A canvasser knocks and talks with you for ten minutes about immigration. They can do two separate things. Researchers switched each one on or off independently, for 1,485 people, so every combination was really run.
The numbers above are d: the shift weighed against how far apart people's answers were to start with.
A real shift, and a modest one. The pooled figure covers 69 interventions and works better with school and college groups than with adults; the door conversation is measured on adults, in the field.
Hearing a concrete story about one person is what moves people; arguments alone showed no detectable effect, and listening added nothing to persuasion, though it raised liking and trust. Every field test so far is American.
Figures from the article's sources: Santoro et al. (2025); Kalla and Broockman (2020); Hsieh, Faulkner and Wickes (2022); Brock-Petroshius and Gilens (2025).