Item-method directed forgetting and perceived truth of news headlines

ABSTRACT

Research on item-method directed forgetting (IMDF) suggests that memories can be voluntarily forgotten. IMDF is however usually examined with relatively simple study materials, such as single words or pictures. In the present study, we examined voluntary forgetting of news headlines from (presumably) untrustworthy sources. Experiment 1 found intact IMDF when to-be-forgotten headlines were characterised as untrustworthy and to-be-remembered headlines were characterised as trustworthy. Experiment 2 separated remember/forget cues and trustworthiness prompts. Forget cues alone had a large effect on memory, but only a small reducing effect on perceived truth. In contrast, trustworthiness prompts alone had essentially no effect on memory, but a large effect on perceived truth. Finally, Experiment 3 fully crossed forget/remember cues and trustworthiness prompts, revealing that forget cues can reduce memory irrespective of whether headlines are characterised as trustworthy or untrustworthy. Moreover, forget cues may bias source attributions, which can explain their small reducing effect on perceived truth. Overall, this work suggests that news headlines can be voluntarily forgotten. At least when people are motivated to forget information from untrustworthy sources, such forgetting may be helpful for curtailing the spread of false information.

KEYWORDS:

• Voluntary forgetting
• item-method directed forgetting
• memory for news headlines
• perceived truth

Disclosure statement

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

Data availability statement

Data and materials for all experiments are available on the Open Science Framework (https://osf.io/csa39/). None of the experiments were preregistered.

Notes

1 G*Power does not allow analyses for within-within interaction effects. At least for 2×2 designs, interactions could however also be conceptualized as main effects for difference scores (e.g., subtracting memory performance after forget cues from memory performance after remember cues to capture voluntary forgetting; this difference score could then for example be compared across two different valence conditions). If conceptualized this way, the sensitivity analyses for main effects might also apply to interaction effects.

2 Originally, additional control conditions without cues were collected for this experiment as well; they are now reported in Supplemental Materials 2 on the OSF project page.

3 This approach to analyzing the free recall data was used in all experiments reported in this manuscript. To make sure that the scoring of partially correct responses with 0.5 points did not skew the results, we examined two alternative ways of coding (namely, strict coding, with all partially correct responses scored as 0s; and lenient coding, with all partially correct responses scored as 1s). For all experiments, the main results for the free recall data stayed the same irrespective of which type of coding was applied. We therefore decided to stick with the coding as reported in the main text, because it most closely captures different types of responses as provided by participants.

4 When the sphericity assumption in repeated-measures ANOVAs was violated, Greenhouse-Geisser corrections were applied.

5 The pattern of results for the ANOVA on truth judgments was however the same when analysis was restricted to participants who completed the truth judgment task first.

6 The pattern of results for the ANOVA on source memory was however the same when analysis was restricted to participants who completed the recognition (plus source memory) test first.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.