"It’s the homophily, stupid!" : A cross-country experiment on how apolitical similarities affect political conversations in Europe

Political conversations are crucial for democracy. Yet, the mechanisms behind the selection of conversation partners remain unclear. We investigate the effect of observing similarities in apolitical characteristics—such as gender, age, personality, or lifestyle choices—on people’s decision to engage in political conversations, and whether ideological expectations or apolitical homophily drive this effect. To this end, we introduce the “dynamic parallel conjoint experiment” and field it in the Czech Republic, France, Italy, and Sweden. We find that observing apolitical similarities in others consistently increases the likelihood of engaging in political conversations. Political expectations engendered by these similarities do not play a relevant role. Rather, similarities motivate political conversations directly through homophily, providing evidence supporting the “incidental model” of political conversations (Minozzi et al., 2020). Our study contributes to debates on the dynamics underlying political conversation and introduces an innovative methodological approach for studying implicit mediation in political behavior.