Recently I seem to be reading a lot of research summaries where my first reaction isn't so much "that methodology is crap" but more "that interpretation is crap". Here's an example, via Futurepundit.

Vindictiveness Doesn't Pay

Vindictiveness doesn't pay. This has been demonstrated by a current study at Bonn and Maastricht Universities. According to this study, a person inclined to deal with inequity on a tit-for-tat basis tends to experience more unemployment than other people. Vindictive people also have less friends and are less satisfied with their lives. The study appears in the current edition of the Economic Journal.

We tend to live by the motto "tit for tat". We repay an invitation to dinner with a counter-invitation; when a friend helps us to move house, we help to move his furniture a few months later. On the other hand, we repay meanness in the same coin. Scientists speak here of reciprocity. A person who repays friendly actions in a like manner is said to behave with positive reciprocity, and one who avenges unfairness acts with negative reciprocity.

Positive and negative reciprocity are interdependent traits: many people incline to positive reciprocity, others more to negative; others, again, incline to both. The researchers from Bonn and Maastricht wanted to discover what influence these traits of character have on parameters such as "success" or "satisfaction with life". For this, they resorted to data from the so-called "socio-economic panel". This contains information gathered by the Deutsche Institut für Wirtschaftsforschung (German Institute for economic Research) in its annual surveys. These involve around 20,000 respondents from all over Germany and cover a diversity of topics.

...

The researchers then related these data to other results of the survey, whereby they stumbled upon a number of interesting correlations: "Thus, positively reciprocal people tend on average to perform more overtime, but only when they find the remuneration fair", declares Professor Dr. Thomas Dohmen of Maastricht University. "As they are very sensitive to incentives, they also tend to earn more money".

This is in stark contrast to vindictive people. With these people, the equation "more money = more work" does not always apply. Even pay cuts are not an effective means of bringing negatively reciprocal people back into line. Ultimately the danger arises that they will take revenge – for example, by refusing to work, or by sabotage. "On the basis of these theoretical considerations it would be natural to expect that negatively reciprocal people are more likely to lose their jobs", Falk explains: "A supposition which coincides with our results. Consequently, negatively reciprocal people experience a significantly higher rate of unemployment".

And in other respects, too, vindictiveness is not a maxim to be recommended. Anyone who prefers to act according to the Old Testament motto of "An eye for an eye, and a tooth for a tooth" has on average less friends – and is clearly less than satisfied with his or her life.

The researchers conclude with a tenant of faith of the liberal orthodoxy: An eye for an eye will make the whole world go blind. The problem with this interpretation of the results is that it's far from obvious that the vindictive do worse because of their vindictiveness. Isn't it more likely that they do worse because of their ingratitude? Let me explain.

The authors of the study use the phrase "tit-for-tat" strategy in a way that is significantly different from the way it is usually used in game theory. In game theory it's usually used to describe a strategy in prisoner's dilemma type games where there are only two possible moves: cooperate or defect. This study appears to be using a model of interaction where there are three possible moves. Players can confer benefit, do nothing, or confer harm. They claim that real people can be divided into groups according to strategy: (1) Those who only reciprocate benefit; (2) Those who only reciprocate harm (the vindictive); and (3) Those who reciprocate both benefit and harm.

The authors seem to think that the poorer performance by group (2) is due to the fact that they retaliate against harm, but it seems to me that there is a much more obvious explanation. Group (2) does worse, not because they are vindictive but because they are ungrateful - they fail to reciprocate benefit. They have fewer friends because they don't reciprocate social invitations. They are less likely to be employed because more money doesn't make them work harder. The problem is not that they respond to harm with "confer harm", but that they respond to benefit with "do nothing".

What I would like to know is how group (3) performed - the group that consistently reciprocates both benefit and harm.