Blind Analysis as a Correction for Confirmatory Bias in Physics and in Psychology

Abstract

Psychologists and other social scientists have increasingly realized the extent to which their data analytic practices are vulnerable to confirmatory biases. In response to similar concerns, physicists have developed various methods of "blind data analysis," in which the analyst is unable to determine the actual substantive results of a study until the main analyses are judged to be complete. We give illustrations from particle physics and cosmology, and offer simulations to illustrate how similar methods might be adapted to common analytic situations in psychology.

Keywords: Confirmation bias; confirmatory bias; data analysis