Academic journal article Economic Inquiry

Fraternity Membership and Drinking Behavior

Academic journal article Economic Inquiry

Fraternity Membership and Drinking Behavior

Article excerpt


Fraternities and sororities are an integral part of life for many college students. In the Harvard College Alcohol Study (CAS), which sampled nationally representative groups of full-time undergraduates at 4-yr colleges on four occasions during 1993-2001, 16% of students reported belonging to a fraternity or sorority. Only 1 of the 125 colleges surveyed had no fraternities. (1) Within the 476 school-year cells (representing 119 schools each year), the prevalence of fraternity membership was 12.4% at the median and over 20% in a quarter of cases.

The activity most associated with fraternities is alcohol consumption. Fraternities are responsible for a substantial portion of campus events at which alcohol is available to students. College itself often conveys an image of fraternity parties with abundant intoxicated students, perhaps engaging in behaviors such as sex and vandalism. CAS data suggest that this stereotype is not baseless. For the 30 pre-survey days, alcohol intoxication was reported by 64% of fraternity members but only 42% of nonmembers. Moreover, the current-school year incidence of unprotected sex and property damage that respondents attributed to drinking was 13% and 14%, respectively, among students belonging to fraternities compared with 8% for each among others. Similarly, 40% of fraternity members, compared with 26% of nonmembers, reported driving after drinking in the past 30 days.

From a policy standpoint, a fundamental question regarding these statistics is whether they reflect causal effects of fraternity membership. A public health argument can be made for intervening in fraternity activities that directly lead to excessive drinking. This is particularly true if alcohol use has peer effects (Kremer and Levy 2003; Lundborg 2006) or there is reason to believe that students make alcohol consumption decisions that are systematically shortsighted. Effects of fraternities on behaviors such as vandalism and unsafe sex would warrant regulation even under the stricter criterion that harmful externalities must be present.

The obvious reason for caution in asserting that the above numbers represent a causal behavioral influence of fraternities is self-selection. Undoubtedly, students choose to join fraternities in part because of preexisting preferences toward behaviors that membership facilitates. It would be erroneous, therefore, to assume that these unconditional prevalence differences accurately portray direct effects of fraternities. Rather, this study investigates whether fraternities are responsible for any of these differences. Put differently, in a counterfactual world without fraternities, would rates of drinking and related behaviors differ just as much between students who, with fraternities present, do and do not belong to them?

Presumably, the factors that confound the observed relationship between drinking and fraternity membership are related to preferences that simultaneously influence joining fraternities and consuming alcohol. The strategy used here to identify causal effects consequently takes advantage of the inclusion in the aforementioned CAS data of a large set of proxies for these preferences that can be directly entered into the regression equations. This approach parallels that of DeSimone (2007), the only related study in the economics literature. But relative to the data in that study, the CAS provides a deeper and more nuanced set of relevant covariates.

These covariates proxy for two types of omitted factors, fixed characteristics determined prior to college enrollment and time-varying preferences that might have since changed. Time-constant heterogeneity is absorbed by college-by-survey year fixed effects, which account for school selection based on prevailing drinking environments, along with parental drinking and respondent high school senior year drinking behaviors. Preference shocks occurring in the meantime are captured by assessments of the importance of parties and various reasons for consuming alcohol. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.