Persistence of Belief Change in the Face of Deception: The Effect of Factual Stories Revealed to Be False

Abstract

Readers typically respond with anger and derogation when they discover that an author has engaged in intentional deception (representing a false story as true). Does this negative response to the author also cause individuals to correct beliefs that may have been changed by the discredited story? In this experiment (N = 160), the alleged truth status of a narrative was manipulated. In one condition, the narrative was presented as fictional (a socially accepted form of untruth). The remaining three conditions initially presented the story as factual. Participants in two of these conditions were informed after reading the story that it was inaccurate due to a) accidental error or b) intentional deception. The story changed attitudes from a no-story control in all conditions. Although readers derogated a deceptive author, they did not correct their attitudes even in the intentional and accidental error conditions. A measure of the parts of the story that the participants rejected (false note circling), suggests that participants were motivated to correct for inaccuracies, but were not able to do so effectively.

Notes

1. Recent studies by Hasson, Simmons, and Todorov (2005) and Richter, Schroeder, and Wohrmann (2009) have shown important limitations on the “comprehension = belief” effect. Specifically, these studies suggest the existence of an epistemic monitoring process that is able to efficiently validate incoming information if individuals have accessible and relevant background knowledge related to the new information. However, Richter et al. (2009) explicitly note that transportation into narrative worlds is “a special case of an epistemic mindset that prevents individuals from epistemic monitoring” (p. 552).

2. Of course, in an autobiography, the author him- or herself also appears as a central character. Although we did not test this case in the current study, we would expect evaluations of the author to have a stronger effect on evaluations of “characters,” or at least the main character, in autobiographies rather than third-person narratives.

3. Another mechanism that may lead to the continued influence of discredited information is the paradoxical effects of thought suppression (Wegner, 1994). Attempting to suppress a particular idea (e.g., “don't think of a white bear”) can lead to increased activation of the concept. However, because narratives typically imply more than just one belief or idea, this process seemed less relevant for the current study.

4. We also included two sets of broader belief items that were related to the story, but less directly implied by the narrative. The first set of general items assessed sympathy toward drug users (e.g., “Other members of society should express sympathy for those people who are addicted to hard drugs like cocaine or heroin”), and the second addressed beliefs about drug policy and drug use outcomes (e.g., “Drugs, including hard drugs like cocaine and heroin, should be legalized” [reverse coded]). However, results revealed that the story did not affect these general items (the experimental conditions were not significantly different from control) and, therefore, these will not be discussed further.

5. The excluded participants were equally distributed across the four experimental conditions. We retained all control condition participants.

6. We also included a measure of transportation into the narrative (Green & Brock, 2000; 16 items, α = .79), which assesses the extent to which individuals are cognitively and emotionally immersed in the story. As expected, transportation did not differ across condition, F(4, 141) = 2.29, p > .06, (with post-hoc tests revealing that the trend was driven by the nonfiction condition being marginally more transporting than the negligence condition), and individuals who were highly transported circled fewer false notes in the story, r(144) = −.16, p = .05. However, transportation was only significantly associated with two belief items in the combined experimental conditions: “the influence of people in a person's neighborhood can lead to that person doing hard drugs,” r(125) = .22, p < .05, and “drug addiction among children is an important problem,” r(124) = .36, p < .01. Examination of these effects by condition revealed that transportation was not a significant predictor of “drug addiction among children is an important problem” in the nonfiction condition, but in the other experimental conditions, more transported individuals were more likely to judge addiction among children as an important problem (marginally significant in the fiction condition). Transportation was a significant predictor of “neighborhood influence” in the intentional deception condition, r(29) = .58, p < .01, but not in the other conditions. While these results provide some suggestive evidence that the influence of transportation may differ depending on the presence of accuracy or correction cues, these findings must be treated with caution because they did not replicate across the set of items.

7. Janet Cooke's story is an exception to the general rule that readers may not know or attend to specific details about narrative inaccuracies; her deception was revealed when “Jimmy” could not be located (and did not exist), so reports recanting the story highlighted this major fact.