Sifting through the data for some clarifications about law student debt figures

Law school debt figures are an important component of any examination of the legal education market. But they are often misunderstood in ways that exaggerate their scope and burden--and using figures that may in other respects conceal the true magnitude of their effect.

To understand, consider statistics on alcohol consumption. The average American consumes 14 alcoholic drinks per week.*

Well, no, the average American doesn't. Among drinkers, the average American consumes 14 alcoholic drinks per week; but 30 percent of Americans don't drink. If you include those Americans, the number drops to 9.8 drinks per week. That's a pretty significant difference--but, of course, either figure is useful, as long as one includes the relevant caveat when disclosing the data.

The same holds true for debt figures. Using one measure, Law School Transparency reports an average debt of $118,670 for law school loans. Well, not exactly--$118,670 among law students who took out law school loans.

Consider, too, the recent New York Times article on the subject: "In 2012, the average law graduate’s debt was $140,000, 59 percent higher than eight years earlier."

That statement is false at one level and deceptive at another.

False, because, as the Wall Street Journal reports, in the very article that was linked, that was the average law graduate's debt among those who borrowed. The Times piece includes no such caveat. Looking at the underlying data, we see that figure is 87%--nearly 1 in 7 have $0 in debt upon graduation.

And deceptive, because that $140,000 figure includes undergraduate debt; certainly, it matters little to the law school graduate where the debt comes from, and law schools should be attuned to the total debt burden of their graduates. But, it's also worth noting that not all law graduate loan debt is attributable to the law school. As the target of the opinion is a critique of the price of law schools, fine-tuning claims with greater precision may be in order.

Back to alcohol. Consider, too, the difference between mean (or average) consumption and median consumption. In the United States, the median American consumes about 3 alcoholic drinks a week. That's not very much, particularly when you compare that to the average figure I cited above of 9.8 drinks a week.

That's because there are a significant number of people who consume zero alcoholic drinks, or hardly anything at all, and a very small number of consumers who imbibe an extraordinary number of drinks. In that figure, then, the average suggests that a "typical" American consumes a lot of alcohol; but, the median, the 50th percentile American, right in the middle of the general population, is only consuming about 3 drinks a week. The average is not a very useful figure on a per-student basis.

Sadly, we lack the more granular data on the median debt loads at schools--despite the fact that we deeply value medians in law school admissions, including LSAT scores and UGPAs. And it might be the case that the debt load picture paints something in the reverse of the alcohol picture. It may well be the case that few students take out "low" debt loads--there may well be a chasm between the no-debt students and the indebtedness students. And if a significant number are slightly below the "average" total, and a handful are quite a bit above that, then the "average" may suggest that the debt situation among graduates is better than it actually is. Of course, the reverse might be true--a few extreme outliers distort the average, and median debt is lower. Given the number of students who enter with zero scholarship-based financial assistance and no personal family support, typical debt loads may be quite different than the "average" figures let on.

In short, the average indebtedness figures are a useful starting point. (Bill Henderson, for instance, notes some of the limitations and complications of such data, and suggests that the problems may be particularly acute at the high end.) But they are hardly a complete picture, and they must be considered in context--on a per-school and a national basis; in conjunction with and distinct from undergraduate debt; taking into account employment outcomes and default rates; and, of course, including careful clarification of what such figures represent, including which graduates they include or exclude.

*Note: this is an extremely quick and dirty analysis using the deciles of data disclosed; I'm sure the actual number is something different.