Hale Simon ’05 asks:

[D]oes anyone else find it wierd that last fall the college was predicting record numbers of students going to the health center for alcohol related problems and that those records would be set by a large margin yet we finished the year with pretty much an average number of these incidents?

No. Last year’s alcohol “crisis” was a lot of hooey, as anyone with basic training in statistics (or even a pinch of common sense) should have recognized. Or, rather, none of the information presented was enough to conclude that there was a real problem. I was going to post about this last spring, but it all sounded so negative and mocking that I put it to one side. Never one to let material go to waste, I provide it below.

Variation happens. Different numbers of students will need treatment at the health center for excessive drinking each year, even if the true (unobservable) rate of alcohol abuse stays the same.

There are many problems at Williams. Excessive alcohol cosumption is not one of them. Or, most of the strategies one might try (and all of those that I have seen publically discussed) to decrease alcohol consumption would probably be worse than the disease.

The best overview of the entire issue is Aiden Finley’s ’04 op-ed, one of the best pieces of writing to appear in the Record in years.

I suspect that more than a few people at Williams would benefit from STAT 101. Evidence for this can be seen in the Record article on this year’s alcohol problems. The author argues that:

the alarming alcohol problems that confronted the College in the fall appear to have subsided

Perhaps. But perhaps the problem was never really that alarming to begin with. The figures cited (52 students required overnight care in the fall semester compared to 72 in all of last year) are, to my mind, merely suggestive. Obvious questions would include:

  • How many students were admitted in the first sememester in the last few years? Those are the numbers that we want to compare to this fall’s 52.
  • What is the break-up among classes for the 52? If we just have a particularly drunk first class, then that is one thing. It too shall pass. If sophomores are getting drunker this year than last year, that would be more worrisome.
  • What are the standards for determining who “needs” overnight care and have those standards changed? Don’t forget that if you run (or work at) a health center whose budget is under threat, you might change your opinion about who needs to spend the night.
  • And so on.

Again, I don’t want to be too caustic, but there was no reason to conclude, based on the public data, that 2003–2004 was different (in a statistically meaningful way) from past years.

Print  •  Email