Alcohol-Related Risk Taking Among Teenagers: an Investigation of Contributing Factors and a Discussion of How Marketing Principles Can Help

William J. Burns, University of Iowa
Sarah E. Hampson, Oregon Research Institute
Herbert H. Severson, Oregon Research Institute
Paul Slovic, Decision Research
ABSTRACT - The role of risk perceptions and personality in adolescents' participation in alcohol-related risk taking (including the consumption of alcohol and alcohol use in connection with driving or riding in a car) was investigated using covariance structure modeling. The modeling analysis indicated that factors pertaining to perceptions that the benefits of using alcohol outweigh the risks, grade level, and desired independence were the only variables to directly affect participation. Personality characteristics (sensation seeking, ego control, independence, and achievement-orientation) were found to indirectly influence participation through their direct influence on adolescents' perceptions of benefits and risks associated with alcohol use. This model applied equally well to boys and girls.
[ to cite ]:
William J. Burns, Sarah E. Hampson, Herbert H. Severson, and Paul Slovic (1993) ,"Alcohol-Related Risk Taking Among Teenagers: an Investigation of Contributing Factors and a Discussion of How Marketing Principles Can Help", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 183-187.

Advances in Consumer Research Volume 20, 1993      Pages 183-187

ALCOHOL-RELATED RISK TAKING AMONG TEENAGERS: AN INVESTIGATION OF CONTRIBUTING FACTORS AND A DISCUSSION OF HOW MARKETING PRINCIPLES CAN HELP

William J. Burns, University of Iowa

Sarah E. Hampson, Oregon Research Institute

Herbert H. Severson, Oregon Research Institute

Paul Slovic, Decision Research

ABSTRACT -

The role of risk perceptions and personality in adolescents' participation in alcohol-related risk taking (including the consumption of alcohol and alcohol use in connection with driving or riding in a car) was investigated using covariance structure modeling. The modeling analysis indicated that factors pertaining to perceptions that the benefits of using alcohol outweigh the risks, grade level, and desired independence were the only variables to directly affect participation. Personality characteristics (sensation seeking, ego control, independence, and achievement-orientation) were found to indirectly influence participation through their direct influence on adolescents' perceptions of benefits and risks associated with alcohol use. This model applied equally well to boys and girls.

These findings are integrated within a social marketing framework and offer strategies to reduce alcohol-related risk taking among adolescents.

INTRODUCTION

Although alcohol use by high-school students has shown an encouraging decline in recent years, teenage alcohol consumption remains at high levels. A 1991 national survey of high school seniors (Johnston, 1992) indicated that 88% of seniors had tried drinking at least once, 54% had tried drinking in the last thirty days, and 30% report drinking 5 or more drinks at a time within the last two weeks. Particularly disturbing was the fact that almost one third did not perceive drinking four or five drinks on a nearly daily basis as a great risk. According to the U.S. Department of Education (1988) drinking and driving is common among teenagers and is the leading cause of death among people aged 16-24 years.

In this paper we begin by investigating a model of alcohol-related risk taking, in which we hypothesize that personality traits, social influences, and grade level contribute to adolescent risk taking largely through their effect on risk perceptions. We then discuss how our findings together with those of other researchers can be used to reduce alcohol-related risk taking by employing basic marketing principles.

METHOD

Subjects and Procedures

We administered our measures to an entire high school located in rural Oregon. The experimenters emphasized that the questionnaires were to be completed anonymously, and encouraged the participants to be completely honest in their responses. The experimenter remained in the classroom to answer questions and to ensure that the students worked independently. However, to protect the confidentiality of subjects' responses, experimenters did not patrol the room. Across the two days, 323 students were assessed, which is 84% of the high school's enrollment of 382 students.

Risk Perception, Social Influence, and Risk Behavior Scales

These three constructs called for ratings of 16 different activities on 12 scales. There were seven activities involving alcohol: drinking wine or wine coolers, drinking beer, drinking hard liquor, drinking five or more alcoholic beverages on a single occasion, riding in a car with an intoxicated driver, and driving a car while intoxicated. The remaining activities included other adolescent problem behaviors such as tobacco and illicit substance use, as well as more socially acceptable risk taking such as driving an all-terrain vehicle. The rating scales assessed a number of aspects of risk perception that we have examined in previous studies (e.g., Slovic, 1987) together with perceptions of social influences and reports of risky behavior: (1) If you did this activity, to what extent do you believe that you would be personally at risk for getting hurt or sick? (Personal Risk), (2) If someone your age did this activity, to what extent do you believe that he/she would be at risk of getting hurt or sick? (Risk to Others), (3) If you did this activity, to what extent would it provide you with pleasure or other benefits? (Benefits), (4) If someone your age did this activity, to what extent could he or she control the risks associated with it? (Controllability), (5) To what extent do you feel pressure from your friends to do this activity? (Peer Pressure), (6) To what extent are people who do this activity admired by their friends? (Peer Admiration), (7) To what extent can a person your age avoid doing this activity? (Avoidability), (8) If you wanted to participate in this activity, how easy would it be to do so? (Ease of Doing), (9) Would your parents approve or disapprove of your doing this? (Parental Approval). Scale 10 asked how often close friends did the activity (never, sometimes, often), scale 11 asked what percentage of same-age peers do this activity, and scale 12 asked for a self-reported frequency of involvement with the activity over the past six months (0=zero, 1=1 or 2 times, 2=3 to 5 times, 3=6 to 10 times, 4=11 or more times).

Personality Variables

Personality scales were administered immediately after completion of the risk-perception, social influence, and risk behavior scales. Sensation seeking, which peaks in adolescence, is associated with participation in a range of risky activities including using alcohol (Zuckerman, 1979). It was assessed by eight items from the Sensation Seeking Scale (Zuckerman, Kolin, Price, & Zoob, 1964). Two items from each of four subscales (thrill and adventure seeking, experience seeking, disinhibition, and boredom susceptibility) were used (alpha = .69). Frequent drug use is associated with poor impulse control (Shedler & Block, 1990) which was assessed by 20 items from Block's Ego Control Scale (alpha = .74). Adolescent problem behavior, including alcohol use, has been associated with low values for academic achievement and high values for independence (Jessor, 1984; Jessor and Jessor, 1977). These variables were assessed using five items each from Jessor's A and I scales (alpha = .88 and .77 respectively).

Structural Model of Alcohol-Related Risk Taking

To investigate how personality characteristics, social influence and risk perception contribute to alcohol-related risk taking we constructed the structural model shown in Figure 1 (the results from this figure are discussed in the next section). We hypothesized that Alcohol-Related Risk Taking could be explained by a combination of constructs representing risk perception, social influence, and personality characteristics. Specifically, Perceived Benefits vs. Risks is thought to positively influence Alcohol-Related Risk Taking. Perceived Benefits vs. Risks was measured by perceptions of personal risk, risk to others, controllability of the risk, parental approval for the activity, and benefits. Those scoring high in terms of Perceived Benefits vs. Risks tend to believe the activity to be controllable, beneficial, and approved of by parents, and also to be of low risk to themselves and others. Thus teenagers scoring higher on this factor should be more likely engage in risky behavior. This construct represents the combination of risks and benefits that Slovic has found typical of intuitive assessments of risk (Slovic, Fischoff, & Lichtenstein, 1986). Likewise, Perceived Social Influence affects positively both Perceived Benefits vs. Risks and Alcohol-Related Risk Taking. Perceived Social Influence was measured by perceptions of peer pressure, peer admiration, avoidability, and ease of doing the activity. As such Perceived Social Influence captures the various perceived contextual factors that favor engagement in risk taking. High scorers on this construct perceive the activity to be admired by their peers, feel under pressure to engage in the activity, see it as difficult to avoid, and easy to do. Hence, adolescents perceiving greater social pressures should be more inclined to view these activities as less risky and to participate more frequently.

FIGURE 1

A PROPOSED STRUCTURAL MODEL DEPICTING FACTORS THAT CONTRIBUTE TO ALCOHOL-RELATED RISK TAKING

Four personality variables and a student's grade level are hypothesized to affect participation in alcohol-related risk taking. Grade Level, and Sensation Seeking are thought to have a direct positive effect on Alcohol-Related Risk Taking while Ego Control (low impulsiveness) is hypothesized to exert a direct negative influence on such behavior. For example, those in higher grades face more opportunities to engage in such risks, those possessing greater sensation seeking needs more often pursue such opportunities, and those exhibiting greater impulsiveness more frequently participate in risky activities without proper reflection.

Personality characteristics and grade level are also hypothesized to have indirect effects on Alcohol-Related Risk Taking. Sensation Seeking positively affects Alcohol-Related Risk Taking through its positive effect on Perceived Benefits vs. Risks. For instance, high sensation seeking needs may encourage teenagers to focus more on the benefits of risky activities than on the dangers. Ego Control affects Alcohol-Related Risk Taking negatively because of its negative influence on Perceived Benefits vs. Risks. Less impulsive individuals for example, are more apt to consider long-run consequences and thus should perceive fewer benefits (short-lived) and more long term risks. Achievement negatively affects Alcohol-Related Risk Taking through its negative influence on Perceived Social Influence. Students who place greater value on academic achievement tend to gravitate more readily toward scholastic activities and receive greater encouragement from teachers and other academically-oriented students. Hence, they should be perceive less social influence to participate in risky activities. Independence positively influences Alcohol-Related Risk Taking through its positive effect on Perceived Benefits vs. Risks. Those with higher values for Independence (i.e., independence from adult authority) are more inclined to view "adult" activities positively, and to reject warnings coming from authorities. Thus, they should perceive more benefits and less risks associated with risk taking activities. Finally, Grade Level exerts a positive influence on Alcohol-Related Risk Taking through its positive effect on Perceived Benefits vs. Risks. As mentioned, students in higher grades have greater opportunities to engage in risky activities. As a result, older students are exposed (in the short term) to more information regarding the benefits than the dangers of such activities. Thus, they should perceive more benefits than risks connected with risky activities such as alcohol use.

RESULTS

The self-reports of alcohol use indicated that 61% of the younger students (i.e., 9th and 10th grades) and 81% of the older students (i.e., 11th and 12th grades) had drunk beer on at least one occasion during the past six months. Similarly, 59% of the younger students and 78% of the older students reported drinking wine, and 46% of the younger students and 70% of the older students reported drinking hard liquor at least once. Moreover, 43% of the younger students and 61% of the older students reported at least one recent incidence of drinking five or more drinks on one occasion. These levels of alcohol involvement are comparable to levels reported in other studies. The correlations between these four alcohol-related activities were high (ranging from .73 to .82) with an alpha value of .93. Hence, these four items were summed to produce an index reflecting a students's frequency and quantity of alcohol use.

For drinking and driving, 31% of the younger students and 44% of the older students reported riding with an intoxicated driver, and 7% of the younger students and 28% of the older students reported having driven while intoxicated. The correlations between these two activities was moderately high (.69). Hence, these two items were summed to form a Drinking & Driving index reflecting the amount of drinking and driving. The two measures of Alcohol-Related Risk Taking were positively skewed because of the relatively high numbers of students reporting little or no alcohol-related risk taking. Accordingly, they were transformed to a natural log scale.

The model in Figure 1 was examined using a covariance structure modeling program called EQS (Bentler, 1989) and estimates were based on elliptical rather than normal distribution theory. The chi-square statistic (c2=185.98, df=89, p<.001) indicated that the model statistically did not fit very well. However, the Bentler-Bonnet Normed Fit Index (BBNFI) was .98 (.90 or higher is considered acceptable) indicating that the model performed well in accounting for covariances among model variables. The largest contribution to the poor model fit appeared to come from improper measurement of the latent constructs Perceived Benefits vs. Risks and Perceived Social Influence. Examining measurement coefficients (all path estimates represent standardized regression coefficients) indicated that only Personal Risk, Risk to Others, and Benefits were good indicators of Perceived Benefits vs. Risks. Likewise, these coefficients suggested that the variables used to measure Perceived Social Influence did not seem to share a common factor. As a result, the model proposed in Figure 1 was modified to better conform to the data but still address the original theoretical propositions.

The revised model in Figure 2 retained the construct of Perceived Benefits vs. Risk, measured now by Personal Risk, Risk to Others, and Benefits. The construct Perceived Social Influence was dropped but two of its measures, Peer Admiration and Avoidability, were retained because of their significant contribution to Perceived Benefits vs. Risks. The model shown in Figure 2 was the product of removing variables and paths that did not make a statistically significant contribution to the prediction of Perceived Benefits vs. Risks or Alcohol-Related Risk Taking. The X2 statistic (c2=66.22, df=40, p=.01) indicated that at best the model fit only marginally well. However, the c2/df ratio was 1.7 (2 or less is considered acceptable) and BBNFI was .99 indicated that the model was reasonably consistent with the data.

As anticipated, Perceived Benefits vs. Risks played a central role in explaining the behavior of Alcohol-Related Risk Taking with a direct effect of .8 (i.e., a one standard deviation change in Perceived Benefits vs. Risks produces a .8 standard deviation change in Alcohol-Related Risk Taking). As expected, Grade Level displayed a direct effect on Alcohol-Related Risk Activities. Independence was originally thought to operate on this construct only indirectly through Perceived Benefits vs. Risks but appeared to exert a direct influence as well. However, there were no other direct effects on Alcohol-Related Risk Taking. Together these three factors (Grade Level, Independence, and Perceived Benefits vs. Risks) accounted for 72% of the variance of Alcohol-Related Risk Taking. Grade Level, Sensation Seeking, Ego Control, Independence, Achievement, Peer Admiration, and Avoidability appear to indirectly affect Alcohol-Related Risk Taking through their direct influence on Perceived Benefits vs. Risks . However, these seven factors explain only a moderate portion of the variance of Perceived Benefits vs. Risks (R2 = .46).

To determine whether the model in Figure 2 applied equally well to either sex, the sample was divided into girls (n=144) and boys (n=148) and a two-group model was estimated under the constraint that corresponding parameters were set equal to each other. Overall the two models behaved in a similar fashion.

The model has limitations that should be mentioned. First, model respecification was guided in part by considerations of fit which limits the generalizability of our findings. Replication of our findings on an independent sample is clearly needed. Second, failing to include factors representing the roles of social and parental influence probably inflated the effect of Perceived Benefits vs. Risks on Alcohol-Related Risk Taking. Lastly, a longitudinal model would better serve to capture the reciprocal relationship between perceived risks and teenage involvement with alcohol. For example, it appears reasonable that frequent involvement with alcohol may contribute to faulty perceptions of the risks that accompany this behavior.

DISCUSSION

Despite its preliminary nature, our model has several implications for interventions to reduce alcohol-related risk taking. It suggests that if we wish to decrease adolescent's involvement in such activities we need, among other things, to influence their perceptions of the risks and benefits associated with alcohol use. Our findings indicate that teaching adolescents to gain control over their impulses, encouraging their willingness to communicate with authority figures (aided by our own willingness and ability to communicate with teenagers), and heightening their motivation to do well in school may alter their perceptions of the risks and benefits of drinking. Likewise, helping teenagers to develop the social skills necessary to realize that such activities are not only avoidable but that such behavior is not admirable also may influence their perceptions of the risks and benefits. Lastly, while grade level and sensation seeking are typically not susceptible to intervention they do help indicate which adolescents are at highest risk and how we might best communicate with them.

The marketing challenge is to implement cost effective programs that influence risk attitudes towards drinking and affect behavioral change. Social marketing provides a framework in which marketing principles can be integrated with social-psychological theory and empirical findings to accomplish goals involving attitudinal or behavior change (Kotler & Roberto, 1989). What the marketing approach brings to the adoption of social change is the ability to plan in terms of the needs of specific target groups. This framework allows efforts to be trained on the needs of particular market segments and goals to be represented in terms of variables that are under our control: designing programs to encourage responsible alcohol consumption (Product), reducing psychological, social, economic, and convenience costs associated with the change of behavior (Price), providing accessible opportunities to learn about and behave in a responsible fashion (Place), and developing messages that communicate the benefits over the costs such as improved health and self-esteem (Promotion).

FIGURE 2

A MODIFIED MODEL DEPICTING FACTORS THAT CONTRIBUTE TO ALCOHOL-RELATED RISK TAKING

Product

Programs designed to influence alcohol-related risk taking need to be comprehensive in their scope and tailored to the needs of specific target groups. Comprehensiveness requires involving multiple publics (e.g., adolescents, parents, school and health officials) and addressing multiple behaviors (e.g., alcohol and drug use, smoking, unsafe sex, poor school performance). Tailoring programs requires recognizing that adolescents differ in their level of risk information, peer pressures, and psychological and social maturity. For example, adolescents at highest risk tend to have greater needs for thrill seeking, lower needs for academic achievement, associate with more deviate peer groups, possess less developed social skills, and often come from families with poorer parenting skills. Hence, media campaigns that focus only on the dangers of substance abuse (e.g., "This is your brain on drugs...") or social interaction (e.g., "Just say no!") ignore the fact that this group often seeks risky activities and saying no to substance abuse may spell rejection from their peer group. Likewise, educational programs that merely provide risk information without providing training and support for behavioral change are likely to produce only short lived results.

Price and Place

Programs must also make it as easy as possible to participate, being mindful of the psychological, social, and time commitments involved with such activities. This is especially true of high-risk adolescents and their parents. For example, these parents may be reluctant to participate because of time conflicts, apprehension about what to expect or failing to see the relevance to their lives. Programs designed to improve parenting skills not only require time and effort but may be psychologically threatening. Often the parents of adolescents having problems in school are not actively involved in school functions and hence may require they develop new relations with teachers and other school officials. Moreover, high-risk adolescents most likely have begun alienating themselves from teachers and associating with peers exhibiting risky behaviors. Their peer groups often condone and even encourage substance abuse. Hence, the physical, psychological, and social costs of quitting addictive behaviors is likely to be very high. Programs must seek to reduce these costs for parents and adolescents by not only facilitating behavioral change but by providing social support for the formation of new reference groups.

Promotion

Lastly, promotional efforts must convince adolescents, parents, and community members that high risk behavior presents a real danger to teenagers and a high cost to families and communities. Promotional messages must also offer definite measures to reduce such risk-related costs. These messages should be crafted to address the individual concerns of the community, parents and teenagers. To motivate adolescents it is especially important to not only present accurate information about such risks but to focus specifically on increasing their sense of personal threat. Adolescents already exhibiting risky behavior may be resistant to this message for a number of reasons. First, members of this group tend to be sensation seekers and hence they typically focus more on the benefits of risk taking than other teenagers. Second, they and their peers frequently engage in high risk activities without any apparent problems thus providing evidence that such behavior is at once exiting, not overly dangerous, and has consequences that are largely controllable. Finally, because substance abuse is physically and psychologically addicting prolonged involvement tends to promote denial of the inherent dangers connected with such activities.

Research is needed to determine effective means to counter resistant beliefs concerning the personal risk of alcohol-related risk taking. Investigators involved with AIDS intervention research have found (Kelly et al., 1991) that getting individuals to acknowledge that high risk behavior places them in genuine danger is accomplished more easily when this message comes from individuals that are highly regarded by their peer group, when they are repeatedly reminded of the value of changing their behavior, and when they have easy access to a social support system promoting behavioral change. Hence, at the national level messages concerning the dangers of substance abuse must come from highly credible sources (e.g., musicians that have used drugs and have quit), must convey personal hardship, must present a realistic way out of the problem, and must be repeated often. Likewise, at the local level peers that are in the process of changing their high risk activities can be effective role models. They can also serve to generate discussion and encouragement among their friends to seek help in changing their high risk activities.

REFERENCES

Bentler, Peter M. (1989), EQS Structural Equations Program Manual. Los Angeles, CA: BMDP Statistical Software.

Jessor, Richard & Shirley L. Jessor (1977), Problem Behavior and Psychosocial Sevelopment: A Longitudinal Study of Youth. New York, NY: Academic Press.

Jessor, Richard (1984), "Adolescent Development and Behavioral Health," in J. D. Matarazzo, S. M. Weiss, J. A. Herd, & N. E. Miller (Eds.), Behavioral Health: A Handbook of Health Enhancement and Disease Prevention. New York, NY: J. Wiley & Sons.

Johnston, Lloyd (1992), Report Issued by the University of Michigan's News and Information Services Regarding Teenage Substance Abuse on January 27, 1992.

Kelly, Jeffrey, Janet S. St. Lawrence, Yolanda E. Diaz, L. Yvonne Stevenson, Allan C. Hauth, Ted L. Brasfield, Seth C. Kalichman, Joseph E. Smith, & Michael E. Andrew (1991), "HIV Risk Behavior Reduction Following Intervention with Key Opinion Leaders of Population: An Experimental Analysis," American Journal of Public Health, 81, 2 168-171.

Kotler, Philip, & Eduardo L. Roberto (1989), Social Marketing: Strategies for Changing Public Behavior. New York: The Free Press.

Severson, Herbert H., Paul Slovic, Sarah E. Hampson, & Linda Schrader (1990), "Adolescent Risk Perception: A Measure to Further our Understanding of Tobacco and Drug Use." Hygie, 9, 27-29.

Slovic, Paul (1987), "Perceptions of Risk." Science, 236, 280-285.

Slovic, Paul, Baruch Fischoff, & Sarah Lichtenstein (1986), "The Psychometric Study of Risk Perception." In V.T. Covello, J. Menkes, & J. Mumpower (Eds.), Risk Evaluation and Management. New York: Plenum.

Shedler, Jonathan, & Jack Block (1990), "Adolescent Drug Use and Psychological Health: A Longitudinal Perspective." American Psychologist, 45, 612-629.

U. S. Department of the Health and Human Services (1990), Seventh Special Report to the U. S. Congress on Alcohol and Health.

Zuckerman, Marvin (1979), Sensation Seeking: Beyond the Optimal Level of Arousal. New York, NY: John Wiley & Sons.

Zuckerman, Marvin, Elizabeth L. Kolin, Leah Price, & Ina Zoob (1964), "Development of a Sensation-Seeking Scale." Journal of Consulting Psychology, 28, 477-482.

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