The Self-Reference Effect in the Visual and Auditory Modalities: Effects of Referent and Valence on Memory Performance

Abstract

People tend to better remember information that has been encoded in reference to the self than information that pertains to someone else, a phenomenon termed the self-reference effect (SRE). It is also believed that this bias for self-relevant information is selective, such that healthy adults prioritize the encoding of self-positive relative to self-negative information (a self-positivity bias). Depressed individuals, on the other hand, are believed to display a self-negativity bias, whereby they remember more negative than positive information about themselves. Previous studies have assessed these two biases using the Self-Referential Encoding Task (SRET). In this task, participants first endorse, using a yes/no judgement, visually presented positive and negative trait adjectives, as either accurately representing themselves or another known character (e.g., Harry Potter). This task is then followed by surprise memory tasks for these adjectives. After the task, when depression is a variable of interest, participants complete a self-report measure of depression. In our study, depression was assessed in a non-clinical sample using the CES-D self-report measure. Participants classified as "depressed" were those who reported levels of depressive symptoms superior to 16, based on the measure’s established cutoff. To our knowledge, no research using the SRET has examined whether these two biases also exist when the information is presented through the auditory modality. Given that self-relevant information is often encountered through spoken language in daily life, it is important to explore how these biases operate in the auditory modality. In the present study, participants were assigned to complete the SRET in either the visual (n=176) or auditory (n=176) modality. Results confirmed a significant SRE in both modalities and did not reveal an interaction between SRE and modality. Contrary to expectations, there was no evidence of a self-positivity bias, nor were there any differences in the pattern of results for depressed (n=186) vs. non-depressed (n=166) participants in the recognition task, although a significant decrease in positivity bias was found for depressed individuals during the endorsement task. Overall, these findings suggest that the SRE is consistent across modalities. However, the absence of both a self-positivity bias in healthy individuals and a self-negativity bias in depressed individuals diverges from previous research. Given the large in-person sample size and highly controlled stimuli in this study, these null effects warrant further investigation into the valence-related memory biases previously reported.