“Take the Middle” – Averaging Prior and Evidenceas Effective Heuristic in Bayesian Reasoning
Abstract
When humans revise their assumptions based on evidence, theyprocess information on the (un)certainties of the situation. Thisprocess can be modeled by a (mathematically optimal) Bayes-ian reasoning strategy. Humans typically deviate from thisnorm and apply heuristic strategies, often by only partially pro-cessing the available information (e.g., neglecting base rates).From a perspective of ecological rationality, such heuristicspossibly constitute viable cognitive strategies in certain situa-tions. We investigate the adequacy of a cognitively plausibleheuristic strategy, which amounts to approximately averagingthe probability information on prior hypotheses and evidence.We compare this strategy to optimal Bayesian reasoning and toinformation-neglecting strategies by exploring the situationalparameter space (number of hypotheses, prior and likelihoodvalues). Finally, we frame this in the context of teachers’ diag-nostic judgments on students’ potential misconceptions (pri-ors) based on students’ solutions (evidence) and interpret theresulting accuracy of decisions within the ecology of informalstudent assessment.