The truth about predictions and emotions: Two meta-analyses of their relationship

Two lines of research converge on the topic of predicting emotions, namely the response expectancy theory and the affective forecasting paradigm. We propose a revised response expectancy model, in which affective forecasts are a subgroup of response expectancies, referring to emotional outcomes. We conducted two meta-analyses in order to assess the effect size of the relationship between predictions and emotions. 106 studies comprising 301 effect sizes were analyzed. Our results showed a medium effect size regarding the association (r = .46, p < .001), and a small effect size regarding the difference (d = .42, p < .001) between predictions and emotions. Valence of emotion, valence of event, and type of design moderated the difference between predictions and emotions. Individuals are both accurate in the relative sense, as indicated by the response expectancy theory, and inaccurate in the absolute sense, as suggested by the affective forecasting paradigm. Thus, our results support the integration of the two paradigms.