This paper examines the use of an idea referred to as expectancy as a predictor of teen alcohol use. Expectancies are concepts that a society reinforces which go on to influence a person's behavior. Current clinical and field studies show that alcohol expectancies are reasonably accurate tools in estimating future drinking patterns. This paper sets out to determine the practical applications of this knowledge in the real classroom.
HISTORY
Prior to the early 1960s, virtually no clinical studies were available on the topic of teen drinking, as literature mostly focused on negative social and moral implications of the activity (Maddox and McCall, 1964). Contrary to somewhat popular notion, however, adolescent drinking is not unique on to the last few decades. In fact, the best indicators show that "drinking among youth has been a longstanding phenomenon" that has shown no significant change over the course of the last 120 years (Barnes, 1982). In the sixties, the issue grew in prominence probably due to the rise of the counterculture and an increase in teen drinking and driving accidents. A number of pioneering social scientists set out to determine basic information about the commonalty and frequency of alcohol use in this age group. Though specific data varied from study to study due to methodology and demographics, a striking picture emerged that "alcohol use is very prevalent among teenagers and young adults." In fact, Barnes (1982) co
n
Once research findings established the basic foundations, further questions soon arose on the psychological reasons behind the increase in consumption. Though the answers are still not definitive by any means, a few commonly accepted theories arose. Teens almost consistently report one of three reasons for drinking: partying, self-expression, and anxiety (Maddox and McCall, 1964). None of this information, however, is of particular alarm. Regardless of the reason, most adolescent drinkers consume only occasionally and generally responsibly (Barnes, 1982; Finn, 1979). In fact, a few authors contend that teenage drinking can be a fairly normal step in the process of identity development (Finn, 1979). "Drinking," claims Maddox and McCall (1964), "is important for validating their self-conceptions as adults or their claims to adult status." A great deal of controversy exists on whether time spent with peers in reckless activities such as drinking is a positive aspect of the socialization process as well.
E
PROBLEM DRINKING
In the late 1970s and into the 1980s researchers begun to realize that they had not designed their studies to examine this much more destructive phenomenon of problem drinking. Differing definitions of problem drinking exist, but virtually all contemporary authors agree that it involves drinking often, and to the point of extreme intoxication. A very complex web of answers began to unfold which included, peer pressure, parental permissiveness, boredom, parental substance abuse and emotional problems (Finn, 1979). Further, it became clear that psychological perceptions of drinking influenced these adolescents' behaviors as much as the physical effects of alcohol (Christiansen et al., 1989).
EXPECTANCIES
Researchers began a continuing effort piece together a single explanation for problem drinking. Given the vast set of variables involved, this goal proved to be a formidable task. They shifted focus to the emerging theory of expectancies as a method of prediction. Goldman et al. (1987) defined alcohol expectancy as referring,
to the anticipation of a systematic relationship between events or objects in some upcoming situation. The relationship is understood to be of an if-then variety: If a certain event or object is registered then a certain event is expected to followÉ
Expectation can be inferred to have causal status in that an individual with his/her own actions, may produce a certain consequence upon noting that an if condition is fulfilled.
More simply stated, drinkers learn certain behavior about drinking from their society. When exposed to alcohol, or alcohol-related cues, they accept and act upon these understandings. Aas (1995) believes, "children learn what to expect of alcohol and those expectations are reinforced throughout our societies." These expectancies develop early in life, perhaps as early as the fourth grade (Miller et al., 1990).
A number of surveys can detect expectancies. The one most commonly used for adolescents is the Alcohol Expectancy Questionnaire- Adolescent Form, (AEQ-A). Other significant surveys include both the Alcohol Effects Scale (AES), and the Effects of Drinking Alcohol (EDA). These questionnaires involve extensive self-reporting that target expectancy in normal populations of adolescents (Aas et al., 1995). The results of these surveys can detect both the transition from non-drinker to drinker and problem drinking when analyzed using two detection constructs, or models.
Baker (1987) outlines the two basic constructs that psychologists know as the Conditioned Withdrawal Model (CWM), and the Conditioned Appetitive-Motivational Model (CAMM). The CWM can detect problem drinking based on the notion that certain expectancies are conditioned responses to avoid or escape signs of withdrawal (Baker et al., 1987). It uses craving and negative affect as indicators of an intention to drink. Users who fit into the CWM model often perceive alcohol as an "antidepressant and a euphoriant (Johnson and Fromme, 1994)." The CAMM, on the other hand, uses positive cues as its main indicator. It "proposes that positive mood states characterize the pursuit of alcohol (Johnson and Fromme, 1994)." This model associates exposure to alcohol or alcohol cues is with positive outcomes, eliciting a response to drink. "Positive alcohol outcome experiences are reliably associated with drinking practices, with heavy drinkers holding stronger expectancies than light drinkers (Johnson and Fromme, 1994).
"
However, several problems have developed in the process of using these constructs as predictive models both in clinical studies and in individual cases. A person develops expectancies in childhood and early adolescence, but they may not influence actual drinking until much later (Christiansen, et al., 1989). In other words, teenagers may have developed a set of expectancies long before, but having never been directly exposed to alcohol or alcohol-cues, they have yet to act upon them. Those unfamiliar with expectancies can think of this effect as a time bomb that the socialization process can trigger any given time. Studies have overcome this barrier by setting up a period between taking the batteries of expectancy surveys and self-reporting studies on actual use. Because of the 'time bomb' effect and study limitations, such long-term tracking can be difficult as an immediate predictor of alcohol use in the individual. Additionally, it has been difficult to find a point at which to discriminate the resp
o
CURRENT RESEARCH
Current research on expectancies is focusing on combining personal and social statistics with current knowledge on the topic as a better indicator. For both expectancy models, high-risk youngsters clearly hold higher than average expectancies (Christiansen et al., 1989.) A number of studies in recent years have focused on differentially equating individual factors and alcohol expectancies to come up with a better model. In essence, these studies are focusing on integrating expectancy theory into the classic model of risk-detection. Initial studies on certain fronts have shown to provide marginally better constructs at predicting both frequency and quantity of alcohol use. However, "a major controversy exists in the literature concerning whether personal and social expectancies predict different types of outcomes (Reese, 1994)."
DISCUSSION
Though the literature and research to date have been very effective at showing the accuracy of adolescent alcohol expectancies, I feel it is quite inadequate as to its practical use in the classroom. A few authors make vague and general references to practical utility, but never discuss the topic. I believe, based on the research findings, however, that the potential for incorporating expectancy theory into prediction method in public school is still great. In no way, would it replace the classic model of risk detection, but would serve as a highly effective tool to enhance it. Furthermore, as new findings accumulate we may determine that classic risk detection and expectancies are not so different in their conclusions.
If schools administered expectancy surveys over the span of an adolescent's academic career, then a much clearer picture of potential problem drinking would develop. Tracking over a long period would overcome the methodological problems encountered in some of the clinical research. If schools track expectancies, known drinkers who may be consuming responsibly and in moderation would be less likely to be mislabeled as problem drinkers. Similarly, the system could more clearly identify the child with the resources to effectively cover up his drinking problems. Since research has shown that most expectancies develop even before adolescence begins, perhaps if the child begins taking these surveys early in development, effective intervention into their perceptions of alcohol might be possible.
A number of problems could also arise if schools were to adopt this theory. For example, surveys or constructs provide no definite answers. If the educational system relies to heavily on this theory alone, it dooms the idea to be no better than traditional detection methods. Additionally, the educational system would have to address the issue of confidentiality. Clinical research relies on a high degree of anonymity to elicit accurate responses (Christiansen, 1989). If children complete expectancy surveys, a guarantee does not exist that they will respond honestly. This effect, however, is likely minimal for two reasons. First, the student is less apt to lie, as expectancy surveys deal with attitudes and perceptions rather then reporting actual drinking behavior. Second, as with similar issues, counselors should maintain a high level of confidentially.
The final complication of long term tracking would not be so simple to solve; few in the educational profession are likely to lack training in expectancy theory. It would be a costly effort to attempt to train counseling staffs to understand and interpret survey results within either the CWM or CAMM constructs, or any new highly specific models that are emerging.
FURTHER STUDY
Further study must begin to explore specifically the practical applications of expectancies in the educational system such as I discuss. Prior research studies have certainly laid the groundwork to show that expectancies can provide a reasonably accurate prediction and detection tool. Now, both the research and educational communities need to come together to make their use a reality. Additionally, current research on differentially equating specific personal and social factors with expectancy constructs must continue; the more accurate and precise that construct models become, the easier it becomes to use them. Finally, scientists must work to better understand "how children's backgrounds, through the socialization process, can influence their drinking patterns (Aas et al., 1995)." In other words, applying expectancy to the real world beckons a better understand of what events in an adolescent's socialization process are likely to set off the "time bomb" expectancies.
CONCLUSIONS
No research has been conducted on the practical applications of what we know about alcohol expectancies. This observation neither nor denies my thesis, but instead merits further investigation. On the basis of on the accumulated knowledge of 30 years of classic model prediction and nearly a decade devoted to expectancy research, I believe that expectancy models do have an appropriate place in today's school systems. Their use may be problematic but the long term benefits of saving the problem drinking adolescents far outweigh the potential negative consequences. More extensive professional investigation is definitely warranted.