What We Should Study

Me a few days ago:

We usually explain human capacity to create and evaluate chains of reasoning in terms of [seeking] truth. … [But] once you give it a bit of thought, you can see many [other] possible and even plausible explanations.

More generally, we humans not only do things, we explain why we do things. Individuals and organizations stand ready to give reasons why we do each of the things we do. While such explanations are often self-serving, they are usually considered the standard default in ordinary conversation, popular media, and in academia.

I have a colleague here at GMU econ who recently expressed to me his feeling that we academics should usually accept such standard explanations unless we see clear strong evidence to the contrary. That is, if an academic journal has a statement of purpose or aim or mission, then we should believe what that statement says about the main social function of that journal in the world — if it says the journal exists to advance knowledge, that is what we should believe. He thinks we should similarly accept official purpose statements of hospitals, universities, charities, and government agencies. (He might not accept mission claims by firms, e.g., “Wal-Mart’s mission is to help people save money so they can live better”; apparently only admired non-profits deserve such deference.)

The most powerful insufficiently-appreciated insight I’ve ever learned is the one intellectual legacy I’d leave, if I could leave only one: we are often wrong about why we do things. Yes it is hardly original, and it might sound trivial, but few appreciate its full depth.

People are way too quick to assume that the main forces shaping the details of common human behaviors and institutions are their standard claimed missions. For example, people assume that the main force shaping doctors and hospitals is their declared mission to make people healthy, that the main force shaping universities and their research patrons is their declared the mission of advancing the frontiers of knowledge, that the main force shaping human capacities to make and evaluate reasons is the estimation of truth, and so on.

Once a social scientist starts to look seriously look for non-standard explanations, however, it is pretty easy to find them. Standard explanations leave many puzzling phenomena poorly explained, phenomena for which non-standard explanations often better account. Yes, there is an unfortunate tendency to latch onto the first plausible non-standard explanation one finds, instead of continuing to search for more possible explanations. I’ve probably been guilty of this myself, such as by perhaps focusing too much on signaling explanations.

But now I understand: today our priority should be a back-to-basics skeptical re-evaluation of human behavior.  That is, we should search for plausible non-standard explanations of our most common behaviors, even those we think “obvious,” and then seek simple matches between the simple robust predictions of each explanation and the puzzling phenomena we need to explain. I’m very interested in participating in such efforts, and uncertain about the best way to proceed.

Within academia, one important obstacle to this project is the tendency of “rigorous” folks like my colleague to insist that non-standard explanations are “extraordinary”, and so require “extraordinary” evidence. They aren’t worth much math modeling until stronger data support is offered, and they aren’t worth collecting much new data to test since they are not yet well supported. (Standard datasets, collected with standard explanations in mind, are usually poorly suited to this task.) Alas the first-cut math models and data analysis appropriate for this first stage of analysis tend to be poor places for academics to signal their math or statistics sophistication.