Doors and Perceptions: Motivations, Beliefs, and Returns to Canvassing (Pre-Recorded)

• In Event: Theoretical Implications of Empirical Models

Abstract

A large empirical literature has documented that canvassing significantly affects voting behaviour. However, returns are heterogeneous across campaigning methods and across countries. This paper proposes a formal model of canvassing to (1) rationalise these differences, (2) suggest ways to optimise campaign strategies, and (3) interpret evidence from field experiments. Activists, who differ in their motivations for engaging in political activities, decide how often to participate in canvassing, and what message to share with voters if they do. Both the participation decision and the message shape the voters' beliefs and their voting behaviour. I show that, in equilibrium, activists with instrumental motivations are more effective at persuading voters but can be less motivated than activists with non-instrumental motivations. Returns to canvassing are therefore non-monotonic in the share of activists with instrumental motivations. As a result, parties can optimise campaigns by recruiting the right types of canvassers. In particular, the optimal recruitment strategy balances different types of activists. Because activists internalise the effect of their participation on their persuasiveness, returns to door-to-door canvassing are always positive, in line with empirical evidence. This arises even when simple communication would generate no returns. Finally, I show that if voters think that the choice of participation carries some information, one should be cautious when extrapolating results from small-scale field experiments that artificially control participation.

Author

• Clement Minaudier, University of Vienna