Presented at the WeberStateUniversity
Undergraduate Research Symposium
March 2005
Risk Taking In Adolescence: Scaling the Problem
JoshuaSchillo and MichaelThorn
Social and Behavioral Sciences
EricAmsel and LeighShaw, Supervisors
Abstract
Frequency and Intention measures of risk taking behavior were distributed to Psychology students. Results replicated previous findings that forms of risk taking are correlated into a single index of risk taking vulnerability, but only for the Frequency scale. The Intention scale revealed more subtle and differentiated reasons for risk taking behavior that were not captured by the Frequency scale. Results suggest that the Intention scale may be the more sensitive and appropriate measure of risk taking.
Introduction
Researchers have typically measured adolescent and young adult risk taking by asking them the frequencywith which they engage in various behaviors, such as unsafe sex (Jessor et al., 1995). Frequency data may be requested directly or recorded on a scale (e.g., never, rarely, moderately, regularly, often). These measures assume that risk taking is a continuous (ratio or interval) scale, which allows for statistical comparisons across risk domains. For example, Jessor et al. (1995) found positive correlations between adolescent and young adult drinking, smoking, drug use, and unsafe sexual practices. Irwin and Millstein (1986) reviewed the research and concluded that a strong relation exists between adolescents’ risky drug use, alcohol use, sexual practices, and dangerous driving.
Evidence of consistency across risk domains may bequestioned for two reasons, each of which addresses the inadequacy of measuring the frequency of risk taking. First, an individual may engage in similar behaviors across risk domains for different reasons. A person may enjoy and seek to engage in certain risk behaviors, but may try other behaviors simply out of curiosity. The frequency of risk behaviors may be correlated, however, the scale fails to capture a difference across domains in the individual’s intention to engage in risk taking.
Second, risk tasking consistency may be artificially highif there are many in the sample who never had an opportunity to engage in risky behavior. An individual may not engage in risky sex or dangerous driving simply because they do not have a partner or a car! Such individuals only inflate the correlation across domains because the measurement scale is based exclusively on the frequency and not the intention of behavior.
The goal of the study was to compare adolescents' and young adults' risk taking across domains (i.e., sex, alcohol, drugs, gambling, driving) by assessing the behavior as a frequency and an intention. The intention measure distinguishes those individuals who seek to avoid risks (Avoidant), lack opportunities to engage in risks (Opportunist), are inquisitive (Curious) and actively seek out risks (Risk-seeker).
Method
Participants
The sample included 110 students (54% Male, 46% Female) in lower- and upper-division psychology courses, who received extra credit for participating. Participants averaged 20 years old (sd = 1.63 years), and most were Freshmen (54%).
Procedure
Participants completed a 10 minute questionnaire in their classroom. The questionnaire requesteddemographic (age, sex, and student status and whether participants spent junior year of High School in Utah or outside) and risk taking information. TimeOrientation of the questionnaire was varied between-subjects. Half of the participants answered questionsabout their risk taking in the “Past Year”; the otherhalf about their “Junior Year in High School.” Therisk domain (i.e., sexual practices, alcohol use, druguse, gambling, and risky driving) was varied within-subjects. In each domain, three risk-taking activitieswere specified (e.g., unprotected intercourse, intercourse with a stranger, sexual activity withsomeone with an unknown history). For each riskactivity in each risk domain, participants were askedto record their frequency of engaging in such activityand to rate that frequency on a 5-point scale (e.g.,none, rarely, multiple times, regularly, frequently).Finally, for each risk domain, participants were askedto select one of the following categories that bestcharacterized their intention with regard to theiractivity in the risk domain.
Avoidant Risk Taker: I did not and I never wouldhave engaged in any of these risk behaviors because I felt that these behaviors were very dangerous for my health, welfare and violated my sense of values.
Opportunistic Risk Taker: I did not engage in any of these risky behaviors, but I might have if the timing or circumstances had been different. I didn’t really think about the consequences of these behaviors, and I wasn’t too worried about them anyway.
Curious Risk Taker: I tried these risky behaviors a few times because I was curious and the opportunity presented itself. I didn’t brag about it to my friends and I wasn’t trying to look cool or anything, I just wanted to know what it felt like.
Risk Seeker: I was a person who participated in these behaviors and I hung out with others who alsodid so as well. I wanted to engage in these behaviorsand I never hid having done them from those who Iwanted to know.
Results
Approximately half of the participants did not record the actual frequency of their risk taking behaviors and, as a result, these scores were dropped from the analyses. The other scales were analyzed separately and then compared.
Frequency Scale: The mean frequency of risky behavior was low (Figure 1). Comparisons across domains showed more risky Driving than Alcohol use and more of the latter than other risk behaviors, F(1,104)=77.43, p<.001.
For each risk domain, students’ mean risk frequency was analyzed by Gender and Time Orientation. There were no effects of Time Orientation. Females engaged in less gambling (Mean=1.12) and risky driving (Mean=2.01) than males (Means=1.32 and 2.31, respectively) (ps<.05).
Correlations across risk domains were positive and mostly significant (Table 1). A factor analysis revealed a single risk taking score for each participant.
Intention Scale: The percentage of participants in each intentional risk category provided insights into the reason for the low risk taking frequency score (Figure 2).
Low frequency risk domains differed in the percentage of risk Avoiders (Gambling = 49%, Sex = 62%, and Drugs= 82%). A sizable minority of Opportunists never engaged in risky Gambling (20%) and Sex (16%) because it was unavailable to them, not because they were disinterested. In higher frequency risk domains, the percentage of Risk-seekers was low. The percentage of those who sought out risky Driving (29%) was actually lower than who were merely Curious (33%). Also, the percentage of Risk-seeking Alcohol users (19%) was not substantially different than those who were just Curious (10%)
The frequency of Intentional Risk Categories were summed over domains. Only a minority (39%) of participants were consistent across risk domains (i.e., selecting the same category for 4 or 5 domains) and most of them (93%) were Avoiders. Each Intentional Risk category was analyzed for the effects of Gender and Time Orientation. For Gambling, there were fewer male Opportunists and more male Risk-seekers this past year than during their junior year of high school.
Scale Comparison: The risk taking factor scores, derived from the Frequency Scale, were correlated with the frequency of each of the four Intentional Risk Taking categories, independent of age, sex and junior year spent outside Utah (Table 2). The risk factor score was strongly correlated with Avoiding (negatively) and Seeking (positively) risks. However, it was not correlated with Opportunist and only weakly correlated with Curious categories, suggesting that these categories of risk taking are not well detected by the frequency measure.
Discussion
The study examined the traditional Frequency measure of risk taking and compared it to a new Intention-based one. The results of the Frequency measure confirmed many of the findings from the adolescent risk taking literature. Although the frequency of risk taking was low, there were significant positive correlations across risk domains (sex, alcohol, drugs, gambling, driving) and each participant could be given a single score to represent their risk taking tendency.
In contrast to the results from the Frequency measure, the Intention measure revealed more subtle and differentiated reasons for participants’ risk taking. Domains in which risk taking was infrequent (sex, drugs, gambling) were comprised of those who avoided (Avoiders) risks and who had no opportunity to engage in them (Opportunist). Domains in which risk taking was more frequent (alcohol, driving) were comprised of those who were seeking risks (Risk Seeker) and merely inquisitive (Curious).
Moreover, there was evidence of more risk taking variability across domains with the Intention than the Frequency measure. There was variability in the percentage of participants in each category across domains. Also, only 39% of participants consistently had the same risk taking intention across domains. Finally, the general risk taking score better predicted participants’ status as risk Avoiders than Opportunists and better predicted their status as risk Seekers than Curious, despite the pairs of categories being similar in risk taking behavior.
The results challenge the assumption (Irwin & Millstein, 1986; Jessor et al., 1995) that risk taking is a general trait underlying and motivating adolescent risk behavior across different risk domains. The results of the intention measure paint a picture of adolescents who have different reasons for the different risk behavior in which they engage. Future research should examine the reasoning of adolescents in each intentional category.
References
Irwin, C., and Millstein, S. (1986) Biopychosocial correlates of risk-taking behaviors during adolescence: Can the physician intervene? Journal of Adolescent Health Care, 7, 82-96.
Jessor, R., Van Den Bos, J., Vanderryn, J., Costa, F.M., & Turbin, M.S. (1995). Protective factors in adolescent problem behavior: Moderator effects and developmental change. Developmental Psychology, 31, 923-933.