Resisting peer pressure: characteristics associated with other-self discrepancies in college students' levels of alcohol consumption.

Abstract

Since college undergraduates tend to increase their use of alcohol to match what they perceive to be normative, the assumption has been that students who believe that others on campus drink more than they do (a common misperception) are in a vulnerable position. Taking a different perspective, we consider large other-self discrepancies in levels of alcohol consumption as indicative of a capacity to resist situational pressures that favor drinking. OLS regression was used to assess the relationship between student background characteristics, self-presentational tendencies, and a gender-specific other-self gap measure. Overall, those individuals who drank closest to what they regarded as typical for same-sex peers at their school were students high in public self-consciousness with a family history of alcohol abuse and males who exhibited a tendency toward cross-situational variability. Students not affiliated with the Greek system who consciously limited their alcohol intake to avoid negative outcomes, on the other hand, drank substantially below what they perceived to be normative for their gender, suggesting that they were the most able to resist peer pressure.

**********

Given the negative consequences associated with the abuse of alcohol on college campuses, many institutions now have specific policies designed to reduce students' levels of alcohol consumption (Wechsler, Kelley, Weitzman, Giovanni & Sebring, 2002). Despite this, the rate of heavy, or binge drinking, has remained relatively stable at around 44%. Moreover, both the percentage of frequent binge drinkers and drinkers who report consuming alcohol for the explicit purpose of becoming intoxicated have increased since the early 1990s (Wechsler, Lee, Kuo, Sebring, Nelson & Lee, 2002).

Although students drink for a variety of reasons (Baer, 2002), peer pressure plays an important role in maintaining these patterns. Peer pressure has three forms: explicit offers of alcohol, role modeling, and social norms (Borsari & Carey, 2001). In this paper we focus on the latter type of social influence.

Across analyses, measures of common campus drinking practices, often constructed by asking survey respondents to estimate how much alcohol the "typical" student at their school drinks (e.g., Baer, Stacy & Larimer, 1991; Wood, Read, Palfai, & Stephenson, 2001), are strongly associated with students' personal drinking habits (see Borsari & Carey, 2001 for a review of this literature). Since they are based on the behaviors of non-intimates whose approval and friendship has yet to be obtained, conceptualizing drinking norms in this manner captures the essence of the concept of peer pressure (Shore, Rivers & Berman, 1983). Students who see heavy drinking as a common activity at their school are likely to increase their levels of alcohol consumption in order to gain social acceptance and avoid negative peer evaluations. Since they tend to overestimate the amount of alcohol consumed by others on campus this serves to perpetuate abusive drinking practices (Baer, Stacy & Larimer, 1991; Perkins & Wechsler, 1996) that may not coincide with their underlying attitudes (Prentice & Miller, 1993; Schroeder & Prentice, 1998).

Given their success in reducing alcohol abuse on some college campuses (Perkins, Haines & Rice, 2005), norm corrective initiatives, which provide students with accurate information about how much other students are drinking, have been the subject of much discussion within the substance abuse literature. This approach is based on the assumption that students who perceive large other-self discrepancies in levels of alcohol consumption are in a vulnerable position and are highly likely to benefit from this type of intervention (Borsari & Carey, 2001). On the other hand, substantial gaps between students' own drinking and what they believe to be normative may reflect a capacity to resist peer influence.

In their meta-analysis of studies on college drinking norms, Borsari and Carey (2003) identify two individual-level characteristics associated with other-self discrepancies in levels of alcohol consumption--Greek membership (non-Greek affiliates > Greek affiliates) and gender (female > male). While Greek participants exhibit smaller other-self discrepancies because they recognize the fact that they drink more than other students (Borsari & Carey, 2001; 2003), the source of the relatively large gap scores observed among women is less clear.

Women may report greater discrepancies between how much they think others are drinking and their own levels of alcohol consumption because they use men as a frame of reference when responding to questions about the typical student's drinking habits (Borsari & Carey, 2003; Korcuska & Thombs, 2003; Lewis & Neighbors, 2004). Alternatively, the gender difference in other-self gap scores may reflect a greater susceptibility to peer pressure among males. In their longitudinal analysis of other-self discrepancies in perceived comfort with campus drinking practices, Prentice and Miller (1993) found that males were more likely than females to adopt attitudes toward alcohol use that matched what they believed to be normative. Women are also more likely than their male counterparts to state that they would be able to resist situational pressures conducive to drinking in a variety of hypothetical situations (Shore et al., 1983). Presumably this is due to the fact that men experience more pressure from others to drink. Students themselves acknowledge this gender difference. They also believe that women are more inclined to suffer negative consequences from excessive drinking (e.g., rape or sexual assault), which may make it easier for females to limit their levels of alcohol consumption, even when they regard doing so as deviant (Suls & Green, 2003).

More generally, college undergraduates who fear alcohol's negative effects may find it easier to resist peer pressure. Many students indicate that they consciously minimize their drinking in order to avoid the risks associated with alcohol intoxication, even on campuses where heavy drinking is common. Frequently given rationales for not drinking to excess include concerns about health, safety and mental alertness; the expense of alcohol (Slicker, 2001); and patterns of familial socialization (Greenfield, Guydish & Temple, 1989).

Consistent with these findings, adolescents who report that their parents abstain from alcohol or drink moderately are less likely than their peers to abuse alcohol (Hawkins, Catalano & Miller, 1992). Moreover, parents may reduce their children's risk for alcohol abuse by making explicit their disapproval of this behavior. Relative to classmates whose parents have less prohibitive attitudes, adolescents with parents who openly oppose youth drinking are less likely to have friends who use alcohol and are more resistant to situational pressures that facilitate drinking (Nash, McQueen & Bray, 2005; Wood, Read, Mitchell & Brand, 2004).

Social-psychological attributes may also influence individuals' abilities to resist peer pressure. Conforming to prevailing social norms increases the likelihood that one will be viewed favorably. The process through which people seek to maximize positive evaluations--by dressing in a particular way, by using props that convey status and/or competence, by effectively fulfilling one's social obligations, and by meeting others' expectations more generally--is referred to as impression management (Goffman, 1959). While we all have a stake in conveying favorable impressions to others and thus avoiding the experience of negative emotions such as embarrassment (Goffman, 1959), people vary in the extent to which they are concerned about how others regard them and in their willingness to modify their behavior to obtain positive feedback (see Buss, 1980 for a review of this literature).

Fenigstein, Scheier and Buss' (1975) public self-consciousness scale is frequently used to measure the degree to which individuals worry about others' reactions to their public performances. This index consists of seven items (e.g., "I am concerned about the way I present myself.") that assess how readily people perceive themselves as the likely objects of scrutiny. Since people high in public self-consciousness are predisposed to view themselves as the objects of others' attentions, they are especially attuned to external, situational standards (Buss, 1980).

Two recent studies indicate that public self-consciousness may affect students' drinking habits. In their analysis of college students' impression-management strategies, Martin & Leary (2001) found that males high in public self-consciousness were especially likely to report engaging in risky behaviors, including alcohol use, for self-presentational reasons. A second study, by Park, Sher and Krull (2006), further suggests that Greek participation may moderate the relationship between gender, public self-consciousness and drinking. In this analysis, public self-consciousness had no effect on levels of alcohol consumption among women or among students unaffiliated with the Greek system, but fraternity members who exhibited this attributional tendency drank significantly more than other students.

Students who habitually alter their behaviors to meet the needs of the immediate situation, who Snyder (1974) originally termed high self-monitors, may also use alcohol to convey favorable impressions in public settings. The self-monitoring scale (Snyder, 1974) has multiple dimensions and includes items that measure social comparison, acting ability, and inconsistency in behavior across social contexts (Lennox & Wolfe, 1984). While self-monitoring and public self-consciousness are correlated (Turner, Scheier, Carver & Ickes, 1978; Santee & Maslach, 1982; Tomarelli & Shaffer, 1985), they are distinct in that the former construct emphasizes behavior over perception by emphasizing people's tendencies to adopt strategic self-presentational strategies, which may include heavy drinking.

Using a variant of Snyder's (1974) self-monitoring scale termed concern for appropriateness (Lennox & Wolfe, 1984), Wolfe, Welch, Lennox, & Cutler (1985) showed that students geared towards gaining social approval drank more than other individuals if they perceived substance use to be common. Novak and Crawford (2001) found a similar moderating relationship between cross-situational variability, a component of concern for appropriateness, perceived campus drinking norms and levels of alcohol consumption. College undergraduates high in concern for appropriateness were also more likely than other students to describe their use of alcohol as being motivated by peer influence (Wolfe, Lennox & Cutler, 1986).

Alcohol restrictive patterns of familial socialization--a factor associated with limited drinking in the study by Greenfield et al. (1989) mentioned earlier--may counteract this tendency. Although they did not include measures of self-presentational concerns in their analysis, Wood et al. (2004) found that parental attitudes non-supportive of drinking reduced the strength of the relationship between perceived peer norms and negative alcohol-related consequences among a sample of recent high school graduates.

We include a comparable measure of parental attitudes in this study, along with indicators of family members' drinking patterns, specific reasons for which students might limit their drinking similar to those identified earlier, measures of self-presentational concerns and strategies, gender, and Greek affiliation. Our purpose is to examine the relationships between these characteristics and students' susceptibility to peer pressure, measured as other-self discrepancies in levels of alcohol consumption. Given the tendency for females to use males as a referent when responding to questions about common campus drinking practices (Korcuska & Thombs, 2003; Lewis & Neighbors, 2004), we assess the relative effects of our independent variables on a gender-specific measure of the gap between perceived campus drinking norms and students' personal drinking behaviors. Drawing on the literature reviewed in the preceding section, we also test for the following moderating influences: self presentational tendencies by gender, self-presentational tendencies by familial socialization, and concerns about alcohol-related sexual violence by gender. Since it seems likely that both males and students affiliated with the Greek system would find it more difficult to limit their drinking, even when that is their intention, we also examine interactions between these variables, a general measure of reasons for not drinking, and OSG scores in a final set of analyses.

METHODS

Participants

During the fall of 2002, the authors administered a comprehensive survey form (including measures of students' demographic characteristics, alcohol use, and a range of social-psychological indicators) in a number of lower-level social science courses at a medium-sized, private Midwestern University. Although all of the students present in the classes in which the survey was given opted to complete the questionnaire, there was the usual rate of absences (about 5-10% of students per session) across classes. This, along with the non-representative nature of our sampling frame, must be taken into consideration when interpreting the results of this survey. In total, 293 undergraduate students of traditional college age (mean = 19.0) completed the survey form. Since students who abstain from alcohol altogether tend to be motivated by different factors than individuals who drink moderately (Greenfield, Guydish, & Temple, 1989; Slicker, 2001), we dropped all nondrinkers (n = 87) from the study, yielding a sample size of 206.

Not surprisingly given our sampling strategy, we over-represented underclassman (77% = freshmen or sophomores). The gender composition of the sample (54% female) was, however, close to that of the undergraduate population at this university. Since the student body at this school is predominantly white, in order to protect the anonymity of any racial or ethnic minorities who completed the questionnaire, we did not ask respondents to report their race. Like most students enrolled at this institution, the majority of the survey respondents (82%) reported that they lived on campus. Another 14% of the sample lived in off-campus apartments, alone or with one or more non-relatives.

Measures

Gender and Greek Affiliation. Gender and Greek participation were measured as the dummy variables female (0 = male, 1 = female) and Greek (0 = not affiliated with Greek system, 1 = Greek affiliate).

Familial Socialization. A measure of parental disapproval of alcohol use was based on respondents' answers to the following question: "My parents disapprove of me drinking." Scores on this variable ranged from 1 "strongly disagree" to 4 "strongly agree." A second family background variable, assessing parents' and other relatives' drinking habits, was constructed by adding students' responses to three variables concerning the drinking behavior of their father, their mother, and other relatives. Each question (e.g., "My father doesn't drink too much") was scored on a four-point scale ranging from 1 "strongly disagree" to 4 "strongly agree", with high scores indicating low levels of family drinking. Scores on this variable ranged from 3 to 12 (Alpha = .82).

Public Self-Consciousness and Self-Presentational Tendencies. Public self-consciousness was measured using the seven relevant items from Fenigstein et al.'s (1975) self-awareness inventory. Diverging from the standard five-point scale used to score these questions (0 = "extremely uncharacteristic" to 4 = "extremely characteristic"), responses to the public self-consciousness items ranged from 1 "strongly disagree" to 4 "strongly agree", with possible scores on this measure ranging from 7 (low PSC) to 28 (PSC). Similar to the values obtained in studies assessing the reliability of this instrument using the standard scoring format (Carver and Glass, 1976; Fenigstein et al., 1975; Turner et al., 1978; Vleeming & Engelse, 1981), the Alpha coefficient for the public self-consciousness scale was .81 among our sample, showing a moderately high degree of internal consistency among the scale items.

Two additional measures, attention to social comparison information (ATSCI) and cross-situational variability (CSV), were constructed using the relevant items from Snyder's (1974) self-monitoring scale. Snyder (1974) defined self-monitoring as the degree to which individuals successfully regulate the impressions they convey to others by modifying their behaviors to meet the demands of the immediate situation. Based on the results of four studies, Lennox and Wolfe (1984) conclude that that the ASCI and CSV measure tendencies distinct from self-monitoring as initially conceptualized by Snyder (1974) in that they are associated with social anxiety and do not necessarily predict successful impression management. The ATSCI subscale captures the degree to which individuals seek social acceptance (e.g., "My behavior often depends on how I feel others wish me to behave.") and root their perceptions in the actions of others (e.g., "It is my feeling that if everyone else in a group is behaving in a certain manner, this must be the proper way to behave."). The measure of CSV, on the other hand, emphasizes variability in behavior across social settings (e.g., "Different situations can make me behave like very different people," and "I am not always the person I appear to be.") and thus a potentially more manipulative and less authentic style of self-presentation. Although these indexes are positively correlated and can be combined into a single measure (the concern for appropriateness scale) that represents a tendency toward conformity (Lennox & Wolfe, 1984), they are somewhat different in focus and can also be treated separately (Lennox & Wolfe, 1984; Wolfe et al., 1985).

We opted to examine the latter two subscales individually so that we could distinguish between the effects of a tendency towards acquiescence (social comparison) and a more strategic and manipulative self-presentational style (cross-situational variability) on reactivity to perceived drinking norms. Interestingly, Lennox (1984) found little evidence of relationship between cross-situational variability and public self-consciousness, suggesting that self-presentational concerns do not necessarily translate into the strategic manipulation of public behavior. Since people concerned about the quality of their public performances are likely to act in accordance with others' expectations, there should be more overlap between measures of public self-consciousness and attention to social comparison information.

Items for both the ATASI and CSV subscales were scored using the four-point scale ranging from "strongly disagree" to "strongly agree" and were coded (or recoded) so that high scores indicated the presence of the particular tendency under consideration. Possible scores on the measure of ATSCI in this analysis ranged from 12 to 48, and scores on the CSV index ranged from 7 to 28.1 Despite our use of a more limited coding scheme (versus the standard six-option response set), each measure yielded an Alpha coefficient above .76 (ATSCI = .80 and CVS = .78), exhibiting a degree of reliability comparable to that obtained in previous research (Lennox & Wolfe, 1984; Novak & Crawford, 2001).

Reasons for not Drinking. In order to test the hypothesis that females find it easier than males to resist peer pressure because they are more likely to fear becoming the victims of sexual violence, we included a measure of the extent to which students perceived sexual assault or rape to be a likely consequence of drinking (1 = "not at all likely" 2 = "somewhat likely" 3 = "very likely"). This question is less than ideal since its interpretation is certain to vary by gender (with women thinking in terms of victimization and men perpetration and/or accusation). It is, however, the only item in our database that addresses the issue of alcohol-related violence.

Drawing on the literature reviewed earlier, we also constructed a composite index focusing on whether students sought to minimize their consumption of alcohol in order to avoid negative outcomes. This measure was based on the sum of respondents' answers to five questions asking them to report the degree to which they limited their drinking: for health reasons, because they fear alcoholism, to avoid getting into trouble, to avoid negative effects on their future careers, and because alcohol is expensive.

Each item was scored using the four-point scale ranging from 1 "strongly disagree" to 4 "strongly agree." Scores on this measure ranged from 5 to 20 (Alpha = .78).

Other-Self Gap Scores. Our dependent variable, a gender-specific measure of other-self discrepancies in levels of alcohol consumption, was constructed in the following manner. Students responses to the a set of three questions concerning how much a "typical" male at their school (for men), or how much "typical" female at their school (for women), drinks during an average week; how many drinks a "typical" male/female consumes at an average sitting; and how frequently a "typical" male/female drank to the point of intoxication during the month prior to the completion of the survey were added together into a composite index. Then a measure of students' drinking behaviors was constructed using the same questions regarding their personal levels of alcohol consumption. In a third step, other-self gap (OSG) scores were computed by subtracting students' composite drinking scores from their scores on the same-sex descriptive norms measure, such that high scores indicated that they perceived that others on campus drank more than they did.

The following hypothetical cases illustrate this procedure. A male reported that he consumed an average of 12 drinks per week, 6 drinks per sitting, and drank to the point of intoxication 8 times during the month prior to completing the survey. He further indicated that the typical male on campus drank an average of 21 drinks per week, 7 drinks per sitting, and drank to the point of intoxication 12 times during the past month. His other-self gap score would then be 40 - 26 = 14. Based on this formula, a female with identical self-reported drinking habits who believed that the typical woman at her school consumed an average of 8 drinks per week, 3 drinks per sitting, and was intoxicated 9 times during the previous month would have an OSG score of -6.

For comparative purposes a second discrepancy measure was computed using a non-gender specific index of campus drinking norms (i.e., drinks per week consumed by the typical student, with no gender specified; drinks consumed by the typical student per sitting, and time the typical student drank to intoxication during the past month).

RESULTS

Descriptive statistics are presented in Table 1. Consistent with previous studies (e.g., Baer, Stacy & Larimer, 1991; Perkins & Wechsler, 1996), the positive mean score on both the gender-specific and general OSG variables indicates that respondents, on average, tended to overestimate other students' use of alcohol. In fact, additional analyses (data not shown) revealed that, across the two measures, 75% of the sample reported that the "typical" student drank more than they did. As expected, as shown in Table 2, women had significantly higher scores on the general, but not the gender-specific, OSG measure.

The relationship between Greek participation and OSG scores was also smaller when a gender-specific measure of drinking norms was used to construct this variable (Table 2), primarily because males affiliated with the Greek system had such high estimates of other men's levels of alcohol consumption (data not shown). Nonetheless, as shown in Table 2, students unaffiliated with the Greek system were, overall, more likely than Greek participants to believe that other students drank more than they did. Additional statistical tests (data not shown) indicated that this difference decreased substantially when students' levels of alcohol consumption were held constant.

Prior to running any higher-level analyses, we examined the degree of overlap between measures of public self-consciousness (PSC), attention to social comparison information (ATSCI) and cross-situational variability (CSV), three of our key independent variables. While measures of ATSCI and CSV were moderately related, the correlation between these two indices (r=.30, p < .001) was somewhat smaller than that observed by Lennox and Wolfe (1984). Moreover, ATSCI, but not CSV, was strongly associated with PSC (r = .04, p = .619; r = .58, p < .001). This indicates that individuals who are concerned about the impressions they convey to others (people high in PSC) are especially likely to gauge others' perspectives and behaviors and succumb to social pressures (tendencies encompassed by the ATSCI subscale), but are no more likely than others to employ a manipulative or deceptive self-presentational style (measured by the CSV subscale).

Multivariate Analyses

Ordinary least squares (OLS) regression was used to assess the relative effects of PSC, ATSCI and CVS, as well as our other independent variables, on perceived other-self discrepancies in levels of alcohol consumption. The results of these analyses are presented in Table 3.2

Column 1 of Table 3 shows the additive effects of each of the independent variables on the gender-specific other-self discrepancy measure. As shown here, measures of familial socialization, as well as concerns about sexual violence and the ATSCI subscale, were not significantly related to OSG scores. Both PSC and CSV were, however, inversely associated with the OS gap measure, which suggests that students who exhibited these tendencies were more likely than other individuals to believe that they drank close to the normative level. Conversely, respondents who reported limiting their drinking to reduce their risk for various negative outcomes had above average OSG gap scores, indicating that they were more likely than other respondents to drink substantially below what they perceived to be typical of other same-sex students on campus.

Since reasons for limiting drinking, in particular, are likely to be related to students' use of alcohol, and overall alcohol consumption affects OSG scores, the latter relationship must be regarded as tentative. In an additional regression (Table 3, Column 2), we added a control for students' personal drinking behaviors. As expected, the magnitude of the effect of the drinking minimization variable was substantially smaller in this second analysis. Still, at every level of alcohol use, students who consciously limited their drinking to avoid negative outcomes had significantly larger OSG scores than other respondents. Moreover, the effects of PSC and CSV on gap scores were virtually the same as they were in the initial model. This is due to the fact that these characteristics were not associated with overall alcohol use (data not shown) and highlights the distinction between focusing on other-self discrepancies in levels of alcohol consumption, versus drinking behavior as an outcome variable. (3)

Moderating Relationships

In a subsequent series of regressions we tested for moderating effects. To avoid problems with multicollinearity, we centered all of our non-dichotomous predictors before computing the cross-product terms of interest, as recommended by Aiken & West (1991).

Perhaps most notable was the significant interaction between gender, CSV and OSG scores, reported in Column 1 of Table 4. Interactions between gender and the other three measures of role-taking/impression management tendencies (PT, PSC, ATSCI), on the other hand, were relatively small in magnitude. These, and all other non-significant higher-order effects, were dropped from the analysis.

Although family members' drinking habits conditioned the relationship between PSC and OSG scores (Table 4, Column 2), the cross product of PSC by parental attitudes toward alcohol use was non-significant. Similarly, gender failed to moderate the effect of the perceived likelihood of sexual violence on the OSG measure. This may indicate that fearing alcohol-related rape or sexual assault does not enhance women's ability to resist peer pressure. More likely though, it is a product of our particular measure, worded in such as way as to imply a concern about perpetration (of relevance to males) as well as victimization. While the coefficient for the limit drinking index by gender was also non-significant, the relationship between this variable and OSG scores did vary by Greek participation (Table 4, Column 3). (4)

We used the procedures outlined by Aiken and West (1991) to determine the nature of this and all other significant interactions, beginning with the effect of CVS on OSG scores by gender. In this case, we calculated predicted gap scores at different levels of CVS for males and then for females. CSV scores were varied from low (one standard deviation below the sample mean) to high (one standard deviation above the sample mean), while other variables were held constant at their mean values (now zero due to the centering of all non-dichotomous independent variables). We used an identical procedure to discern the nature of the other two significant interaction terms.

Predicted OSG scores across levels of CVS by gender are presented in Figure 1. Once again, high other-self discrepancy scores suggest that students are drinking substantially below the norm, while low values on this variable indicate a convergence to the elevated level of drinking that is believed to be common on campus. Given this, the pattern depicted in Figure 1 suggests that males may be more susceptible than females to peer influence. While CSV had little impact on OSG scores among the females in our sample, the males who exhibited this behavioral tendency (a manipulative and inauthentic self-presentational style) were more likely than other individuals to drink close to what they perceived to be typical for same-sex peers at their school. Still, even those males who were high in CSV (one standard deviation above the mean value), and had among the lowest OSG scores in our sample, were drinking at a level that they perceived to be below the norm.

[FIGURE 1 OMITTED]

Figure 2 shows the relationship between PSC and predicted OSG scores by family members' drinking patterns. As indicated here, PSC decreased OS discrepancies in levels of alcohol consumption primarily among respondents who had a family history of alcohol abuse, indicating that early socialization, or "upbringing", does counteract the susceptibility to peer pressure that accompanies heightened public self-consciousness. It is also interesting to note that students low in PSC from families within which heavy drinking was common had above average OSG scores, suggesting that they were some of the individuals most able to resist conforming to what they perceived to be normative campus drinking practices.

[FIGURE 2 OMITTED]

The third interaction, between the limited drinking index, Greek standing and OSG scores is represented in Figure 3. As shown here, conscious attempts to minimize one's drinking increased OSG scores primarily among students unaffiliated with the Greek system. Non-Greek respondents who reported that they limited their use of alcohol to avoid negative outcomes drank substantially less than what they perceived to be typical for same sex peers at their school. This was not the case for Greek participants. Despite efforts to limit their use of alcohol for health, financial and other reasons, these individuals continued to drink at a level that was close to what they perceived to be normative.

[FIGURE 3 OMITTED]

DISCUSSION

Consistent with earlier studies showing that women tend to use males as a referent when responding to questions about other students' drinking habits (Korcuska & Thombs, 2003; Lewis & Neighbors, 2004), gender differences in OSG scores were substantially smaller (and non-significant) when we used the sex-specific, versus a general, measure. This confirms the notion that the relatively large other-self discrepancies in levels of alcohol consumption observed among women across analyses are likely to be a reflection of methodological issues (Borsari & Carey, 2003) and not gender differences in susceptibility to peer influence.

Nonetheless, our analyses did provide some evidence that males may be more reactive than females to social pressures supportive of heavy drinking. While cross-situational variability was unrelated to OSG scores among females, males who exhibited this tendency were more likely than other individuals to drink at a level close to what they perceived to be normative. This finding consistent with previous studies showing that college men are more readily affected than women by others' perceived attitudes towards alcohol use on campus (Prentice & Miller, 1993) because heavy drinking is believed to be an integral part of the male role (Suls & Green, 2003).

It is curious though that the cross-situational variability subscale was the only one of the three impression-management variables that moderated the effect of gender on OSG scores. Cross-situational variability, characterized by a manipulative and duplicitous self-presentational style, implies a more superficial type of conformity than attention to social comparison information, which reflects attitudinal as well as a behavioral shifts in response to prevailing social cues. Insofar as it is the former self-presentational strategy that underlies males' acquiescence to what they believe to be normative drinking practices, their commitment to binge drinking is likely to be especially short-lived. While this issue lies beyond the scope of this study, the fact that levels of alcohol consumption drop drastically post college graduation, even among some of the heaviest drinkers (Bartholow, Sher & Krull, 2003; Donovan, Jessor & Jessor, 1983), is in line with this interpretation.

Although attention to social comparison information was not associated with OSG scores, students high in public self-consciousness with a family history of alcohol abuse were more likely than other respondents to drink at a level close to what they perceived to be normative for same-sex peers at their school. This suggests that it is an explicit concern about others' evaluations, as well as a willingness to alter one's behavior to meet situational demands (cross-situational variability), rather than changes in perception, that puts certain individuals at risk for succumbing to situational influences. While this may seem counterintuitive, it is consistent with prior research showing that students' drinking patterns may not reflect their beliefs about alcohol (Prentice & Miller, 1993) and that their abuse of this substance is often transitory.

Even so, some individuals who abuse alcohol in college do go on to develop more long-term drinking problems (Jackson, Sher, Gotham & Wood, 2001). Excessive drinking on college campuses also has a number of more immediate negative consequences (Wechsler, Lee, Kuo, Sebring, Nelson & Lee et al., 2002), which makes the identification of factors likely to reduce student alcohol abuse a potentially important endeavor. As hypothesized, our findings suggest that coming from a family within which alcohol abuse is rare may serve as a protective factor, decreasing the risk for drinking in response to situational pressures among individuals predisposed to conformity because of a heightened concern with the evaluations of others.

Nonetheless, it was respondents from families within which alcohol abuse was common who were low in public self-consciousness who had the highest OSG scores. Why these individuals exhibited the greatest resistance to peer pressure is not immediately clear. Perhaps observing the negative consequences of alcohol abuse firsthand made these individuals regard binge drinking as a risky proposition, and this, combined with their lack of concern with how others viewed them, enhanced their capacities to resist the influence of their peers. This latter interpretation is speculative and bears further investigation. The fact that parental attitudes towards their children's drinking did not moderate the PSC-OSG relationship does, however, indicate that students learn to incorporate alcohol into their social repertoires through observation of parents and other family members, and that it is these acquired patterns, rather than parental beliefs, that shape their tendency to use this substance for self-presentational reasons after they leave home.

From a prevention standpoint, this reinforces the adage that actions speak louder than words and that parents might reduce the risk for excessive drinking during the college years among children sensitive to self-evaluative feedback through early modeling of more appropriate drinking habits. When this has not been the case, students high in public self-consciousness should be especially sensitive to norm corrections that challenge their perceptions of how much others are drinking.

Another finding with policy implications concerns the relationship between other-self discrepancies in levels of alcohol consumption and affiliation with the Greek system. The low OSG scores observed among Greek participants were due primarily to their heavy drinking. Despite this, the interaction between Greek standing and attempts to limit drinking suggests that members of fraternities and sororities find it difficult to resist conforming to prevailing social norms even when they are worried about the negative effects of alcohol intoxication, and that interventions emphasizing the potential costs of alcohol abuse are likely to have little impact among this population.

Students unaffiliated with the Greek system should benefit more from this type of information. As predicted, non-Greeks who consciously sought to minimize their use of alcohol to reduce the likelihood of a variety of negative outcomes had among the largest OSG scores, overall and at each level of drinking, suggesting that they were the individuals best able to resist perceived situational pressures that ran counter to this aim. While one could argue that the high OS discrepancies observed among non-Greeks who are concerned about alcohol's negative effects might reflect their selection of reference groups, and a vulnerability to subsequent increases in drinking, a capacity to resist peer pressure seems the more plausible explanation.

The cross-sectional nature of our data makes it impossible to test this assumption. Furthermore, our use of a non-representative sample potentially undermines the generalizability of the study findings. Additional research is needed to address these issues. Future analyses might also focus on less distal sources of feedback about what is appropriate drinking by considering friends' use of alcohol as well as broader campus norms.

Although variables associated with other-self discrepancies do not necessarily predict students' overall levels of alcohol consumption, they provide insight into who is more, or less, likely to drink for social acceptance. Given the importance of understanding students' motivations for using alcohol in the development and implementation of effective interventions, identifying characteristics associated with susceptibility to peer influence itself should have a variety of applications.

REFERENCES

Aiken, L. S. & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage Press.

Baer, J. S. (2002). Student factors: Understanding individual variations in college drinking. Journal of Studies on Alcohol, Vol. Supp. 14, 40-53.

Baer, J. S., Stacy, A., & Larimer, M. (1991). Biases in the perception of drinking norms among college students. Journal of Studies on Alcohol, 52, 580-586.

Bartholow, B. D., Sher, K. J., & Krull, J. L. (2003). Changes in heavy drinking over the third decade of life as a function of collegiate fraternity and sorority involvement: A prospective, multilevel analysis. Health Psychology, 22, 161-626.

Borsari, B. & Carey, K. B. (2001). Peer influences on college drinking: A review of the Research. Journal of Substance Abuse, 13, 391-424.

Borsari, B. & Carey, K. B. (2003). Descriptive and injunctive norms in college drinking: A meta-analytic integration. Journal of Studies on Alcohol, 64, 331-341.

Buss, A. H. (1980). Self-Consciousness and Social Anxiety. San Francisco: W. H. Freeman.

Carver, C. S. & Glass, D. C. (1976). The self-consciousness scale: A discriminant validity study. Journal of Personality Assessment, 40, 169-172.

Donovan, J. E., Jessor, R., & Jessor, L. (1983). Problem drinking in adolescence and young adulthood. Journal of Studies on Alcohol, 44, 109-137.

Fenigstein, A., Scheier, M. F., & Buss, A. H. (1975). Public and private self-consciousness: Assessment and theory. Journal of Consulting and Clinical Psychology, 43, 522-527.

Goffman, E. (1959). The Presentation of Self in Everyday Life. Garden City, New York: Anchor Books.

Greenfield, T. K., Guydish, J., & Temple, M. T. (1989). Reasons students give for limiting drinking: A factor analysis with implications for research and practice. Journal of Studies on Alcohol, 50, 108-115.

Hawkins, J. D., Catalano, R. R., & Miller, J. Y. (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: Implications for substance abuse prevention. Psychological Bulletin, 112, 64-105.

Jackson, K. M., SHer, K. J., Gotham, H. J., & Wood, P. (2001). Transitioning into and out of large-effect drinking in young adulthood. Journal of Abnormal Psychology, 110, 378-391.

Korcuska, J. S. & Thombs, D. L. (2003). Gender role conflict and sex-specific drinking norms: Relationships to alcohol use in undergraduate men and women. Journal of College Student Development, 44, 204-216.

Lennox, R. (1984). Public self-consciousness, social anxiety, and variability of social behavior. Psychological Reports, 54, 911-914.

Lennox, R. D., & Wolfe, R.N. (1984). Revision of the self-monitoring scale. Journal of Personality and Social Psychology, 46, 1349-1364.

Lewis, M. A., & Neighbors, C. (2004). Gender-specific misperceptions of college student drinking norms. Psychology of Addictive Behaviors, 18, 334-339.

Martin, K. A. & Leary, M. R. (2001). Self-presentational determinants of health risk behavior among college freshmen. Psychology and Health, 16, 17-27.

Nash, S. G., McQueen, A., & Bray, J. H. (2005). Pathways to adolescent alcohol use: Family environment, peer influence, and parental expectations. Journal of Adolescent Health, 37, 19-28.

Novak, K. B. & Crawford, L. A. (2001). Perceived drinking norms, attention to social comparison information, and alcohol use among college students. Journal of Alcohol and Drug Education, 46, 18-33.

Park, A., Sher, K. J., & Krull, J. L. (2006). Individual differences in the "Greek effect" on risky drinking: The role of self-consciousness. Psychology of Addictive Behaviors, 20, 85-90.

Perkins, H. W., Haines, M. P., & Rice, R. (2005). Misperceiving the college drinking norm and related problems: A nationwide study of exposure to preventative information, perceived norms and student alcohol misuse. Journal of Studies on Alcohol 66, 470-478.

Perkins, H. W. & Wechsler, H. (1996). Variation in perceived college drinking norms and its impact on alcohol abuse: A nationwide study. Journal of Drug Issues, 26, 961-974.

Prentice, D. A. & Miller, D. T. (1993). Pluralistic ignorance and alcohol use on campus: Some consequences of misperceiving the norm. Journal of Personality and Social Psychology, 64, 243-245.

Santee, R. T. & Maslach, C. (1982). To agree or not to agree: Personal dissent amid social pressures to conform. Journal of Personality and Social Psychology, 42, 690-700.

Schroeder, C. M. & Prentice, D. A. (1998). Exposing pluralistic ignorance to reduce alcohol use among college students. Journal of Applied Social Psychology, 28, 2150-2180.

Shore, E. R., Rivers, C., & Berman, J. J. (1983). Resistance by college students to peer pressure to drink. Journal of Studies on Alcohol 44, 352-361.

Slicker, E. K. (2001). University students' reasons for NOT drinking: Relationship to alcohol consumption level. Journal of Alcohol and Drug Education, 42, 83-102.

Snyder, M. (1974). Self-monitoring as expressive behavior. Journal of Personality and Social Behavior, 30, 526-537.

Suls, J. & Green, P. (2003). Pluralistic ignorance and college student perceptions of gender-specific alcohol norms. Health Psychology, 22, 479-486

Tomarelli, M. M. & Shaffer, D. R. (1985). What aspects of self to self-monitors monitor? Bulletin of the Psychonomic Society, 23, 135-138.

Turner, R. G., Scheier, M. F., Carver, C. S., & Ickes, W. (1978). Correlates of self consciousness. Journal of Personality Assessment 42, 285-289.

Vleeming, R. G. & Engelse, J. A. (1981). Assessment of private and public self-consciousness: A Dutch replication. Journal of Personality Assessment, 45, 385-389.

Wechsler, H., Kelley, K., Weitzman, E. R., Giovanni, J. P., & Seibring, M. (2002). What colleges are doing about student binge drinking: A survey of college administrators. Journal of American College Health, 48, 219-226.

Wechsler, H., Lee, J. E., Kuo, M., Seibring, M., Nelson, T., & Lee, H. (2002). Trends in binge drinking during and era of increased prevention efforts. Journal of American College Health, 50, 203-217.

Wolfe, R. N., Lennox R. D., & Cutler, B. L. (1986). Getting along and getting ahead: Empirical support for a theory of protective and acquisitive self-presentation. Journal of Personality and Social Psychology, 50, 356-361.

Wolfe, R. N., Welch, L. K., Lennox, R. D., & Cutler, B. L. (1985). Concern for appropriateness as a moderator variable in the statistical explanation of self-reported use of alcohol and marijuana. Journal of Personality, 53, 1-16.

Wood, M. D., Read, J. P., & Mitchell, R. E., Brand (2004). Do parents still matter? Parental and peer influences on alcohol involvement among recent high school graduates. Psychology of Addictive Behaviors, 81, 19-30.

Wood, M. D., Read, J. P., Palfai, T. P., & Stevenson, J. F. (2001). Social influence processes and college student drinking: The mediation role of alcohol outcome expectancies. Journal of Studies on Alcohol, 62, 32-43.

Correspondence concerning this article should be addressed to Katherine B. Novak, Department of Sociology, Butler University, 4600 Sunset Ave., Indianapolis Indiana, 46208. Phone: [phone]; E-mail: [e-mail]; Fax: [phone]

A previous version of this paper was presented at the annual meeting of the Midwest Sociological Society, Omaha, NE, April 2006.

(1) One of the 13 indicators of ATSCI was inadvertently omitted when constructing the undergraduate survey.

(2) Mean substitution was used in this and all subsequent analyses. Additional regressions excluding all cases with missing data on any variable yielded results that were virtually identical to those obtained using this procedure.

(3) Our contention is that the former measure reflects students' tendency to use alcohol for a particular purpose, namely social acceptance. Making a similar point regarding their independent variable, Wolfe and associates (1986) emphasize the fact that the concern for appropriateness scale, described earlier, is not itself associated with alcohol use. Rather, it predicts the extent to which respondents' drinking is motivated by peer influence.

(4) Adding levels of alcohol consumption into the statistical model had little effect on the coefficients for CSV by gender or PSC by family members' drinking. For this reason, we included the control for drinking behavior only in the analysis presented in columns 3 of Table 4.

Lizabeth A. Crawford

Bradley University 

Katherine B. Novak

Butler University

TABLE 1
Descriptive Statistics

                                  Mean    Std Dev   Range     n

Female                              .54       .50   0-1      206
Greek Participant                   .31       .46   0-1      206

Parents Disapprove of Alcohol      2.52       .86   1-4      205
Family Little Drinking             9.42      2.12   3-12     205

Public Self-Consciousness         20.00      4.03   7-28     204
Attention Social Comparison       29.83      4.24   12-48    202
Cross-situational Variability     18.14      3.23   7-28     205

Sexual Assault/Rape Likely         1.69       .67   1-3      203
Limit Drinking                    10.68      2.75   5-20     205

Gender-Specific Other--Self Gap    5.33     11.14   -33-43   196
General Other--Self Gap            4.90     13.58   -38-39   193

TABLE 2
Mean Other-Self Gap Scores by Gender and Greek Participation

                      OSG         OSG
                    Specific    General

Male (n=88)          4.43      -0.67
Female (n=103)       5.49       9.05

Difference          -1.05      -9.57 ***

Non-Greek (n=129)    6.11       6.46
Greek (n=62)         2.68        .65

Difference           3.43 *     5.81 **

*** p < .001, ** p < .01, * p < .05

TABLE 3
OLS Regressions Predicting Other-Self Gap Scores (n=206)

                            Column 1              Column 2

                         b         Beta        b         Beta

Constant                2.81                20.28 **

Female                   .90        .04     -3.27 *      -.15
Greek                  -2.83       -.12     -2.47        -.11

Parents' Disapprove     -.69       -.05      -.69        -.05
Family Little Drink      .15        .03      -.06        -.01

PSC                     -.52 *     -.19      -.46 *      -.17
ATSCI                    .13        .05       .11         .04
CSV                     -.51 *     -.15      -.42 *      -.13

Limit Drinking          1.53 ***    .39       .74 **      .19
Assault/Rape Likely     1.63        .10       .82         .05

Alcohol Use                                  -.31 ***    -.46

Adj. R-Square               .213 ***             .337 ***

*** p < .001, ** p < .01, * p < .05