Journal of Pediatric Psychology Advance Access originally published online on February 23, 2005
Journal of Pediatric Psychology 2005 30(5):425-435; doi:10.1093/jpepsy/jsi066
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Predictors of Children’s and Adolescents’ Risk Perception
1 University of Alabama, 2 University of Mississippi Medical Center, and 3 Washington State University
All correspondence concerning this article should be addressed to L. Greening, Department of Psychology, The University of Alabama, Tuscaloosa, Alabama 35487-0348. E-mail: green{at}bama.ua.edu.
Received January 11, 2004; revisions received May 8, 2004, and June 25; accepted August 25, 2004
 |
Abstract |
|---|
 Top Abstract Cognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Objective To test cognitive-developmental, social-cognitive, motivational, and cognitive hypotheses about the psychological mechanisms underlying children’s risk perception. Method Youth (N = 1315) ranging from 9 to 17 years of age completed measures assessing adolescent egocentrism, personal experience with four negative health events, how much they worried about the health events, and their perceived skill for event-related activities. The measures were completed twice, 12 months apart. Results Lacking personal experience with and worrying less about health threats were significant predictors of more optimistically biased risk perception a year later. Conclusions The lack of experience with and not worrying about serious health consequences may desensitize children to potential health risks. Clinical applications for health education programs are discussed.
Key words: risk perception; children; adolescents; unrealistic optimism; health beliefs.
Many health education programs are predicated on the cognitive theory that health beliefs guide health behaviors. As outlined in the often-cited health belief model (HBM; Janz & Becker, 1984
), people are more likely to engage in health behaviors if they perceive (a) vulnerability to health threats, (b) that the consequences of the threat are severe, and (c) that treatment or taking preventive measures will be effective. Subsequent modifications to this model include the addition of perceived social or monetary barriers to the adaptive response. A cue to action which can be internal (e.g., symptoms) or external (e.g., health communication) is hypothesized to trigger these cognitive processes. However, demographic and sociopsychological variables may influence perceptions and thereby indirectly affect the likelihood of the response. Although there are variants to this framework, the different models proposed share many of the same elements (van der Pligt, 1994).
Cognitive models for health behaviors are potentially appealing to health professionals targeting children’s and adolescents’ health because as children become more autonomous from adults they tend to rely on their own cognitive processes to assess health risks. Although HBM has been found to predict children’s expectations to use medicines to treat illnesses (Bush & Iannotti, 1990), there is limited support for the model and for components of the model predicting children’s health behaviors. This is especially evident for perceived vulnerability to health threats. Early research revealed low and mostly negative relations between children’s health behaviors and perceived vulnerability to risks (Gochman & Saucier, 1982). These initial findings were extended to adolescents (Greening & Stoppelbein, 2000), thus challenging the hypothesis postulated by HBM that greater-perceived risk is related to adaptive health behaviors in children and adolescents. Janz and Becker (1984) suggested that the negative correlations may be explained by respondents perceiving that they have lowered their risk of harm by taking preventive measures. At the very least, the findings suggest that risk perception is a complex process that warrants a deeper understanding from both health educators and researchers alike.
Research on risk perception confirms that people do not simply process risk data like computers, as evidenced by a pervasive tendency for people to perceive their risk of harm to be below average (Cohn, Macfarlane, Yanez, & Imai, 1995; Gochman & Saucier, 1982; Quadrel, Fischhoff, & Davis, 1993; Weinstein, 1980, 1984, 1989; Whalen et al., 1994). Weinstein (1980) coined this cognitive bias as unrealistic optimism because although some people’s risk may be below average, it is unrealistic for everyone’s risk to fall below average. Children and adolescents are just as likely as adults to exhibit this bias; however, adults may show stronger biases.
Unrealistic optimism has direct implications for health promotion including impeding health education efforts. Hence, our goal was to unveil the mechanisms underlying risk perception early in development and to offer guidance for health education programs. Our hypotheses were derived from four theoretical frameworks commonly discussed in the risk-perception literature including the cognitive-developmental, social–cognitive, motivational, and cognitive theories. These theories are discussed below to set the foundation for the hypotheses for this study.
|
Cognitive-Developmental Theory of Risk Perception |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Developmental theorists hypothesize that adolescents are prone to unrealistic optimism because as adolescents graduate from concrete to abstract thinking, their thinking becomes more self-focused. One manifestation of this developmental phenomenon, commonly referred to as adolescent egocentrism, is adolescents’ tendency to think that they are unique; hence, they cannot be vulnerable to health risks like the typical average person (Elkind, 1967
). Dolcini et al. (1989) tested this hypothesis and found that adolescents who scored high on egocentrism perceived greater risk of harm and not lower risk as hypothesized by the theory. The authors speculated that the adolescent-egocentrism measure they used might have measured self-reflection and that the findings indicate that more reflective youth think more about the risks of dangerous activities. Nevertheless, the cross-sectional design precluded any definitive conclusions. Hence, adolescent egocentrism was tested in this study as a predictor of unrealistic optimism in children and adolescents using a longitudinal design. It was hypothesized that youth who became more egocentric within a year would become more optimistically biased about their risk of harm than adolescents who did not become more egocentric. The magnitude of the effect was compared to factors derived from three other theoretical perspectives to compare their relative contributions to risk perception. |
Social-Cognitive Theory of Risk Perception |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
According to social-cognitive theorists, children expect greater risk of injury after experiencing negative consequences, subsequently lessening the likelihood of future participation in the risky activities (Bandura, 1986
; Peterson, Gillies, Cook, Schick, & Little, 1994). Likewise, children who do not experience negative consequences would be desensitized to expectations of future risk. Morrongiello and Rennie (1998) provided empirical support for this hypothesis with a sample of risk-taking children, who reported lower risk of negative consequences if they had not experienced serious injuries. Similar findings were observed with adolescent survivors of weather disasters (Greening, Dollinger, & Pitz, 1996). These cross-sectional findings were expanded upon by testing the hypothesis that experience with negative consequences predicts children’s and adolescents’ risk perception. Youth, who had personally experienced serious consequences from engaging in risky activities were expected to be less optimistically biased about their risk of future harm from these events a year later. |
Motivational Theory of Risk Perception |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Although often conceptualized as a rational cognitive process, risk perception does not necessarily occur in an emotionally neutral context. Threats of injury or harm often elicit uncomfortable cognitions and emotions including fear, anxiety, and worry. Hence, people minimize their risk of harm so as to assuage and avoid such distressing thoughts and feelings. Risk judgments and negative cognitions are, therefore, hypothesized to be positively related according to the motivational theory. Although Peterson et al. (1994)
reported a relation between fear and children’s ratings for the severity of injuries expected from a simulated bicycle accident, worry has not been examined as a predictor of children’s and adolescents’ risk perception to date. This gap in the literature was addressed in this study by testing the magnitude of the effect for increased worrying about a health threat on unrealistic optimism in comparison to variables derived from three other theories about risk perception. |
Cognitive Theory of Risk Perception |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
According to cognitive theorists, people exhibit unrealistic optimism because of faulty cognitive processes or perceptions. Self-serving biases about event-related skills, for example, can cause people to believe that they possess the necessary skills to successfully avoid health risks (Millstein, 1993
). This is supported by evidence of people exhibiting self-serving biases about a range of personal characteristics (Chandler, Greening, Robison, & Stoppelbein, 1999; McKenna, Stanier, & Lewis, 1991) and perceiving lower risks for events involving some element of personal control or error versus for random, unpredictable events (Gerrard, Gibbons, & Warner, 1991; Greening & Chandler, 1997; Quadrel et al., 1993; Whalen et al., 1994). However, little is known about the role of self-serving biases in children’s risk perception. We, therefore, tested the cognitive hypothesis that children and adolescents who report an improvement in their skill level during a year would show an increase in their unrealistic optimism for respective health events.
To summarize, the hypotheses for this study were that (a) children and adolescents who show an increase in adolescent egocentrism over the course of a year would show an increase in unrealistic optimism, (b) youth who experienced negative consequences from engaging in risky behaviors during the course of a year would report less optimistically biased risk of harm from the events, (c) youth who show an increase in worrying about health threats over the course of a year would report less optimistically biased personal risk of harm from the events, and (d) youth who exhibited stronger self-serving biases about event-related skills over the course of a year would report more optimistically biased risk of harm from the events. We examined the relative proportion of variance that adolescent egocentrism, personal experience, worrying, and self-serving biases contributed to risk perception so as to compare their relative contribution to unrealistic optimism early in development when health beliefs are typically established. Gender was also included as a covariate in analyses because according to social psychological theory, girls are socialized to be wary of risks whereas boys are socialized to be undaunted by potential risks (Harris & Miller, 2000).
Examining predictors in childhood and adolescence could provide a developmental perspective on a fundamental component of HBMs and prove especially useful for conceptualizing a comprehensive model of risk perception. Although risk perception does not necessarily effect changes in behavior, acknowledging one’s personal risk may be the first in a chain of events toward health promotion (van der Pligt, 1994). We focused specifically on risk perception and not health behavior per se because health educators often utilize risk communication without a sound foundation in the underlying mechanisms, thus possibly impeding their progress.
|
Method |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Participants
Participants included 1,799 students attending public schools in a suburban community [Mdn household income, $70,623; range, <$10,000 (4.2%) to >$200,000 (10.5%)]. Parental consent to participate was obtained for 1,783 students, of which 43 were absent from school (absentee rates = 3% across schools) and 13 declined to participate. An additional 36 students were excluded from analyses because of missing data, leaving 1,691 participants and yielding a 94% participation rate. The students ranged from 9 to 17 years of age (M = 12.51, SD = 1.85). Nineteen percent were 5th graders in elementary school; 61% were 6–8th graders in middle school; and 20% were 9–11th graders in high school. More males (53%) participated and most were Caucasian (89.5%). The remaining students were African American (5%), Asian American (3.5%), or "Other" (2%). The ethnic composition of the sample was representative of the community.
Data were collected again 12 months later from the same students, with the exception of the 11th graders (n = 115) because they were scheduled to take a national achievement test. Thirty-nine students were absent at follow up, 8 declined to participate, 192 students no longer attended school in the area, and 22 were excluded because of missing data, leaving 1,315 participants at Time 2. Most of the original sample (91%) participated at Time 2, excluding the 11th graders and 78% participated when including the 11th graders. Drop outs, excluding the 11th graders, were older, F(1, 1574) = 30.73, p < .0001, in higher grades, F(1, 1574) = 22.63, p < .0001, were more often African American, 
2(4) = 20.30, p < .0001, reported that they flossed more often at Time 1, rode a bicycle less frequently, drove more often, had been in more accidents as a passenger, rated their swimming skill lower, and worried more about tooth decay, Fs(1, 1574) = 4.50–11.48, p < .05. Sixty-seven percent of the participants were in the 6–8th grade at Time 2 and 33% were in the 9–11th grade. The students ranged from 10 to 17 years of age, with a mean age of 13.13 years (SD = 1.56). More males (53%) and Caucasian (91%) students participated, followed by African Americans (4%), Asian Americans (4%), and "Other" (1%).
Measures
Demographic Questionnaire.
Students indicated their age, gender, grade level, and ethnicity on a brief demographic questionnaire.
Risk Ratings.
Perceived risk was assessed using Weinstein’s (1984) comparative risk rating procedure for assessing unrealistic optimism. Respondents rated on a 5-point scale their risk of experiencing four negative events compared to the average person their age [1, "it is impossible I will ever have" (event occur); 2, "my chance is less than the average kid my age"; 3, "my chance is the same as the average kid my age"; 4, "my chance is more than the average kid my age, but not 100%"; and 5, "I will definitely" (experience the event)]. The events included tooth decay, being in a motor vehicle accident while driving, and sustaining bicycle and watersport-related injuries that warranted medical attention (e.g., required emergency room or doctor visit). Students who reported higher risk were described as less optimistically biased, whereas students who reported lower risk were more optimistically biased. Conditional risk assessments (e.g., risk if you were to engage in a risky activity) were indicated for motor vehicle accident, and sustaining bicycle and watersport-related injuries because conditional risk assessments have been found to yield more accurate risk judgments (Ronis, 1992). The events were selected because (a) they were used in past research (e.g., Cohn et al., 1995; DiLillo, Potts, & Himes, 1998) and allowed for comparisons of findings, (b) they had high enough base rates for minors in the region to have the opportunity for negative outcomes, (c) they were among the leading causes of morbidity and mortality for minors (Minino & Smith, 2001) and are often the subject of health promotion campaigns, and (d) youth may not have had extensive negative experiences yet, thus allowing for an examination of the impact of early negative experiences. Although driving a motor vehicle did not meet the latter criterion, it was included to examine the variance explained by predictors for an event when experience is limited; whereas including dental care allowed for a test of the hypotheses with an event that requires daily attention. Internal consistency for the items was observed to be modest, Cronbach’s a = .46 and .53 at Times 1 and 2, respectively, supporting the decision to examine risk ratings for each event individually.
Skill Ratings.
Respondents rated on a 5-point scale their skill level (1, "the worst of kids my age"; 2, "less than the average kid my age"; 3, "same as the average kid my age"; 4, "better than the average kid my age"; and 5, "the best of kids my age") compared to an average peer for event-related behaviors including dental hygiene, riding a bicycle, driving, swimming, and riding a Jet Ski. Internal consistency for the items was .60 and .61 at Times 1 and 2, respectively, suggesting that skill ratings be examined separately for each event.
Worry.
The students rated on a 5-point scale (1, "not all"; 2, "a little"; 3, "many times"; 4, "most of the time"; and 5, "all the time") how much they worried about experiencing the four health events. Internal consistency was .61 and .62 for the items at Times 1 and 2, respectively. The magnitude of the coefficients supported examining worry ratings individually for each event to evaluate event-specific predictors.
Past Experience.
Participants rated on a 5-point scale (1, "never"; 2, "a little"; 3, "sometimes"; 4, "many times"; and 5, "every day") how often they engaged in behaviors that could increase or decrease their risk of the four health events including tooth brushing, flossing, bicycle riding, driving, and swimming or riding a boat or Jet Ski. Rating options for the two dental hygiene questions were quantified (1, "less than once a week"; 2, "2–5 times a week"; 3, "once a day"; 4, "2–3 times a day"; and 5, "every time after I eat"). Internal consistency for the items was low at Times 1 and 2, Cronbach’s 
= .20 and .23, respectively. Although there is evidence for correlations among health behaviors, the findings have been found to be inconsistent (Elliott, 1993). The low intercorrelations supported examining the effects for the frequency of activities on risk perception separately for each event.
Past Negative Experience.
Students indicated on a 7-point scale how often they had experienced tooth decay, a motor vehicle accident as a passenger and then as a driver, and bicycle or watersport-related injuries that warranted medical attention (e.g., emergency room or doctor visit). Ratings ranging from 0 to 5 reflected the number of times they experienced the event. An event experienced more than 5 times was rated as 6. Internal consistency was .44 at Time 1 and .48 at Time 2, thus supporting an examination of event-specific experiences.
Adolescent Egocentrism Scale.
The adolescent egocentrism scale (Enright, Shukla, & Lapsley, 1980) is a 15-item measure of adolescent egocentrism. Respondents rated on a 5-point Likert-type scale ranging from 1 to 5 the importance of each statement. A principal factor analysis revealed one common factor. Internal consistency is moderately high (Cronbach’s = .78–.86) and was high for this sample, = .85 and .86 at Times 1 and 2, respectively. The scale correlates positively with measures of self-consciousness and correlates weakly with a measure of empathic concern, suggesting evidence for discriminant validity (Enright et al., 1980).
Design and Procedure
After obtaining approval from the school board and the Institutional Review Board, students were recruited to participate in a 20 to 25-minute study on health perceptions. No incentives were offered for participation. Students with parental consent and who assented to participate were tested in groups of 15–30 students. Graduate students administered the measures in counterbalanced order and repeated the procedure again a year later for a total of two data collection points. All measures were coded by random numbers to ensure the confidentiality of responses. A master list of the participants’ names and their corresponding numbers was maintained by the researchers to match Times 1 and 2 data.
|
Results |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Time 1
Students reported that their risk of harm was significantly below average at Time 1, ts = –26.39 to –16.89, p < .0001 (see Table I for means). Most of the students (73–91%) reported that they did not worry or worried a little about each event,
2s (4, N = 1691) = 1162.36–2477.12, p < .0001. Forty-one percent reported no experience with tooth decay; 69% reported no watersport-related injuries; and 80% and 94% had not been in a motor vehicle accident as a passenger or driver, respectively, 2s (6, N = 1691) = 1218.70–7271.88, p < .0001. Most (71%) reported sustaining bicycle injuries three or fewer times, 2 (6, N = 1691) = 526.92, p < .0001. The majority (71%) had never driven or else drove "a little", 2 (4, N = 1691) = 1616.71, p < .0001, whereas most (75%) reported engaging in watersports sometimes or many times, 2 (4, N = 1691) = 987.15, p < .0001. The distribution for bicycle riding was normal. Most brushed their teeth 2–3 times a day (71%), 2 (4, N = 1691) = 2828.32, p < .0001, and flossed 2–5 times a week or less often (59%), 2 (4, N = 1691) = 986.05, p < .0001. Self-report skill ratings were above average across all the events, ts = 4.37–35.67, p < .0001 (see Table I). A log transformation of the scores was used in analyses because of the skewed distributions.
|
Females scored significantly higher (M = 48.60, SD = 9.42) than males (M = 46.57, SD = 10.84) on the measure of egocentrism, F(1, 1689) = 16.53, p < .0001. No significant differences were observed between grade levels for egocentrism when using Scheffé’s post-hoc test.
Time 2
The students reported below average risks for all four events at Time 2, ts = –26.94 to –13.15, p < .0001 (see Table I for means). Paired t tests revealed that they perceived more optimistic risk for tooth decay from Time 1 to Time 2, but were less optimistically biased about motor vehicle accidents. There were no significant changes in risk ratings for bicycle or watersport-related injuries. However, the students were significantly less worried about experiencing tooth decay and bicycle injuries at Time 2. There were no significant differences in the magnitude of these changes between the different age levels.
Students reported more motor vehicle accidents as a driver, as well as more bicycle injuries (see Table I). Tests of age differences revealed that 10- to 14-year-old students experienced significantly more bicycle injuries since Time 1, ts = –16.92 to –4.73, p < .0001. No significant changes were observed in the sample for the number of tooth decay incidents or watersport-related injuries. The students reported riding bicycles and engaging in watersport-related activities less frequently since Time 1, but they were driving more often. Tests of age differences revealed that 13- to 15-year-old students reported driving more often since Time 1 than younger age levels, ts = –11.84 to –5.29, p < .0001. There were no other significant age differences nor were there significant changes in the frequency of tooth brushing or flossing for the sample.
The students reported above average skill for all event-related activities, ts = 6.13–31.55, p < .0001, with significant increases in perceived driving skill and dental care since Time 1 (see Table I). There were no significant age differences in the magnitude of these changes. Finally, the students’ mean adolescent egocentrism score declined from Time 1 to Time 2 with no significant age differences in the magnitude of the decline.
Hierarchical Regression Analyses
Zero order correlations were performed to assess for multicollinearity among the variables. Only results significant at the .0001 level of probability were considered in correlational and regression analyses because of the number of tests conducted and the large sample size. Significant intercorrelations ranged from .10 to .40, indicating that multicollinearity would not be an issue for regression analyses.
Hierarchical regression analyses were performed to examine the proportion of variance in changes in the students’ risk perception that could be explained by changes in adolescent egocentrism, negative experience, worry, and biased skill ratings. Age and gender were included as covariates in analyses because of their significant relations with some of the variables. Grade level and age were highly correlated (r = .96), thus only age was included in regression analyses. Residualized change scores were used instead of raw score changes because the former allows for the statistical control of differences among individuals at Time 1 (Cohen, Cohen, West, & Aiken, 2003). Residualized change scores were derived by subtracting the predicted value (via regressing Time 2 score on the Time 1 score) from the observed value. The resulting score is an estimate of the incremental change after statistically controlling for between-subject variance at baseline.
As shown in Table II, an increase in the number of tooth decay incidents within the year predicted a decline in unrealistic optimism for future decay; the semipartial correlation also was statistically significant, sr2 = .04, p < .0001. Changes in worrying about tooth decay did not contribute additional variance to perceived risk. However, an increase in the frequency of dental care predicted more optimistically biased risk ratings, sr2 = .04, p < .0001. Perceiving improvement in one’s dental care skill since Time 1 also contributed significantly to an increase in unrealistic optimism for tooth decay after controlling for the other variables, sr2 = .08, p < .0001.
|
Experiencing more bicycle and watersport-related injuries during the year predicted a decline in unrealistic optimism (sr2bicycle injury = .03, p < .001; sr2watersport-related injury = .05, p < .0001). An increase in worrying about injuries accounted for additional variance (sr2worry, bicycle injury = .05, p < .0001; sr2worry, watersport-related injury = .01, p < .01). Only one variable, perceiving improvement in one’s driving skill, predicted an increase in unrealistic optimism for motor vehicle accident, sr2 = .02, p < .01.
Hierarchical regression analyses were repeated for individual grade levels and again for individual age levels; both of which revealed the same pattern of results as for the total sample. Tests of quadratic relations were performed to assess if a curvilinear relation might be a better fit than a linear relation for the data. Tests of quadratic relations were performed within individual grade and at age levels separately and yielded statistically nonsignificant results.
Exploratory Analyses
Tests of mediational effects were performed on an exploratory basis because the concurrent assessment of variables at Time 2 precluded a true test of mediation. Partial mediation was observed only for predicting the perceived risk of tooth decay (p < .0001). Perceiving improved dental care skills was a significant mediator for the effect of increased frequency of tooth brushing on more optimistically biased risk for tooth decay from Time 1 to Time 2, z = –7.09.
Change in risk perception was also examined as a predictor of change in dental hygiene for exploratory purposes. Changes in perceived risk of tooth decay, number of tooth decay incidents, worrying about tooth decay, and perceived dental care skill were evaluated as predictors of changes in tooth brushing and flossing in two separate models. Using hierarchical regression analyses, each variable was added sequentially to the model with gender, age, and egocentrism included as covariates. All the equations were statistically significant, Fs = 15.79–34.68, p < .0001. Male students and students who reported a decline in unrealistic optimism for tooth decay were brushing their teeth significantly less frequently at Time 2 than at Time 1, ßs = .15 to .18 and –.25 to –.14, p < .0001, respectively; and students who perceived their dental care skill had improved were brushing their teeth more often, ß = .28, p < .0001. Analyses for predicting flossing revealed that a decline in unrealistic optimism for tooth decay predicted a decline in flossing, F = 10.01, p < .0001, R2 = .03, ß = –.14. Only perceiving improved dental care skill predicted increased flossing in the full model, F = 12.12, p < .0001, R2 = .07, ß = .19.
|
Discussion |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Health educators often rely on risk communication to heighten the public’s awareness of their risk of harm. However, these efforts are typically applied without a clear understanding of the mechanisms motivating people’s risk perceptions, which generally tend to be biased. As replicated in this study, for example, children and adolescents, like adults, show a pervasive phenomenon called unrealistic optimism when evaluating their risk of harm. Tests of four theoretical explanations for this cognitive bias yielded support for the social-cognitive and motivational theories that lacking personal experience with negative health events and not worrying about health threats, respectively, account for a statistically significant proportion of the variance in this pervasive bias. Worrying did not predict perceived risk of tooth decay among participants in this study; however, tooth decay is generally perceived as less physically threatening than the other events studied, which may have attenuated possible effects for worry.
Our finding for personal experience is consistent with research involving young adults (Weinstein, 1989), but contradicts DiLillo et al.’s (1998) failure to find a relation between experience with injuries and young children’s risk estimates for hypothetical risky situations. Few respondents in the latter study had actually experienced event-related injuries, which may have attenuated possible effects. Participants’ limited experience with serious consequences also might have reduced the impact for personal experience in this study. Although further research is recommended with survivors who have experienced a broader range of consequences to examine how the severity of one’s experience affects risk perception, an effect for personal experience was observed early in development and before the children in this study had acquired extensive negative experiences with the risky events.
Although experimental and cross-sectional studies have revealed effects for worry on adults’ risk perception, this is the first study known to examine worry in relation to children’s and adolescents’ risk perception. Some have argued that the relation between worry and risk perception actually supports conceptualizing risk perception as a stable personality trait rather than as a cognitive process (Gochman & Saucier, 1982). Although it makes intuitive sense that anxious people, for example, might be characteristically prone to perceive greater risks of harm, Johnson and Tversky (1983) were able to manipulate adults’ risk perception by experimentally inducing a state of worry. Furthermore, the low intercorrelations among risk ratings in this study do not support a general sense of vulnerability to threats.
Personal experience and worry were not observed to be significant predictors of the students’ perceived risk of being in a motor vehicle accident. In fact, only perceived driving skill predicted risk perception for this event. This finding suggests that until young people experience the risks of driving, they may not have a schema for their risk of harm. Further, longitudinal research is recommended with teenagers to assess if personal experience and negative cognitions influence their risk perception as they encounter more opportunities to be in an accident. Research is also recommended to evaluate the longevity and magnitude of the effects for all the variables over a longer duration of time. It may be that the effects observed dissipate with more opportunities to experience inconsequential outcomes from engaging in risky activities.
Support for the cognitive-developmental hypothesis that adolescent egocentrism causes biased risk perception was not observed in this study. There were no age differences in the students’ perceived risk of harm from the different events nor did an increase in egocentrism predict changes in risk perception. There has been only one other study to date that has tested this hypothesis and it revealed a small but positive relation between adolescent egocentrism and perceived risk of harm (Dolcini et al., 1989). The relation, however, was only significant for girls.
Exploratory analyses examining mediating effects for risk perception revealed that practicing preventive dental care predicted more optimistically biased risk perceptions for tooth decay when the students perceived improvement in their dental care skill. This finding is similar to reports of correlations between perceptions about the efficacy of protective efforts and risk judgments (Gerrard et al., 1991). Underlying both findings is people’s perceived control over preventing negative consequences when they practice preventive behaviors (Quadrel et al., 1993; Whalen et al., 1994). Further research is warranted to extend this mediating effect to other health events.
Although predicting health behavior was not the focus of this study, the proportion of variance that risk perception explained in dental hygiene was examined in this study for exploratory purposes. Interestingly, perceiving greater personal risk predicted lower frequency of tooth brushing and not increased brushing as hypothesized in HBM. Nevertheless, this finding is consistent with other studies (Gochman & Saucier, 1982; Greening, 1997; Greening & Stoppelbein, 2000) and supports the contention that children who engage in protective behaviors may believe that they are reducing their risk of harm (Janz & Becker, 1984). A longitudinal study that involves time-series analyses with multiple data points is necessary to ascertain if risk perception and health behaviors produce bidirectional effects on these respective variables. Gerrard, Gibbons, Benthin, and Hessling (1996) demonstrated such reciprocity between risk behaviors and health attitudes in a longitudinal study with adolescents. Hence, a similar pattern may apply to adaptive health behaviors and suggests that risk perception might need to be conceptualized as a dynamic rather as a static process in health belief and behavior models.
At the present time, evidence for gender differences in risk perception appears to be equivocal. Like Whalen et al. (1994), Gender differences were not observed across health events in this study. Interestingly, others have reported that girls perceive greater risks (e.g., Morrongiello & Rennie, 1998), whereas some report the opposite pattern (Gladis, Michela, Walter, & Vaughan, 1992). Perhaps, past reports of gender differences and the social psychological theory that girls are socialized to be wary of risks and boys are socialized to embrace risks (Harris & Miller, 2000) no longer apply to present day society. Changes in gender roles, including the increasing number of females playing on coeducation school football teams and serving as soldiers in the military, may have reduced the magnitude of previous differences.
Limitations
Generalizations of the findings are largely limited to Caucasians and upper-middle to upper socioeconomic status groups. Replications are needed with more diverse populations to extend generalizations to minority and lower socioeconomic groups. Generalizations are also limited to 5–10th graders because 11th graders did not participate in the follow up, owing to a scheduling conflict. Furthermore, the students reported on their personal experience with negative health consequences and on activities that they do not engage in regularly, with the exception of dental hygiene, thus precluding generalizations to indirect exposure to negative consequences and to daily activities that pose risks. Findings from cross-sectional research suggest, however, that witnessing but not directly experiencing lethal events may also heighten adolescents’ perceived risk of future harm (Greening & Dollinger, 1992). Nevertheless, it is recommended that future studies include daily risky activities, such as competitive sports to assess if extensive experiences that do not result in negative consequences might counteract the effect of that one atypical accident that could affect perceptions. Although, there is a debate about conceptualizing risk perception as a personality trait, further longitudinal studies that include measures of trait anxiety or general vulnerability are warranted to test this hypothesis. Finally, other factors not assessed, such as engaging in preventive behaviors and the onset of puberty may be assessed in future research to evaluate the contribution of possible behavioral and developmental factors in youth’s risk perceptions.
Conclusions about the effect of negative experience may be challenged by questions about the validity of the children’s reports of their risky experiences. However, research suggests that children and adolescents can provide reports that are accurate and consistent with parents’ reports (Gerrard et al., 1996; Potts, Martinez, & Dedmon, 1995). Nevertheless, future research might include documentation of injuries (e.g., medical records) to substantiate self-reports.
The relatively low internal consistency of the risk, skill, worry, and experience measures raises questions about using sum scores for these variables. However, low reliability estimates for sum scores have been observed across health events by others (e.g., Elliott, 1993; Gladis et al., 1992; Greening & Dollinger, 1992) and suggest that it would be more useful to examine event-specific variables rather than sum scores across a range of risky events when investigating risk perception. Weinstein (1980) has also recommended examining health events individually because each event presents with unique qualities that tend to be overshadowed in aggregated data. Several studies have, therefore, utilized single item measures of risk appraisals and skill ratings (e.g., Cohn et al., 1995; Gerrard et al., 1996; Gladis et al., 1992). Nevertheless, further research is recommended to develop more psychometrically sound approaches to assessing event-specific risks and related variables. Although a limitation, the findings observed in this study with event-specific items are generally similar to research using aggregated scores across events (Weinstein, 1980).
Clinical Applications
Finding empirical support for the social-cognitive hypothesis that personal experience influences risk perception suggests that exposing youth to negative consequences may help heighten their risk perception. Such approaches, even if through virtual reality, pose potential ethical concerns. However, research with witnesses of near fatal events, who show a heightened sense of vulnerability to future events (Greening & Dollinger, 1992) suggests that indirect exposure may be a viable alternative to directly exposing youth to the risks of health threats. Community-based programs whereby victims of serious negative events share their experience with young people with the goal of heightening their risk perception might have some merit here. As revealed by Bandura’s (1986) seminal research on modeling, children can learn from people who suffer negative consequences for their actions. Such exposure might also elicit negative reactions, such as worrying, which might further facilitate the effect of exposure on risk perception. Empirical research is needed, however, to determine the process of change and if utilizing an experiential approach is more cost effective than standard risk communication practices.
|
Conclusions |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Although there is evidence that risk perception influences children’s and adults’ health behavior (Morrongiello & Rennie, 1998
; van der Pligt, 1994), many question the clinical utility of these findings given the small magnitude of the effect and that risk perception may not be necessary to elicit health behaviors. However, risk perception is included in many health behavior models and is recognized as a vital component of health promotion (van der Pligt, 1994). Acknowledging the personal risk of potential health hazards is one step among a series toward adopting adaptive health-care behaviors. The adaptive response will likely not occur immediately after risk awareness, as behavioral changes tend to occur in stages. It is also important to recognize that other variables in health belief and behavior models are hypothesized to influence behavior along with risk perception and can include perceptions about response costs, availability of resources, and social support (van der Pligt, 1994). Hence, even though people may appreciate their personal risk of harm, they may not exhibit an adaptive health response because of social and monetary barriers, and so forth. But people may still need to acknowledge their risk of harm before they can start to consider these other factors that eventually lead to the adaptive response. These other factors likely account for the relatively small magnitude of the direct effect for risk perception on health behaviors, but do not minimize the importance of risk perception for health promotion.
It is important from a public health perspective for us to understand the factors that contribute to children’s and adolescents’ risk perception before they start to develop biased perceptions and long-standing maladaptive health practices. The high rate of preventable accidents in childhood and adolescence (Minino & Smith, 2001) further underscores the importance of continuing with this line of research. Early intervention and prevention can help with the long-range goal of reducing the incidence and prevalence of preventable health problems that can seriously impair one’s psychosocial functioning and add to the ever rising costs of medical care. The present findings may prove especially useful toward this goal because this is the only longitudinal study known to date that compared competing hypotheses about significant predictors of risk perception early in development. Further longitudinal research with multiple data points is recommended to continue learning about the predictors of risk perception as well as the process of change over time.
|
References |
|---|
|
TopAbstractCognitive-Developmental Theory...Social-Cognitive Theory of Risk...Motivational Theory of Risk...Cognitive Theory of Risk...MethodResultsDiscussionConclusionsReferences |
|---|
Bandura, A. (1986). Social Foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Bush, P. J., & Iannotti, R. J. (1990). A children’s health belief model. Medical Care, 28, 69–86.[CrossRef][Web of Science][Medline]
Chandler, C. C., Greening, L., Robison, L. J., & Stoppelbein, L. (1999). It can’t happen to me ... or can it? Conditional base rates affect subjective probability judgments. Journal of Experimental Psychology: Applied, 5, 361–378.
Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum.
Cohn, L. D., Macfarlane, S., Yanez, C., & Imai, W. K. (1995). Risk perception: Differences between adolescents and adults. Health Psychology, 14, 217–222.[CrossRef][Web of Science][Medline]
DiLillo, D., Potts, R., & Himes, S. (1998). Predictors of children’s risk appraisals. Journal of Applied Developmental Psychology, 19, 415–427.[CrossRef]
Dolcini, M. M., Cohn, L. D., Adler, N. E., Millstein, S. G., Irwin, C. E., Kegeles, S. M., et al. (1989). Adolescent egocentrism and feelings of invulnerability: Are they related? Journal of Early Adolescence, 9, 409–418.[Abstract]
Elkind, D. (1967). Egocentrism in adolescence. Child Development, 38, 1025–1034.[CrossRef][Web of Science][Medline]
Elliott, D. S. (1993). Health-enhancing and health-compromising lifestyles. In S. G. Millstein, A. C. Petersen, & E. O. Nightingale (Eds.), Promoting the health of adolescents: New directions for the twenty-first century (pp. 119–145). New York: Oxford University.
Enright, R. D., Shukla, D. G., & Lapsley, D. K. (1980). Adolescent egocentrism-sociocentrism and self-consciousness. Journal of Youth and Adolescence, 9, 101–116.
Gerrard, M., Gibbons, F. X., Benthin, A. C., & Hessling, R. M. (1996). A longitudinal study of the reciprocal nature of risk behaviors and cognitions in adolescents: What do you think shapes what you think, and vice versa. Health Psychology, 15, 344–354.[CrossRef][Web of Science][Medline]
Gerrard, M., Gibbons, F. X., & Warner, T. D. (1991). Effects of reviewing risk-relevant behavior on perceived vulnerability among women marines. Health Psychology, 10, 173–179.[CrossRef][Web of Science][Medline]
Gladis, M. M., Michela, J. L., Walter, H. J., & Vaughan, R. D. (1992). High school students’ perceptions of AIDS risk: Realistic appraisal or motivated denial? Health Psychology, 11, 307–316.[CrossRef][Web of Science][Medline]
Gochman, D. S., & Saucier, J. F. (1982). Perceived vulnerability in children and adolescents. Health Education Quarterly, 9, 142–155.[Web of Science][Medline]
Greening, L. (1997). Adolescents’ cognitive appraisals of cigarette smoking: An application of the protection motivation theory. Journal of Applied Social Psychology, 27, 1972–1985.[CrossRef]
Greening, L., & Chandler, C. C. (1997). Why it can’t happen to me: The base rate matters, but overestimating skill leads to underestimating risk. Journal of Applied Social Psychology, 27, 760–780.[CrossRef]
Greening, L., & Dollinger, S. J. (1992). Illusions (and shattered illusions) of invulnerability: Adolescents in natural disaster. Journal of Traumatic Stress, 5, 63–75.
Greening, L., Dollinger, S. J., & Pitz, G. (1996). Adolescents’ perceived risk and personal experience with natural disasters: An evaluation of cognitive heuristics. Acta Psychologica, 91, 27–38.[CrossRef][Medline]
Greening, L., & Stoppelbein, L. (2000). Young drivers’ health attitudes and intentions to drink and drive. Journal of Adolescent Health, 27, 94–101.[CrossRef][Web of Science][Medline]
Harris, M. B., & Miller, K. C. (2000). Gender and perceptions of danger. Sex Roles, 43, 843–863.[CrossRef]
Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education. Quarterly, 11, 1–47.[Web of Science][Medline]
Johnson, E. J., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of Personality and Social Psychology, 45, 20–31.[CrossRef][Web of Science]
McKenna, F. P., Stanier, R. A., & Lewis, C. (1991). Factors underlying illusory self-assessment of driving skill in males and females. Accident Analysis and Prevention, 23, 45–52.
Millstein, S. G. (1993). A view of health from the adolescent’s perspective. In S. G. Millstein, A. C. Petersen, & E. O. Nightingale (Eds.), Promoting the health of adolescents: New directions for the twenty-first century (pp. 97–118). New York: Oxford University.
Minino, A. M., & Smith, B. S. (2001). Deaths: Preliminary data for 2000. National Statistics Reports, 49, 1–40.
Morrongiello, B. A., & Rennie, H. (1998). Why do boys engage in more risk taking than girls? The role of attributions, beliefs, and risk appraisals. Journal of Pediatric Psychology, 23, 33–43.
Peterson, L., Gillies, R., Cook, S. C., Schick, B., & Little, T. (1994). Developmental patterns of expected consequences for simulated bicycle injury events. Health Psychology, 13, 218–223.[CrossRef][Web of Science][Medline]
van der Pligt, J. (1994). Risk appraisal and health behavior. In D. R. Rutter & L. Quine (Eds.), Social psychology and health: European perspectives (pp. 131–151). Brookfield VT: Avebury.
Potts, R., Martinez, I. G., & Dedmon, A. (1995). Childhood risk taking and injury: Self-report and informant measures. Journal of Pediatric Psychology, 20, 5–12.
Quadrel, M. J., Fischhoff, B., & Davis, W. (1993). Adolescent (in)vulnerability. American Psychologist, 48, 102–116.[CrossRef][Medline]
Ronis, D. L. (1992). Conditional health threats: Health beliefs, decisions, and behaviors among adults. Health Psychology, 11, 127–134.[CrossRef][Web of Science][Medline]
Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39, 806–820.[CrossRef][Web of Science]
Weinstein, N. D. (1984). Why it won’t happen to me: Perceptions of risk factors and susceptibility. Health Psychology, 3, 431–457.[CrossRef][Web of Science][Medline]
Weinstein, N. D. (1989). Effects of personal experience on self-protective behavior. Psychological Bulletin, 105, 31–50.[CrossRef][Web of Science][Medline]
Whalen, C. K., Henker, B., O’Neil, R., Hollingshead, J., Holman, A., & Malone, B. (1994). Optimism in children’s judgments of health and environmental risks. Health Psychology, 13, 319–325.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||