Stringham’s piece demonstrating that social drinking builds social capital is fun and intelligent. It makes complete sense that when the alternative to having a couple of drinks at the bar with co-workers or potential clients, is sitting at home watching TV, the drinking man is king. If anyone needed a little bit of convincing that drinking was for their own good, now they’ve got it and they’ve got Ed Stringham to thank.
This related piece, linked to by Tyler Cowen has me a little more sceptical. At first glance, I can’t help but think it suffers from an availability bias. It states in the abstract that they hold constant for drug use. I’m curious what proportion of the sample was elliminated by this. My guess would be a huge chunk. What we really have to ask is, what kind of tenth grade binge drinker doesn’t recreationally use drugs as well? My guess would be somone destined to earn a significantly higher wage than the rest of his peers.

Here’s my availability bias:
First, I remember always hearing survey reports of drinking stats in high school and college and always being shocked at how low they were (think what you will of me or the people I associated with, but I stand by my observations at the very least representing the institutions I attended). So things like more than 6, 12 ounce beers in a night classifies you as a binge drinker. I’m not sure what the specific criterea for this survey was, but I’d bet dollars to donuts if you’d show’d it to me in high school or college I’d have scoffed at its low figure. In this classification almost everyone I remember in high school and college was a binge drinker.
Here’s an empirical claim you might not agree with. I would think that the majority of binge drinking high schoolers are also recreationally using some form of drugs, which would elliminate them from this sample. Once they’re elliminated you’re left with a rather small set. My availability bias probably has the correlation already in it. Almost everyone I knew in high school and college was a binge drinker and I’d consent that they probably earn higher than average wages over the course of their lifetime, but that probably would have been true no matter what. Since everyone I knew in high school and college is much more well to do than someone who makes an avergae wage level.
But even holding income levels constant. The remainder of kids in the sample after taking out drug users leaves a very small subset of nerdy goodie goodies. Isn’t the thesis that abstaining from recreational drug use while being a binge drinker leads to higher wages than taking drugs while binge drinking more reasonable than claiming that being a binge drinker leads to higher wages?
My last concern is that if you control for drug use, as measured by a self reported survey v. recognizable issues like arrest or school discipline related to drug use, you would get drastically different results. When self reporting I think a lot of people would clam up about recreational drug habits, anonymous or not. While using the proxy of people who got caught using drugs, has a selection bias. Rich kids don’t get caught, or at least don’t get in trouble. In this sense, isn’t the wage result obvious?