Since the NY Times didn’t provide a place for comments in the article itself, I thought I should provide some space (as a public good :) The article did provide a link to the published paper, but those who can’t get through the pay-wall may prefer the working paper version (from 2009, so it’s not the same as the published version). FYI, here are some comments from Greg Mankiw and Steve Landsburg (plus an exchange I had with Landsburg).
PS. Of course, unless and until somebody puts a comment here I’m left with the tentative conclusion that maybe this particular public good is non-rival because nobody demands it!!
Update Dec 17 5pm: Oh good, some comments :) And here are some older comments on the working paper when I posted it back in 2009.
Update Dec 18 8am: Based on comments so far (plus emails I received and will post anonymously below) I think I have four observations:
1) Regarding our “story”: The job of an academic paper like this is to tell a “story” about some data. We’re not claiming that it’s The Only Possible Truth, just that it’s one plausible story. And ideally our story should be viewed in the context of previous stories, i.e., previous research in this area. (That’s why academic papers often begin with a literature review.) The previous literature in our area includes all sorts of analysis, from public goods experiments-by-mail (Marwell and Ames 1981; similar experiments are now done by email, see e.g., Rubinstein 2006) to a look at “cheating” on dues-payments to professional associations (Laband and Beil 1999) to a “letter drop experiment” in which fully-addressed envelopes containing cash were left in college classrooms to see whether the ones left in economics classes were more or less likely to reach their intended destination (Yezer et al. 1996). Most of these previous studies (but not all of them) lend weight to our one-sentence lit review in the NYT: “Academic research suggests that there’s a good deal of truth to the stereotype.”
2) Regarding our statistical work: Obviously we cannot explain it all in 700-word NY Times article, so I encourage you to read the working paper version (linked above). It’s worth noting that the paper passed peer-review, meaning that the statistical work should not be dismissed off-hand. My co-author Elaina Rose and I were very careful in trying to focus solely on the effects of economics classes on giving behavior. Of course, our article is only one story about one data set, but I think it’s a reasonable story with a solid analysis of an interesting data set. Again, we had a one-sentence summary in the NYT: “Although students remained anonymous, we could look at all of the 8,743 members of our data set and determine what their majors were, when they took economics classes (if at all) and whether or not they donated to ATN or WashPIRG during each quarter of our study period.” (Incidentally we also knew age, race, and gender, so our analysis took those into account too.)
3) Regarding the nature of public goods: Some comments express doubts about whether WashPIRG and ATN count as public goods. This is definitely fair for WashPIRG but less so for ATN. And, again, it’s worth looking at our research in the context of the previous literature. Frey and Meier (2004) look at a nearly identical situation at the University of Zurich, where students are asked if they want to contribute to scholarship funds for needy students and for foreign students; almost 60% of students donate (!) and again economists donate less than the rest. Perhaps you want to argue that these scholarship funds are also not public goods, or perhaps you want to argue that “all taxation is theft” and therefore that there is a fundamental flaw in all public goods experiments. I stand by the claim in the NYT article that all this is “mostly beside the point. Regardless of the groups’ actual social value, a purely self-interested individual would choose to free-ride rather than contribute.”
4) Regarding politics &etc: Politics comes to the fore surprisingly quickly in many comments, and arguably I started it by closing the NYT piece with the claim that “Our research suggests that economics education could do a better job of providing balance.” Perhaps an economist with an even greater respect for free markets than me would have told a different story (“Look at how economics education is getting students to stop wasting their money on worthless causes!”) but I would like to at least note that if our statistical analysis had come out differently—if we had found that economists give more than others and that taking economics classes increased giving—then we would have still been excited to publish our work. In other words, we didn’t go into this trying to tell a political story; we went into this trying to see what we could learn from a cool data set, and that was the main driver of the whole effort.