, .40, and
.35 (p values for all three <.001).  There were only 61(8.4%) people of the
sample had talked about radon with neighbors or experts.  Thus for the case of
the radon, a dummy variable was created with  0 for those who never talked with
neighbors nor experts, and 1 for those who had talked in the last month.
                Community involvement was measured with four different questions
asking about various community group activities (Chronbach's alpha was .59).
They were combined and recoded as a four-point scale, 0 through 3 with 0
representing no community involvement, 2 moderate involvement, and 3 high
involvement.
                Personal experience relevant to each of the four health problems
was asked about self, parents, friends and acquaintances.  These measures were
coded as 1 through 5.  A score of 5 represents "self experience" with the health
problem (i.e., the respondent has heart disease, AIDS, etc.).  The smaller the
score, the further the risk experience is from the respondent.  A score of 2,
for example, would be that the respondent has a friend with heart disease or
AIDS, whereas 1 means "no experience."
                Self protective behaviors (e.g., installing water filters and
blood cholesterol level tests) should affect individuals' risk perception.  Few
studies have taken past self-protective behaviors into consideration.  Self
protective behaviors for each health problem except AIDS were measured by dummy
variables.  Respondents were asked if they had installed water filters or if
they tested their drinking water for contaminants; if they had tested their
blood for cholesterol; if they tested their home's radon level.  In the case of
heart disease, we also controlled for the respondents' exercise (a dummy
variable) and diet habits.  These variables were included in the regression as
control variables.  In addition, four demographic characteristicsDage, income,
education and gender were also controlled in the regression analyses.
 
              Results of the data analysis
                The first step in the data analysis was to check the mean
differences in people's personal level and societal level judgments for each of
the four topics.  Table 2 shows the results of the paired t-test of the mean
differences across the four topics.  It appears that people, in general, have
more concern at the societal level than at the personal level for each topic.
As we expected in hypothesis 4a, AIDS is the case where the most discrepancy
between the two levels of judgment was found: Personal concern about AIDS was
the lowest of all four health issues, but the highest of all at the societal
level (means are 2.24 and 5.20 respectively).  With regard to radon, people are
not concerned at personal level and only slightly concerned at the societal
level (means are 2.88 and 3.74 respectively).
 
              Table 2. Paired t-test for the mean differences across personal
and societal level concerns for each topics.
 
 
 
              Personal level
              Societal level
              Paired t-test
 
 
              Mean
              sd
              Mean
              sd
              t
              df
              Water chemical
              3.56
              1.28
              4.70
              1.06
              24.97**
              731
              Radon
              2.88
              1.06
              3.74
              1.15
              18.91**
              723
              AIDS
              2.24
              1.34
              5.20
              1.13
              48.91**
              736
              Heart disease
              3.83
              1.26
              4.92
              1.11
              20.79**
              740
                     ** p<.001
 
 
                   The peculiarity of AIDS in terms of the discrepancy between
the two levels of judgment also is presented in Figure 1, which shows the
percent of respondents for each value of discrepancy between societal and
personal levels.  The discrepancy values were computed by subtracting the values
of personal level concern from the values of societal level concern for each
respondent.  Mean discrepancy values with standard deviations in parentheses for
each issues are 1.09 (1.43) for water chemicals; 2.96 (1.64) for radon;  0.85
(1.21) for AIDS; and 1.15 (1.24) for heart disease.  As shown in figure 1, more
people have positive scores on the discrepancy measure, suggesting that in
general, they have a pattern of personal optimism and impersonal pessimism
across all four topics.
 
 
              Figure 1. Percentages of the respondents for each level of
discrepancy between the two levels of judgments.
 
                   To test hypotheses 1 through 4, two regressions, using
personal level concern and societal level concern as the dependent variables,
were run for each of the four topics (see Table 3).  The model fit measures (R2)
show significant p values across all of the regressions.
                Hypothesis 1 predicted that personal level risk judgment would be
distinct from societal level risk judgment.  The regression results show that
people's personal and societal level judgments are distinct from each other.
That is, people's concern about various health problems is a function of
different factors or different functions of the same factor across the two
levels.  For example, in the case of the chemical contaminants in drinking
water, community involvement, personal experience and mass media were
significant only for people's societal concern.  Education, age and
interpersonal communication were significant factors only for personal concern
about radon.
                   For AIDS, age and magazine reading were significant for
personal concern, whereas gender and personal experience were significant for
societal concern.  On the other hand, age, gender and community involvement were
significant factors for societal concern for heart disease but not for personal
concern.  The most interesting distinction was found in the relationship of
education to concerns about AIDS.  At the personal level, the more educated a
person is, the less concerned s/he becomes (b=-.13, p<.01); at the societal
level, the direction is reversedDthe more educated a person is, the more
concerned s/he becomes (b=.10, p<.05).  These results support the first
hypothesis that personal level risk judgment is distinct from the societal level
risk judgment.
                   The second hypotheses stated that mass media exposure affects
people's societal level judgment more than their personal level judgment.  This
hypothesis was only partly supported.  For the case of the chemical contaminants
in drinking water, mass media strongly influenced societal judgment (b=.14,
p<.01 for TV; b=.15, p<.001 for magazine) compared to personal level judgment,
in which both TV and magazines showed border line p values (.069 and .047
respectively).  But the function of mass media varies across the four topics.
The predicted pattern was apparent only in the case of the chemical contaminants
and radon case, but not in AIDS or in heart disease.
                   The third hypothesis that interpersonal communication affects
personal level judgment more than societal level judgment was partially
supported.  In the case of radon, interpersonal communication shows the pattern
predicted: a strong influence only on the personal level.  In terms of water
contaminants, interpersonal communication showed a strong influence on the
personal level, but a border line p value (.041) for the societal level concern.
For the other two health issuesDAIDS and heart diseaseDinterpersonal
communication was not a significant predictor either for the personal or the
societal level risk judgments.
                Hypothesis 4a suggests that the discrepancy between people's
personal and societal level judgments will be greater for AIDS.  As shown in
Table 1, this hypothesis was supported.  But hypothesis 4b about the impersonal
impact of mass media was not supported by the regression analysis.  That is, the
mass media showed no significant influence on either of the personal or societal
level risk judgments[2].  This finding could be a function of a ceiling effect.
That is, the mass media have been saturated with AIDS coverage and most people
are aware and know how it is transmitted.  As a result, there is little new
information about AIDS in the mass media which might affect people's personal
and societal level of concern.  But, as shown in Table 1, AIDS has the greatest
discrepancy between personal and societal concern among the four topics.
              The relationship of personal experience with the four risk factors
was inconsistent across personal and societal judgments.  For AIDS and chemical
contaminants in drinking water, personal experience showed a significant
relationship to only societal concern.  It may be that with AIDS, some
combination of personal control and a self-serving bias allowed individuals to
reason that they personally were not at risk, but that overall society was at
risk.  Even when people knew and had heard of somebody with AIDS, they still
felt safe because most people believe that they know how to prevent the disease.
Personal  experience with AIDS increased societal level concern but not personal
concern.
                The case of water contaminants, however, is more puzzling.  One
possible explanation for the lack of influence of personal experience on
personal level risk judgment might be that there are fairly easy ways of
controlling most chemical contaminants in water, such as installing water
filters.  Ease of control may have alleviated the impact of personal experience
on people's personal level of concern.
                   The fifth hypothesis states that the discrepancy between
personal level and societal level concerns for each topic will be smaller for
the people who are more involved in community service.  To test this hypothesis,
an absolute value of the difference between each respondent's personal level
concern and societal level concern for each topic was regressed on the same
independent variables as in the prior regressions.
                        The overall fit of the model was significant for two of the
four issues: AIDS and heart disease (see Table 4).  The coefficients for
involvement are all negative and significant at .05 except for the case of water
contaminants and heart disease.  But, in the case of heart disease, the p value
for the coefficient is .089, which becomes significant if a one-tailed p value
was applied since the hypothesis is directional.  The insignificant result in
the case of water chemicals is surprising especially because the issue is often
a community-wide problem.  That is, if one individual has problems with drinking
water, it is highly likely that the whole neighborhood shares the same problem.
                   Several demographic factors, such as education, age and
gender, account for the observed discrepancy in risk judgment about AIDS, which
has the highest mean value of the dependent variable among the four issues.
Both age and education are positively related to the discrepancy of personal and
societal concern for AIDS.  Women also showed greater discrepancy in their
judgment about AIDS than did men.
                   For heart disease, personal experience decreases the
difference between personal and societal levels of concern.  Even though mass
media was not a significant predictor for either personal or societal concern,
TV exposure appears to decrease the difference between personal and societal
levels of concern (b=-.09, p<.05).
 
              Table 4. Results of regression of the absolute differences between
personal and societal level risk judgments.
 
 
               Water Contaminants
 
               Radon
 
               AIDS
 
               Heart Disease
 
 
               (
 
               (R2
 
               (
 
               (R2
 
               (
 
               (R2
 
               (
 
               (R2
 
               Demographics
 
 
               1.01
 
 
               1.79*
 
 
               3.28**
 
 
               3.14**
                  Age
               .09
 
               .10*
 
               .12**
 
               -.33
 
                  Sexa
               .01
 
               .08
 
               .12**
 
               .06
 
                  Education
               .00
 
               .09
 
               .17**
 
               .07
 
                  Income
               .07
 
               .08
 
               -.00
 
               .04
 
               Behaviors
 
               0.28
 
               0.01
               N/A
 
               0.96
               Real-Life factors
 
               0.15
 
               1.26*
 
               0.95*
 
               4.20**
                  Involvement
             -.01
 
               -.12**
 
               -.09*
 
               -.07
 
                  Experience
               .05
 
               -.03
 
               .04
 
               -.21**
 
               Interpersonal
 
               0.49
 
               0.03
 
               0.00
 
               0.00
                  Talking
              -.08
 
               .01
 
               .01
 
               .01
 
               Mass Comm.
 
               0.12
 
               0.05
 
               0.43
 
               0.65
                 TV
 
               .03
 
               .01
 
               -.05
 
               -.09*
 
                 Magazine
               .02
 
               .02
 
               -.04
 
               .04
 
                 Paper
              -.01
 
               .00
 
               .02
 
               .01
 
               Total R2 (%)
               2.04
               3.15
               4.66**
               8.95**
 
                 a. 1=male; 2=female    * p<.05  ** p<.01
 
 
              Summary and discussion
                Traditionally studies on risk perception have come from
geography, sociology, political science, anthorpology and psychology (Slovic,
1987).  Many of those studies approached risk perception as a unidimensional
concept that can be captured unidimensionally in the ubiquitous "risk estimate"
(Dunwoody & Neuwirth, 1991).  Following a group of scholars who argued for the
multidimensional aspects of risk perception (i.e., Dunwoody & Neuwirth, 1991;
Slovic, 1987; Coleman, 1993; Tyler & Cook, 1984), this study investigated two
different referential components of risk judgmentsDpersonal and societal level
judgments.  This study has replicated and expanded the impersonal impact studies
by using survey data, which provide a more natural picture of people's risk
judgments about various health issues, and by using issue-specific risk
judgments, which allow to investigate issue-specific functions of various
factors.
                In an effort to illuminate "the multivariate nature of risky
situations," as Dunwoody and Neuwirth (1991) point out, this study has
incorporated a number of relevant factors which may be related to risk
perceptions, such as community involvement, personal experience and
self-protective behaviors, as well as communication variables (both mass media
and interpersonal).
                One finding is that personal level and societal level risk
judgments appear to be distinct and should be dealt with separately in risk
perception studies to illuminate the complex nature of risk perception.  In
addition,  interpersonal communication appears to influence people's concern
about health problems primarily at the personal level.  The mass media,
especially television and magazine exposures, on the other hand, have
significant impacts at the societal level for some topics but not others.  These
differences need further exploration.
                Community involvement appears to be an important predictor for
the differences in concern about health topics at the personal and societal
levels.  It seems that if people get involved with their community, they hold
less discrepant visions between social and personal risks.
                In addition to the relatively consistent findings across the four
issues, this study found that the functions of factors vary across health
issues.  AIDS, for example, appeared to be very different from other issues in
many ways.  In the case of AIDS, the discrepancy between people's personal level
and societal level concern might not be the "unrealistic" optimism, but rather
quite "realistic" optimism, considering the specific route of the virus
transmission.  The current data did not allow a test of this speculation.
                Different issues may have different degrees of desirability,
perceived probability, personal experience, perceived controllability and
stereotype salience.  Studies in optimistic bias in "self-other" risk
comparisons found, in general, that optimism is greater for risks rated low in
probability and for risks judged to be controllable by personal action
(Weinstein, 1989).  These issue characteristic factors found in unrealistic
optimism studies should be incorporated in the future studies of the
personal-societal risk comparisons.
                Among issue characteristic factors, perceived controllability
deserves more attention from researchers.  Some studies on unrealistic optimism
have pointed to the importance of perceived controllability in the operation of
unrealistic optimism: the greater the perceived controllability of a negative
event, the greater the tendency for people to believe that their own chances to
have the problem are less than average (McKenna, 1993; Weinstein, 1980).  Thus,
we could expect that perceived controllability of health problems would increase
the discrepancy between risk judgments at the personal and societal levels.
                Mass media variables used in this study were not issue specific
but did focus on news, science and health information.  Even though we measured
them by specifically asking about some health/medical or scientific issues for
newspapers and magazines, we do not know how often the respondents were exposed
to each specific health issue, such as water contaminants, radon, AIDS or heart
disease; neither do we know about the content of the media coverage for those
issues.  Studies with multiple-method approach using both content analysis and
survey data analysis are required for a better understanding of the functions of
communication factors in risk judgments.  Also attention to mass media in
addition to a simple measure of exposure might illuminate the rather subtle
functions of mass media in risk perception.
                The measurement of the societal risk judgment of this study is
different from the usual operationalizations used in similar studies.  This was
an attempt to overcome the tendency to judge society based on smaller segments.
There are always segments of society that are at greater risk.  Thus, asking
whether an individual is at greater risk than others opens the door to making
that judgment based on a perceived small group which may actually be at greater
risk.  The attempt in this study was to operationalize the societal level
variable in such a way that requires respondents to generalize more broadly.
Future studies may wish to re-operationalize this variable taking into account
such issues.  As Slovic (1987) argued, risk perception means more than expected
probabilities of getting involved with a health problem.  Attempts to understand
risk perceptions should utilize broader conception of risk, incorporating not
only the quantitative but also the qualitative characteristics, such as
familiarity, control, catastrophic potential, equity and level of knowledge.
                References
 
              Coleman, C. (1993). The influence of mass media and interpersoanl
                   communication on societal and personal risk judgments.
Communication Research,
                   20, 611-628.
 
              Comstock, G. (1982). Violence in television content: An overview.
                   In D. Pearl, L. Bouthilet and J. Lazar (Eds.), Television and
behavior: Ten
                   years of scientific progress and implications for the
eighties (Vol. 2, pp.
                   108-125). Rockville, MD:N.I.M.H.
 
              Culbertson, H. M. & Stempel, G. H. (1985). Media malaise:
                   Explaining personal optimism and societal pessimism about
health care. Journal
                   of Communication, 35 (2), 180-190.
 
              Dunwoody, S. & Neuwirth, K. (1991). Coming to terms with the
                   impact of communication on scientific and technological risk
judgments. In L.
                   Wilkins and P. Patterson (Eds.), Risky Business (pp. 11-30).
New York:
                   Greenwood.
 
              Gerbner, G. , Morgan, M. & Signorielli, N. (1982). Programing
                   health portrayals: What viewers see, say and do. In D. Pearl,
L. Bouthilet & J.
                   Lazar (Eds.), Television and behavior (Vol. 2, pp.291-307).
Rockville,
                   MD:N.I.M.H.
 
              Glynn, C. G., Ostman, R. E. & McDonald, D. G. (1995). Opinions,
                   perception, and social reality. In T. Glasser and C. Salmon
(Eds.), Public
                   opinion and the communication of consent (pp. 149-280). New
York: Guilford.
 
              Gunther, A. C. & Muncy, P. (1993). Biased optimism and the
                   third-person effect. Journalism Quarterly, 70,  58-67.
 
              Harris, D. M. & Guten, S. (1979). Health protective behavior: An
                   exploratory study. Journal of Health and Social Behavior, 20,
17-29.
 
              Halpern, M. T. & Warner, K. E. (1994). Radon risk perception and
                   testing: Sociodemographic correlates. Journal of
Environmental Health, 56 (7),
                   31-35.
 
              Hawkins, R. P. & Pingree, S. (1982). Television's influence on
                   social reality. In D. Pearl, L. Bouthilet and J. Lazar
(Eds.), Television and
                   behavior: Ten years of scientific progress and implications
for the eighties
                   (Vol. 2, pp. 224-247). Rockville, MD:N.I.M.H.
 
              Kirscht, J. P., Haefner, D. P., Kegeles, S. & Rosenstock, I.
                   (1966). A national study of health beliefs. Journal of Health
and Human
                   Behavior, 7,  248-254.
 
              Langer, E. J. & Roth, J. (1975). Heads I win, tails it's chance:
                   The illusion of control a as function of the sequence of
outcomes in a purely
                   chance task. Journal of Personality and Social Psychology,
32,  951-955.
 
              LeJeune, R. & Alex, N. (1973). On being mugged: The event and its
                   aftermath. Urban Life and Culture, 2,  259-287.
 
              Mays, V. M. & Cochran, S. D. (1988). Issues in the perception of
                   AIDS risk and risk reduction activities by Black and
Hispanic/Latina women.
                   American Psychologist, 43, 949-957.
 
              McKenna, F. P. (1993). It won't happen to me: Unrealistic
                   optimism or illusion of control? British Journal of
Psychology, 84 (1), 39-50.
 
              Peltu, M. (1991). Risk perception and the real world. New
                   Scientist, 131 (1782), 4.
 
              Perloff, R. M. (1993). Third-person effect research 1983-1992: A
                   review and synthesis. International Journal of Public Opinion
Research, 5 (2),
                   167-184.
 
              Perloff, L. S. & Fetzer, B. K. (1986). Self-other judgments and
                   perceived vulnerability to victimization. Journal of
Personality and Social
                   Psychology, 50, 502-510.
 
              Robertson, L. S. (1977). Car crashes: perceived vulnerability and
                   willingness to pay for crash protection. Journal of Community
Health, 3,
                   136-141.
 
              Slovic, P. (1987). Perception of risk. Science, 236, 280-285.
 
              Tversky, A. & Kahneman, D. (1973). Availability: A heuristic for
                   judging frequency and probability. Cognitive Psychology, 5,
207-232.
 
              Tversky, A. & Kahneman, D. (1979). Causal schemata in judgments
                   under uncertainty. In M. Fishbein (Ed.), Progress in social
psychology.
                   Hillsdale, NJ: Erlbaum.
 
              Tyler, T. R. (1984). Assessing the risk of crime victimization:
                   The integration of personal victimization experience and
socially transmitted
                   information. Journal of Social Issues, 40, 27-38.
 
              Tyler, T. R. (1980).  Impact of directly and indirectly
                   experience events: The origin of crime-related judgments and
behaviors. Journal
                   of Personality and Social Psychology, 39, 13-28.
 
              Tyler, T. R. & Cook, F. L. (1984). The mass media and judgments
                   of risk: Distinguishing impact on personal and societal level
judgments.
                   Journal of Personality and Social Psychology, 47,  693-708.
 
              Weinstein, N. D. (1989). Optimistic biases about personal risks.
                   Science, 246, 1232-1233.
 
              Weinstein, N. D. (1980). Unrealistic optimism about future life
                   events. Journal of Personality and Social Psychology, 39,
806-820.
 
              Weinstein, N. D. & Klein, W. (1995). Resistance of Personal Risk
                   Perceptions to Debiasing Interventions.  Health Psychology,
14(2), 132-140.
 
              Weinstein, N. D., Klotz, M. L. & Sandman, P. M. (1988).
                   Optimistic biases in public perceptions of the risk from
radon. American
                   Journal of Public Health, 78 , 796-800.
 
              Weinstein, N. D. & Lachendro, E. (1982). Egocentrism as a source
                   of unrealistic optimism. Personality and Social Psychology
Bulletin, 8 (2),
                   195-200.
 
              Wildavsky, A. & Dake, K. (1990). Theories of risk perception: Who
                   fears what and why? Daedalus, 119 (4), 41-60.
                 [1]  Questions and response categories for each concepts are
reported more fully in the Appendix A.
                 [2]  Mass media were not significant in a regression in which
the difference between the two levels of concern was predicted (See table 4).
 
            "Personal-societal" risk comparison
 
             "Personal-societal" risk comparison
 
            "Personal-societal" risk comparison
 
              Appendix A
 
              Personal risk judgment:
              How likely do you think it is that chemical contaminants in your
drinking water (radon in your home; AIDS; heart disease) will cause you health
problems?
 
              Societal risk judgment:
              How likely do you think it is that chemical contaminants in your
drinking water (radon in your home; AIDS; heart disease) will be a very serious
problem for our country in the future?
 
              For both of the risk judgments, response categories were "not al
all likely," " not very likely," "not sure," "somewhat likely," " very likely,"
and "has already caused problems" (Each was coded as 1 through 6 respectively).
 
              Mass media:
              How often do you read articles in your newspaper about
[environment; health/medical; food and nutrition]?
 
              How often do you watch the following kinds of daily or weekly
television programs [national network news; local news; news documentaries;
science programs]?
 
              How often do you read the following types of magazines [health or
medical magazines or newsletters; science magazines]?
 
              For each of the mass media questions, response categories were
"almost never," "once or twice/wk," "three or four times/wk," and "Five or more
times/wk" (Each was coded as 1 through 4 respectively).
 
              Interpersonal communication
              Within the past month, about how many times would you say you have
talked with your neighbors about the following topics [chemical contaminants in
drinking water; heart disease caused by diet and lack of exercise; lung cancer
from radon in your home; AIDS]?
 
              Within the past month, about how many times would you say you have
talked with experts, professionals, or other highly knowledgeable people about
the following topics [chemical contaminants in drinking water; heart disease
caused by diet and lack of exercise; lung cancer from radon in your home; AIDS]?
 
              For each of the interpersonal communication questions, response
categories were "never," "1-2 times," "3-4 times," and "5 or more times" (Each
was coded as 1 through 4 respectively).
 
              Experience
              Have you or others you know experienced any of the following
health problems [heart disease; lung cancer without smoking; AIDS]?
 
              For each of the personal experience questions, response categories
were "no one I know," "acquaintances," "friends/relatives," "mother/father," and
"I have" (Each was coded as 1 through 5 respectively).
 
              For chemicals in the water, we used two questions to create a
comparable scale to the ones for the other problems above:
              Who do you know who has or has had chemical contaminants in their
drinking water?
              no one I know     2.  acquaintances       3.  friends, relatives  4.
                neighbors
 
              For those who said that they tested for chemical contaminants, we
asked about the test result and coded as 5 if it showed contamination and added
that to the scale came out from the above question.
 
              Involvement
              About how much time each month are you involved in activities of
the following groups [neighborhood groups; service, professional or youth
oriented groups; religious organizations; other organized groups such as
recreational/social groups]
 
              For each of the involvement questions, response categories used
were "none," "less than one hour," "one to five hours," and "more than five
hours" (Each was coded as 1 through 4).
 
              Demographics
              Which income category best describes the annual gross income
(before taxes) for all wage earners in your immediate family living at this
address last year?  The response categories used were "less than $10,000"
"$10,000-$19,999" "$20,000-$29,999" "$30,000-$39,999" "$40,000-$49,999"
"$50,000-$59,999" and "$60,000 or more."  Each was coded as 1 through 7
respectively.
 
              How many years did you attend school?  Response categories were
"1. elementary school (1-9 years)," "2. attended high school (9-11 years)," "3.
graduated high school (12 years)," "4. attended college," and "5. graduated
college."
 
              Are you Male____ or Female ____ (coded male as 1; female as 2)
 
              How old were you on your last birthday? _____years
 
 
              Behavior:
              Please indicate the extent you have done or plan to do the
following.
                   installed water filter
                   tested my home for indoor radon
                   had your blood cholesterol level tested
                   tested my water for chemical contaminants
 
              For each of the questions above, response categories were "I see
no need to do this," "I may consider doing this," "I plan to do this," and "I
have done this."  For radon, the last category was coded as 1 and all the others
as 0.  For blood cholesterol test, if they tested and the results showed problem
they were coded as 2; if they tested and the results were okay coded as 1, and
coded as 0 if they haven't tested.  Behaviors related to water contaminants were
coded as 1 if respondents either installed water filter or tested for water
contaminants and 0 for the others.
 
 
 
 
 
 
           Appendix B.  The unstandardized regression coefficients and p values
for the regression of the personal and societal level risk judgments.
 
 
 
            Chemicals in water
            Radon
            AIDS
            Heart Disease
 
            Personal
            Societal
            Personal
            Societal
            Personal
            Societal
            Personal
            Societal
 
 
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            Income
            -.06
            (.03)
            .051
            -.01
            (.03)
            .734
            -.05
            (.03)
            .070
            -.02
            (.03)
            .463
            -.03
            (.03)
            .421
            -.03
            (.03)
            .235
            -.09
            (.03)
            .001
            -.05
            (.03)
            .046
            Education
            -.05
            (.06)
            .365
            -.07
            (.05)
            .136
            -.10
            (.05)
            .042
            .02
            (.05)
            .684
            -.16
            (.06)
            .006
            .10
            (.05)
            .032
            -.01
            (.05)
            .830
            .09
            (.05)
            .074
            Sex
 
            .26
            (.11)
            .014
            .27
            (.09)
            .002
            .29
            (.09)
            .001
            .52
            (.10)
            .000
            ..01
            (.11)
            .918
            .38
            (.09)
            .000
            -.07
            (.10)
            .479
            .23
            (.10)
            .017
            Age
 
            -.01
            (.00)
            .000
            -.01
            (.00)
            .008
            -.01
            (.00)
            .021
            -.00
            (.00)
            .301
            -.01
            (.00)
            .001
            -.00
            (.00)
            .702
            -.00
            (.00)
            .508
            -.01
            (.00)
            .002
            Involve
            -.05
            (.05)
            .327
            -.10
            (.04)
            .018
            .05
            (.05)
            .242
            -.07
            (.05)
            .184
            .07
            (.06)
            .180
            -.05
            (.05)
            .228
            .02
            (.05)
            .649
            -.09
            (.05)
            .050
            Experience
            .03
            (.06)
            .630
            .14
            (.05)
            .007
            .04
            (.05)
            .465
            -.03
            (.06)
            .543
            .10
            (.10)
            .316
            .19
            (.08)
            .022
            .43
            (.04)
            .000
            .08
            (.05)
            .049
            Talking
            .28
            (.09)
            .001
            .15
            (.07)
            .041
            .43
            (.16)
            .009
            .21
            (.18)
            .242
            .03
            (.07)
            .707
            .05
            (.06)
            .397
            .11
            (.06)
            .069
            .08
            (.06)
            .175
            TV
 
            .04
            (.02)
            .069
            .05
            (.02)
            .003
            .04
            (.02)
            .372
            .03
            (.02)
            .065
            .02
            (.02)
            .165
            .00
            (.02)
            .855
            .02
            (.02)
            .177
            -.02
            (.02)
            .192
            Magazine
            .08
            (.04)
            .047
            .12
            (.03)
            .000
            .01
            (.04)
            .844
            .04
            (.04)
            .322
            .08
            (.04)
            .047
            .01
            (.03)
            .834
            -.02
            (.04)
            .532
            .00
            (.04)
            .947
            Paper
 
            -.01
            (.02)
            .612
            -.01
            (.02)
            .604
            -.01
            (.02)
            .551
            -.02
            (.02)
            .332
            -.00
            (.02)
            .859
            .02
            (.02)
            .357
            -.01
            (.02)
            .442
            .02
            (.02)
            .198
 
 
 
 
 
 
           Appendix C. The unstandardized regression coefficients and p values
for the regression of the absolute discrepancies between personal and societal
level risk judgments.
 
 
 
            Water Contaminants
 
            Radon
 
            AIDS
 
            Heart Disease
 
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            b
            (se)
            p
            Income
            .04
            (.03)
            .172
            .04
            (.03)
            .090
            -.00
            (.04)
            .936
            .02
            (.03)
            .447
            Education
            .00
            (.05)
            .976
            .09
            (.05)
            .052
            .23
            (.07)
            .001
            .09
            (.06)
            .122
            Sex
 
            .01
            (.10)
            .886
            .16
            (.09)
            .081
            .37
            (.13)
            .004
            .14
            (.10)
            .175
            Age
 
            .01
            (.00)
            .061
            .01
            (.00)
            .035
            .01
            (.01)
            .007
            -.00
            (.00)
            .495
            Involve
            -.01
            (.05)
            .859
            -.12
            (.05)
            .008
            -.13
            (.06)
            .038
            -.09
            (.05)
            .089
            Experience
            .07
            (.06)
            .254
            -.03
            (.05)
            .555
            .11
            (.12)
            .353
            -.23
            (.05)
            .000
            Talking
            -.12
            (.09)
            .076
            .06
            (.17)
            .735
            .02
            (.08)
            .839
            -.01
            (.06)
            .848
            TV
 
            .01
            (.02)
            .499
            .00
            (.02)
            .858
            -.03
            (.02)
            .273
            -.04
            (.02)
            .049
            Magazine
            .01
            (.04)
            .735
            .02
            (.04)
            .650
            -.05
            (.05)
            .326
            .04
            (.04)
            .381
            Paper
 
            -.01
            (.02)
            .747
            -.00
            (.02)
            .981
            .01
            (.02)
            .642
            .00
            (.02)
            .886