Perceptions of Risk and Vulnerability
Susan G. Millstein and Bonnie L. Halpern-Felsher
Why Are Perceptions of Risk and Vulnerability Important?
Individuals’ judgments about risk are viewed as a fundamental element of most theoretical models of health and risk behavior, including Social Cognitive Theory (Bandura, 1994), the Health Belief Model (Rosenstock, 1974), the Theory of Reasoned Action (Fishbein and Ajzen, 1975), the Theory of Planned Behavior (Ajzen, 1985), Self-Regulation Theory (Kanfer, 1970), and Subjective Culture and Interpersonal Relations Theory (Triandis, 1977). All of these theories posit that individuals’ beliefs about the consequences of their actions and perceptions of their vulnerability to those consequences play a key role in behavior. Although we will later question whether existing studies address these hypotheses adequately, the strength of the logical association between risk perceptions and behavior is compelling. As a result, risk perceptions play a fundamental role in behavioral intervention programs, which try to get adolescents to recognize and acknowledge their own vulnerability to negative outcomes. The ability to judge risks also is considered to be an essential element of decision-making competence, according to theorists, researchers, and practitioners in the behavioral sciences, medicine, social work, law, and social policy (Gittler et al., 1990; Hodne, 1995).
Adults have speculated about adolescents’ lack of competence in recog-
nizing and assessing risk since the time of Aristotle. Adolescents typically are viewed as being unable to judge risk appropriately, and as having strong beliefs in their invulnerability to harm. In recent years, the question of adolescents’ competence has emerged as a result of efforts to regulate the legal rights of adolescents to make decisions in the realms of medical and mental health treatment, including their rights to refuse treatment or to obtain treatment without parental knowledge and/or consent, as well as their rights to participate in research, including experimental clinical trials. Additionally, adolescents’ capacity to exercise existing rights is of fundamental interest to the juvenile justice system (Butterfield, 1996).
Much of the interest in adolescents’ perceptions of risk and vulnerability is motivated by the desire to understand why youth engage in potentially threatening behaviors, with an aim toward guiding the development of interventions that will be successful in preventing their onset. Relevant questions for gaining such understanding include the following:
What skills are needed for assessing risk?
Do adolescents have these skills?
How competent are adolescents in identifying and assessing risk?
What kinds of factors influence adolescents’ ability to judge risk?
How do adolescents’ perceptions compare to those of adults?
Do adolescents’ perceptions of risk influence their decisions?
In this paper, we review existing data to address these questions. We acknowledge that answers to these questions will not give a complete picture of why adolescents engage in risky behavior—other crucial questions remain, such as whether adolescents are competent decision makers or able to apply their decision-making skills in all situations. Nevertheless, a focus on risk perception is a reasonable vantage point from which to consider adolescent risk and vulnerability. We will begin our discussion by giving the reader a sense of the size of the risks themselves. That is, how big are the risks that adolescents face? This will provide a context for later assessing the adequacy of adolescents’ judgments concerning those risks.
How Big Are the Risks?
Some of the threats to adolescents’ well-being pose sizable risks. For example, 40 percent of Latino youth fail to complete high school or the equivalent, such as the General Education Development Tests (GED) (Fed-
eral Interagency Forum on Child and Family Statistics, 2000). However, for many risks, the actual chance of a negative outcome occurring is relatively small. For example, Fischhoff et al. (2000) reported that for adolescents, the actual probability of a female getting pregnant within one year is less than 6 percent, and the probability of a male getting someone pregnant in the next year is less than 3 percent. The probability of being the victim of a violent crime (e.g., homicide, rape, robbery) is less than 10 percent. Even smaller is the probability of an adolescent dying from any cause in the next year (.08 percent) or by the time an adolescent turns 20 (.04 percent). Of course, the probability of experiencing these outcomes is highly dependent on one’s behavior and environment. Risks for acquiring sexually transmitted diseases (STDs) are quite high among youth who have highly connected sexual networks involving people who live in areas of high infectivity. The risks are far lower for adolescents who live in geographic areas of low disease rates, and are essentially nil for sexually inactive youth.
But small risks are not unimportant. Although many of the actual risks adolescents face are numerically small, their potential outcomes can be severe and life altering. For example, STDs such as gonorrhea and chlamydia are associated with subsequent rates of pelvic inflammatory disease (PID) as high as 10 to 40 percent. It is estimated that of the approximately 200,000 to 400,000 adolescent females who develop PID each year, 40,000 to 84,000 of them eventually will find themselves infertile. The fiscal costs are also high; excluding costs related to HIV and AIDS, we spend an estimated $882 million yearly treating STDs in adolescents (Gans et al., 1995). Furthermore, most causes of adolescent morbidity and mortality are preventable, thus behooving us to make attempts to reduce them.
Adults’ Perspectives on Adolescent Vulnerability
Although few empirical data speak directly to adults’ perceptions concerning risks to adolescents, other indicators point to the sources of adults’ concerns. Reading the popular press, listening to parents of teenagers, and examining the content of preventive programs makes it clear that adults’ concerns focus primarily on the major causes of morbidity and mortality and adolescents’ involvement in behaviors that are associated with these negative health and social outcomes. In a series of focus groups with high-risk youth, their parents and grandparents, a report from the Office of Disease Prevention and Health Promotion (1993) found that parents were concerned about the lack of adult supervision for their children and the
failure of schools and communities to meet the needs of adolescents. Other areas of concern included peer influences and poor schools. Ferguson and Williams (1996) found that among a group of parents of 17 year olds, 38 percent expressed concerns about their children’s driving; 43 to 97 percent of the parents supported additional restrictions to protect their children.
Parents of adolescents also view adolescents as more vulnerable than do adolescents themselves. Beyth-Marom et al. (1993) found that across a series of risky behaviors, adolescents’ perceptions of the risks to themselves were significantly lower than the risks their parents perceived for them. Similar findings are reported by Cohn et al. (1995). For example, although 31 percent of the adolescents believed there was little or no harm in getting drunk once or twice, only 9 percent of the parents believed this. These formal comparisons of adults’ perceptions of adolescents’ risks with adolescent perceptions of their own risks is confounded by the fact that the adult is judging risks for another person—and such judgments are typically higher than personal risk judgments (Weinstein, 1980, 1983, 1984; Whalen et al., 1994).
Conceptualizing and Measuring Perceptions of Risk and Vulnerability
There are many ways to conceptualize and measure perceptions of risk and vulnerability. We can examine the content of individuals’ risk and vulnerability beliefs—identifying those things that worry or concern them, as well as the degree of anxiety generated by these concerns. We can observe whether people recognize the risks inherent in a given situation, or we can look at how accurately someone judges a specific risk. Risk judgments may focus on situations (e.g., is having unprotected sex dangerous?) or on their potential outcomes (e.g., what is the chance that you will get an STD?). Personal risk can be viewed in absolute terms (e.g., what is your chance . . . ?) or relative terms (e.g., how does your risk compare to others?). For any given individual, we can also examine his or her relative ranking of the importance of various “risks” to assess his or her risk perceptions.
Each of the many ways of assessing individuals’ perceptions has something to tell us about their sense of risk and vulnerability. But they also appear to measure different aspects of this construct we call vulnerability. The literature reflects this conceptual diversity, making it difficult to compare across studies. In our review, we will consider them separately. We will not attempt to tell the reader which of these approaches are best. Rather, we
will comment on the use and limitations of each. To avoid confusion, we will use the following terms throughout this paper.
We use the term risk judgment to reflect magnitude assessments of risk. A risk judgment that focuses solely on an outcome (e.g., how likely are you to get an STD?) would be considered to be a nonconditional risk judgment. In contrast, a situation-specific or conditional risk judgment is one in which explicit mention of an antecedent condition such as a situation or a behavior is made (e.g., how likely are you to get an STD if you have unprotected sex?). When assessments of risk do not involve magnitude estimates, we use the term risk identification. These assessments may focus solely on a situation (e.g., is it risky to have unprotected sex?) or may include identification of specific consequences as well (e.g., what might happen if you have unprotected sex?).
Although risk identification and judgment may be the most direct ways to tap assessments of risk, individuals’ affective responses to specific situations and/or outcomes also can be informative. When assessments focus on the degree of anxiety or concern individuals have about particular situations (e.g., how worried would you be if you had unprotected sex?) or outcomes (e.g., do you worry about getting an STD?), we refer to them as feelings or perceptions of vulnerability. Like the more cognitive aspects of risk perception, these affective manifestations can be conditional or unconditional. Asking people to identify the things that concern them (e.g., what kinds of health problems do you worry about?) points to content areas where perceptions of vulnerability may exist; we refer to these simply as concerns. We acknowledge that individual differences such as generalized anxiety or pessimism may influence the degree to which situations or outcomes are identified, judged, or experienced as risky. However, we do not consider these generalized states or their measurement to reflect perceptions of risk or vulnerability.
We will now review what we know about the content of adolescents’ concerns, their perceptions of vulnerability, their ability to identify risk, and their ability to judge risk. In each section, we will first give the reader a sense of how adolescents as a group perform, followed by an examination of how these capacities vary by the age of the adolescent. Other sources of variation in risk perception and assessment, such as gender, race/ethnicity, and economic status, are discussed later.
Sources of Concern
One approach to understanding adolescents’ perceptions of vulnerability is to identify those issues about which adolescents express concern. Numerous surveys have documented these concerns. Common adolescent concerns include those related to appearance (height, weight, acne), emotional states (depression, anxiety), interpersonal relationships (parents, friends, and other adults), school (schoolwork, school problems, and career), environmental threats (air pollution and nuclear war), and health and physical complaints (headaches, stomachaches, vision problems, dental problems). Adolescents also acknowledge the importance of the health issues most frequently identified by health professionals, including substance use, sexual behavior, birth control, sexually transmitted disease, and pregnancy (Alexander, 1989; Benedict et al., 1981; Brunswick, 1969; Giblin and Poland, 1985; Marks et al., 1983; Millstein and Irwin, 1985; Parcel et al., 1977; Sternlieb and Munan, 1972).
Although adolescents clearly acknowledge a wide range of concerns, only a few issues consistently rank high. These include issues pertaining to school, dental health, acne, interpersonal relationships, and mental health (American School Health Association et al., 1989; Sternlieb and Munan, 1972). Concerns related to substance use, sexual behavior, nutrition, and exercise rank lower in most adolescent samples (Eme et al., 1979; Feldman et al., 1986; Sobal et al., 1988).
Age Differences in Adolescents’ Concerns
Observed age differences in health concerns are consistent with the developmental tasks faced by adolescents. Younger adolescents (between 11 and 13) generally are more concerned with physical development, including puberty (Byler et al., 1969), and with how one is viewed by members of the opposite sex. Middle adolescents (i.e., about 14–15 years old) are more concerned with appearance (especially among female adolescents), interpersonal relationships with peers and members of the opposite sex, and self-esteem. Older adolescents are more concerned with school, grades, and their future career plans (Eme et al., 1979; Violato and Holden, 1988) as well as their emotional health. A study of more than 5,000 children and adolescents found increasing interest among youth in topics such as growth and development, preventive health behaviors, mental health, and social-emotional development as children moved into the early adolescent years (Byler et al., 1969).
Adults’ Perceptions About Adolescents’ Concerns
Adults underestimate the degree of concern adolescents report about their health (Sobal et al., 1988) and appear to have misperceptions about adolescents’ level of knowledge about specific topics such as AIDS (Manning and Balson, 1989). Adults also fail to recognize some of the specific health concerns of adolescents. An example occurs in regard to dental and oral health. Across studies, adolescents consistently rank dental concerns of being of great importance (Parcel et al., 1977; Sobal et al., 1988; Sternlieb and Munan, 1972). Yet this topic rarely receives attention in discussions of adolescent health. Other areas in which adults fail to recognize adolescents’ concerns include school problems and teens’ relationships with adults (Sobal et al., 1988).
To summarize, existing data indicate that adolescents do express concerns about negative effects that can result from volitional behaviors as well as from environmental hazards such as natural disasters, technological risks, and violence. Adults often underestimate these concerns. Although examination of adolescents’ concerns gives us some indication of the sources of concern, it does not inform us about the degree to which these concerns are accompanied by feelings of vulnerability. For example, to what degree do concerns about oral health translate into adolescents’ feelings of personal vulnerability to caries or gum disease? For this reason, perceptions of vulnerability may be a closer reflection of adolescents’ beliefs about their vulnerability.
Perceptions of Vulnerability
Data from numerous studies indicate that adolescents feel vulnerable to experiencing negative outcomes. Adolescents’ worry and feelings of vulnerability to AIDS have been well documented (DiClemente et al., 1987; Pleck et al., 1990; Price et al., 1985; Strunin, 1991). Other behavior-related risks, such as getting sick from drinking alcohol and acquiring an STD also appear to generate feelings of vulnerability. In a recent study, we found that a majority of sexually inexperienced seventh and ninth graders reported that they would be worried and concerned about getting an STD if they had sex without a condom; 53 percent said they would be very worried and 26 percent reported feeling somewhat worried. Similar percentages were obtained for adolescents’ ratings of concern. Feelings of vulnerability appear to generalize beyond behavior-related risks. When asked to imagine being at a picnic when a tornado strikes, 36 percent of adoles-
cents in seventh and ninth grade reported that they would be very worried that they would die, and 22 percent reported being somewhat worried about dying from the tornado. The corresponding values for being concerned with dying were 43 percent and 20 percent. Bachman (1983) showed that 30 percent of high school seniors reported worrying frequently about the threat of nuclear war.
Age Differences Among Adolescents
In our recent study (Millstein and Halpern-Felsher, 2001) cross-sectional analyses showed a negative relationship between age and feelings of vulnerability to alcohol (r=–.30) and sex risks (–.35). Feelings of vulnerability to alcohol-caused illness were significantly higher in fifth and seventh graders than ninth graders and adults. Perceptions of vulnerability to STDs were significantly higher in seventh and ninth graders than adults. Longitudinal analyses (in progress) showed that among sexually inexperienced seventh and ninth graders, worry over getting an STD as a result of unprotected sex decreased significantly over a 6-month period. However, reported concern for STDs did not. Additional data supporting these findings are provided by data from the National Survey of Adolescent Males (Pleck et al., 1993), which showed that adolescent males’ (ages 15–19) worry about AIDS decreased over a two-year period.
Thus, in contrast to popular views that adolescents do not worry or concern themselves about risks, the data indicate that many, and in some cases most, adolescents report feeling vulnerable to negative outcomes. It is not entirely clear, however, what these assessments really mean. Do they reflect true anxiety and/or perceptions of potential harm, or are they primarily cognitive expressions, meant to acknowledge that these are things, in general, to worry about? Who, after all, does not worry (at least academically) about negative outcomes such as AIDS?
Studies of adolescents’ capacity for considering consequences have shown that even young adolescents have the ability to identify negative consequences associated with medical procedures (Kaser-Boyd et al., 1985; Lewis, 1981; Weithorn and Campbell, 1982) and with engaging in risky behaviors (Beyth-Marom et al., 1993; Finn and Brown, 1981; Furby et al., 1997). Typically, these studies ask participants to imagine themselves in a
hypothetical situation, and to identify any risks that they perceive in that situation. For example, Beyth-Marom et al. (1993) asked adolescents to consider what might happen if they engaged in six different risky activities. They found that adolescents identified, on average, four to seven consequences for each activity. A broad range of consequences were mentioned, including physical effects, psychological effects, and social reactions from family, other authority figures, and peers.
A few studies have examined the extent to which adolescents spontaneously mention or consider risks, which may give a more realistic picture of adolescents’ ability to recognize risk in real-life situations. Lewis (1981) asked 108 adolescents in grades 7, 8, 10, and 12 about the advice they would give to peers facing cosmetic surgery and to peers considering participation in a clinical trial of a new acne medicine. Sixty percent of the adolescents mentioned possible risks associated with the situations, and 26 percent mentioned potential future consequences. Halpern-Felsher and Cauffman (2001) used the same scenarios in their study of 190 adolescents (grades 6, 8, 10, and 12). They found that 12 to 32 percent of teens mentioned risks and 10 to 13 percent mentioned future consequences in the cosmetic surgery scenario. In the informed consent scenario, 42 to 63 percent of the adolescents recognized risks and 7 to 16 percent recognized future consequences.
Age Differences in Risk Identification
Studies examining age differences in adolescents’ ability to identify risks report conflicting findings. Lewis (1981) found dramatic increases in awareness of risks between grades 10 (50 percent of subjects mentioning) and 12 (83 percent). Mention of future consequences showed steady increases over grades 7–8 (11 percent), 10 (25 percent), and 12 (42 percent). However, a replication of the Lewis study using a somewhat larger sample (Halpern-Felsher and Cauffman, 2001) reported no significant age differences in adolescents’ consideration of risks and long-term consequences. Two other studies also failed to find age differences. In a small sample (N=62) of people with learning and behavior problems, Kaser-Boyd et al. (1985) found younger adolescents to be as competent as older adolescents in their ability to consider consequences. Ambuel (1992) studied 13 to 21 year olds (N=75) who suspected an unplanned pregnancy and were seeking a pregnancy test. Comparisons between younger minors (ages 13–15 years) and
older minors (16-17) showed no differences in their consideration of consequences.
In four studies focusing on risk identification that compared adolescents and adults, two found adults to be more competent in identifying risk. Beyth-Marom et al. (1993) reported small age-related differences in subjects’ ability to generate consequences associated with six risky behaviors, with adults spontaneously mentioning more consequences than adolescents on one-third of the behaviors examined. Halpern-Felsher and Cauffman (2001) found that in comparison to adolescents, adults considered a greater number of risks as well as long-term consequences. Both of these studies used hypothetical situations to estimate participants’ capacity for identifying risk. In contrast, the other two studies focusing on risk identification examined perceptions of risk in real life decision-making situations and failed to find differences between adolescents and adults in their ability to consider consequences (Ambuel, 1992; Kaser-Boyd et al., 1985). Both samples were relatively small, including fewer than 80 participants.
To summarize, it is clear that many adolescents, especially older adolescents, are capable of recognizing and identifying risks. Adolescents also are able to identify risks spontaneously, in response to hypothetical scenarios as well as in real-life decision situations. However, the overall level of competence among adolescents is not exceptionally high. In the Lewis (1981) study, few of the younger adolescents and less than half of the tenth graders could be considered competent in identifying important potential risks. Halpern-Felsher and Cauffman (2001) also reported lower than expected levels of awareness of risks among both adolescents and adults. Furthermore, most studies that have examined age differences in risk identification report age-related increases in individuals’ awareness and consideration of risks (Ambuel, 1992; Beyth-Marom et al., 1993; Lewis, 1981); fewer have failed to find such differences (Lewis, 1980; Weithorn and Campbell, 1982). Although it is not clear whether there are absolute points at which one should be considered competent in identifying risks, the finding of age-related increases in risk identification does call into question the degree to which we should consider adolescents, particularly younger adolescents, competent. The extent to which the ability to identify risks changes over time has not been studied using longitudinal data, nor has the link between risk identification and risk behavior been studied.
In addressing whether adolescents are able to judge risks accurately, one can use a number of different indicators. One indicator might be to look at whether adolescents view themselves as “invulnerable,” a common attribution that would render them unable to judge risks adequately. If we interpret invulnerability as meaning that the individual judges risk as nonexistent (i.e., risk estimates of zero-percent chance), research does not support this characterization of adolescents. Both Quadrel et al. (1993) and Millstein and Halpern-Felsher (2001) found perceptions of invulnerability to characterize a minority of the adolescents they sampled.
Ideally, we would judge adolescents’ competence in assessing risk by comparing their perceptions to their actual risk status. But determining an individual’s risk for experiencing a specific negative outcome is difficult because it depends on so many individual and environmental factors. Two people drinking the same amount of alcohol and getting into their car to drive can have very different probabilities of having an accident, depending on their body weight, food consumption, level of tolerance for alcohol, weather conditions, and so on. Similarly, the risk of pregnancy depends on factors, such as the individual’s age, history of sexually transmitted disease, stage in the menstrual cycle, and type and extent of contraceptive use.
Because of the complexity in ascertaining risk status, estimates of individuals’ risks often are based on aggregated risk. Thus, for example, we might judge an adolescent’s risk of acquiring an STD as a function of the incidence of STDs in sexually active adolescents. Using this approach, Fischhoff et al. (2000) compared adolescents’ risk judgments with data for estimating actual risk. Their sample included more than 3,500 15 to 16 year old adolescents from the 1997 National Longitudinal Study of Youth. The adolescents were asked to judge the probability that they would experience specific outcomes (i.e., risk judgments were nonconditional). The comparisons indicated that, as a group, adolescents’ estimates were fairly accurate for some events (such as being in school one year later or becoming pregnant over the next year). For example, adolescents estimated their chance of experiencing or causing pregnancy within the year as 6.3 percent, which is very close to the actual probability of less than 6 percent. They were slightly optimistic about their chances of obtaining a high school diploma and getting a 4-year college degree. They were pessimistic about their estimates concerning the probability of serving time in jail or prison,
judging the likelihood at 5 percent, which is nearly 10 times higher than the actual probability of 0.6 percent. They greatly overestimated their chances of dying in the next year or by their twentieth birthday, judging the probability as 18.6 percent, while in actuality it is less than 1 percent. Of course, as averages, these estimates do not inform us about the proportion of adolescents who are accurate. Examination of the distribution of percentage estimates indicates that although adolescents as a group appeared to have reasonable estimates in many areas, few individuals in the sample actually demonstrated an accurate sense of risk.
Even in the absence of comparisons such as these, there are indications that adolescents are quite inaccurate in estimating risk. When we examine adolescents’ quantitative estimates and compare them with even approximate probability estimates, we find that adolescents overestimate risk. For example, Halpern-Felsher et al. (2001) reported that nonsexually active adolescents and young adults estimated their chance of getting an STD if they had unprotected sex once as 44 percent, and the chance of contracting HIV/AIDS given the same situation as 38 percent. Participants who had never used alcohol estimated a 79-percent chance of getting into an accident if they drove with a drunk driver. It is possible that adolescents’ high risk judgments can be explained by their inability to understand and use quantitative percentages. However, analyses controlling for adolescents’ skill in understanding percentages continue to yield high estimates (Millstein and Halpern-Felsher, 2001).
Methods for estimating adolescents’ competence also have included the use of comparative risk assessments. These assessments ask people to estimate whether their chance of experiencing an outcome is higher, the same, or lower than other people like themselves. The rationale is that, in a given population, some people’s risk will be higher than others, some will be lower than others, and some will be the same as others. Not everyone can be at lower risk. Given this, the mean comparative risk assessment at the aggregate level should be normally distributed. However, studies generally find that, as a group, adolescents (as well as adults) bias their assessments in the direction of viewing their risk as lower than the risk for similar others (Whalen et al., 1994).
Age Differences in Risk Judgments
Earlier, we noted that risk judgments could be nonconditional or conditional. When we move to review age differences in judgments of personal
risk, the distinction between these types of risk assessments becomes particularly important. To illustrate the fundamental difference between these kinds of assessments, let us consider risk judgments concerning lung cancer. A nonconditional risk judgment asks the individual to judge his or her risk of developing lung cancer, without specifying any potential situational factors that might be relevant. If we ask a nonsmoker to judge his or her risk of developing lung cancer, the response would reflect his or her assessment of the likelihood that a nonsmoker will develop lung cancer. But if we ask the same question of a smoker, his or her assessment would reflect a judgment about the likelihood that a smoker will develop lung cancer. Nonconditional assessments thus pose a problem of interpretation because they assess different things as a function of the respondent’s behavioral experience.
An alternative method for eliciting risk judgments is to have individuals judge their risk under specific situations or conditions. Asking individuals to judge their risk of developing lung cancer if they smoked cigarettes represents what we call a conditional risk judgment (see Halpern-Felsher et al., 2001). Regardless of one’s own smoking status, individuals are responding to the same question and are asked to make the same assessment. These risk judgments are more useful as their meaning is less dependent on factors such as the behavioral characteristics of the respondent (Ronis, 1992; Van der Velde et al., 1996).
The use of nonconditional risk judgments is especially troubling in studies examining age differences in risk perception because experience is a known source of variation in risk judgment (Gerrard et al., 1996b; Van der Plight, 1998) that varies by age as well. For this reason, we will limit our review and commentary to studies that either elicit conditional risk judgments or control for behavioral experience if using nonconditional risk judgments. Thus, we will not comment on some frequently cited studies such as those reported by Gochman and Saucier (1982).
Three investigations have studied age differences in adolescents using conditional assessments of risk. Of them, only Cohn et al. (1995) failed to find age differences among adolescents. Urberg and Robbins (1984) found that perceptions of smoking-related risks had a curvilinear relationship to age in a sample of adolescents in grades 6 through 12. A strong inverse relationship was characteristic of adolescents in grades 6 through 8; and a smaller, positive relationship was found among adolescents in grades 8 through 12. Results from our recent work (Millstein and Halpern-Felsher, 2001), which examined risk judgments across a range of domains, suggest a
more linear, negative relationship with age, with older adolescents’ judgments of risk significantly lower than those of younger adolescents.
Although these cross-sectional studies suggest the possibility of age differences in risk judgment, only longitudinal studies that examine changes within subjects over time can tell us whether these differences are truly developmentally based or due to cohort differences. Only our study offers such longitudinal data using conditional risk judgments. Analyses are still in progress. However, preliminary analyses of adolescents’ risk judgment for STDs indicate that among sexually inexperienced ninth graders, there is a significant decrease over a one-year period in their perceived risk of personally getting an STD if they have unprotected sex (from a mean risk judgment of 43 percent to a mean of 36 percent). We saw no significant changes in the seventh graders’ risk judgments over the one-year period.
A number of differences among these studies are worth noting. Although all of the studies specified particular situational conditions, their specificity differed. Urberg and Robbins (1984) queried subjects about risks “if you smoked regularly.” Cohn et al. (1995) asked about risks associated with engaging in a behavior once or twice. We used highly detailed scenarios to minimize variability in how participants would interpret underspecified risk situations (Fischhoff, 1996) and to assure that risk judgment differences were not a function of such interpretive differences (Biehl and Halpern-Felsher, 2001; Ellen et al., 1998; Fischoff, 1996). For example, in the scenario designed to elicit judgments about risks for STD, the risk situation (having unprotected sex) specified both the type of sexual partner as well as the number of unprotected episodes of sex.
Comparisons of Adolescents and Adults
A reasonable indicator for judging adolescents’ competence in assessing risks is to compare their performance with that of adults because adults generally are considered competent in the eyes of the law. It can also shed light on potential developmental differences in the ability to judge risks. A small number of studies provide such adolescent-adult comparisons. Cohn et al. (1995) found adolescents to rate their risk of harm as lower than did their parents. But two other studies support the idea that adolescents perceive greater risk than adults. Quadrel et al. (1993) had adolescents and their parents assess the probability of experiencing a variety of behavior-linked negative outcomes. Defining absolute invulnerability as the belief that one faces no risk (i.e., zero-percent probability) of experiencing par-
ticular outcomes, they found that adolescents were less likely to judge themselves as invulnerable than were their parents. Similarly, Millstein and Halpern-Felsher (2001) found that a significantly greater proportion of adults demonstrated perceptions of absolute invulnerability (34 percent) than did adolescents (14 percent). Additionally, we found that adolescents’ risk judgments were significantly higher than those of adults’ judgments across a range of natural hazards and behavior-linked outcomes. The magnitude and direction of the findings remained consistent across different types of risk judgment measures. The differences also remained significant after controlling for experience with the behavior and the negative outcomes.
A number of additional studies provide information about differences in risk perceptions between adolescents and adults. Perceptions of risks to other children (not personal risk assessments) were examined by McClure-Martinez and Cohn (1996), who found perceptions of risk of childhood injury to be higher among adolescent mothers than older mothers. Sastre et al. (1999) found adolescents (ages 15–20) to be more accurate than adults in perceiving a linear dose-risk relationship for smoking cigarettes. Goldberg et al. (2001a) found adolescents to make more heuristically based errors than adults in their judgments of risk. Quadrel et al. (1993) had adolescents and their parents assess the probability of experiencing a variety of behavior-linked negative outcomes. They found that adolescents were less likely to judge themselves as absolutely invulnerable than were their parents. Similarly, Millstein and Halpern-Felsher (2001) found that a significantly greater proportion of adults demonstrated perceptions of absolute invulnerability (34 percent) than did adolescents (14 percent).
Overall then, research to date shows that many, and in some cases most, adolescents report feeling some degree of vulnerability to negative outcomes and few evidence perceptions that they are invulnerable to harm. In fact, most studies show perceptions of decreased risk with age (Bernstein and Woodall, 1987; Millstein and Halpern-Felsher, 2001; Quadrel et al., 1993); judgments of risk are greatest in younger adolescents, and greater in adolescents than in adults. Although we do not yet have longitudinal data to inform us about whether these differences represent actual developmental phenomena, we believe they may well be, given what we know about other aspects of development. In the section of this paper entitled “Risk Judgment: A Developmental-Ecological Perspective,” we discuss how cognitive and psychosocial development, coupled with social experiences, could suggest that this is the case. However, in interpreting these findings, it is
essential to consider what judgments of risk actually represent. If we construe them to represent perceptions of vulnerability, research points to a heightened sense of vulnerability in adolescents compared to adults. If we view risk judgments as literal expressions of risk status, a different picture emerges—one of adolescents as far less accurate than adults.
Demographic Correlates of Risk and Vulnerability Perceptions
Several individual-level demographic factors have been hypothesized to influence perceptions of risk and vulnerability, including gender, race/ ethnicity, and socioeconomic status. With the exception of gender, few studies have been conducted; these are reviewed in the following paragraphs.
There are fairly consistent gender differences in adolescents’ health concerns, perceptions of vulnerability, and perceptions of risk. Across studies representing a diversity of samples, females consistently report thinking more about their health and having more health concerns than do males (Alexander, 1989; Brunswick and Josephson, 1972; Feldman et al., 1986; Parcel et al., 1977; Porteous, 1979; Radius et al., 1980; Sobal et al., 1988; University of Minnesota, 1989; Violato and Holden, 1988).
Females also judge risks as being more likely than do males. Adolescent females (17–20) perceived greater risks and fewer benefits associated with drug use, alcohol use, and sexual behavior than did adolescent males (Parsons et al., 1997). Across a number of well-specified driving situations, including situations involving the use of alcohol, Mundt et al. (1992) found that older adolescent females (18–20) rated the probability of getting into a serious driving accident as more likely than did the males. These differences appear to persist into adulthood. In a study of parents of 17 year olds, mothers were more concerned about driving-related safety issues for their children than were fathers (Ferguson and Williams, 1996).
Gender differences also emerge in cross-cultural studies. Kassinove and Sukhodolsky (1995) compared American and Russian adolescents (ages 10 to 18) concerning the degree to which they worry about 13 outcomes, such as developing cancer, being a victim of a violent crime, poor grades, and dying, as well 6 items concerning global worry, such as world hunger, overpopulation of the planet, and environmental pollution. In both the Russian and American samples, adolescent females worried more about both
personal and global issues than did their male counterparts. Similar gender differences have been found in Swedish studies (Drottz-Sjoberg and Sjoberg, 1991).
Few studies have examined differences in health concerns and perceptions of risk as a function of race or ethnicity, and even fewer have disentangled the effects of social class or economic status in their analyses. The limited research available generally shows that in comparison with white adolescents, black adolescents think more about their health (American Cancer Society, 1979; Sobal et al., 1988), have more health concerns (Sobal et al., 1988), are more concerned about future illness (American Cancer Society, 1979), and believe they are more susceptible to specific health outcomes, such as cancer (Price et al., 1988). There are some exceptions, however. Ey et al. (2000) found that Caucasian adolescents (11–19 year olds) perceived themselves at more risk for stroke, cancer, heart attack, and motor vehicle accident than did their African American counterparts.
Differences also emerge in the nature of the specific concerns. Black adolescents have been found to be more concerned about substance use than white adolescents (Alexander, 1989). Concerns about mental health may show a different pattern, with white adolescents reporting mental health concerns more often than Hispanic or black adolescents (Parcel et al., 1977).
Actual differences in health status among white, black, and Hispanic youth could explain some of these perceptions, particularly about threats to health. But this is unlikely to explain all of the differences. For example, Strunin (1991) found Asian adolescents to be more worried about getting AIDS than were Caucasian adolescents, despite their far lower rates of sexual activity and infection. Futhermore, because health status is tied so closely to economic status and minorities are generally less economically advantaged, it is not clear whether these racial/ethnic differences would persist or new ones would emerge in studies controlling for economic status.
Surprisingly little is known about how economic conditions influence adolescents’ perceptions about the world (see Bloom-Feshbach et al., 1982,
for a review). Two studies reported no differences in perceptions of risk and vulnerability to health problems as a function of socioeconomic status (Gochman and Saucier, 1982; Michielutte and Diseker, 1982). Yet we would expect the broad environmental context of adolescents’ lives to influence their general perceptions of the world, including their perceptions of risk. In their comparison of American and Russian adolescents, Kassinove and Sukhodolsky (1995) found that the Russian adolescents were significantly more worried than the American students on both the personal and global worry scales. Similarly, the concerns that predominate in adolescents’ lives have been shown to vary over generations as a function of social and economic factors that members of different cohorts experience (Natapoff and Essoka, 1989; Porteous, 1979).
THE RELATIONSHIP OF BELIEFS AND BEHAVIOR
Do Perceptions of Risk and Vulnerability Influence Behavior?
Perceptions of risk are viewed as playing a central role in motivating adolescents’ behavior. To validate this hypothesis, one would want to study people before they began to engage in risk behavior. We would examine their beliefs about the risks they would face if they engaged in the behavior (i.e., conditional risk assessments), and then would follow them longitudinally to see whether they eventually engage in the specific behavior. Theoretically, those people who perceived less risk and/or felt less vulnerable to negative outcomes would be more likely to end up engaging in the behavior than those who initially perceived higher risk.
Given the broad implications of a causal link between risk judgment and risky behavior, one is struck by the absence of such studies in the literature. Instead, a typical research paradigm has been to look at differences in risk judgment between people who engage in risky behavior (“engagers”) and those who do not (“nonengagers”). For example, studies have examined differences between smokers’ and nonsmokers’ judgments of risk for getting lung cancer or have compared personal risk estimates of contracting HIV among individuals who do and do not engage in unsafe sex. These studies found that individuals with behavioral experience rate their risk of experiencing negative outcomes as higher than do nonengagers (Cohn et al., 1995; Gerrard et al., 1996b; Gladis et al., 1992; McKenna et al., 1993; Moore and Rosenthal, 1991, 1992). Given the use of nonconditional assessments (e.g., “What is your risk of developing lung cancer?”)
and a cross-sectional design, it is no surprise to find that smokers perceived themselves as being at higher risk for developing lung cancer than do nonsmokers. But when we ask people about their personal risk, were they to engage in a specific behavior (e.g., “What is your risk of developing lung cancer if you smoke?”), we find lower judgments of risk among people who are engaging in risky behaviors (Benthin et al., 1993; Halpern-Felsher et al., 2001; Urberg and Robbins, 1984). In other words, engagers perceive the risks of engaging in the behavior as lower.
Together, this set of studies suggests that people who are engaging in risky behavior recognize that these behaviors entail risk, but view the risks as less significant than do people who do not engage in the risks. The cross-sectional nature of these studies means, of course, that we do not know whether those who engage in risky behavior perceived lower risk prior to the onset of the behaviors. Thus, they cannot answer the question of whether risk judgments influenced the behavior, as is hypothesized by models of health behavior, or whether the risk judgments being assessed in these studies reflect behavioral experiences—a plausible alternative hypothesis that we will entertain later (Finn and Brown, 1981; Gerrard et al., 1996a, 1996b; Halpern-Felsher et al., 2001; Urberg and Robbins, 1984). Of course, if the relationship between perceptions of risk and involvement in risky behavior is a reciprocal one (which we believe is the case), both hypotheses could be true, with judgments of risk influencing behavior as well as experience playing a role in how people view risk. Showing these kinds of effects would require fairly lengthy longitudinal studies, as one would have to recruit people prior to the onset of risky behaviors, and follow them long enough for negative outcomes1 to have occurred to them or at least experienced vicariously. Ideally, these studies would begin early enough to investigate the role that cognitive development plays (i.e., in very early adolescence, around the age of 10) and be followed into early adulthood, with frequent assessments of adolescents’ behaviors, experiences with negative outcomes, and risk judgments. If such studies were con-
ducted, we would expect to find a reciprocal relationship between risk judgments and experience.
It is important to recognize that risk perceptions are but one of many factors influencing behavior. Intervention programs invest a great deal of energy attempting to influence adolescents’ perceptions of risk. The assumption is that given the knowledge that specific behaviors entail risk, adolescents will avoid those behaviors. Yet the relatively high overestimates of risk we see in adolescents should alert us to the suspicion that this explanation may be an overly simplistic one. A more reasonable way to view perceptions of risk is that they are necessary for motivating protective behavior, but they are not sufficient. It is unlikely that someone purposely will avoid a potentially pleasurable activity if he or she perceives absolutely no risk in doing so. But simply perceiving risk may not be sufficient. Indeed, although behavioral and decision-making models propose a key role for risk perception, they also articulate other critical influences on behavior, such as perceptions of benefits.
The Role of Perceived Benefits
The theoretical importance of benefits is recognized in the decision-making literature, which posits that individuals consider both risks and benefits (e.g., a cost-benefit model) when making decisions (compare to Baron, 1988; Weinstein and Fineberg, 1980). Benefits also are highlighted in theories of health-related behaviors, and have been shown to be an important predictor of drinking behavior (Christiansen et al., 1989; Smith et al., 1995).
Goldberg et al. (2001b) found that with increasing age and experience, adolescent respondents perceived the benefits of alcohol to be more likely and the risks to be less likely. The perceived benefits of alcohol became a more important predictor than risks in adolescents’ intentions to drink alcohol across age groups and levels of experience. Among respondents who drank, the vast majority reported experiencing consequences that were positive. Furthermore, as an indication of the robust nature of this finding, this pattern of results was replicated with respondents’ self-generated responses about alcohol and in the context of another health-threatening behavior, cigarette smoking. These results are in contrast to the messages adolescents typically receive about risk behavior, which are mainly about the negative, and often fatal, results. The authors argue that the failure to experience
even minor negative outcomes, combined with the unexpected experience of positive outcomes, may have caused the benefits to loom larger in adolescents’ decision making.
Goldberg et al. (2001b) also argue that rather than interpreting the relatively smaller risk estimates of older respondents as biased perceptions of “invulnerability,” a better explanation is that they are adjusting their perceptions on the basis of both their positive experiences and the failure to experience negative outcomes. They present data to support this “adjustment” interpretation. Examining open-ended responses concerning the bad and good things that could happen from drinking alcohol, they found that from fifth graders through the adults, there were progressively fewer respondents who said there was “nothing” good about drinking alcohol (28 percent of fifth graders, 16 percent of seventh graders, 9 percent of ninth graders, and 2 percent of adults). If one includes the “missing” responses, which may well indicate that the respondents did not think there was any good that could result from drinking, the adjustment is even more dramatic: 38 percent of fifth graders, 24 percent of seventh graders, 12 percent of ninth graders, and 3 percent of adults.
Other studies have found similar results. Urberg and Robbins (1981) found significant differences in both perceived costs and perceived benefits between 12 to 15 year old nonsmokers who intended to smoke and those with no smoking intentions; adolescents with intentions to smoke perceived more benefits and fewer costs than their counterparts. Covington and Omelich (1992) examined perceived risks and benefits to smoking between regular smokers and nonsmoking sixth, eighth, and tenth graders and also found that regular smokers perceived fewer costs and more benefits than did nonsmokers. Furthermore, they reported that the perception of risks decreased with age.
We can see from this discussion concerning benefits that it is important to understand the kinds of outcomes valued by adolescents. Health concerns may not rank as high as social concerns. Understanding the interplay between perceived risks and benefits highlights the importance of recognizing that people value different things. All other things being equal, individuals are more likely to take risks if they stand to gain highly valued outcomes (benefits). Similarly, they are more likely to avoid risks when they stand to lose highly valued outcomes. There are, of course, individual differences in determining what it is that people value.
Summary of the Relationship Between Beliefs and Behavior
Studies assessing whether risk judgments are sensitive to behavioral experiences have yielded different results depending on the type of risk assessment measure used. The majority of studies have used nonconditional measures and find that individuals’ estimates concerning the likelihood that they will experience a particular negative outcome (e.g., STDs) is higher among those who engage in risky behaviors linked to the outcome (i.e., unprotected sex) than among those who do not engage in the risky behavior. The few studies that have used conditional risk assessments find that adolescents who have engaged in a risk behavior perceive less risk than do nonengagers. The degree to which these associations are causal or the direction of the associations cannot be determined due to the cross-sectional designs used.
Furthermore, few studies have empirically examined other factors that might be playing a role in adolescents’ decisions to engage in risk behavior, such as perceived benefits or the value, importance, or severity of different negative and positive outcomes.
RISK JUDGMENT: A DEVELOPMENTAL-ECOLOGICAL PERSPECTIVE
We have described how judgments of risk can vary among individuals, as well as how behavioral experiences and the social environment can influence these perceptions. But the information we have presented is static; it does not integrate what we know about the enormous changes that take place during the adolescent years as a function of cognitive development, psychosocial development, and changes in the nature of the social environment. Only by bringing in a developmental-ecological perspective can we integrate what is currently known to help us understand how risk perception may change over time, the effects of experience at different periods, and the role that behavioral experiences may play.
Judging risks involves an assessment of the degree to which a given antecedent is causally linked to a particular consequence or outcome. Understanding and assessing correlational evidence and making judgments about causal relationships requires the ability to process a large array of data (e.g., processing information about all possible combinations of antecedents and outcomes), as well as metacognitive skills needed to integrate new information relevant to one’s own theories about causal relationships. Young
adolescents have a limited ability to coordinate information, attend to smaller portions of available data (Byrnes et al., 1999; Inhelder and Piaget, 1958; Shaklee and Goldston, 1989), and tend to think in fewer dimensions than older adolescents (Piaget, 1971). Additionally, although younger adolescents are able to use theories about causal relationships, they generally are not equipped to reflect on those theories. As a result, younger adolescents are less able to consider the possibility that a contingency is false, or to consider that alternative causal relations are possible (Kuhn et al., 1988). This would suggest that younger adolescents would be more likely than older adolescents or adults to believe what they have been taught about causal relationships between risky behaviors and negative outcomes—namely, that engaging in these behaviors entails significant risk.
As adolescents mature, they become better able to entertain the possibility that a particular contingency is false. Through observation, they also learn about the relationship of risk behaviors and negative outcomes. Even if an adolescent has not engaged in risk behavior, the exposure to risky behavior increases dramatically over time as a result of the number of peers who engage in these behaviors (Centers for Disease Control and Prevention, 1998). Because most experiences with risk behaviors do not lead to negative outcomes, few examples of these outcomes are likely to have been observed. Adolescents who have just begun to think about theories of causality have great difficulty integrating such information and recognizing that there can be exceptions to the rule (Kuhn et al., 1988). Psychosocial development may also provide adolescents with further impetus for questioning what they have been taught. They have greater needs for autonomy, which often translate into their desires for autonomous decision making (Connell and Halpern-Felsher, 1997; Midgley and Feldlaufer, 1987; Steinberg and Silverberg, 1986). Challenging adults’ teachings about the risks of particular behaviors could be one way in which such needs are expressed.
The development of thinking skills and changes in how information is processed appears to continue throughout the adolescent years and into young adulthood. However, not all of these developmental changes point to increased rationality. Jacobs and Potenza (1991), Davidson (1995), and Reyna and Ellis (1994) show that classical decision-making biases, such as the use of the representativeness heuristic, increase between childhood and adulthood. The ways in which people deal with evidence of noncovariation appears to improve with subsequent development (Kuhn et al., 1988), but remains suboptimal and shows increasing evidence of motivational bias.
People process outcomes and evidence in ways that reflect their underlying theories (Klaczynski and Narasimham, 1998; Kuhn, 1992). In particular, they uncritically accept evidence in favor of their views while spending a considerable amount of time finding flaws in evidence that is contrary to their views. This suggests that even when people confront evidence that risk behaviors can lead to negative outcomes (e.g., through increases in their vicarious exposure to negative outcomes), they would maintain their theories of low perceived risk rather than to react with judgments of increased risk.
In the absence of personal experience with risk behaviors, cognitive and psychosocial development, along with changes in the social environment, could thus explain why younger adolescents perceive risks as being so high, and why judgments of risk appear lower among older adolescents and young adults. If the age group differences we see are indeed developmental in nature, it would suggest a natural tendency for risk judgments to decrease over time.
Against such a backdrop, now let us imagine what might happen if an adolescent began to engage in risky behavior. The probability that he or she actually would experience a negative outcome is relatively low. Additionally, the probability of experiencing positive effects and benefits would be relatively high. In the absence of experiencing negative outcomes, we would expect to see adolescents’ perceptions of risk show even more dramatic decreases. Preliminary data from our longitudinal study hint at such an effect. Among adolescents who were not sexually active, we saw significant decreases over a one-year period in their perceptions of personal risk for STDs and HIV (decreases of 8 percent and 9 percent, respectively). However, the corresponding decreases among adolescents who became sexually active during the same time period were even greater (15 percent and 21 percent, respectively). Therefore, although both groups demonstrated a decrease in risk judgments over the one-year period, the amount of change was significantly larger among adolescents who began to engage in a risky behavior and did not experience a negative outcome.
It is possible the lower judgments of risk in adults get their impetus from age-related increases in susceptibility to inferential biases. Investigators who have found increases in decision-making biases between childhood and adulthood suggest that this increase is due to their acquisition of certain knowledge structures. For example, Davidson (1995) shows that young people are more likely to show the representativeness bias (as it applies to scenarios about the elderly) after they become aware of stereotypes
about the elderly. Decreasing judgments of risk also may reflect greater awareness of actual probabilities (which are generally low) and a move toward greater accuracy in older individuals. These are questions to be explored in further research.
CONCLUSIONS AND NEXT STEPS
Ultimately, we are interested in understanding why adolescents make the decisions they do and their competence in making these decisions. Why do adolescents engage in potentially threatening behaviors? Are they able to make informed decisions about undergoing or foregoing potentially dangerous medical treatments? Should we be granting them more legal rights or should they be more restricted? In considering adolescents’ perceptions of risk and vulnerability, we are acknowledging the importance of these perceptions in addressing these larger questions. Existing research has offered us insights into how adolescents view risk and vulnerability, identified some of the important correlates of these constructs, and raised interesting questions concerning the effects of these perceptions on adolescents’ behavior. It also has suggested to us some important next steps for a comprehensive research agenda and for program development, and some implications for a broad range of social policies concerning adolescents.
Despite strong beliefs to the contrary, there is little in the scientific literature to indicate that adolescents see themselves as invulnerable to harm. If anything, they appear to overestimate many of the risks around them. Given this, we believe that efforts to decrease public and scientific perceptions of “the invulnerable adolescent” may be warranted. As noted by others (Beyth-Marom et al., 1993; Hamburg et al., 1993), we may be doing adolescents a disservice by perpetuating this myth. Perceptions of “the invulnerable adolescent” could, for example, have negative influences on policy or funding priorities relating to adolescents. It becomes much easier for policy makers to believe that adolescent interventions are futile if they have been convinced that adolescents are destined to see themselves as invulnerable.
Although we see little evidence to suggest that adolescents perceive themselves as invulnerable to harm, existing data are more ambiguous when it comes to the question of adolescents’ actual competence in identifying and judging risk. Most studies report age-related increases in individuals’ awareness and consideration of risks, with adults showing greater awareness than adolescents. Depending on the standards used, adolescents’ compe-
tence in identifying risk does not appear to be exceptionally high. Furthermore, if we consider risk judgments to be measures of accuracy, age is associated with greater accuracy, and adults are more accurate (and thus more competent) than adolescents. Given that adults are generally more knowledgeable than adolescents, such an interpretation is reasonable.
On the other hand, if we choose to interpret risk judgments as reflections of individuals’ sense of vulnerability, it would suggest that the young feel more vulnerable than older adolescents, and adolescents feel more vulnerable than adults. Given the consistency of age differences across different types of risk, it seems plausible to suggest the possibility that individuals’ risk judgments do reflect generalized feelings of vulnerability or anxiety, and that such feelings are greater among the young. Such an interpretation is reasonable in light of changes in cognitive development and exposure to a changing social environment and also suggests to us that more attention be paid to the more affective dimensions of risk perception.
Caution is warranted in interpreting these results, however, because of serious concerns about whether the research paradigms used to study risk perception are able to give us an adequate picture of how adolescents judge risks in real situations. Most studies utilize hypothetical scenarios, which carry little of the emotion-arousing potential of real situations. Real-world studies have the advantage of offering contextually accurate settings but are often of limited generalizability due to their use of highly selected samples. Laboratory-based studies offer control and the ability to look for generalizable processes, but lack ecological validity. As a result, there has been a call for efforts to create more realistic, “real-world” simulations that would offer the advantages of laboratory research while maintaining salience and ecological validity (e.g., Ebbesen and Konecni, 1980). Such studies also could provide a means of more accurately measuring the affective components of risk perception.
The need for definitive studies concerning the risk perception-risk behavior relationship cannot be overstated. Because of the theoretical importance of the perceived risk construct to behavior, prevention and intervention programs frequently invest a great deal of effort trying to get adolescents to recognize and acknowledge their own vulnerability to negative outcomes. Such approaches are supported by experimental studies that have shown it is possible to reduce risk perception biases (Weinstein, 1983). However, these efforts have taken place without having conducted the essential longitudinal studies. Such studies would take a large sample of young adolescents before they began to engage in risky behaviors, and then follow
them long enough to observe their perceptions of risk over time, as they engage in the behaviors and possibly experience negative outcomes. It would be most desirable to see such studies conducted in populations representing different cultural backgrounds and levels of economic advantage. In addition to addressing some of the fundamental questions concerning the developmental trajectory of perceptions of risk and vulnerability, longitudinal studies also could inform us about some of their hypothesized developmental correlates such as future perspective, perspective taking, autonomy needs, and impulse control.
If risk judgments play a prominent role in the onset, continuation, or cessation of risky behavior, then perceptions about risk should continue to be incorporated into programs trying to prevent adolescent risk behavior and/or increase strategies to prevent such harm. Of course, doing so presents some real challenges. The actual risks posed by many of the behaviors we caution against are relatively small. Yet few educators would suggest providing adolescents with information about the actual risks, as doing so would (one fears) lead adolescents to minimize their importance. On the other hand, continuing to emphasize the likelihood of negative outcomes seems counterproductive if young people already feel a sense of heightened vulnerability, particularly in view of the inhibiting effects of excessive anxiety on preventive health behaviors (Leventhal, 1971). Furthermore, we have suggested that such an emphasis may backfire as adolescents become aware of the reality that most experiences with risky behaviors do not lead to negative outcomes and are in fact experienced as positive. A more appropriate goal in educating youth about health risks may be to find ways to translate small probabilities into real possibilities, without raising anxiety to unproductive levels.
Theoretical models can help inform us about potentially useful approaches to intervention, whether or not perceptions of risk prove to be crucial. For example, the Theory of Reasoned Action (Fishbein and Ajzen, 1975) includes a role for beliefs about the value attached to potential outcomes, expectations about social norms, and perceived benefits.
Programs that attempt to personalize and make vivid the reality of negative outcomes, such as those that expose adolescents to individuals who have AIDS, emphasize the meaning and impact of the outcomes rather than the probability of their occurrence (Sutton, 1982). To further develop such an emphasis would require not only that we understand what adolescents value, but also whether we, as a society, are providing adolescents with access to those outcomes. For many adolescents, such access is severely
limited as a result of living in impoverished, racist, or violent environments. An even more challenging task would be to create opportunities that youth would find highly desirable, and which are simultaneously incompatible with risky behavior (Leventhal and Keeshan, 1993). Learning more about adolescents’ perceptions of the benefits associated with engaging in risky behaviors could help in this endeavor. Perceived benefits typically are not studied, perhaps because of adults’ understandable reluctance to acknowledge those benefits. Yet by helping us to understand better what draws adolescents to risky behavior, such research could assist us in identifying desirable alternatives to those behaviors.
The social normative component has been utilized in preventive interventions that attempt to correct adolescents’ misperceptions about the number of teens engaging in risky behaviors. Adolescents who perceive higher levels of risk behavior in their peers view the risks of those behaviors as lower (Urberg and Robbins, 1984), and perceptions that peers condone more (or less) risky behavior are associated with more (or less) engagement in these behaviors (Boyer et al., 2000; DiClemente, 1991; Kinsman et al., 1998; Romer et al., 1994). Given that adolescents spend approximately twice as much time with their peers as they spend with parents or other adults (Brown, 1990; Savin-Williams and Berndt, 1990), peers and perceived peer norms thus become a major source of socialization and development during this period and a potentially effective intervention tool.
Understanding the relationship between risk judgment and behavior also will require that we turn attention to contextual factors that may influence the relationship. We tend to ask whether risk judgment influences behavior, rather than asking about the conditions under which it does (or does not). A person may agree that driving drunk is risky, but may decide to do so when under the influence of alcohol. Judgments concerning the risks of HIV may do little to influence the sexual behavior of someone who believes he or she will not live past the age of 30. These kinds of contextual factors are crucial but have received little attention. An example of a neglected and potentially important contextual factor is emotion. Emotion is known to influence adults’ perceptions of risk (Isen, 1993; Nygren et al., 1996) as well as their risk tolerance. We would expect emotional states to influence adolescents’ perceptions as well. Important questions include: How do specific emotions influence adolescents’ judgments of risk? Do the effects of emotion on risk perception differ for adolescents and adults or as a function of development? For example, does fear play a more important role in adults’ risk judgments than for adolescents’? In a given situation, do
we see developmental differences in the degree to which emotion is experienced? To the degree that emotion is important, it would also have implications for the ways in which we study risk perception, suggesting that we create study environments that mimic the kinds of emotionally arousing situations in which judgments about risk are typically made.
The importance of risk perceptions—theoretically, in program development, and in defining standards of decision-making competence—warrants rigorous study. Existing literature is fraught with problems, primarily stemming from issues concerning the measurement of risk perceptions and from the lack of longitudinal studies. But the problems we have identified are solvable, and we are optimistic that future studies will be able to answer many of the pressing questions we have raised.
Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckman (Eds.), Control: From cognition to behaviors (pp. 11-39). New York: Springer-Verlag.
Alexander, C. S. (1989). Gender differences in adolescent health concerns and self-assessed health. Journal of Early Adolescence, 9, 467-479.
Ambuel, B. (1992). Social policy of adolescent abortion. The Child, Youth, and Family Services Quarterly, 15(2), 5-9.
American Cancer Society. (1979). A study of health-related awareness and concerns among 4th, 5th and 6th graders (C.R.S. No. 5649). New York: Author.
American School Health Association, Association for the Advancement of Health Education, & Society for Publication Health Education, Inc. (1989). The national adolescent student health survey: A report on the health of America’s youth. Oakland, CA: Third Party.
Bachman, J. G. (1983). Schooling as a credential: Some suggestions for change. International Review of Applied Psychology, 32(4), 347-360.
Bandura, A. (1994). Social cognitive theory and exercise of control over HIV infection. In R. J. DiClemente & J. L. Peterson (Eds.), Preventing AIDS: Theories and methods of behavioral interventions (pp. 25-59). New York: Plenum Press.
Baron, R. M. (1988). An ecological framework for establishing a dual-mode theory of social knowing. In D. Bar-Tal & A. W. Kruglanski (Eds.), The social psychology of knowledge (pp. 48-82). Cambridge, England: Cambridge University Press.
Benedict, V., Lundeen, K. W., & Morr, B. D. (1981). Self-assessment by adolescents of their health status and perceived health needs. Health Values: Achieving High Level Wellness, 5, 239-245.
Benthin, A., Slovic, P., & Severson, H. (1993). Psychometric study of adolescent risk perception. Journal of Adolescence, 16, 153-168.
Bernstein, E., & Woodall, W. G. (1987). Changing perceptions of riskiness in drinking, drugs and driving: An emergency department-based alcohol and substance abuse prevention program. Annals of Emergency Medicine, 16, 1350-1354.
Beyth-Marom, R., Austin, L., Fischhoff, B., Palmgren, C., & Jacobs-Quadrel, M. (1993). Perceived consequences of risky behaviors: Adults and adolescents. Developmental Psychology, 29(3), 549-563.
Biehl, M., & Halpern-Felsher, B. L. (2001). Adolescents’ and adults’ understanding of probability expressions. Journal of Adolescent Health, 28, 30-35.
Bloom-Feshbach, S., Bloom-Feshbach, J., & Heller, K. A. (1982). Work, family, and children’s perceptions of the world. In S. B. Kamerman & C. D. Hayes (Eds.), Families that work: Children in a changing world (pp. 268-307). Washington, DC: National Academy Press.
Boyer, C. B., Shafer, M. A., Wibbelsman, C. J., Seeberg, D., Teitle, E., & Lovell, N. (2000). Associations of sociodemographic, psychosocial, and behavioral factors with sexual risk and sexually transmitted diseases in teen clinic patients. Journal of Adolescent Health, 27, 102-111.
Brown, B. B. (1990). Peer groups and peer cultures. In S. S. Feldman & G. R. Elliott (Eds.), At the threshold: The developing adolescent (pp. 171-196). Cambridge, MA: Harvard University Press.
Brunswick, A. F. (1969). Health needs of adolescence: How the adolescent sees them. American Journal of Public Health, 59, 1730-1745.
Brunswick, A. F., & Josephson, E. (1972). Adolescent health in Harlem. American Journal of Public Health , 62(2) (Suppl. 2), 1-62.
Butterfield, F. (1996, May 12). States revamping laws on juveniles as felonies soar. The New York Times, p. 1.
Byler, R., Lewis, G. M., & Totman, R. J. (1969). Teach us what we want to know. New York, NY: Mental Health Materials Center.
Byrnes, J., Miller, D. C., & Reynolds, M. (1999). Learning to make good decisions: A self-regulation perspective. Child Development, 70(5), 1121-1140.
Centers for Disease Control and Prevention (1998). Youth risk behavior surveillance—United States, 1997. Morbidity and Mortality Weekly Report (MMWR), 47(SS-3).
Christiansen, B., Roehling, P., Smith, G., & Goldman, M. (1989). Using alcohol expectancies to predict adolescent drinking behavior after one year. Journal of Consulting and Clinical Psychology, 57(1), 93-99.
Cohn, L., Macfarlane, S., Yanez, C., & Imai, W. (1995). Risk-perception: Differences between adolescents and adults. Health Psychology, 14(3), 217-222.
Connell, J. P., & Halpern-Felsher, B. L. (1997). How neighborhoods affect educational outcomes in middle childhood and adolescence: Conceptual issues and an empirical example. In G. D. J. Brooks-Gunn & J. L. Aber (Eds.), Neighborhood poverty Volume I: Context and consequences for children (pp. 174-199). New York, NY: Russell Sage Foundation.
Covington, M. V., & Omelich, C. L. (1992). Perceived costs and benefits of cigarette smoking among adolescents: Need instrumentality, self-anger, and anxiety factors. In D. G. Forgays & D. K. Forgays (Eds.), Anxiety: Recent developments in cognitive, psychophysiological, and health research (pp. 245-261). Washington, DC: Hemisphere.
Davidson, A. R. (1995). From attitudes to actions to attitude change: The effects of amount and accuracy of information. In R. E. Petty and J. A. Krosnick (Eds.), Attitude strength: Antecedents and consequences (pp. 315-336). Mahwah, NJ: Lawrence Erlbaum Associates.
DiClemente, R. J. (1991). Predictors of HIV-preventive sexual behavior in a high-risk adolescent population: The influence of perceived peer norms and sexual communication on incarcerated adolescents’ consistent use of condoms. Journal of Adolescent Health, 12, 385-390.
DiClemente, R. J., Zorn, J., & Temoshok, L. (1987). The association of gender, ethnicity, and length of residence in the Bay area to adolescents’ knowledge and attitudes about Acquired Immune Deficiency Syndrome. Journal of Applied Social Psychology, 17(3), 216-230.
Drottz-Sjoberg, B.-M., & Sjoberg, L. (1991). Adolescents’ attitudes to nuclear power and radioactive wastes. Journal of Applied Social Psychology, 21, 2007-2036.
Ebbesen, E. B., & Konecni, V. J. (1980). On the external validity of decision-making research: What do we know about decisions in the real world? In T.S. Wallsten (Ed.), Cognitive processes in choice and decision behavior (pp. 21-45). Hillsdale, NJ: Lawrence Erlbaum Associates.
Ellen, J. M., Adler, N. E., Dunlop, M. B., Gurvey, J. E., Millstein, S. G., & Tschann, J. M. (1998). Perceived risk for STDs and partner-specific condom intentions. Pediatric Research, 43, 4A.
Eme, R., Maisiak, R., & Goodale, W. (1979). Seriousness of adolescent problems. Adolescence, 14(53), 93-99.
Ey, S., Klesges, L. M., Patterson, S. M., Hadley, W., Barnard, M., & Alpert, B. S. (2000). Racial differences in adolescents’ perceived vulnerability to disease and injury. Journal of Behavioral Medicine, 23(5), 421-435.
Federal Interagency Forum on Child and Family Statistics (2000). America’s children: Key national indicators of well-being. Washington, DC: Author.
Feldman, W., Hodgson, C., Corber, S., & Quinn, A. (1986). Health concerns and health-related behaviours of adolescents. Canadian Medical Association Journal, 134(5), 489-493.
Ferguson, S. A., & Williams, A. F. (1996). Parents’ views of driver licensing practices in the United States. Journal of Safety Research, 27(2), 171-180.
Finn, P., & Brown, J. (1981). Risks entailed in teenage intoxication as perceived by junior and senior high school students. Journal of Youth and Adolescence, 10(1), 61-76.
Fischhoff, B. (1996). The real world: What good is it? Organizational Behavior and Human Decision Processes, 65, 232-248.
Fischhoff, B., Parker, A. M., Bruine de Bruin, W., Downs, J., Palmgren, C., Dawes, R., & Manski, C. F. (2000). Teen expectations for significant life events. Public Opinion Quarterly, 64, 189-205.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Furby, L., Ochs, L. M., & Thomas, C. W. (1997). Sexually transmitted disease prevention: Adolescents’ perceptions of possible side effects. Adolescence, 32(128), 781-809.
Gans, J. E., Alexander, B., Chu, R. C., & Elster, A. B. (1995). The cost of comprehensive preventive medical services for adolescents. Archives of Pediatrics and Adolescent Medicine, 149(11), 1226-1234.
Gerrard, M., Gibbons, F., Benthin, A., & Hessling, R. (1996a). A longitudinal study of the reciprocal nature of risk behaviors and cognitions in adolescents: What you do shapes what you think, and vice versa. Health Psychology, 15(5), 344-354.
Gerrard, M., Gibbons, F., & Bushman, B. (1996b). Relation between perceived vulnerability to HIV and precautionary sexual behavior. Psychological Bulletin, 119(3), 390-409.
Giblin, P. T., & Poland, M. L. (1985). Health needs of high school students in Detroit. Journal of School Health, 55(10), 407-410.
Gittler, J., Quigley-Rick, M., & Saks, M. J. (1990). Adolescent health care decision making: The law and public policy. Washington, DC: Carnegie Council on Adolescent Development.
Gladis, M., Michela, J., Walter, H., & Vaughan, R. (1992). High school students’ perceptions of AIDS risk: Realistic appraisal or motivated denial? Health Psychology, 11(5), 307-316.
Gochman, D., & Saucier, J. F. (1982). Perceived vulnerability in children and adolescents. Health Education Quarterly, 9(2&3), 46/142-58/154.
Goldberg, J. H., Halpern-Felsher, B. L., & Millstein, S. G. (2001a). Adolescent and adult errors in health risk judgments due to the representativeness heuristic. Manuscript submitted for publication.
Goldberg, J. H., Halpern-Felsher, B. L., & Millstein, S. G. (2001b). Beyond invulnerability: The importance of benefits in adolescents’ decision to drink alcohol. Manuscript submitted for publication.
Halpern-Felsher, B. L., & Cauffman, E. (2001). Costs and benefits of a decision: Decision-making competence in adolescents and adults. Journal of Applied Developmental Psychology, 22, 257-273.
Halpern-Felsher, B. L., Millstein, S. G., Ellen, J. M., Adler, N. E., Tschann, J. M., & Biehl, M. (2001). The role of behavioral experience in judging risks. Health Psychology, 20, 120-126.
Hamburg, D. A., Millstein, S. G., Mortimer, A. M., Nightingale, E. O., & Petersen, A. C. (1993). Adolescent health promotion in the twenty-first century: Current frontiers and future directions. In S. G. Millstein, A. C. Petersen, & E. O. Nightingale (Eds.), Promoting the health of adolescents: New directions for the twenty-first century (pp. 375-388). New York: Oxford University.
Hodne, C. J. (1995). Medical decision making. In M. W. O’Hara & R. C. Reiter (Eds.), Psychological aspects of women’s reproductive health (pp. 267-290). New York: Springer.
Inhelder, B., & Piaget, J. (1958). The growth of logical thinking from childhood to adolescence: An essay on the construction of formal operation structures. New York: Basic Books.
Isen, A. M. (1993). Positive affect and decision making. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 261-277). New York: Guilford Press.
Jacobs, J. E., & Potenza, M. (1991). The use of judgment heuristics to make social and object decisions: A developmental perspective. Child Development, 62, 166-178.
Kanfer, F. H. (1970). Self regulation: Research, issues and speculations. New York: Appleton-Century-Crofts.
Kaser-Boyd, N., Adelman, H. S., & Taylor, L. (1985). Minors’ ability to identify risks and benefits of therapy. Professional Psychology: Research & Practice, 16(3), 411-417.
Kassinove, H., & Sukhodolsky, D. G. (1995). Optimism, pessimism and worry in Russian and American children and adolescents. Journal of Social Behavior and Personality, 10(1), 157-168.
Kinsman, S. B., Romer, D., Furstenberg, F. F., & Schwarz, D. F. (1998). Early sexual initiation: The role of peer norms. Pediatrics, 102, 1185-1192.
Klaczynski, P. A., & Narasimham, G. (1998). Representations as mediators of adolescent deductive reasoning. Developmental Psychology, 34(5), 865-881.
Kuhn, D. (1992). Thinking as an argument. Harvard Educational Review, 62(2), 155-178.
Kuhn, D., Amsel, E., O’Loughlin, M., Schauble, L., Leadbeater, B., & Yotive, W. (1988). The development of scientific thinking skills. San Diego, CA: Academic Press.
Leventhal, H. (1971). Fear appeals and persuasions: The differentiation of a motivational construct. American Journal of Public Health, 61, 1208-1224.
Leventhal, H., & Keeshan, P. (1993). Promoting healthy alternatives to substance abuse. In S. G. Millstein, A. C. Petersen, & E. O. Nightingale (Eds.), Promoting the health of adolescents: New directions for the twenty-first century (pp. 260-284). New York: Oxford University Press.
Lewis, C. C. (1980). A comparison of minors’ and adults’ pregnancy decisions. American Journal of Orthopsychiatry, 50, 446-453.
Lewis, C. C. (1981). How adolescents approach decisions: Changes over grades seven to twelve and policy implications. Child Development, 52, 538-544.
Manning, D. T., & Balson, P. M. (1989). Teenagers’ beliefs about AIDS education and physicians perceptions about them. Journal of Family Practice, 29, 173-177.
Marks, A., Malizio, J., Hoch, J., Brody, R., & Fisher, M. (1983). Assessment of health needs and willingness to utilize health care resources of adolescents in a suburban population. Journal of Pediatrics, 102(3), 456-460.
McClure-Martinez, K., & Cohn, L. D. (1996). Adolescent and adult mothers’ perceptions of hazardous situations for their children. Journal of Adolescent Health, 18(3), 227-231.
McKenna, F. P., Warburton, D. M., & Winwood, M. (1993). Exploring the limits of optimism: The case of smokers’ decision making. Bristish Journal of Psychology, 84, 389-394.
Michielutte, R., & Diseker, R. (1982). Children’s perceptions of cancer in comparison to other chronic illnesses. Journal of Chronic Diseases, 35, 843-852.
Midgley, C., & Feldlaufer, H. (1987). Students’ and teachers’ decision-making fit before and after the transition to junior high school. Journal of Early Adolescence, 7(2), 225-241.
Millstein, S. G., & Halpern-Felsher, B. L. (2001). Judgments about risk and perceived invulnerability in adolescents and young adults. Manuscript submitted for publication.
Millstein, S. G., & Irwin, C. E., Jr. (1985). Adolescents’ assessments of behavioral risk: Sex differences and maturation effects. Pediatric Research, 19, 112A.
Moore, S., & Rosenthal, D. (1991). Adolescent invulnerability and perceptions of AIDS risk. Journal of Adolescent Research, 6(2), 164-180.
Moore, S., & Rosenthal, D. (1992). Australian adolescents’ perceptions of health-related risks. Journal of Adolescent Research, 7(2), 177-191.
Mundt, J. C., Ross, L. E., & Harrington, H. L. (1992). A modeling analysis of young drivers’ judgments of accident risk due to alcohol use and other driving conditions. Journal of Studies on Alcohol, 53(3), 239-248.
Natapoff, J. N., & Essoka, G. C. (1989). Handicapped and able-bodied children’s ideas of health. Journal of School Health, 59(10), 436-440.
Nygren, T. E., Isen, A. M., Taylor, P. J., & Dulin, J. (1996). The influence of positive affect on the decision rule in risk situations: Focus on outcome (and especially avoidance of loss) rather than probability. Organizational Behavior & Human Decision Processes, 66(1), 59-72.
Office of Disease Prevention and Health Promotion. (1993). Recent reports on high-risk adolescents. Public Health Reports, 108, 68-77.
Parcel, G. S., Nader, P., & Meyer, M. (1977). Adolescent health concerns, problems, and patterns of utilization in a triethnic urban population. Pediatrics, 60(2), 157-164.
Parsons, J. T., Siegel, A. W., & Cousins, J. H. (1997). Late adolescent risk-taking: Effects of perceived benefits and perceived risks on behavioral intentions and behavioral change. Journal of Adolescence, 20, 381-392.
Piaget, J. (1971). The theory of stages in cognitive development. In D. R. Green, M. P. Ford, & G. B. Flamer, (Eds.), Measurement and Piaget (pp. 1-11). New York: McGraw-Hill.
Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1990). Contraceptive attitudes and intention to use condoms in sexually experienced and inexperienced adolescent males. Journal of Family Issues, 11(3), 294-312.
Pleck, J. H., Sonenstein, F. L., & Ku, L. C. (1993). Masculinity ideology: Its impact on adolescent males’ heterosexual relationships. Journal of Social Issues, 49(3), 11-29.
Porteous, M. A. (1979). A survey of the problems of normal 15-year-olds. Journal of Adolescence, 2, 307-323.
Price, J. H., Desmond, S., & Kukulka, G. (1985). High school students’ perceptions and misperceptions of AIDS. Journal of School Health, 55(3), 107-109.
Price, J. H., Desmond, S. M., Wallace, M., Smith, D., & Stewart, P. (1988). Differences on black and white adolescents’ perceptions about cancer. Journal of School Health, 58(2), 66-70.
Quadrel, M. J., Fischhoff, B., & Davis, W. (1993). Adolescent (in)vulnerability. American Psychologist, 48(2), 102-116.
Radius, S. M., Dillman, T. E., Becker, M. H., Rosenstock, I., & Horvath, W. J. (1980). Adolescent perspectives on health and illness. Adolescence, 25(58), 375-384.
Reyna, V. F., & Ellis, S. C. (1994). Fuzzy-trace theory and framing effects in children’s risky decision making. American Psychological Society, 5(5), 275-279.
Roe-Berning, S., & Straker, G. (1997). The association between illusions of invulnerability and exposure to trauma. Journal of Traumatic Stress, 10(2), 319-327.
Romer, D., Black, M., Ricardo, I., & Feigelman, S. (1994). Social influences on the sexual behavior of youth at risk for HIV exposure. American Journal of Public Health, 84, 977-985.
Ronis, D. L. (1992). Conditional health threats: Health beliefs, decisions, and behaviors among adults. Health Psychology, 11(2), 127-134.
Rosenstock, I. M. (1974). Historical origins of the Health Belief Model. Health Education Monographs, 2, 1-9.
Sastre, M. T. M., Mullet, E., & Sorum, P. C. (1999). Relationship between cigarette dose and perceived risk of lung cancer. Preventive Medicine: An International Journal Devoted to Practice & Theory, 28(6), 566-571.
Savin-Williams, R. C., & Berndt, T. J. (1990). Friendship and peer relations. In S. S. Feldman & G. R. Elliott (Eds.), At the threshold: The developing adolescent (pp. 277-307). Cambridge, MA: Harvard University Press.
Shaklee, H., & Goldston, D. (1989). Development in causal reasoning: Information sampling and judgment rule. Cognitive Development, 4, 269-281.
Smith, G., Goldman, M., Greenbaum, P., & Christiansen, B. (1995). Expectancy for social
facilitation from drinking: The divergent paths of high-expectancy and low-expectancy adolescents. Journal of Abnormal Psychology, 104(1), 32-40.
Sobal, J., Klein, H., Graham, D., & Black, J. (1988). Health concerns of high school students and teachers’ beliefs about student health concerns. Pediatrics, 81(2), 218-223.
Steinberg, L., & Silverberg, S. (1986). The vicissitudes of autonomy in early adolescence. Child Development, 57, 841-851.
Sternlieb, J. J., & Munan, L. (1972). A survey of health problems, practices, and needs of youth. Pediatrics, 49(2), 177-186.
Strunin, L. (1991). Adolescents’ perceptions of risk for HIV infection: Implications for future research. Social Science & Medicine, 32(2), 221-228.
Sutton, S. R. (1982). Fear arousing communications: A critical examination of theory and research. In J. R. Eiser (Ed.), Social psychology and behavioral medicine (pp. 303-337). New York: John Wiley.
Triandis, H. C. (1977). Interpersonal behavior. Monterey, CA: Brooks Cole.
University of Minnesota. (1989). The state of adolescent health in Minnesota. Minneapolis: University of Minnesota Press.
Urberg, K., & Robbins, R. L. (1981). Adolescents’ perceptions of the costs and benefits associated with cigarette smoking: Sex differences and peer influence. Journal of Youth and Adolescence, 10(5), 353-361.
Urberg, K., & Robbins, R. L. (1984). Perceived vulnerability in adolescents to the health consequences of cigarette smoking. Preventive Medicine, 13, 367-376.
Van der Plight, J. (1998). Perceived risk and vulnerability as predictors of precautionary behaviour. British Journal of Health Psychology, 3, 1-14.
Van der Velde, F. W., Hooykaas, C., & Van der Plight, J. (1996). Conditional versus unconditional risk estimated in models of AIDS-related risk behavior. Psychology and Health, 12, 87-100
Violato, C., & Holden, W. B. (1988). A confirmatory factor analysis of a four-factor model of adolescent concerns. Journal of Youth & Adolescence, 17(1), 101-113.
Weinstein, M. C., & Fineberg, H. (1980). Clinical decision analysis. Philadelphia, PA: Saunders.
Weinstein, N. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5), 806-820.
Weinstein, N. D. (1983). Reducing unrealistic optimism about illness susceptibility. Health Psychology, 2(1), 11-20.
Weinstein, N. D. (1984). Why it won’t happen to me: Perceptions of risk factors and susceptibility. Health Psychology, 3(5), 431-457.
Weinstein, N. D. (1989). Effects of personal experience on self-protective behavior. Psychological Bulletin, 105(1), 31-50.
Weithorn, L. A., & Campbell, S. B. (1982). The competency of children and adolescents to make informed treatment decisions. Child Development, 53, 1589-1598.
Whalen, C. K., Henker, B., O’Neil, R., Hollingshead, J., Holman, A., & Moore, B. (1994). Optimism in children’s judgments of health and environmental risks. Health Psychology, 13(4), 319-325.