Rationally Biased Learning

When we assess pros and cons in decision making, we weigh losses more than gains (Kahneman and Tversky (1979)). We are more frightened by a snake or a spider than by a passing car or an electrical shuffle. Such human assessments are qualified of biases, because they depart from physical measurements or objective statistical estimates. Thus, there is "bias" when a behavior is not aligned with a given "rationality benchmark" (like expected utility theory), as documented in the "heuristics and biases" literature (Kahneman et al. (1982); Gilovich et al. (2002)). However, if such biases are found consistently in human behavior, they must certainly have a reason. Some scholars (see (Gigerenzer (2004, 2008); Hutchinson and Gigerenzer (2005))) claim that those"so-called bias" were in