Skip to main content

Table 4 Results from Bayesian analyses across Experiments 1–3

From: Can corrections spread misinformation to new audiences? Testing for the elusive familiarity backfire effect

dV Effect direction BF10
Experiment 1
 FIS NE < FCO (familiarity backfire) 2.801
 FBR NE = FCO (no familiarity backfire) 0.154a
Experiment 2
 FISd NE = FCO (no familiarity backfire) 0.135a
 FBRd NE = FCO (no familiarity backfire) 0.363
Experiment 3
 FISl- NE > FCO (corrective effect) 11.757b
 FBRl- NE > FCO (corrective effect) 3.065a
 FISl+ NE = FCO (no familiarity backfire) 0.135a
 FBRl+ NE = FCO (no familiarity backfire) 0.774
Experiments 1–3
 FIS(l-) NE = FCO (no familiarity backfire) 0.104a
 FBR(l-) NE > FCO (corrective effect) 1.799
 FIS(l+) NE = FCO (no familiarity backfire) 0.112a
 FBR(l+) NE = FCO (no familiarity backfire) 0.760
  1. Note. FIS and FBR: false-claim inference scores and belief ratings from the delayed test. As test delay was manipulated in Experiment 2, only the delayed-test variables (FISd and FBRd) were entered into analysis. No-load (FISl-; FBRl-) and load (FISl+; FBRl+) conditions of Experiment 3 were included in separate analysis of Experiment 3, and also in separate conjoint analyses. The condition factor includes only conditions NE (no-exposure) and FCO (fact-check-only). aindicates substantial and bindicates strong evidence for or against the null