Does Drinking in College Affect Your Grades?

To some people, the following conclusion should be filed under “Duh.” But even they might appreciate the empirical rigor undertaken by Scott E. Carrell, Mark Hoekstra, and James E. West in a new working paper called “Does Drinking Impair College Performance? Evidence from a Regression Discontinuity Approach” (abstract here; PDF here). The authors point out that although the effect of alcohol on low-frequency but high-risk events – mortality and teen pregnancy, e.g. – is widely studied, “little is known about the effect of drinking on the majority of users.” The summary:

This paper examines the effect of alcohol consumption on student achievement. To do so, we exploit the discontinuity in drinking at age 21 at a college in which the minimum legal drinking age is strictly enforced. We find that drinking causes significant reductions in academic performance, particularly for the highest-performing students. This suggests that the negative consequences of alcohol consumption extend beyond the narrow segment of the population at risk of more severe, low-frequency, outcomes.

The choice of data set is interesting, and the explanation thereof is interesting in its own right:

We use administrative data on 3,884 students at the United States Air Force Academy (USAFA) [where one of the authors, West, teaches] between 2000 and 2006. This educational setting offers two distinct advantages for our analysis. First, in contrast to many college campuses, the explicit ban on underage drinking at the USAFA is strictly enforced; violations can lead to expulsion. As a result, in an anonymous survey of underage students in our sample, only 37 to 39 percent of students reported drinking any alcohol since arriving at the academy. By comparison, the National Institute on Alcohol Abuse and Alcoholism (NIAAA) reports that 83% of college students nationwide drink, and that 41% of college students reported consuming five or more drinks on a single occasion within the past two weeks (NIAAA, 2010). The second advantage of using USAFA data is a consistent measurement of academic achievement.


They could try my college, a dry college, as a natural experiment, though I would guess there are many other contributing variables that affect any result variable one might test.


While drinking surely didn't help in my math and physics studies in college, it sure made it easier to discuss math, physics, chess and economics with the psych-, English- and history-major co-eds.

Ian Kemmish

I can't read the paper because I don't have a subscription, but how rigourously does it distinguish between "drinking" and "starting to drink"?

In this country, for example, one starts to drink at 18 (at the latest....) and goes up to Cambridge at 19. Cambridge does not have a reputation as an underperforming institution, so it may be that, once you've got the initial binge out of your system, everything settles down again.

David L

Not that I doubt the conclusion, I would think that limiting your sample population to one institution, and to such a highly specialized and institution at that, would make these results suspect...


I'm not sure you can generalize the results from this particular population to college students at large, given that it's a self-selected group who have particular values systems and behaviors in common that differ widely from the general population.

Also, you have to wonder if the sudden access to alcohol at age 21 makes their use of it different from those who have been drinking with discretion for years.


Rampant grade inflation offsets whatever negative correlation of grades and drinking. I can vouch as a recent grad of Northwestern, where the school seal says "nobody gets a C" in Greek or something.


This is probably true of all social behaviors especially "risky" ones.

I bet if you just take the kids who fill the weekends with social engagements, verses the kids who don't, you'll find a correlation there as well.

Since drinking often goes along with being social, blammo, there's your correlation.

And I can't get over expelling a kid for drinking when signing up means he's risking his life. To each his own I guess.

Joe D

I also question the result that would seem to compare performance in the fourth year (or even second half of the fourth year) to the three previous years. "Senioritis" is a well-known malady; perhaps it doesn't affect AF cadets?

Eric M. Jones

I graduated with a 4.0 BAC.


I agree with the commenters above. Having attended university at both the US and Canada I can attest to very different alcohol related cultures in each. While drinking too much is not unusual for college life, for most people there is a bad few years when you first start drinking that doesn't begin to calm down till you've been able to drink legally for a year or two. I often remarked in my undergrad years how much American 21 year olds acted like Canadian 19 and 20 year olds when alcohol was involved. In other areas attitudes were similar but specifically in regards to alcohol there were significant differences.

I'd like to see a comparative study tracking students through 4 (or 5) years of college in comparable schools in countries with different drinking age laws.


I went to school with some people whose grades increased as their drinking increased. Got to relieve the stress of studying somehow.


Guess I'm a lucky outlier ;O


I am in the US. I drank more in high school than I did in college. Partly because I went through that phase of life being new to alcohol in HS. The other part was that I met my to be wife and we were serious for the last two years. Therefore I did not need to deal with going to parties and bars and drinking to find girls. I had one and we had plenty of fun ourselves. Like rabbits.


Is it the alcohol that creates the issue or the social aspects that revolve around alcohol? Maybe this is a chemical problem with alcohol or maybe it is due to late nights out with lack of sleep.


perhaps some of what is being shown, when studying "a college in which the minimum legal drinking age is strictly enforced" is that there is a letting-loose, or over-reaction in some students, when they finally, suddenly become free of the prohibition. I'm not in favor of lowering the drinking age--I went to MSU when the Michigan drinking age was 18, and do I have stories!--but I don't think you should overlook this effect either, when studying this, or thinking about alcohol policy.


I can't think of many student bodies that would have less of a "majority of users" representation than that of a service academy. They're 100% academic outliers (it's just as difficult to get into the AFA as any Ivy League school, if not more), there isn't the same concern about needing to find a job upon graduation (cadets don't need student loans and they're committed to Uncle Sam for 5 years), women (the majority of college students in the US) are a huge minority, and I imagine that their rules would deter many personalities from wanting to attend in the first place.

Ed B

air force academy does NOT seem representative of either average or superior students. It is *highly* selective and an awesome school - its just not representative of the rest of the country - at all...

David Cohen

Dan...don't drink :) even when you're 21? The study was authored by a usafa professor.


There is no value in sampling the alcohol consumption at a dry military school and applying the data to normal colleges.

Someone should do a study on the income of binge drinkers, in normal colleges, 5-10 years after college compared to their dry peers.

Panem et Circanses

WATCH YOUR CAUSALITY!!! You have NO idea whether drinking causes bad gradesor the other way around!!!