Content-Type: text/html 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.