Optimistic update bias holds firm: Three tests of robustness following Shah et al.

- Three tests reveal the robustness of optimistic belief updating.
- Optimistic belief updating is observed for both positive and negative life events.
- Belief updating is more Bayesian in response to good news than bad.
- It has been claimed that optimistic belief updating does not exist.
- We show that the methodologies used to make this claim are flawed.

A diverse body of research has demonstrated that people update their beliefs to a greater extent when receiving good news compared to bad news. Recently, a paper by Shah et al. claimed that this asymmetry does not exist. Here we carefully examine the experiments and simulations described in Shah et al. and follow their analytic approach on our data sets. After correcting for confounds we identify in the experiments of Shah et al., an optimistic update bias for positive life events is revealed. Contrary to claims made by Shah et al., we observe that participants update their beliefs in a more Bayesian manner after receiving good news than bad. Finally, we show that the parameters Shah et al. pre-selected for simulations are at odds with participants' data, making these simulations irrelevant to the question asked. Together this report makes a strong case for a true optimistic asymmetry in belief updating.