This research examined the applicability of the Theory of Planned Behavior for understanding drinking behaviors among sorority females. The noted increase in drinking among females in college and the significant influence of the Greek system on heavy drinking provide rationale for this longitudinal study which offers insight for identifying factors associated with alcohol consumption. Findings suggest that our sample of sorority females do indeed drink more heavily than general college females. For example, recent research by LaBrie and colleagues (in press)
on college females found 23.2% met the qualification for “binge drinker” (compared to 35.3% of our sorority sample) and 15.5% met the qualification for “frequent binge drinker” (compared to 20.3% of our sample). Findings suggest that Greek-affiliated females drink more than general college females.
Our findings suggest that there may be two main pathways predicting alcohol consumption among sorority members. Primarily, intentions to drink mediated the relationship between both attitudes and subjective norms on drinking behavior. Perceived behavioral control, however, had a direct path to drinking behavior, suggesting that the perception of difficulty in refraining from problematic drinking among our sample of sorority members influenced their drinking behaviors. Thus, while it is well documented that intentions predict behaviors, it appears that among Greek-affiliated sororities, where alcohol use has been found to be more prevalent and accepted by members, lower perceived control over the behavior does not predict the intention to drink but does influence actual alcohol consumption. Perceived control therefore has a direct effect on drinking and is not mediated by intentions.
To our knowledge, very few studies to date have examined sorority females independently from fraternities and no studies have applied the TPB to sorority drinking. To test the applicability of TPB in understanding sorority drinking, a latent structural equation model revealed that intentions to drink over the next 30 days significantly predicted drinking behaviors (average drinks and maximum drinks per occasion) one month later. This key finding is consistent with research on TPB and college drinking that underscores the significant role of intentions on behavior (e.g., Bentler & Speckart, 1979
; Conner, et al., 1999; Norman & Conner, 2006
; Schlegel, et al., 1992
). Thus, TPB is an effective and appropriate model for predicting sorority drinking in college. Yet, even with the longitudinal design, it is still possible that the relationship between intentions and behavior found among our sample reflects the stability of drinking over time. Future research is encouraged to examine the role of previous drinking within this model to clarify the strength of intentions on behavior.
An examination of the constructs of the TPB revealed that all three predictor variables—attitudes toward drinking, subjective norms, and perceived behavioral control—were significantly and positively correlated. These inter-factor correlations suggest that as attitudes toward drinking increase, so do subjective norms about drinking, as well as a perceived lack of control over drinking related problems. Furthermore, intentions to drink mediated the relationship between attitudes and norms on drinking behaviors. Perceived control defined as the inability to refrain from drinking, on the other hand, directly predicted drinking behaviors. Taken together, the predictive model accounted for 44.7% of the variance in intentions and 73.4% of the variance in behavior.
Interestingly, while some studies suggest that attitudes are better predictors of intentions to drink than are subjective norms (Trafimow, 1996
), among our sample of sorority members, norms were a stronger predictor of intentions than attitudes and control. This may be due to variations in the operationalization of constructs, but more likely, this finding is due to the important role of peer influence among sorority members and the pressure for sorority women to fit in with their peer referent group. In fact, our findings support recent research that examined many predictors of alcohol use and found social norms to be among the greatest predictors of college student use (Neighbors, Lee, Lewis, Fossos, & Larimer, in press
). Peer associations have been found to significantly predict drinking behaviors among college students (Durkin et al., 2005
). Among sorority members, these peer associations may have an even stronger impact when coupled with the strong perception of alcohol use as highly acceptable among Greek-affiliated students. Thus, sorority women may be forming intentions to drink based not only on their personal attitude toward drinking but more importantly to match the perceived behavior of their sorority sisters.
In the current model, we specified PBC to be the inability to refrain from problematic drinking. While norms and attitudes significantly predicted intentions to drink, PBC did not predict intentions, yet did predict drinking behaviors 30 days later. This finding is logical considering the intention
to engage in a behavior assumes a rational and planned decision. Experiencing difficulty in refraining from drinking or a low sense of control over the performance of a behavior suggests that the behavior is no longer a rational choice where intentions to act are formed. The significant direct path between PBC and drinking behavior indicates that perhaps among sorority women who have low perceived behavioral control, it does not predict their intentions, but it does directly predict their drinking behavior. Our finding supports similar evidence found by Schlegel and colleagues (1992)
, suggesting that among heavier drinkers, intentions and perceived control both influence drinking.
While this study contributes to the understanding of female drinking patterns, deemed necessary by the national rising trend in female drinking, all Greek-affiliated students are at-risk for heavy consumption. As such, examination of male fraternity members is also necessary. Further, this study examined drinking behaviors among a selection of sorority members from one midsized, private, Western university. The sorority system at larger universities may be different and as such, our results would need to be replicated in other contexts for greater external validity. Also, the current study utilized retrospective, self-report measures to assess drinking behaviors. While retrospective data collection has its limitations, applying TPB necessitates using a longitudinal design in order to examine future intentions and then assess actual drinking behaviors during that timeframe. Studies applying TRA or TPB in the past have assessed intentions without a follow-up assessment of actual behaviors (see studies 1 & 2 of Trafimow, 1996
;Williams & Hine, 2002) and many use a weekly timeframe between intentions and behavior (e.g., Armitage, Conner, Loach, & Willetts, 1999
; Norman & Conner, 2006
; see study 3 of Trafimow, 1996
). Thus retrospective data was necessary and allowed for the prediction of behavior over an entire month to be examined. Moreover, students were assured of confidentiality and no penalties existed for reporting drinking behavior. Still, future studies may wish to examine female drinking patterns in vivo, examining actual drinking situations among sororities. Additionally, the study assessed attitudes toward drinking by using the DMQ (Cooper, 1994
). While not the traditional approach to assessing attitudes, the DMQ is still an appropriate and beneficial scale to measure attitudes towards drinking because responses to items indicate that the participant endorses alcohol use positively, thereby providing a positive evaluation toward performing the behavior. The traditional approach to measuring attitudes within the TPB framework is to assess an overall evaluation of the behavior in question. Measures of the attitude construct include the belief that the behavioral performance is associated with certain outcomes. The DMQ assesses the reasons why students drink (e.g., to be sociable), suggesting that the student engages in that behavior because of its association with that outcome (e.g., being sociable). Higher scores on the DMQ refer to a more positive endorsement of drinking for that reason, suggesting an underlying positive evaluation of drinking because of its association with the desired outcome. Therefore, it is logical to use the DMQ as an assessment of a student's attitude toward drinking.
In conclusion, this study provides valuable information on drinking among college female students. While females are at increased risk for negative consequences, sorority members may be at even greater risk for excessive drinking due to the strong influence of the Greek system. This study establishes the TPB as an appropriate model for understanding factors that may predict drinking behaviors, especially female drinking within a social group. It appears as though among sorority members, drinking can be both a rational and planned decision, guided by intentions, but yet also not controlled by volition. Perhaps the culture of alcohol as acceptable among Greek-affiliated students allows sorority members to plan their drinking, but at the same time increases the perception that they have little control over the actual behavior. As both pathways appear to be significant in predicting alcohol consumption, interventions aimed at reducing drinking may wish to address both a healthy, mindful, and planned decision when it comes to drinking, as well as build skills to encourage control over the actual drinking situation. Specifically, interventions aimed at reducing heavy drinking among sororities would benefit from targeting the constructs of subjective norms and behavioral control. Effective interventions applying the social norms approach are already ongoing on many college campuses and challenge perceptions of peer alcohol consumption (Task Force on College Drinking, NIAAA, 2002
). Such efforts may also consider targeting behavioral control, allowing women to realize that drinking is a choice in which they do have control over their drinking behavior. Activities that role play how to increase control and that encourage goal setting, may help women to form rational, planned decisions about their drinking behaviors. Increasing the perception of control may positively impact college campuses and reduce risk associated with drinking. Thus, by encouraging mindful decision-making and providing skills to control drinking behavior, a reduction in drinking and associated risk may occur.