Dis)continuing the continued influence effect of misinformation

Misinformation can often have a lasting impact on the causal inferences people make about events even after it is unambiguously corrected. This is known as the continued influence effect of misinformation. Understanding the underlying cognitive processes involved in correcting misinformation is important for developing effective counter-misinformation strategies. A review of continued influence studies suggests that methodological factors, such as the scenario in which misinformation appears, can affect the strength of a correction independently of experimental manipulations. This thesis’ primary aim was to advance understanding the continued influence effect and the conditions that give rise to it, by addressing issues with the methodological approach. Experiments 1A-2B developed and validated a methodology for web-based testing of the continued influence effect in order to target larger and more diverse samples. Two key claims from the continued influence literature were replicated; and the introduction of a novel control condition showed that misinformation is referred to even if it is only mentioned as part of the correction. Experiments 3-5 re-examined the claim that corrections, which explain how misinformation occurred, reduce reliance on misinformation. Findings showed no evidence that explanatory corrections reduce misinformation reliance in multiple scenarios; and continued influence of misinformation was observed in some scenarios but not others. Experiments 6-7 revisited the claim that misinformation, which implies a likely cause of an outcome is more impervious to correction than explicitly stated misinformation. Findings showed no evidence for this claim in three scenarios. I argue that the continued influence effect is a tenuous claim predicated on findings from a small set of scenarios that are unrepresentative of real-world situations in which misinformation is encountered. I propose that the conditions under which continued influence of misinformation occurs are poorly understood and recommend formally modelling the continued influence effect to gain a better understanding of this phenomenon.