Investigating the replicability and boundary conditions of the mnemonic advantage for disgust

ABSTRACT

Research has demonstrated that people remember emotional information better than neutral information. However, such research has almost exclusively defined emotion in terms of valence and arousal. Discrete emotions may affect memory above and beyond such dimensions, with recent research indicating that disgusting information is better remembered than frightening information. We initially sought to determine whether participants are sensitive to the effects of discrete emotions when predicting their future memory performance. Participants in Experiment 1 were more confident in their memory for emotional (both frightening and disgusting) images relative to neutral images, but confidence did not differ between frightening and disgusting images. However, because we did not replicate the mnemonic advantage of disgust, subsequent experiments were concerned with testing the replicability of this effect. Because metamemorial judgments sometimes eliminate memory effects, participants in Experiment 2 did not make such judgments. Even so, the effect did not replicate. The disgust advantage was ultimately replicated in Experiment 3, where participants completed a secondary task at encoding. The disgust advantage is replicable but appears less robust than previously recognised. A single-paper meta-analysis indicated that the effect is more likely under divided attention, perhaps because the mechanisms which mediate disgust-memory are relatively automatic.

KEYWORDS:

• Disgust
• fear
• memory
• metamemory
• replication

Disclosure statement

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

Notes

1 The authors would like to thank Hanah Chapman for providing the images used in the present investigation.

2 The results of Experiment 3 also argue against this explanation in that the disgust advantage was observed despite a significant difference between frightening and disgusting images in terms of arousal.

3 One participant in Experiment 2 became distressed upon seeing the emotional images during the encoding phase and terminated the experiment. This participant was replaced.

4 In D’Argembeau and Van der Linden’s (2004) studies, encoding instructions manipulated whether participants attempted to memorise the colour of the word, rather than the word itself.

5 Preregistration materials for Experiment 3 are available at: https://aspredicted.org/g58ve.pdf.

6 Due to a computer error, one participant had to restart the LDT from the beginning midway through the task. As such, this participant received additional exposure to some of the studied images. Because analyses indicated that the recall results did not change when this participant was excluded, this participant's data were retained in the analysis of recall data.

7 One participant had an accuracy of 0% because this participant never made a response. Another participant with an accuracy of 0% appears to have reversed the keys associated with the correct responses for the corresponding line positions. Both participants were excluded from all LDT analyses. Because the recall results did not change when excluding these participants, their data were retained in the recall analyses.

8 One participant who reported frequently reversing the rating scale for valence was excluded from analyses of stimulus ratings. Additionally, two participants reported distress and left prior to the stimulus rating phase (but after the recall phase).

9 Note, however, that in Marchewka et al.’s (2016) study of directed forgetting, disgusting and frightening images did not differ in recognition.

10 Alternatively, it may be possible to overcome such confounds between emotional categories by taking a different approach to manipulating emotion altogether. Whereas nearly all laboratory studies of emotional memory to date have used different images or words to represent different emotional categories, it may be possible to instead manipulate emotion by varying the context within which a stimulus is experienced, thereby eliminating a variety of confounds which arise due to differences between items. For example, a piece of food would likely be appraised as being more disgusting if it was placed in a garbage can as opposed to on a plate. We would like to thank an anonymous reviewer for pointing this possibility out.