Record
How the score worksWhat controlled tests found

617 campaign ads, each tested against a control group

Swayable archive · 132 randomized experiments, 2018 and 2020
FavorabilityVote choiceVote choice (yes/no)
+1.41 ptsAverage ad
1.10 ptsSpread between ads
76%Ads above zero
0%Of a new study explained by style
Average ad
+1.4
One spread above
+2.5

Points of vote choice (yes/no).

0%of the next study explained, fitted on all the others

So every script is a test to run, and the chart below is a prior, not a rule.

helps unclear hurtsVolunteer
Helps · 1
  1. Plain-folks appealVolunteer+1.18 ±0.7
Unclear · 26
  1. Veteran as messenger+0.81 ±1.1
  2. Explicit ask to vote forVolunteer+0.56 ±0.6
  3. Contrast+0.42 ±0.6
  4. Amateur, phone-shot styleVolunteer+0.36 ±0.5
  5. Anger+0.32 ±0.6
  6. Negative testimonial+0.31 ±0.5
  7. Pride+0.26 ±0.8
  8. High production valueVolunteer+0.19 ±0.6
  9. Somber+0.15 ±0.6
  10. Explains factsVolunteer+0.12 ±0.6
  11. Humor+0.10 ±0.9
  12. Everyday people as messengersVolunteer+0.06 ±0.5
  13. Sounds scripted+0.06 ±0.5
  14. Advocates an opinion+0.06 ±0.5
  15. Policy facts+0.05 ±0.5
  16. Politician as messenger+0.03 ±0.6
  17. Positive tone-0.03 ±0.4
  18. Personal storyVolunteer-0.04 ±0.6
  19. Enthusiasm-0.05 ±0.4
  20. Positive testimonialVolunteer-0.06 ±0.5
  21. Celebrity as messenger-0.12 ±1.4
  22. Offers a fact the viewer likely hasn't heardVolunteer-0.13 ±0.5
  23. Negative tone-0.17 ±0.4
  24. Cites a source for its facts-0.18 ±0.5
  25. Fear-0.26 ±0.9
  26. Animation-0.63 ±1.0
Hurts · 1
  1. Text on screenVolunteer-1.34 ±1.1

Points of vote choice (yes/no), ads with the feature against ads without, with the 95% interval. Random-effects meta-regression, one feature at a time. Hewitt, Broockman, Coppock, Tappin, Slezak, Coffman, Lubin, Hamidian. How Experiments Help Campaigns Persuade Voters. APSR 2024. Dataverse doi:10.7910/DVN/LBPSSV (CC0).