Sunday, January 10, 2010

Disconnected statistics about colleges

This summary is interesting (and the graphics make it easy to digest) but as with all such endeavors, e.g. the Harper's Index, statistics disconnected from context can mislead as well as illuminate.

Some quick examples:

1. Does this include 2-year colleges or just 4-year? It never says and it matters.

2. Comparing starting salaries of UConn and Yale grads ignores differences in completion rates between the two schools - starting rather than finishing is probably the treatment of interest - it ignores different earnings paths over the lifecycle, which may accentuate the initial differences, and it ignores other benefits of going to a better school, such as getting, on average, a higher earning spouse.

3. It is hard to tell whether a class is good or not by the title. One could learn a lot about popular notions of justice by critically engaging with Judge Judy. One can learn a lot of labor economics and a lot of industrial organization under the cover of an economics of sports class.

4. What is wrong with a wine-making degree? That is a fine business and one that many have done well in.

5. Economics majors (included in the social science category that supposedly all ends up working at Starbucks) actually do better than majors in some of the hard sciences. Perhaps the authors of this chart should, well, read the literature, in particular, this paper by my friends, Dan, Seth and Lowell.

Understanding the real world, sadly, often requires more than clever graphics and a few statistics presented without context.

Via the Goods**t blog

No comments: