Special Session Summary Revise and Resubmit: Biases in Perceptions of Health Risk

Suresh Ramanathan, New York University
Geeta Menon, New York University
[ to cite ]:
Suresh Ramanathan and Geeta Menon (2002) ,"Special Session Summary Revise and Resubmit: Biases in Perceptions of Health Risk", in NA - Advances in Consumer Research Volume 29, eds. Susan M. Broniarczyk and Kent Nakamoto, Valdosta, GA : Association for Consumer Research, Pages: 182-184.

Advances in Consumer Research Volume 29, 2002     Pages 182-184

SPECIAL SESSION SUMMARY

REVISE AND RESUBMIT: BIASES IN PERCEPTIONS OF HEALTH RISK

Suresh Ramanathan, New York University

Geeta Menon, New York University

OVERVIEW OF SESSION

There has been considerable research in both the social psychology and the consumer behavior literature on the topic of motivated reasoning and perceived risk (e.g., Kunda 1990, Perloff and Fetzer 1986, Luce and Kahn 1999, Raghubir and Menon 1998, Block and Williams 2001). Studies in this domain have documented the phenomenon of "unrealistic optimism" (Weinstein 1980, 1984) whereby people tend to believe that they are more special than the average person is and thus suffer from an illusion of imperviousness. More recent work has shown that such a bias leads to lower intentions to get tested or alter behavior (Luce and Kahn 1999, Block and Williams 2001), and therefore, can result in very self-destructive behavior. This bias can however be decreased by facilitating easier retrieval of risk behaviors from memory (Raghubir and Menon 1998) and by framing messages negatively in terms of what one could lose by lack of compliance with the recommended behavior (Block and Keller 1995).

The broad purpose of this proposed session was to present work that adds to this growing body of research, and to delineate conditions under which perceived health risks can be affected in such a way as to ensure more positive attitudes and intentions towards testing. In doing so, we sought to identify several new dimensions that could potentially facilitate or impede compliance. Specifically we aimed to show that the content of health-related messages and the way they are phrased could have a powerful impact on attitdes and behavior (Kahn, Luce and Miller’s paper; Menon, Block and Ramanathan’s paper). We also showed that there are significant individual differences in the way people make judgments of risk, based on their innate traits (Keller, Lipkus and Rimer’s paper). Together, these studies demonstrated the importance of understanding how best to communicate health-risk issues to people who may have different response mechanisms in such a way as to affect their intentions to get tested and compliance behavior positively.

Specifically, the focus of the first paper by Barbara Kahn, Mary Frances Luce and Elizabeth Gelfland Miller was on how different types of test result feedback could result in differences in intentions to get re-tested, and on the moderating role of the information conveyed in the feedback. The research, conducted among people actually getting tested for breast cancer, showed how ambiguity in false positive results could result in decreased compliance for testing. The second paper by Geeta Menon, Lauren Block and Suresh Ramanathan looked at how cues in a health message influence health risk estimates and subsequent intentions and behavior via a process of confirmatory hypothesis testing. Across three studies, cues that highlighted a larger number of frequent risk-behaviors were found to work best in breaking through the wall of imperviousness, thus ensuring greater intentions to get tested. In contrast, enumerating a large number of infrequent behaviors had a backfire effect, resulting in even weaker estimates of self-risk than for a control group with no information at all. In the third paper, Punam Anand Keller, Isaac Lipkus and Barbara Rimer focused on individual differences among people in their response to feedback on risk, showing that people with a tendency towards depressive realism were more likely to be accurate in their estimates of risk. Across two studies, they showed that depressives were more likely to come up with better estimates even if they were required to be optimistic about their risks. They demonstrated that this effect was due to their having generally lower confidence in their beliefs and being more likely to make internal attributions for their failures.

 

"TESTING INTENTIONS FOLLOWING FALSE POSITIVE MAMMOGRAPHY RESULTS: THE MODERATING EFFECT OF INFORMATION"

Barbara E. Kahn, University of Pennsylvania

Mary Frances Luce, University of Pennsylvania

Elizabeth Gelfland Miller, University of Pennsylvania

The substantial amount of media attention devoted to cancer and other screening tests is likely to encourage initial screening. However, there is reason to believe that regular adherence to testing guidelines is relatively rare, even after consumers have been persuaded to initially adopt testing. For instance, in one study, 37% of women surveyed who had one prior mammogram decided not to get a second one (Jepson and Rimer 1993). We argue that the testing experience itself is likely to influence this regular testing over time.

In the context of mammography, we address women’s intentions to continue to get routine, regular screening, even after an initial or early testing experience indicates that the disease is not present. We consider the effects of prior results on testing intentions and we investigate whether these effects are moderated by information provided in the context of initial test results. We test our hypotheses in two between-subject experiments conducted on women in their hospital gowns waiting to get their mammogram in a University hospital waiting room.. Although we focus on the mammography context, we believe our results are likely to generalize to any stressful testing situation, where screening must be done on a regular basis without the prior presence of symptoms as a motivator.

It is important to study initial medical test results as screening tests are typically calibrated to err on the side of false positive results. In our first experiment, we address this issue by experimentally manipulating three test result conditions: (1) control, or the normal est result, (2) a false alarm result where patients are told that their tests results are abnormal but are later told that the malady (breast cancer) is not indicated at this time, and (3) a false alarm + condition, where patients are told that their test results are abnormal, and are later told that the malady (breast cancer) is not indicated at this time, but that they are now at an increased risk of getting the malady (breast cancer) in the future. Our findings show that either type of false alarm result is stressful and can reduce confidence in the screening test. While both stress and lack of confidence could be associated with a decrease in testing, we find that the false alarm + result is not associated with a decrease in planned future compliance. However, the false alarm result is associated with decreased planned compliance. We further find that reaction to this relatively more ambiguous false alarm condition is moderated by information provided to subjects.

We consider three information conditions: (i) BSE condition (BSE): in which an explanation of the efficacy of breast-self exams is provided, highlighting the role the patient can play in the screening process; (ii) Mammogram as Non-Definitive condition (MND): in which a statement is provided explaining how a positive mammography result does not necessarily mean the patient has breast cancer, highlighting the fact that mammography is only a first step in the screening process; (iii) Control condition: in which the explanation of the mammogram process is repeated and no further information is provided. We find that both BSE and MND information mitigate the detrimental effects of a false positive result on planned compliance, but have no effect within the comparatively unambiguous conditions of a clear negative or a false alarm + result. Thus, we conclude that information provided at the time of receiving results can increase mammography intentions for the patients with false alarm experiences who are not at increased risk of the disease. We further investigate these processes in our second experiment, where the relevant information is found to moderate the effect of a prior false positive result in a patient’s actual medical history on planned compliance. We attribute these effects to the fact that BSE information functions to suggest the woman has control and therefore she can engage in problem-focused coping, while MND information lessens the stress that is typically associated with a false positive result.

 

"WE ARE AT AS MUCH RISK AS WE’RE LED TO BELIEVE: EFFECTS OF MESSAGE CUES ON JUDGMENTS OF HEALTH RISK"

Geeta Menon, New York University

Lauren Block, Baruch College, CUNY

Suresh Ramanathan, New York University

One of the greatest challenges in advertising health-related information is overcoming the target audiences’ self-positivity bias (i.e., the tendency for people to believe that they are invulnerable to disease). Understanding how self-risk estimates are affected by various specifications of risk factors is crucial since recent theories of consumer health behavior stress that an individual’s perceived risk is a necessary precursor to behavioral change (Luce and Kahn 1999). Moreover, a growing body of literature suggests that people have a tendency to assume that they are special (Weinstein 1980, 1984, 1987; Perloff and Fetzer 1986), and hence impervious to the issues being advertised (Raghubir and Menon 1998). Therefore, overcoming this "self-positivity bias" is the first step in increasing behavioral compliance.

In the current paper, we examine the effects of message cues on reducing the self-positivity bias, subsequently improving attitudes and increasing behavioral intentions and actual compliance. We chose hepatitis C because of the prevalence of the disease and because our student population engages in many activities that can cause it such as body piercing, sharing razors and toothbrushes, and getting tattoos. In Study 1 we manipulated the type of message cue by making the risk of contracting hepatiis C seem easier or more difficult. Results show that the self-positivity bias is lower when participants are presented with risk-behaviors that are frequently engaged. These effects, in turn, translate into higher concern levels and intentions to get tested for the disease. We replicated and extended this finding in Study 2, in which we examined the interactive effects of the type of and the number of message cues used in an ad, such that self-risk estimates are reduced when more risk-behaviors that make the contraction of hepatitis C seem easy are enumerated. However, this effect is reversed with risk-behaviors that make the contraction of hepatitis C seem difficult. Finally, in Study 3, we delineate the underlying process by which these effects manifest. Results of this study provide evidence for a confirmatory hypothesis testing process by which people process risk behavior information in a health message. The operation of the self-positivity bias acts as an a priori hypothesis. Incoming information is interpreted as either confirming (infrequent risk behaviors) or disconfirming evidence (frequent risk behaviors). Confirming risk-factor information, or infrequent risk-behaviors, result in less effortful processing, less discounting of the evidence, and more confidence in one’s judgments than disconfirming or frequent risk-behaviors does.

The major contributions of this paper are two-fold: (a) From a theoretical standpoint, we identify the process by which message cues affect judgments of risk, reports of intentions to behave in a certain way in the future, and the depth of information processing. We demonstrate that the type of message cues (Study 1) and the number of message cues (Study 2) interact to affect the self-positivity bias, and consequently people’s attitudes, intentions and behavior. Most importantly, we delineate the process by which these effects manifest (Study 3). We show that people hold an a priori hypothesis that they are invulnerable to diseases, and that providing message cues that shake this belief manifest in greater cognitive effort in processing the information and lower confidence with which people hold their beliefs. In doing so, we elaborate on the theoretical understanding of how memory and context factors work in opposing ways and suggest a more comprehensive framework. (b) From a practical standpoint, we provide tangible methods that can be used by health-care marketers to grab the attention of the audience. Since we demonstrate that the effectiveness of the message depends on the interactive effect of the number of risk-factors and the specific risk-factors chosen, we underscore the importance of the actual risk-behaviors used in the ad. The use of the "wrong" risk-behaviors may do more damage than good.

 

"DEPRESSIVE REALISM AND HEALTH RISK ACCURACY: SADDER, WISER...AND SCARED"

Punam Anand Keller, Dartmouth College

Isaac Lipkus, Duke University Medical Center

Barbara Rimer, National Cancer Institute

A common goal of PSAs (public service announcements) has been to communicate risk information for a particular outcome. Recent research indicates that the relationship between perceived risk and behavior may be moderated by individual differences (e.g., Weinstein and Nocolich 1993; Boney-McCoy, Gibbons and Gerard 1999). This recognition has led to the identification of the role of individual differences, such as personality traits, on people’s reactions to risk communication (Jessor, Turbin and Costa 1998).

Although personality traits have long been implicated in the formation of risk perceptions, we have not seen a recent study on personality in the marketing literature (Bettman 1971; Cox 1967; Jacoby 1961; Popielarz 1967; Schaninger 1976). The link between PSAs and personality is based on the premise that acknowledging health risks may involve threat not only to one’s physical welfare, but also to one’s self regard. Therefore, with some form of health threat, particularly those associated with risk-increasing behaviors, reactions may be in part directed to blstering self-regard by reducing perceived inconsistency. Consistent with this view, Schaninger (1976) found that self-esteem was negatively correlated with perceived risk. More recent research in social psychology indicates that people with high self-esteem feel less vulnerable and show greater resistance to health threats than do people with low self-esteem (e.g., Boney-McCoy, Gibbons and Gerard 1999).

The present study explores the association between another personality trait, depression, and perceived risk. In particular, we were interested in the phenomenon known as depressive realism (Alloy and Abramson 1988). Support for the depressive realism effect is obtained when depressives (the label we use for people who are sub-clinically or mildly depressed) outperform people who are not depressed. In a risk context, the depressive realism effect would be manifested in more accurate risk estimates among depressives as compared to non-depressives who typically believe they are at lower risk (Perloff and Fetzer 1986) and are more likely to predict socially desirable rather than realistic futures (Sherman 1980).

Two studies provide support for the depressive realism view that depressives outperform non-depressives. Depressives and non-depressives women were required to reduce their prior breast cancer risk estimate to be more accurate. Participants provided their risk of getting breast cancer before (baseline) and after (follow-up) receiving personalized (Experiment 1) or standard (Experiment 2) medical risk feedback. Although there were no significant differences in risk estimates at baseline, the follow-up risk estimates indicate that depressives adjusted their personal risk such that they were more accurate or closer to the medical estimates provided in the risk feedback than were the estimates provided by non-depressive women. Given the pessimistic personal risk estimate of breast cancer prior to the risk feedback (25% versus the medical risk of approximately 3%), an increase in accuracy required participants to be more optimistic.

These findings support the depressive realism view that the depressives are more accurate risk estimators even when they are required to lower risk estimates. We also provide evidence for why depressive realism occurs. As compared to non-depressives, depressives have lower confidence in their prior beliefs, make internal attributions for failure to comply and have higher fear of being evaluated negatively.

REFERENCES

Alloy, L. B. and L. Y. Abramson (1979), "Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology: General, 108, 441-485.

Bettman, Jim (1971), "The structure of consumer choice processes," Journal of Marketing Research, 8, 465-471.

Block, Lauren and Punam Anand Keller (1995), "When to Accentuate the Negative: The Effects of Perceived Efficacy and Message Framing on Intentions to Perform a Health Related Behavior," Journal of Marketing Research, 32(May), 192-203.

Block, Lauren and Patti Williams (2001), "Undoing the Effects of Seizing and Freezing: Decreased Defensive Processing of Personally Relevant Messages," Journal of Applied Social Psychology, forthcoming.

Boney-McCoy, Sue, Gibbons, Frederick, X. and Meg Gerrard (1999), "Self-esteem, compensatory self enhancement, and the consideration of health risk," Personality and Social Psychology Bulletin, 25, 954-965.

Cox, Donald. F. (1967) "SynthesisBPerceived Risk and Information Handling," in D.F.Cox (ed.) Risk Taking and Information Handling in Consumer Behavior, Boston, Harvard University Press, 82-108.

Jacoby, Jack and L Kaplan (1971), "Personality and Innovation Proneness," Journal of Marketing Research, 8, 244-247.

Jepson, Christopher and Barbara K. Rimer (1993), "Determinants of Mammography Intentions Among Prior Screenees and Nonscreenees," Journal of Applied Social Psychology, 23(1), 40-51

Jessor, Richard, Turbin, Mark, S. and Frances M. Costa (1998), "Protective Factors in Adolescent Health Behavior," Journal of Personality and Social Psychology,75, 788-800.

Kunda, Ziva (1990), "The Case for Motivated Reasoning," Psychological Bulletin, 108, 480-498.

Luce, Mary Frances and Barbara E. Kahn (1999), "Avoidance or Vigilance? The Psychology of False Positive Test Results," Journal of Consumer Research, 26 (December), 242-259

Perloff, Linda S. and Barbara K. Fetzer (1986), "Self-other judgments and perceived vulnerability of victimization," Journal of Personality and Social Psychology, 50, 502-510.

Popielarz, B. T. (1967), "An explanation of perceived risk and willingness to try a new product," Journal of Marketing Research, 4, 368-373.

Raghubir, Priya, and Geeta Menon (1998), "AIDS and Me, Never the Twain Shall Meet: The Effects of Information Accessibility on Judgments of Risk and Advertising Effectiveness, Journal of Consumer Research, 25 (June), 52-63.

Schaninger, Charles M. (1976), "Perceived risk and personality," Journal of Consumer Research, 3 (September), 95-100.

Sherman, Steven J. (1980), "On the self-erasing nature of errors of prediction," Journal of Personality and Social Psychology, 39, 211-221.

Weinstein, Neil D.(1980), "Unrealistic Optimism about Future Life Events," Journal of Personality and Social Psychology, 39 (October), 806-820.

Weinstein, Neil D. (1984), "Why it Won’t Happen To Me: Perceptions of Risk Factors and Illness Susceptibility," Health Psychology, 3, 431-457.

Weinstein, Neil D. (1987), "Unrealistic Optimism About Susceptibility to Health Problems: Conclusions from a Community-Wide Sample," Journal of Behavioral Medicine, 10 (5), 481-500.

Weinstein, Neil, D. and M. M. Nicolich (1993), "Correct and Incorrect interpretations of correlations between risk perceptions and risk behaviors," Health Psychology, 12 235-245.

----------------------------------------