Misremembering Brexit: partisan bias and individual predictors of false memories for fake news stories among Brexit voters

ABSTRACT

Exposure to fake news stories can result in false memories for the events portrayed, and this effect can be enhanced if the stories conform to the reader's ideological position. We exposed 1299 UK residents to fabricated news stories about Brexit. 44% of participants reported a false memory for at least one fabricated story, with a higher rate of false memories for stories that reflected poorly on the opposing side. This effect of ideological congruency was somewhat greater among participants who were exposed to a threat to their social identity as a Leave or Remain supporter; however, this moderating effect was only statistically significant in exploratory analyses using a more conservative definition of false memory. Participants with higher cognitive ability and analytical reasoning scores were less susceptible to false memories. Individuals with better knowledge about Brexit showed better discrimination between true and false stories, while self-reported engagement with the Brexit debate was associated with an increased tendency to “remember” any story, regardless of its truth. These results implicate a combination of social and individual factors in the development of false memories from fake news, and suggest that exposure to social identity threats may enhance the polarising effects of fake news.

KEYWORDS:

• False memory
• fake news
• motivated reasoning
• analytical reasoning
• engagement

Data availability statement

The data and materials associated with this paper may be found at https://osf.io/rn4ae/.

Disclosure statement

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

Notes

1 The exclusion of these participants slightly reduces the power of the two-way ANOVA described below relative to the target sample size in our preregistration, however the change is negligible. The power of the remaining analyses is not affected by this change.

2 Some participants indicated a change in voting preference when asked how they would vote tomorrow compared with how they voted in 2016; 46 participants switched from Leave to Remain, while 36 switched from Remain to Leave. The results reported here did not change significantly if ideological congruency and threat condition were defined based on current (rather than 2016) position on Brexit.

3 The ideological congruency and threat condition variables were not included in our preregistered plan for this analysis; however in hindsight we believed it important to account for the impact of these variables rather than collapsing across them. The inclusion of these variables does not materially affect the outcome for the other predictor variables, and the new analysis remains adequately powered, with 95% power to detect small effects. The preregistered analysis (containing only the variables labelled 1–5 above) can be found in supplemental materials.