Hello everybody. I am an academic economist, and I have a question.
There is a "common" type of error I see in some of my colleagues, and I would like to see what epistemologists have said about it.
Here are two examples.
From the 60s to the 80s, wages rose and at the same time males worked less while females worked more.
When wages rise, there are two competing forces in the determination of hours worked: on the one hand working becomes more attractive (or leisure more expensive; this is the substitution effect); on the other hand, if you worked the same number of hours, you would be richer and would like to consume more leisure (that is the income effect).
If somebody asks "why did males work less and females more in response to rising wages?", an incomplete answer would be "because for males the income effect dominates, while for females the substitution effect dominates." That is almost tautological.
A "real" explanation (there are some out there) would explain why that is the case. One explanation, for example, works with fertility decisions and says that since women spent a lot of time with kids, having kids became more expensive, and that reduced the number of kids families had.
The "informal fallacy" in this case would be "I have explained why males work less and females more in response to rising wages, by arguing that males have a stronger income effect, while females a stronger substitution effect".

The second example has to do with polarization as studied by Lord Ross and Lepper (J. Personality and Social Psychology, 1979, "Biased Assimilation and Attitude Polarization: The Effects of Prior Theories on Subsequently Considered Evidence"). In those experiments two groups of people who disagree about an issue (say, the effectiveness of a policy, like the death penalty to deter crime) are shown a common piece of evidence, and their beliefs polarize further (those who believed the policy was effective become more convinced of their initial views; those who believed it was ineffective become more convinced of their initial views).

For economists trying to explain the phenomenon with a "rational" framework (where people apply Bayes rule to form their beliefs) the first puzzle was "how can two people's beliefs move in opposite directions, when faced with common evidence, if they are processing information in a rational (Bayesian) way? how can the same evidence be "good news for the effectiveness of the policy" for some, and "bad news" for others?".

The possibility of two people polarizing was first pointed out by Walley (1991), and studied further by Teddy Seidenfeld and L. Wasserman (1993), "Dilation for Sets of Probabilities," Annals of Statistics 21(3), pp. 1139-54.

Then, an informal fallacy in this case is the following. "A person with a high belief that a policy is effective can raise her beliefs about its effectiveness when presented with a given piece of evidence, while at the same time a person with low initial beliefs might revise her beliefs downward after being shown the same piece of evidence. Hence we have explained why people with high beliefs increase their beliefs, while people with low beliefs decrease their beliefs."

I think there is some error in that conclusion as it has not been explained why it is that those with high beliefs increase, and those with low beliefs decrease.

The analogy with the wages case is the following: males are people with high beliefs and females those with low beliefs; just pointing out that males can have a larger income effect (people with high beliefs can see their beliefs increase) is not enough to explain why we observed the systematic pattern (of males working less overall, or people with high beliefs increasing their confidence in the policy).

The bottom line of this post is that I want to know if this kind of "faulty reasoning" or "faulty explanation" has been studied or been given a name. Can anybody point to a reference?

Much appreciated

Juan