Adult memory for instances of a repeated emotionally stressful event: does retention interval matter?

ABSTRACT

This pre-registered study examined the impact of different retention intervals on remembering each instance of an emotionally stressful, repeated event. Eighty-nine adult female participants imagined being a victim of four similar domestic violence instances over a four-week period. Participants then completed recall and recognition memory questions about each instance either immediately, one-week, or three-weeks after the final instance. Overall, the findings showed that memory performance was often most accurate for the first and last instance compared to the middle instances. That is, participants reported more correct information, made fewer memory errors, and had better quality memory reports for the first and last instances compared to the middle instances. However, following a short delay (i.e., no delay and one-week), participants reported more correct information and were better at discriminating between correct and false details for the last instance relative to the others instances, while at a longer delay (i.e., three-weeks), more correct information was recalled for the first instance compared to other instances (there was no effect for memory discrimination). These findings suggest that memory for instances of a repeated event can depend on the position of an instance, and under some circumstances, the retention interval.

KEYWORDS:

• Retention interval
• repeated events
• victim memory
• emotional stress
• repeated abuse

Acknowledgement

This research was conducted for the doctoral dissertation of Natali Dilevski.

This study is pre-registered on the open science framework: https://osf.io/gb9z7.

Study materials can be viewed on the open science framework: https://osf.io/b79ad/. The data for the study is available upon request from the first author.

We are especially grateful to Christianne Evasco for their assistance with scoring the memory narratives. We are also grateful to all the participants that took part in this research.

Disclosure statement

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

Notes

1 In the repeated-event literature memory accuracy can be defined “narrowly” and/or “broadly” (Dilevski et al., 2020; Price et al., 2016; Woiwod et al., 2019). In this study, we are primarily interested in assessing a narrow memory accuracy ability. This measure reflects an individuals’ ability to attribute experienced details to the correct instance. Two measures are coded to assess a narrow memory accuracy ability–correct details (i.e., a reported detail that occurred in the target instance) and internal intrusion errors (i.e., a detail reported as occurring in the target instance, but actually occurred in a nontarget instance). A broad definition of accuracy is calculated by summing together correct details and internal intrusion errors. These measures combined reflect an individuals’ ability to remember what has happened across multiple instances (i.e., experienced details).

2 In the pre-registration, we noted that we would ask participants to complete a second memory report three-months after the first memory report. We did not collect data for a second memory report because the majority of participants indicated (after the first memory report) that they would not be available in three-months.

3 Due to a technical error, recognition memory data from one participant was not recorded. Thus, this data was not included in the recognition analyses.

4 For example, post-hoc analyses revealed that for instance 3, the three-week delay group recalled fewer correct details (defined narrowly) than the no delay group, t₅₇ = −3.77, p = .001, 95% CI [−17.12, −3.63]. A similar effect between the three-week delay and no delay groups was found for instance 2, though not significant, t₅₇ = −2.46, p = .053, 95% CI [−10.85, .06]. No other differences were significant. This suggests that for instance 2 and 3, more memory decay occurred following a longer retention interval than a short retention interval.

5 Indeed, a post-hoc analysis revealed that for instance 1, recall of correct details (defined narrowly) was equivalent between retention interval groups, F_{2, 86} = 1.57, p = .22, ${\mu }^2$ = .04.