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

Cognitive Research: Principles and Implications

Table 1 Contrasts run in Experiment 1

dV/hypothesis	Effect tested	F(1,367)	P
False-claim inference scores
H1_FIS: NE < FCO	Familiarity backfire effect	5.31	.022^a
H2_FIS: NE < CO	Illusory truth effect	2.72	.100
H3_FIS: CFC < CO	Effect of claim+correction vs. claim-only	2.73	.099
H4_FIS: NE > CFC	Effect of claim+correction vs. baseline	0.01	.941
False-claim belief ratings
H1_FBR: NE < FCO	Familiarity backfire effect	< 0.01	.971
H2_FBR: NE < CO	Illusory truth effect	3.03	.082
H3_FBR: CFC < CO	Effect of claim+correction vs. claim-only	13.75	<.001^a
H4_FBR: NE > CFC	Effect of claim+correction vs. baseline	4.78	.029
True-claim inference scores
H1_TIS: NE < FCO	Effect of affirmation vs. baseline	36.09	<.001^a
H2_TIS: NE < CO	Illusory truth effect	4.23	.041^a
H3_TIS: CFC > CO	Effect of claim+affirmation vs. claim-only	9.40	.002^a
True-claim belief ratings
H1_TBR: NE < FCO	Effect of affirmation vs. baseline	82.84	<.001^a
H2_TBR: NE < CO	Illusory truth effect	5.32	.022^a
H3_TBR: CFC > CO	Effect of claim+affirmation vs. claim-only	30.95	<.001^a

Note. Hypotheses are numbered H1–4 (primary hypothesis in bold; see text for details); subscripts FIS, TIS, FBR, and TBR refer to false-claim and true-claim inference scores and belief ratings, respectively. Conditions are NE no-exposure; CO claim-only; FCO fact-check-only; CFC claim-plus-fact-check. ^aindicates statistical significance (for secondary contrasts: after Holm-Bonferroni correction)