Correcting eyewitness suggestibility: does explanatory role predict resistance to correction?

ABSTRACT

Many studies have documented that exposure to post event misinformation can lead eyewitnesses to misremember witnessing events they did not see and do so with high confidence. The goal of the present study was to investigate whether reporting of suggested misinformation can be reversed following a correction, and if so, whether misinformation would be more resistant to correction when it serves an explanatory function than when it does not. In two experiments participants witnessed an event, were exposed to a blatantly false suggestion(s) and one week later received a correction followed by a test of their memory for the witnessed event. We found evidence for both the persistence of misinformation following a correction (E1) and the complete reversibility of misinformation effects following a highly salient correction (E2). Although false reporting of the misinformation doubled when it served an explanatory function relative to when it did not (E1 and E2), in both experiments we found no evidence that resistance to correction varied as a function of the misinformation’s explanatory role. Our findings suggest that, with a salient correction provided by a credible source, people are capable of updating their knowledge with new information that reverses what they previously thought.

KEYWORDS:

• Eyewitness suggestibility
• correcting misinformation
• false memory
• explanatory role
• explanatory coherence

Acknowledgements

The research reported in this article is based on a thesis conducted by Blair Braun in partial fulfilment of the requirements for the master’s degree at Kent State University. We thank Zoe Beard, Erin Satola, Liza Kiefer, Lauren Urban, Johanna Stimmel, Jacob Montgomery, Rachel Dimberio, Jo Wyatt, Gabrielle Norcross, Madeline Leonard, Michael Loeffler, Samantha Aguridakis, Michaela Broadnax, Allison Leon, Alison Rossey, Christie Franks, Taylor Petti, Kathyrn Molle, Allison Smith, A’Lea Yonker, Morgan Raube, Ashley Xoing, Casey Wilson, Christina Ravenscraft, Marjorie Peterson, Jenna Nourse, Tamiqua Handschke, Brittany Burgess, Sedona Sieg, and Jamie Luehring for assistance with data collection and scoring.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In both Experiments 1 and 2, the data were collected at two Midwestern state universities concurrently, with approximately half the participants from each site (N’s = 134 and 141 for E1, and N’s = 144 and 144 for E2). In both experiments, performance did not vary, nor interact, with campus location (all p’s >.10). Hence, all results are reported collapsed across campus location.

2 We gave participants the option of selecting “remember” or “believe” because there is some evidence that explanatory role can have differential effects on these measures of recollective experience (Rindal et al., 2017). In the present study we collapsed across “remember in video” and “believe in video” because separate log linear analyses conducted on “remember in video” and “believe in video” responses revealed a similar pattern of results for both response types (see Appendix E in the supplementary materials for the results). Any misattribution of the suggestion to the video (whether remember or believe in video) is an error, because the video never depicted the suggested toilet papering incident.

3 Although control participants were never exposed to the suggestion, control participants serving in the negation condition were nevertheless exposed to the correction; it was an empirical question whether the correction would influence the base rate. Analyses of control performance showed that in most cases exposure to the correction did not influence the base rate of reporting (see Appendix F of supplementary materials).

4 The prank and spider items were analysed separately because all misled participants received both suggestions and the correction were manipulated within subjects. Analysing both items together in a log linear analysis was not possible without violating the assumption of independent observations.

5 Preliminary analyses of the control items showed that, across all measures, the correction did not influence the base rate of falsely reporting of the suggestions. Explanatory role had no effect for the prank item, but for the spider suggestion, the base rate of false assents and false recall was higher in the explanatory condition relative to the non-explanatory condition (results for analyses of control performance can be found in Appendix F of supplementary materials). For all conditions, the base rate is illustrated in the figures with a horizontal line.

6 This possibility occurred to us after getting the results and not prior to designing the study.