Availability Transitions in College Drinking
Principal Investigator: Paul J. Gruenewald, Ph.D.

Return to Video Presentation


The availability component of the college drinking study was designed to examine two issues: (1) the measurement of drinking patterns and problems, and (2) the impact of availability transitions on college drinking. Since this research component began in year three of the current center award, data collection and analysis is still underway with the last data collection phase to take place in the fall of 2001. At this time, preliminary results of the first studies of drinking patterns and problems are available and final analyses for publication are being actively pursued. Examination of the impacts of availability transitions on college drinking will be pursued in years four and five of the current center cycle.

Changing the way we talk about drinking:

For much of the history of alcohol studies, the challenge for survey researchers has been how to measure drinking in a sensible and comprehensive way. Two general approaches have been taken to this problem: (1) Ask people how much they drink and how often. (2) Ask people how often they drink at a certain level (say four or more drinks). Answers to these questions are really summary measures of drinking patterns, provided to us by our respondents.

Drinking patterns can be best captured by determining how often drinkers drink at any given level. Asking the questions, "How often did you drink five drinks?" or "How often did you drink only one drink?" and so on, across some period of time. While we cannot ask these sorts of questions directly on surveys, we have developed a survey questionnaire and mathematical models of drinking adequate to derive this information from survey self-reports.

The drinking distributions we get from these analyses we call "drinking exposures". In the same way you would think of a person being exposed to, say, radon in the environment, we view drinkers as exposing themselves to alcohol. So, the following individual exposes himself to drinking three drinks on about 2 1/2 days per month.

Using this approach we can see, very graphically, how drinking exposures differ between drinkers.

And we can see how drinkers differ in their patterns between groups. For example, differences between older and younger college students, and men and women.

....

We can also use distributions of drinking exposures to understand what we mean by drinking types like "heavy" or "binge" drinkers. What we see is that such types are really not types at all, but rather loosely defined collections of drinkers who happen to drink "heavily" or "binge" during the time under study. This observation belies the validity of drinking typologies in general.

Changing the way we talk about drinking problems:

Once we see that traditional survey measures of drinking are really different measures of individual exposure distributions we can also see how to properly go about measuring risks related to drinking: Regress drinking problems over drinking exposures. For various reasons this turns out to be extremely complicated, but boils down to estimating risks associated with drinking at each level of exposure. These analyses give us our first ever look at properly defined dose-response functions related to drinking.

Some thoughts about availability transitions:

The second major goal of the availability component of the college drinking study is to examine the impacts of changes in availability upon college drinking patterns. With the fully developed model of drinking patterns provided in the first phase of study presented here, we are armed to explore these changes across the different places where college students reside and may drink.

[Return to Top]