LAUSR

Theorists have long argued that discussing public affairs with others increases citizens’ knowledge of politics. Yet, empirical tests of this claim reach contradictory results, with some studies reporting large effects of discussion on knowledge while others report small effects or fail to confirm the hypothesis. To account for this inconsistency, the current study meta-analyzes this literature. The results, based on 163 research findings from 134 independent studies (N = 412,933), indicate positive and significant mean effect sizes of r = .15 for discussion frequency, r = .1 for discussion heterogeneity, and r = .18 for discussion network size. While all three effects are statistically significant, a meta-analytic relative weights analysis reveals that discussion heterogeneity explains little variance in political knowledge once discussion frequency and network size have been accounted for. In other words, how much citizens talk about politics matters much more than whom they talk to.