Research in behavioral science is hard.
Self-report is notoriously unreliable; rigorous observation of actual behavior is very difficult to get right; creating comparable experimental and control groups takes more than a little ingenuity – and then you have the problem of replicable results. Researchers who find methods that even partially steer around these known difficulties become heroes in their fields; their methods become standards, and genuine scientific knowledge can result.
But sometimes this ingenuity results in denying the obvious, and then we get silly science leading to bad theory. Some behavioral economics experiments, alas, fall into that category. They embody the Wizard of Oz fallacy: insisting that experimental subjects “Pay no attention to the man behind the curtain” when the curtain is standing open for all to see. Continue reading