EDIT: I think I was also probably premature with the whole idiot thing, although I think Olly’s point is right on.

Fair point. But I don’t think that contradicts my point – which I would like to clarify: the conclusions derrived from the experiment are faulty. Taxes, by definition, come from an earned income. The experiment can apply to money given but it cannot apply to taxes (some exceptions like inheritance taxes could be loosely tied to the research but even that is highly dubious).

Tim: No one is calling anyone an idiot. I am calling a particular aspect of the experiment and the conclusions being drawn “idiotic” – which they are. However, I am excited to read the full paper and possibly have my views amended.

PS: nice to talk with you all again. I hope you are both doing well.

Olly: If you haven’t already emailed Prof. Harbaugh about it I can send it along, just let me know. And IIRC, you can draw significant conclusions from small samples, but you can’t always assume its mean is the true mean. Anyway, that’s why I wonder about the small sample size. After I have a chance to read the paper I may be of a completely different opinion.

Colin: You can call Bill Harbaugh a lot of things, but I think “idiot” is pretty far off the mark. I mean, as I stated in my post I have not read the paper yet and this post is based on the Science synopsis, so I could be pretty off-base and I am certainly not an expert on neuroeconomics. Mostly I thought it was an interesting idea, that it’s cool a professor in the UO department is in Science, and that it might be worth reading the whole thing to see what they actually claim.

Yours, Bill Harbaugh.

Steady on, Colin, jeez. In order for an experiment to be “scientific”, it has to be replicable: that is, independent researchers should be able to do the same thing you did and get comparable results. For this to happen, external factors have to be controlled as much as possible. So, sending people out to earn their seed money before conducting the experiment would be pretty much the least scientific thing you could possibly do in this situation, although I’m sure it would result in some nice anecdotes.

Tim: “As the ODE story points out, only 19 people were sampled for this study. That’s a pretty small sample to draw conclusions from…”

It’s not statistically significant, but are they claiming it is? (Any chance you could email me that PDF?)