Sensitivity to misinformation retractions in the continued influence paradigm: Evidence for stability

Abstract

Research has consistently shown that misinformation can continue to affect inferential reasoning after a correction. This phenomenon is known as the continued influence effect (CIE). Recent studies have demonstrated that CIE susceptibility can be predicted by individual differences in stable cognitive abilities. Based on this, it was reasoned that CIE susceptibility ought to have some degree of stability itself; however, this has never been tested. The current study aimed to investigate the temporal stability of retraction sensitivity, arguably a major determinant of CIE susceptibility. Participants were given parallel forms of a standard CIE task four weeks apart, and the association between testing points was assessed with an intra-class correlation coefficient and confirmatory factor analysis. Results suggested that retraction sensitivity is relatively stable and can be predicted as an individual-differences variable. These results encourage continued individual-differences research on the CIE and have implications for real-world CIE intervention.

Executive function and the continued influence of misinformation: A latent-variable analysis

“…Brydges et al [ 17 ] and Sanderson et al [ 18 ] reported estimates of α = .65 and α = .46, respectively. Similarly, McIlhiney et al [ 32 ], who used two parallel CIE tasks in a test-retest format, reported estimates of ω time1 = .53 and ω time2 = .60. While these findings could suggest that the CIE paradigm suffers from similar issues found when using other experimental tasks in individual-differences research [e.g., 73 ], McIlhiney et al’s [ 32 ] findings suggested that the CIE paradigm showed acceptable stability in individual-differences variation.…”

“…Similarly, McIlhiney et al [ 32 ], who used two parallel CIE tasks in a test-retest format, reported estimates of ω time1 = .53 and ω time2 = .60. While these findings could suggest that the CIE paradigm suffers from similar issues found when using other experimental tasks in individual-differences research [e.g., 73 ], McIlhiney et al’s [ 32 ] findings suggested that the CIE paradigm showed acceptable stability in individual-differences variation. Following recommendations cited by McIlhiney et al, we attempted to alleviate reliability issues by incorporating additional retraction and control items; however, our estimated internal consistency of ω = .57 indicated no improvement to reliability.…”

“…Regarding measurement of the CIE, the current study provides further psychometric data on the CIE paradigm, which has been hitherto limited. [ 78 ] In previous individual-differences research using the CIE paradigm [e.g., 17 , 18 , 32 ] the reported estimates of internal-consistency reliability have been problematic. Brydges et al [ 17 ] and Sanderson et al [ 18 ] reported estimates of α = .65 and α = .46, respectively.…”

“…There were three possible responses, namely: (i) yes, I put in a reasonable effort; (ii) maybe, I was a little distracted; and (iii) no, I really was not paying attention. Participants selecting option 3 were excluded ( n = 3), following precedent [e.g., 32 ].…”

“…The current, exploratory, study aimed to investigate whether individual differences in executive function (EF) were predictive of individual differences in CIE susceptibility. To do this, participants were given three measures of each EF subcomponent from Miyake et al’s [ 19 ] model (i.e., working-memory updating, prepotent-response inhibition, and mental-set shifting), followed by a CIE-paradigm task [e.g., 32 ]. While we did not expect an effect of shifting, we included it for completeness.…”