Social approval and facilitation in predicting modeling effects in alcohol consumption

Abstract

Purpose: An important question for alcohol abuse prevention and treatment is whether individuals with high needs for social approval, or those who drink heavily in social contexts, are particularly vulnerable to modeling effects in alcohol consumption. Methods: Male and female heavy social drinkers (N=202), as distinguished by these cognitive and situational variables, participated in a multisession dyadic modeling effects study along with a same-sex confederate model who exhibited alternating patterns of heavy and light consumption in an experimental barroom. Results: Subjects with high needs for social approval, and those who tend to drink heavily in social contexts, were particularly vulnerable to imitating directional changes in modeled drinking levels across heavy and light consumption experimental sessions. Additionally, modeling effects were revealed, including reductions in drinking levels, regardless of individual characteristics such as demographics or levels of intoxication achieved on “usual drinking occasions.” Implications: Findings suggest that individuals exhibiting high needs for social approval, and those who tend to drink heavily in social contexts, may benefit from (1) befriending lower risk models and (2) prevention and/or intervention efforts to reduce risk for substance use by reducing excessive needs for social approval and/or reducing exposure to social contexts where heavy drinking and related risk behavior is normative.

Introduction

Since the advent of behavioral psychology (Bandura, 1969) and the latter development of cognitive–behavioral and social learning theory Bandura, 1977, Bandura, 1986, modeling effects, or what has often been described as observational learning, have been referred to as the central mechanism through which individuals learn personal and social behaviors. Observational learning, or modeling effects, have also been deemed to play a major role in the development and maintenance of addictive behaviors. Additionally, since the consumption of alcohol in Western societies most often occurs in social contexts, the potential for reinforcing social consequences from alcohol consumption to contribute towards the development of problem drinking behavior has been an integral part of behavioral theory Bandura, 1986, Cox, 1990, Galizio & Maisto, 1985.

A social learning approach towards understanding the problem of alcoholism or chemically dependent behavior has encouraged experimental studies to identify reinforcers surrounding the use of alcohol. This approach has included studying events that occurred concurrently with the consumption of alcohol, how the immediate consequences of consumption might serve some reinforcing function, and what the situational determinants of drinking may include. If the parameters of reinforcement for the normal use of alcohol could be experimentally isolated and observed, steps could then be taken to prevent the development of dependencies, or to substitute other forms of reinforcement for addicted individuals.

With this evolving theory in mind, an early empirical study was conducted in an experimental “alcohol research laboratory” to examine modeling effects in social drinking that might directly impact the drinking behavior of heavy social drinkers (Caudill & Marlatt, 1975). At that time, since peer influence effects had just been identified as a leading cause of relapse to alcohol use in recovering alcoholics (Marlatt & Gordon, 1980), the study was intended to document a direct effect of modeling influences on alcohol consumption in a controlled laboratory setting. The demonstration of such influences could potentially offer information to be used in the development of strategies for relapse prevention based on a social skills training approach (Marlatt & Gordon, 1985). This modeling study (Caudill & Marlatt, 1975) demonstrated a robust modeling effect. Over the next several decades, additional modeling studies demonstrated that this effect was replicable across a wide range of individuals and differing demographic subgroups Caudill & Goldberg, 1992, Collins & Marlatt, 1981, Collins et al., 1985, Quigley & Collins, 1999. A recent meta-analysis of modeling effects research (Quigley & Collins, 1999) indicated consistent and robust modeling effects with large effect sizes. The results of the meta-analysis also suggested that the modeling effects paradigm may be a useful context to test various mediator or moderator models of modeling effects in high-risk drinkers. We suggest that identifying individuals with heightened susceptibilities to modeling effects in alcohol consumption is particularly important. According to predictions derived from social learning theory (Bandura, 1986), susceptible individuals may imitate the drinking of low- as well as high-consumption models (Caudill & Lipscomb, 1980). Robust modeling effects demonstrated in studies that compared heavy versus light consumption modeling with at-risk drinkers support this suggestion (Quigley & Collins, 1999). The current study was designed to further test this theory with a repeated measurement (heavy and light consumption) within-subjects modeling effects design.

In 1988, Chipperfield and Vogel-Sprott extended the modeling effects paradigm by examining the potential for an individual difference variable, namely one's family history of alcoholism, to mediate the modeling effect. In this study, male drinkers with a family history of alcoholism responded more to modeling effects in alcohol consumption than did those with no such history. Surprisingly, no studies since then have pursued this issue further, or examined other potential mediators of modeling effects.

The current study was designed to test whether several individual difference variables mediate the modeling effect in drinking. These variables were derived from the general modeling literature, which suggests that individuals vary in their susceptibility to modeling effects Bandura, 1969, Bandura, 1986. Perhaps one of the most widely described mediators of modeling is a strong need for social approval. Another social-cognitive and/or social-situational variable linked to increased risk behavior in higher risk drinkers is drinking for social facilitation. Social facilitation has been defined as drinking to improve one's communication with others (Goldman, Del Boca, & Darkes, 1999), or as the extent that individuals drink, or drink heavily, in social contexts (Beck, Thombs, Mahoney, & Fingar, 1995). This belief in alcohol as a social facilitator has been linked with an array of substance abuse problems, including later development of alcohol problems in adolescents, heightened levels of alcohol use and related risk behaviors in high school and college student, more adverse consequences from drinking in current drinkers, and a heightened risk for relapse in treated alcohol abusers Beck et al., 1995, Beck et al., 1993, Christiansen et al., 1989, Goldman et al., 1999, Kilbey et al., 1998, Thombs et al., 1993. For these reasons, the current study was designed to determine if at-risk heavy social drinkers who are high in their need for social approval or who report drinking for social facilitation are particularly susceptible to modeling effects in alcohol consumption.

The current study included four innovations in its experimental design. First, subjects were exposed to alternating heavy and light consumption levels across four modeling effects sessions. Second, we used a community-based sample of heavy drinkers instead of college students. Third, the current study defined the high-risk population using information on blood alcohol concentration (BAC) levels typically achieved when drinking instead of quantity by frequency measures, since these latter indices often ignore the amount of time taken to consume the number of drinks reported per typical occasion. Fourth, drinkers completed taste comparisons after each drinking session instead of during the session to enhance the naturalistic nature of the dyadic coaction social drinking context examined.