Optimistic Update for Positive Life Events? An Unbiased Test

Abstract

A large body of research has documented a valence dependent asymmetry in belief updating. Specifically, participants update their beliefs to a greater extent when receiving desirable information compared to undesirable information. Here, we ask whether asymmetric belief updating is observed for predictions regarding everyday positive and negative life events. First, using simulations, we show that an artificial bias in updating would be observed if a skewed distribution of base rates is utilized without controlling for estimation errors. This artificial bias is flipped for positive and negative life events. We then go on to use normally distributed base rates, for which the simulation does not produce an artificial bias. Our human data reveal an optimistic update bias when predicting everyday life events and provides evidence that the pattern is similar regardless if one is estimating a positive or negative event.