Journal of Personality and Social Psychology 1986, Vol. 50, No. 3, 502-510 Copyright 1986 by the American Psychological Association, Inc. 0022-35l4/86/$00.75 Self-Other Judgments and Perceived Vulnerability to Victimization Linda S. Perloff and Barbara K. Fetzer University of Illinois at Chicago People who have not been victimized by negative life events tend to perceive themselves as less vulnerable than others (i.e., as "uniquely invulnerable") to victimization. We examined the conditions under which people's judgments of others' vulnerability differ from judgments of their own vulnerability. In Study 1, subjects saw vague comparison targets (either the "average person" or the "average college student") as more vulnerable than themselves to 10 negative events. In contrast, subjects perceived a specific target (their closest friend, sibling, or same-sex parent) as equally invulnerable as themselves. In Study 2, subjects who were instructed to consider a vague, abstract target (either the "average college student" or "one of your friends") made downward comparisons, choosing a real or hypothetical other who was especially vulnerable to a particular event. On the other hand, subjects who were instructed to consider a specific, concrete target (their closest friend) perceived no self-other differences in risk status. We concluded that when given the opportunity, subjects actively engage in downward comparisons, thereby seeing themselves as relatively invulnerable. Whereas vague targets facilitate downward comparisons, specific targets make such comparisons more difficult. Both cognitive and motivational mechanisms underlying such downward comparisons are discussed. Nonvictims, individuals who have not been victimized by un- desirable life events, often underestimate their own personal vul- nerability to victimization in relation to other people's vulner- ability. In other words, nonvictims appear to have an "illusion of unique invulnerability," in which they see themselves as less vulnerable to victimization than they seemost other people (Per- loff, 1983). This biased perception is reflected in the common saying, "It won't happen to me," a statement that generallyim- plies that it will happen to others instead. Although many studies have demonstrated the existence of these illusions of unique in- vulnerability, we still know relatively little about the mediating processes underlying these perceptions. In our two studies, we explore the social comparison processes that nonvictimsuse when estimating their own and others' risk. Specifically, we examine the conditions under which people's judgments of others' vul- nerability differ from judgments of their own vulnerability. Numerous studies have documented the existence of illusions of unique invulnerability among nonvictims.For example, people have judged themselves as less likely than others to be victims of diseases such as cancer, heart attack, pneumonia, alcoholism, and venereal disease (Harris & Guten, 1979 Kirscht, Haefner, Kegeles, & Rosenstock, 1966 Perloff, 1982 Weinstein, 1980, 1982, 1984) Los Angeles residents perceived themselves as healthier than the average person and as having fewer illnesses per year (Larwood, 1978) and college students estimated that they would live 10 years longerthan the actuarial average provided We are grateful to Fred Bryant, Reid Hastie, David McKirnan, and three anonymous reviewers for insightful comments on earlier versions of this manuscript. The authors are grateful to Fred Bryant, Janice Schreckengost, and Sue Yellen for rating the open-ended responses in Experiment 2. Correspondence concerning this article should be addressed to Linda S. Perloff, Department of Psychology, University of Illinois at Chicago, Box 4348, Chicago, Illinois 60680. by "insurance company longevity" data (Snyder, 1978). Lastly, college students saw themselves as less likely than the typical person to get divorced and as more likely to work hard at making their marriage a success (Perloff & Farbisz, 1985). Thus there is considerable evidence that nonvictims' appraisals of personal risk are often systematicallybiased. One potential danger of such biased appraisals is that they may discourage people from taking adequate precautions to avoid victimization (Perloff, 1983 Weinstein, 1980). Theorists have offered several possible underlying mechanisms to explain these biased perceptions of risk. For example, it has been proposed that perceptions of unique invulnerabilitymay stem from a lack of information about others (Weinstein, 1983) or from a need for personal control (Janoff-Bulman & Lang- Gunn, in press Perloff, 1983). (For a more detailed discussion of underlying mechanisms, see Perloff, 1983.) The focus in this article is on a related factor, namely, nonvictims' choice of social comparison others. We explore how people go about judging their own versusothers' vulnerabilityto future negativelife events. The Role of Social ComparisonProcesses Whether we view ourselves as more or less vulnerable than others to future victimization may depend, in part, on who those others are. To the extent that we make "downward comparisons" (Wills, 1981), comparing ourselves with others who are less for- tunate and more at risk than we are, we may be apt to see our- selves as uniquely invulnerable. Both motivational and cognitive mechanisms may underlie these downward comparisons. From a motivational perspective, downward comparisons with vul- nerable others may serve an ego-defensive function by reducing nonvictims' anxiety and enhancing feelings of personal control (cf. Perloff, 1983 Wills, 1981). As Taylor, Wood, and Lichtman (1983) noted, "downward comparisons have the psychological advantage of making one feel good about one's situation relative to the comparison other" (p. 27). 502

PERCEIVED VULNERABILITY 503 In contrast to this motivational framework, Weinstein (1980) suggested a cognitive explanation for howdownward social com- parisons facilitate perceptions of unique invulnerability. Ac- cording to Weinstein (1980), people harbor illusions of invul- nerability in part because they compare themselves to an inap- propriate standard, "an unrealistic stereotype of a person who does nothing to improve his or her chances or even engages in counterproductive activity" (p. 819). Weinstein suggested that such inappropriate comparisons stem from people's reliance on a "representativeness" heuristic (Kahneman & Tversky, 1973). Use of the representativeness heuristic involves assigning an in- dividual to a particular category on the basis of whether his or her attributes resemble the characteristic features (or the modal member) of this category. According to Weinstein (1980), For many events���contracting lung cancer and becoming an alco- holic, for example���people may have a stereotyped conception of the kind of person to whom this event happens. If they 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 expe- rience the event may actually fit the stereotype, (p. 808) In support of this reasoning, college students who had a clear picture in their minds of a particular type of person to whom a negative event was likely to happen saw their chances of expe- riencing the event as less than average (Weinstein, 1980, Study 1). However, no researchers have yet directly explored whether nonvictims actually choose comparison others who represent their stereotype of a person at risk. We attempt to address this issue. One unaddressed question concerns the specific conditions under which people engage in downward social comparisons when estimating their vulnerability to negative events. In most previous studies of perceived risk, subjects have been asked to compare their chances of encountering a given negative event with the chances of the "average" or "typical" person (e.g., Harris &Guten, 1979 Lang, 1980 Larwood, 1978 Perloff, 1982 Rob- ertson, 1977 Svenson, 1981 Weinstein, 1980, 1983, 1984). To the extent that people tend to view themselves as better than average (Myers & Ridl, 1979) and as more intelligent than their average peer (Wylie, 1979), the "average person" may be seen as someone who is, almost by definition, less advantaged, less in- telligent, and generally worse off than oneself. Consequently, it is perhaps not surprising that subjects judge themselves as more immune to negative events than the average person is. Although in most previous research on perceived vulnerability the investigators have used the average person as the target for social comparison, other comparison targets maywell inhibit the commonly reported illusion of unique invulnerability. More specifically, although people tend to see themselves as relatively invulnerable to undesirable life events when they compare them- selves with the average person, it is not clear whether people also see themselves as relatively invulnerable when they compare themselves with their friends or relatives. There are two reasons to expect subjects to see themselves and their friends or relatives as equally vulnerable, although seeing themselves as less vul- nerable than the average person. First, the "average person"may be a sufficiently vague and ambiguous target that it leads people to visualize an abstract stereotype or prototype (Cantor & Mis- chel, 1977) of a person at risk. Having this prototype in mind may make it easier to construe the average person as the kind of person who is vulnerable to victimization. In contrast, when the target is a specific individual, it may be more difficult either to manufacture a hypothetical other who is highly vulnerable or to distort the risk-relevant characteristics of the specific target. Sec- ond, we might expect people to be motivated to resist acknowl- edging the vulnerability of their friends or family for the same reasons that they resist acknowledging their own personal risk: to reduce anxiety and fear. Indeed, other researchers have found that self-serving biases extend to include perceptions of one's close friends and loved ones (e.g., Burger, 1981 Schlenker & Miller, 1977). Study 1 was designed to enable us to examine the conditions under which nonvictims rate others as more vulnerable than themselves. Specifically, do people see themselves as uniquely invulnerable, regardless of the specific comparison other? Or is the ubiquitous perception of unique invulnerability a bias that emerges only when subjects make judgments about relatively impersonal, vague comparison targets, such as the average person or the average college student? On the basis of the preceding discussion, we hypothesized that subjects' ratings of their own vulnerability would be lower than their ratings of the average person's vulnerability (i.e., subjects would show perceptions of unique invulnerability). In contrast, wehypothesized that subjects would rate themselves and a close friend or family member as equally invulnerable. Study 1 Method Subjects. Subjects in Study 1 were 101 introductory psychology stu- dents (48 male, 53 female) at the University of Illinois at Chicago, who participated in order to fulfill a course requirement. Materials. Subjects were asked to estimate both their own vulnerability and another person's vulnerability to 10 negative life events: cancer, heart attack, hypertension, drinking problem, venereal disease, diabetes, injury in a car accident, nervous breakdown, mugging, and divorce.1 Weselected these 10 events because past research had shown that they evoke percep- tions of unique invulnerability in college students (Lang, 1980 Perloff, 1982 Weinstein, 1980). We assessed estimates of personal vulnerability by asking subjects how likely they thought it was that they would expe- rience the particular event sometime duringtheir life. Subjects also rated the vulnerability of one of several different comparison targets. Weassessed estimates of the comparison target's vulnerability by asking subjects how likely they thought it was that the comparison other would experience the event sometime during his or her life. Responses were made on 7- point Likert scales, ranging from 1 (not at all likely) to 7 (extremely likely). Procedure. Subjects filled out a questionnaire in same-sex groups of 8. The experiment was described as a study of the beliefs people have about themselves and others. Within each experimental session, subjects were randomly assigned to one of five conditions that differed only by comparison target. In all five conditions, subjects estimated how likely they themselves were to experience each of the 10 events. In addition, subjects were asked to estimate the likelihood that one of five different 1 Six subjects indicated that they had been divorced, had been injured in a car accident, or had been mugged, and their data were therefore discarded from the sample.