works for different college drinkers. Results showed that all subjects tended to view alcohol's effects positively, but that heavier drinkers viewed positive effects as arousing, whereas light drinkers viewed positive effects as sedating. The study by Rather et al. (1992) represents an example of possible future trends in expectancy research. The extent to which expectancies are linked to important interpersonal relationships, such as peer networks, is another area in need of study. Moreover, there may be important differences between people and within individuals over time in the cognitive processes mediating belief and behavior (cf. Stacy et al. 1991), such as semantic structure, accessibility, encoding systems, and retrieval biases. Thus, examination of individual differences is a likely future trend. In terms of other future directions of expectancy research, more rigorous design, measurement, and theoretical models are needed (Leigh 1989; Leigh and Stacy 1991). From a general perspective, children's understanding of alcohol and alcohol effects is an area particularly ripe for future research because of the developmental and prevention implications of recent findings.

Adolescent Problem Behavior

Human development involves a process of "individuation," that is, the process of becoming independent both physically and psychologically from one's parents and family (Jessor 1984). This process involves the gradual and sometimes problematic assumption of adult behaviors (Jessor 1982, 1984; Jessor et al. 1991). Alcohol use and abuse have been associated with problems in the transition from adolescence to adulthood through a pattern in which some young people who drink early and with greater frequency are also likely to be characterized by truancy, other drug use, and precocious sexual activity (Jessor and Jessor 1977).

Since its identification, the problem behavior syndrome has been replicated, and research supports the presence of a single underlying dimension (Donovan et al. 1988). This pattern of problem behavior has been described in different psychological terms. For example, Newcomb (1987) suggested that this pattern of behavior can be conceptualized as a process of precocious development, in which young people take on adult behaviors early. This pattern results in early detachment from educational pursuits, early entry into employment, and early marriage

and childbearing. Recent data also suggest that drinking earlier in adolescence, particularly in childhood, may be associated with more problematic adjustments later in life: Typically, alcohol is first used during adolescence, although almost one-third of heavy drinkers first used alcohol in preadolescence (Windle 1991).

Personality Factors in Drinking

A problem-prone behavior pattern in adolescence has also been described as part of a broader personality style. This personality style is variously described as sensation-seeking, impulsive, unable to delay gratification, undercontrolled, or thrill-seeking, and is perhaps the only personality pattern prospectively linked to alcohol problems (Cox 1987). It has not yet been determined whether these personality patterns refer to a single syndrome or to a constellation of different personality traits. For example, Windle (1990) concluded that undersocialized aggressive behavior patterns, rather than hyperactivity per se, predate alcohol problems among young people. He suggested that a broad personality factor of behavioral undercontrol has not yet been adequately defined and that more research is necessary to determine the boundaries of different personality styles.

In a study by Sher et al. (1991), structural equation modeling was used to explore the effect on alcohol involvement of family history, alcohol expectancies, school achievement, psychopathology, cognitive functioning, and personality. As shown in figure 2, family history was not directly related to alcohol involvement, but rather exerted indirect effects on drinking through its relationship with "behavioral undercontrol" (a measure of personality that included such traits as hyperactivity, sensation-seeking, antisocial behavior, and impulsivity) and with alcohol expectancies. In addition to its indirect path to alcohol involvement through expectancies, behavioral undercontrol exerted its own direct effect. These findings are important because they represent an integration of cognitive, personality, and genetic factors that relate to alcohol

use.

Individual differences in response to alcohol, such as alcohol sensitivity, tolerance (Newlin and Thomson 1990), and stress-response dampening (Sher 1987), may also be related to undercontrolled personality. For example, Sher and Levenson (1982) demonstrated increased stressresponse dampening with low scores on the

California Personality Inventory Socialization scale (indicative of antisocial or undercontrolled behavioral tendencies). Thus, alcohol may be more reinforcing for persons with a more sensation-seeking or impulsive personality profile (Sher 1991).

Moreover, from a developmental perspective, variants of the behavioral undercontrol construct in adolescence have been consistently associated with risk for alcohol problems (Zucker 1987; Zucker in press; Zucker and Gomberg 1986). Preschoolers at risk by virtue of positive family history have been found to show hyperactivity, aggressivity, and impulsivity (Zucker and Fitzgerald 1991). Thus, the personality factor of behavioral undercontrol may be quite important in understanding alcohol use by young people (Sher et al. 1991) as well as other populations.

Social Processes That
Influence Drinking

Social influence is typically defined as the impact of the environment, where environment is broadly interpreted to include not only the physical and social setting, but also the interpersonal behavior of others, including peers and family members. Social influence can take many forms, and, as with psychological processes, different factors have different effects on different people

at different times. Also like psychological processes, social processes do not occur in isolation from other influences that motivate drinking. Social influences can be more direct, as through role modeling/imitation and social reinforcement of behavior (i.e., the defining of appropriate conduct by significant others), or more indirect, where influences on the development of values, attitudes, and other behaviors affect drinking behaviors (Kandel and Andrews 1987). It is likely that young people reinforce each others' beliefs about alcohol's effects. Moreover, genetic, psychological, and sociodemographic/family structure factors might also exert influence on the socialization process itself, a process essential in the development of behavior, including drinking behavior (Barnes 1990). Recent research has focused on the role of the family in terms of familial relationships, parental behavior, and parenting styles; the role of peer influences; and the effect of media and advertising on drinking behaviors.

The Role of the Family

Family relationships are perhaps the most frequently studied social influence on the use of alcohol. The observation that alcoholism tends to run in families has led to many genetics studies (see chapter 3) and to studies of family influences from an environmental perspective. This interest in the role of the family stems from the fact that the family represents the primary socialization agent of children in our society. Indeed, in the developmental progression that characterizes socialization of children in our society, relationships between parent and child are key during early development and may provide a basis for the selection of peer groups that become influential during adolescence (Barnes 1990).

Studies of the relationship of parents' drinking patterns to alcohol use among their adolescent offspring have yielded mixed results. A recent study by Chassin et al. (1991) investigated the relationship between parental alcoholism and alcohol or other drug use in adolescence in a community sample of teens. The relationship was also examined between parental alcoholism and the expression of symptoms by adolescents externally (as by acting out behavior) or internally (as by experiencing depression or anxiety). Adolescent children of alcoholics were found to have significantly higher levels of alcohol use and of both types of symptoms than children from nonalcoholic families. Interesting findings

were also revealed in the analysis comparing control adolescents with children of recent1 alcoholics and children of recovering alcoholics. Conduct problems in their children were more frequent among recent alcoholics than among controls, but no differences were found among recovering alcoholics versus controls. Adolescents in both recent and recovering alcoholic families, compared separately with controls, showed significantly higher tendencies to use alcohol and to have experienced an alcohol- or drug-related consequence.

Kandel et al. (1978) found that adolescents' perceptions of parental alcohol use were related to their own use and that adolescents tended to imitate the behavior of significant others (Kandel 1983). Harburg et al. (1982), however, found that adolescents were less likely to imitate parental drinking when it became extreme. Barnes et al. (1986), while finding support for a parentmodeling explanation of adolescent drinking patterns, particularly among heavy- and moderatedrinking parents, also found evidence that adolescents of abstaining mothers experienced higher rates of heavy drinking. Thus, offspring drinking does not always parallel parental alcohol use; rather, some parental drinking patterns appear to be rejected by teens as models for their own behavior.

There is also evidence to suggest that the family is a primary agent in the social control of drinking. In the analysis of a 1984 nationwide survey, Room (1989) reported on experiences of respondents relative to pressure to cut down or curtail drinking. Thirty percent of respondents reported having applied such pressure in the past year alone, and more than 20 percent of respondents said they had been recipients of such pressure at some time in their lives. These control efforts appeared primarily within the family and flowed predictably from older generations to younger and more commonly from women to men. This pattern of social control has been documented in Finnish and Soviet households as well (Holmila 1988). More recent analysis of American survey data from 1990 (Room et al. in press) suggested that, in the past decade, many more individuals were willing to pressure others to control their drinking. Although families are sometimes cited for allowing or denying alcohol

use, this research shows that the family can also function to restrain and control drinking practices.

Interpersonal influences of parents and peers were investigated by Kandel and Andrews (1987) using causal modeling techniques. They found that peer influences greatly exceeded parental influences in explaining the frequency of adolescent alcohol use. In initiation of alcohol use, however, parents were found to play a significant role relative to defining drinking norms, both through their own alcohol use and through their attitudes about the harmfulness of alcohol. Distinct racial differences in parental influence have been identified. In a study comparing attitudes and beliefs of heavy-drinking black and white teens, Ringwalt and Palmer (1990) found that black adolescents were significantly more apt than whites to appreciate the health consequences of drinking and to be concerned about the disapproval of drinking by adults in their lives. In particular, there was a tendency for black students to consider parental disapproval as an important reason to curtail drinking, while for white students, the disapproval of friends was important.

Parenting practices that can be especially potent in the development of adolescent behaviors are support (nurturance) and control (discipline). The quality of the parent-child relationship and the manner in which parents socialize children have been related to many childhood and adolescent adjustment difficulties, including alcohol use and abuse.

Parenting practices that can be especially potent in the development of adolescent behaviors are support (nurturance) and control (discipline) (Bames 1990). The quality of the parent-child relationship and the manner in which parents socialize children have been related to many childhood and adolescent adjustment difficulties, including alcohol use and abuse (for reviews, see Snyder and Huntley 1990 and Patterson et al. in press). With respect to frequency of alcohol

use, Kandel and Andrews (1987) found that perceived closeness to parents discouraged teen drinking. Barnes et al. (1986) interviewed 12- to 17-year-olds and their parents in a study of parenting factors and adolescent drinking behaviors. They found that adolescents whose parents provided high levels of nurturance and support reported fewer alcohol problem behaviors than youths who received low levels of nurturance and support. Holmes and Robins (1987) found that disciplinary practices by parents that were described as unfair, harsh, and inconsistent were predictive of alcohol and depressive disorders in adults. Thus, it is possible that parenting practices play a role in problematic drinking throughout the lifespan.

Catalano et al. (1992) reviewed the literature and analyzed data collected as part of a longitudinal study to assess racial/ethnic differences among black, white, and Asian youth in alcohol and drug use relative to family factors that included parenting practices. While family factors were significantly associated with the variety of substances used (i.e., alcohol, marijuana, cigarettes, and chewing tobacco) among fifth graders, different factors emerged as important in different ethnic groups. Predictors of fewer substances used among white students included parenting style, involving such factors as praise and clear rule setting; practices regarding monitoring and punishment; and absence of drug abuse or delinquency among siblings. Among blacks, predictors were attachment to parents and parental tendency to select the child's friends; among Asians, parental disapproval of alcohol use by the child, child residing with both parents, and absence of drug use or delinquency among siblings. Thus, consideration of ethnic differences is important in understanding the influence of parenting and other family factors on alcohol use.

Peer influences are perhaps the most powerful predictor of alcohol use among young people.

The role of parenting practices in antisocial behavior is important because of the consistent link between such behavior and adolescent alcohol use (Donovan et al. 1988). Dishion et al. (1991) examined antecedents of involvement with antisocial peers in two cohorts of boys assessed at 10 and 12 years of age. Although both

poor monitoring and lack of discipline by parents were associated with involvement with antisocial peers in both age groups, poor monitoring was found to be a significant antecedent for involvement with antisocial peers at age 12, even when the child's own antisocial behavior was taken into account.

Relationships within the family also affect the drinking patterns of adults. Bennett and Wolin (1990) investigated the family for clues to environmental factors that may increase risk or confer protection from alcoholism. Their research suggested that alcoholism becomes more harmful to family functioning, and hence is more likely to be passed to the next generation, if drinking interferes with the family's participation in such activities as dinner times, holidays, and vacations-activities referred to as “family rituals." Defined as "symbolic forms of communication between family members" (page 198), family rituals, as conceptualized by Bennett and Wolin, are repetitive actions that take on special meaning to the family. Conversely, they found that maintenance of family rituals, even through years characterized by heavy drinking, appeared to be protective in terms of alcoholism transmission. Moreover, Bennett and Wolin (1990) found that risk for alcohol problems in the adult offspring was reduced when parents deliberately engaged in family rituals. This concept has been supported by detailed studies in two generations of families in which one or both parents in the first generation were alcoholic (Bennett and Wolin 1990). The concept, however, awaits testing with larger samples and other designs.

Studies reviewed by Steinglass et al. (1987) suggest that drinking also can play a role in the family's communication and problem-solving capacity. There is some evidence that facets of marital relationships might be affected by patterns of alcohol consumption: Husbands who were episodic alcoholics were reported by Leonard (1990) to display hostile interactions when drinking. This was interpreted as suggesting a pattern of coercive control in the relationship between husband and wife.

Peer Influences

Peer influences are perhaps the most powerful predictor of alcohol use among young people (Jessor and Jessor 1977; Kandel and Andrews 1987). Kandel (1986) noted that studies of adolescent development relating to peer influence and friendships show that adolescent friends first

have in common demographic factors such as age and gender, followed by common behaviors, particularly alcohol and other drug use.

Although the manner by which peers influence one another is not entirely clear, there are two different factors that might be involved: Influence may occur as a result of similarities, such as similar alcohol use attitudes and behaviors, that develop as youths spend time together (socialization); peer influence may also occur through a process by which young people choose their friends from among other young people who also use or want to use alcohol or other drugs (selection) (Kandel 1986). Kandel's work suggested that both processes are important, and recent research programs have sought to specify peer relationships that promote or minimize alcohol and other drug use.

Swaim et al. (1989) employed path analytic techniques to assess emotional distress and peer drug associations as predictors of alcohol and other drug use in 11th and 12th graders. The set of variables measuring aspects of emotional distress (self-esteem, anxiety, depression, anger, and blame/alienation) accounted for only 4.8 percent of the variance in alcohol and other drug use, but when peer drug associations (e.g., friends using alcohol and other drugs, and friends' actions in persuading or preventing the student from using them) were added, the total variance accounted for was 43.4 percent. Further analysis revealed that peer drug associations mediated the effect of distress on alcohol and other drug use. That is, contrary to theories accounting for adolescent use of alcohol and other drugs on the basis of direct stress reduction, this research suggests that peer factors play a primary role.

Differentiating among best friends, other good friends, and same-age peers in a large sample of Irish teens, Morgan and Grube (1991) investigated peer influences in the prediction, initiation, and maintenance of alcohol and other drug use. Students were surveyed twice, with 1 year between surveys. Of the peer-related factors, the best predictor of alcohol use at the time of the second survey was the student's report of alcohol use by the best friend. Also significant, but not as strongly predictive, were students' reports of use by other friends and students' perception of approval of drinking by the best friend. Reported use by same-age peers was not a significant predictor. Morgan and Grube also looked at peer-related factors in the initiation and maintenance of drinking and found that maintenance was best predicted by reported

drinking of the best friend. Reported drinking by other friends had significant, but less, influence. Initiation of drinking, on the other hand, was best predicted only by reported use by other good friends. Use by same-age peers was not a significant predictor in either analysis. The authors concluded that the closeness of the relationship is an important factor in peer influence of drinking behavior and that different factors appear to influence initiation versus maintenance of alcohol use.

The Effects of Media and
Advertising

A number of studies have addressed the content of the mass media and media portrayals of alcohol in both advertising and fictional programming. Research in this area is difficult to assess due to the diversity of methods used to address hypothesized media effects. Although the magnitude of advertising in everyday life is large, summaries of research findings have been inconclusive. This may be due at least partially to the methodological problems posed by such research. For example, can a control group be found that has not seen alcohol advertising?

Adlaf and Kohn (1989) reanalyzed Strickland's (1983) data on students from grades 7, 9, and 11 who drank; they reconfirmed that for these youth a common factor of alcohol abuse was present, frequent intoxication contributed to abuse, and peer association had greater effects than advertising. However, a New Zealand study by Wylie et al. (1989) of 9- to 14-year-old boys concluded that "television advertising was creating strong positive associations between the companies' products and significant cultural and emotional elements, thus contributing to a climate that was not supportive of alcohol education" (Wylie et al. 1989, page 639).

Smart (1988), in a careful review of available data, concluded that the relationship between advertising and overall consumption has yet to be demonstrated. His summary of studies of advertising bans in certain communities, advertising expenditures, and experimental research showed no effects on total consumption. Reports of changes in advertising laws in Saskatchewan (Makowsy and Whitehead 1991) are consistent with this conclusion: A recent lifting of advertising restrictions had no overall effect on alcohol consumption, although drinking shifted from spirits to beer.