Does Good News Affect Motivated Reasoning?
This graph plots the treatment effect of seeing good news (vs. bad news) on motivated reasoning. Motivated reasoning is measured by how like subjects believe the news is true. Bayesians' estimates of P(True News) would not be affected by good news or bad news (solid red line). Long-dashed green line: treatment effect of pro-party news. Short-dashed purple line: treatment effect of pro-performance news.
This graph shows that there is no evidence for positivity-motivated reasoning, and that moderate effect sizes, such as those that are as large as the effect of motivated reasoning in politics or performance domains, can be ruled out. There is no evidence that subjects' priors, subjects' demographics, or issue type affect this null result.
Error bars indicate 95% confidence intervals. Religious group: subject affiliates with any religion. Age and income cutoffs are at the median. Prior Good/Bad: initial guess was biased in the good/bad direction.