This paper was presented at the Association for Education in Journalism
and Mass Communication in Toronto, Canada, August 2004.
If you have questions about this paper, please contact the author
directly. If you have questions about the archives, email
[log in to unmask] For an explanation of the subject line, send email to
[log in to unmask] with just the four words, "get help info aejmc," in the
body (drop the "").
(Oct 2004)
Thank you.
Elliott Parker
************************************************************************
Optimistic Bias about Cancer Risk and Information Sources in Appalachia
Hong Ji
Doctoral student
[log in to unmask]
and
Daniel Riffe
Professor
and
Presidential Research Scholar
(740-593-2597)
[log in to unmask]
E. W. Scripps School of Journalism
Ohio University
Athens, OH 45701
Paper submitted to the Science Communication Interest Group of the
Association for Education in Journalism and Mass Communication for its
annual conference, Toronto, Canada, August 2004.
Abstract
This study examined relationships of Appalachian residents'
optimistic bias to their knowledge about cancer and information source. A
significant but slight relationship between optimistic bias and cancer
knowledge was found. Optimistic bias was not significantly related to the
ease of access to health information, or to number and types of sources
identified.
Optimistic Bias about Cancer Risk and Information Sources in Appalachia
Introduction
Cancer is the second primary cause of death in the United
States (Cancer death rates—Appalachia, 1994-1998, 2002). The Appalachian
area, with a population of about 21 million, and with 399 counties in 13
states (Rosswurm et al., 1996, p.442), is a region known for high cancer
risk because of its higher rates of poverty, lower education and limited
health care access (Cancer death rates—Appalachia, 1994-1998, 2002). In
addition, research suggests that environmental threats in the area may be
related to perceived cancer risk (Riffe, 2003).
Yet, personal cancer risk perception does not always reflect
reality. In fact, optimistic bias is common: what Weinstein identified
(1980) as the belief that other people are more easily caught in negative
events than one's self. Riffe found optimistic bias existed when people
judged how cancer risk was linked to environmental threats (Riffe, 2003).
In addition, use of media for science and health-related news was related
to optimistic bias (Riffe, 2003).
Although public health campaigns for years focused on cancer
risks providing cancer-related information, people continue to engage in
the unsafe practices such as smoking, poor diet and excessive drinking. Why
would campaigns and efforts fail to change same people's knowledge about
their lifestyle risks? Are they not being reached by the campaign? Are they
unable to access information and gain knowledge about cancer risks? These
questions are somewhat simplistic, of course; many people choose to engage
in risky behaviors and unrealistic optimism often operates among these
people, enabling them to accept the risk which, they believe, is less for
them. Presumably, knowledge about the causes of cancer should be related to
one's risk perception and likelihood of behavioral change and therefore,
the existence of any optimistic bias. On the one hand, more knowledge might
lead to greater self risk perception due to the greater concerns about many
causes of the disease, or it may conversely lead to less self risk
perception because of awareness that particular causes are not part of
one's lifestyle.
Thus, this study explores how people living in Appalachian
Southeast Ohio perceive cancer risk, relating it to their knowledge about
cancer, their access to information, and their information use.
Literature review
Optimistic bias: Weinstein (1980) defined optimistic bias
as a person's belief that "his or her chances of experiencing a negative
event are less than average" (p. 806), such as having a drinking problem,
attempting suicide, or getting lung cancer (p. 810). Among other
explanations, this unrealistic judgment sometimes originates with people's
belief that they are better than others (Weinstein & Klein, 1996, p.1).
Cancer optimistic bias: Fontaine and Smith (1995) compared
optimistic bias in cancer risk perceptions among Americans and British
people, finding that British respondents exhibit greater optimistic bias
about cancer than American respondents. Optimistic bias was not
significantly related to gender.
Optimistic bias vs. media use/information channel: The
relationship among optimistic bias and media use is mixed. Coleman (1993)
examined impacts of mass media, interpersonal communication and
self-efficacy on risk judgments, and argued that mass media affected
personal-level risk to a certain degree, while interpersonal channels and
self-efficacy influenced societal-level risk judgments. Her study suggested
that the factors impacting people's risk perception were complicated
(p.623), and "[r]eliance on one medium—television— may somewhat limit the
overall risk picture one obtains" (p.625). Chapin (2000) proposed that
increased media use would not predict increased optimistic bias. Salwen and
Dupagne (2003) argued that optimistic bias was significantly related to
"amount of television viewing and frequency of newspaper reading" (p.70).
Riffe (2003) argued that optimistic bias was significantly related to
medical and science research news reading, and not significantly related to
other general media use.
Demographics vs. optimistic bias: Weinstein (1987) argued
that optimistic bias was not significantly different by age, sex, job
status or education. However, Salwen and Dupagne (2003) found that age was
a primary predictor for optimistic bias about the so-called "Y2K" problem.
Younger people thought other people were more easily affected by Y2K than
themselves. Race was also a significant negative predictor. Non-White
people believe that other people were more likely to experience Y2K
problems (p.69-70). Chapin (2000) reported that there was no gender
difference existing in optimistic bias about the likelihood of HIV
infection in his study. Moreover, his study did not find relationships
among optimistic bias and educational level and HIV/AIDS knowledge. In his
study on the environmental hazards, cancer risk judgment, and media use in
Appalachia, Riffe (2003) found that optimistic bias was significantly
related to people's educational level, but not significantly related to
age, gender or people's household income.
This study explores the relationship of optimistic bias to
people's cancer knowledge and the health information sources people have
available. The assumption here is that risky perception might be related to
people's knowledge about cancer; to health information sources (the number
of sources and the type of sources such as news media, internet, and
personal communication) people refer to; and to the ease of access to
health information.
Research questions
Based on previous studies, four research questions were
presented regarding to the relationship of Appalachian residents'
optimistic bias to their cancer knowledge and identified sources:
RQ1: How is perceived risk and optimistic bias about cancer related to
people's cancer knowledge?
RQ2: How is perceived risk and optimistic bias about cancer related to the
perceived ease of access to health information?
RQ3: How is people's perceived risk and optimistic bias about cancer
related to the number of information sources they identify?
RQ4: How is people's perceived risk and optimistic bias about cancer
related to the types of information channels they identify?
Method
The study, funded by the American Cancer Society's Ohio
Division, was conducted in July 2002. Telephone interviews were made to
random sample of residents living in 13 counties in Southeast Ohio,
including Adams, Gallia, Jackson, Lawrence, Meigs, Monroe, Morgan, Noble,
Pike, Ross, Scioto, Vinton, and Washington counties. The interviewees were
asked about their risk of cancer and the cancer risk for other people
living in the same area or same town. These two variables were measured by
a 10-point scale with 1 representing "no risk" and 10 representing
"extremely high risk." They were also asked about the ease of access to
information about health concerns (1 to 4 scoring: very difficult, somewhat
difficult, somewhat easy, very easy); about what they thought were major
causes of cancer (open-ended responses, recoded to give each respondent an
overall score for number of causes named); and about the sources they would
refer to for health information. Demographic information was also
collected, including gender, age, educational level, marital status, race,
household income in 2001 and county.
The sources variable was based on open-ended responses to
the question of "what sources would you use for information about health
issues," and with responses recoded to yield information on type and number
of sources. Ultimately, the types of sources were broadly categorized as
news media (such as newspapers, television, magazines, etc.); internet/web;
and personal communication (such as doctor, family and friends, agency
etc.). Thus, the primary source they would identify, the total number of
sources they named, whether the array of sources included digital devices
(such as computer, web) and whether they used news media could be
ascertained, all from the open-ended responses.
Findings
Table 1 provides descriptive information on the sample as
well as descriptive results for some key knowledge and source variables.
The respondent ages ranged from 18 to 99 (mean= 47). Few respondents had
not completed high school, 34.7% were high school graduates, 32.9% had some
college education, and 15.5% had completed college education. Thirty three
point four percent of respondents were male, and nearly 65% were married.
Nearly two thirds (60.2%) of respondents indicated that they
felt it was "very difficult" to access health information for the family
and for themselves and about 26% of said it was "somewhat difficult,"
meaning a total of 86% reported at lease some difficult!
Cancer knowledge was varied (mean number of causes named
=2.81): 5% of could not name any cancer cause, 13% could name one cause,
23% could name two causes, 25.5% could name 3 causes, and 23.5% could name
4 causes. In total, nearly three-fourths named from two to five causes.
The number of sources that respondents identified ranged from
zero to five (mean=1.62) when asked where they would get information for
health issues: 52.1% identified only one source, while 30.6% named two sources.
News media were seldom named by respondents for information
about health issues. Nine of ten respondents identified no news media.
Digital devices were used by 44.9% of respondents.
The primary or first source identified was personal
communication (35.2%), followed by Web or internet (32.5%), others (28.6%),
and media (3.7%).
Even though 86% had said it was difficult to get information,
45.5% of respondents identified multiple sources for health information.
The perception of other people's cancer risk averaged 6.54,
while the perception of self risk averaged 5.24. Optimistic bias (the
self-other difference) mean was 1.30 (6.54-5.24).
RQ1: How is perceived risk and optimistic bias about cancer related to
people's cancer knowledge?
People's cancer knowledge, which was calculated as the
number of cancer causes respondents named, is slightly but significantly
and positively related to optimistic bias (See table 2). Pearson's r, the
correlation coefficient between cancer knowledge and optimistic bias, is
0.09 (p=0.05). However, while optimistic bias is a function of the other's
risk perception and one's own risk perception, no significant correlation
coefficient existed between self risk and cancer knowledge. Instead, a
significant correlation was found between other's cancer risk score and
knowledge (r=0.13, p=0.003). The latter explains the slight but significant
relationship of optimistic bias to knowledge.
RQ2: How is perceived risk and optimistic bias about cancer related to the
perceived ease of access to health information?
The greater optimistic bias "gap" is among those who find
it "very difficult" to access health information. But overall, the ease of
access to health information respondents reported is not significantly
related to optimistic bias scores (See table 3). However, recall that 86%
of respondent indicated that information access was at least some difficult.
RQ3: How is people's perceived risk and optimistic bias about cancer
related to the number of information sources they identify?
The number of information sources respondents used is not
significantly related to optimistic bias (See table 4). The number of
identified sources is also not significantly related to the two optimistic
components: self risk perception and perception of other's risk.
RQ4: How is people's perceived risk and optimistic bias about cancer
related to the types of information channels they identify?
The primary source respondents identified (F=0.31, p=0.816), identification
of multiple sources (t=-0.34, p=0.736), identification of digital devices
(t=-0.75, p=0.454), and number of news media identified (r=0.01, p=0.788)
are not significantly related to optimistic bias (See table 5). However,
both others' risk perception and self risk perception are individually
significantly related to the primary source identification. Those who
identify news media as primary media have higher self risk perception
(6.28) and others' (7.11) risk ratings.
Discussion and Conclusion
This study found optimistic bias about cancer among residents
of Appalachia. The optimistic bias about cancer is significantly related to
people's cancer knowledge. The positive relationship between optimistic
bias and cancer knowledge indicates that people who could name more cancer
causes tend to believe that they have less cancer risk than others do.
Although optimistic bias is a function of both other's risk perception and
self risk perception, cancer knowledge is related to perception of other's
risk, but not to self.
This study found little of any relationship of risk to
number and types of sources identified for health information.
One of the goals of health campaigns is to identify causes
of problems and educate the public about them. Presumably, such knowledge
will lead to change. This study's finding that knowledge is related to
optimistic bias through an elevated risk attached to others, leaves many
questions unanswered. Optimistic bias, according to many, a function of
self-other distancing, is often defensive. It may be that knowledge of
multiple causes—many of which are not within one's own behaviors a
lifestyle—makes it for easier to assign high risk to others, who may engage
in such behaviors. Similarly, such knowledge enables one to rate his/her
risk as lower.
Ultimately, knowledge about cancer is related to a biased
optimism about cancer.
Table 1: Demographics, Ease of Access to Heath Information, Cancer
Knowledge, Source Information, and Cancer Risk Perceptions
Age 18 (Minimum) 99 (Maximum) 47 (Mean) n=502
1-20 4.8%
24
21-30 15.3%
77
31-40 17.7%
89
41-50 21.3%
107
51-60 19.3%
97
61-70 10.8%
54
71-80 7.2%
36
81 and
over 3.6% 18
Education n=498
- high school graduate 7.6% 38
High school graduate 34.7% 173
Some college 32.9% 164
College graduate 15.5% 77
Postgraduate 9.2% 46
Gender n=503
Female 66.6% 168
Male 33.4% 335
Marital status n=503
Married 65% 325
Unmarried 35% 178
Race n=503
White 96.2% 484
Others 3.8% 19
Income n=434
No more than $10,000 10.1% 44
$10,001-$25,000 23% 100
$25,001-$40,000 20.7% 90
$40,001-$60,000 24.2% 105
$60,001 and over 21.9% 95
Ease of access to health information n=502
Very difficult 60.2% 302
Somewhat difficult 26.1% 131
Somewhat easy 7.4% 37
Very easy 5.0% 25
Don't know 1.4% 7
(Continued)
Cancer knowledge: the cancer causes named (mean=2.81) n=490
0 5.1%
25
1 13.3%
65
2 22.9%
112
3 25.5%
125
4 23.5%
115
5 7.6%
37
6 1.8%
9
8 0.4%
2
Number of sources named (Mean=1.62) n=497
1 2.2% 11
2 52.1% 259
3 30.6% 152
4 12.3% 61
5 2.4% 12
6 0.4% 2
Number of types of news media identified (Mean=0.09) n=497
1 90.9% 452
2 8.7% 43
3 0.4% 2
Digital device sources identified n=497
Yes 44.9% 223
No 55.1% 274
Primary source identified n=486
Personal communication 35.2% 171
Web/internet 32.5% 158
Others 28.6% 139
Media 3.7% 18
Multiple source identified n=497
Yes 45.5% 226
No 54.5% 271
Mean
n
Other's cancer risk 6.54 502
Self cancer
risk 5.24 502
Optimistic
bias 1.30 502
Table 2: Optimistic Bias by Cancer Knowledge
Mean Sd. N
Cancer cause number 2.81 1.41 490
Optimistic bias 1.30 2.82 502
(r=0.09, p=0.05*)
Cancer cause number 2.81 1.41 490
Self
risk 5.24 2.40 502
(r=0.04, p=0.43)
Cancer cause number 2.81 1.41 490
Other's
risk 6.54 2.52 502
(r=0.13, p=0.003**)
*Significant level is 0.05.
** Significant level is 0.01.
Table 3: Optimistic Bias by the Ease of Access to Health Information
Ease of access to health information Optimistic bias (Mean) n=502
Very
difficult 1.51 302
Somewhat
difficult 0.95 131
Somewhat
easy 1.35 37
Very
easy 0.56 25
Don't
know 1.00 7
(F=1.37, p=0.244)
Table 4: Optimistic Bias by the Number of the Identified Sources
Mean Sd. N
Source number 1.62 0.85 497
Optimistic bias 1.30 2.82 502
(r=0.06, p=0.192)
Source number 1.62 0.85 497
Self risk 5.24 2.40 502
(r=-0.025, p=0.576)
Source number 1.62 0.85 497
Other's risk 6.54 2.52 502
(r=0.042, p=0.351)
Table 5: Optimistic Bias by Types of the Identified Sources
Optimistic bias (Mean) Other's
risk Self risk n
Primary source
News
media 0.83 7.11 6.28
18
Internet/web 1.25
6.49 5.25 158
Personal
communication 1.34 6.20 4.87 171
Others 1.45
7.01 5.56 139
(F=0.31, p=0.816) (F=3.00,
p=0.030*) (F=3.47, p=0.016*)
Multiple sources
More than one
source 1.34 6.59 5.25 226
One source or
less 1.25 6.52 5.27 271
(t=-0.34,
p=0.736) (t=-0.303, p=0.762) (t=0.08, p=0.937)
Digital device source
Use digital
source 1.40 6.53 5.13 223
Non-digital
source 1.21 6.58 5.37 274
(t=-0.75,
p=0.454) (t=0.208, p=0.835) (t=1.108, p=0.269)
Number of news media
Mean
Sd. N
Optimistic
bias 1.30 2.82 502
Number of news
media 0.09 0.31 497
(r=0.01, p=0.788)
Other's
risk 6.54 2.52 502
Number of news
media 0.09 0.31 497
(r=0.078, p=0.083)
Self
risk 5.24 2.40 502
Number of news
media 0.09 0.31 497
(r=0.068, p=0.130)
*Significant level is 0.05.
References
Cancer death rates---Appalachia, 1994-1998 (2002). MMWR: Morbidity & Mortality
Weekly Report, 51 (24), 527-529.
Chapin, J. R. (2000). Third-Person perception and optimistic bias among urban
minority at-risk youth. Communication Research, 27 (1),
51-81.
Coleman, C. (1993). The influence of mass media and interpersonal
communication
on societal and personal risk judgments. Communication
Research, 20 (4),
611-628.
Fontaine, K. R. & Smith, S. (1995). Optimistic bias in cancer risk
perception: a cross-
national study. Psychological Reports, 77, 143-146.
Salwen, M. B. & Dupagne, M. (2003). News of Y2K and experiencing Y2K:
exploring the relationship between the third-person effect
and optimistic bias.
Media Psychology, 5, 57-82.
Riffe, D. (2003). Both sides of the digital divide in Appalachia: uses and
perceived
benefits of Internet access. Paper presented to the
Association for Education
in Journalism and Mass Communication, Kansas City,
MO.
Riffe, D. (2003). Environmental hazards, cancer risk judgments, and media
use in
Appalachia. Paper presented to the Association for Education
in Journalism and Mass Communication, Kansas City,
MO.
Riffe, D. & Knight, J. (2002). Environmental threats, information sources and
optimistic bias: environmental risk in Appalachia. Paper
presented to the
Association for Education in Journalism and Mass
Communication, Miami,
FL.
Rosswurm, M. A. et al. (1996). Illness experiences and health recovery
behaviors of
patients in Southern Appalachia. Western Journal of Nursing
Research, 18
(4), 441-459.
Weinstein, N. D. (1980). Unrealistic optimistic about future life events.
Journal of
Personality and Social Psychology, 39 (5), 806-820.
Weinstein, N. D. (1987). Unrealistic optimism about susceptibility to
health problems:
conclusions from a community-wide sample. Journal of
Behavioral
Medicine, 10 (5), 481-500.
Weinstein, N. D. & Klein, W. M. (1996). Unrealistic optimistic: present and
future.
Journal of Social and Clinical Psychology, 15 (1), 1-8.
|
