Christensen, T. C., Wood, J. V., & Barrett, L. F. (2003). Remembering everyday experience
through the prism of self-esteem. Personality & Social Psychology Bulletin, 29, 51 – 62.
Prior research shows that there is a tendency for people to revise history by altering their memories in self-aggrandizing ways. (e.g., John Dean gave an accurate description of how the Watergate cover-up had unfolded, but he changed his recollection to make himself look better by saying that he had the courage to warn Nixon, etc.). Christensen, Wood, and Barrett claim that whether this type of revision occurs depends on a person’s level of self-esteem.
The authors are proposing that self-esteem works to bias recall by changing autobiographical memories to be congruent with one’s self-theory or self-beliefs. Thus, people who are high in self-esteem (HSE) may recall autobiographical memories in self-aggrandizing ways and people low in self-esteem (LSE) may recall their autobiographical memories in self-denigrating ways. The purpose of this research was to build on the existing literature on self-esteem and memory bias. This was done in six ways
n Differences in participants’ initial subject ratings of experience were controlled for.
n Memory for one’s own reactions to event (instead of memory for the events) was studied.
n Memory for experience in response to specific events was studied.
n Memory for a variety of reactions was studied (state self-esteem, positive & negative emotions, and event valence).
n Bias rather than accuracy was focused on.
n Everyday memories (rather than laboratory-based events) were studied.
Method
In two experiments, participants took home a packet of self-report measures including the Rosenberg Self-Esteem Scale. Three days later they returned the packet and received seven diary records and were asked to complete one record each night for seven nights. In addition to the diary record, the participants were asked to indicate their state self-esteem, their positive and negative emotions, and the event valence.
Twice after the end of the week, participants completed surprise recall measures for a specific pleasant and unpleasant event (in study 1 this was done 9 and 13 days after the week and in study 2 this was done 9 and 21 days after the week). The participants were given new diary forms that contained their original narrative followed by blank rating scales. The participants were supposed to answer the scales “exactly as they believed they had answered them the night the event occurred.” This protocol was the same across the two studies, except in study 2, the participants completed a mood measure immediately before the recall session.
Results
In study 1, HSE and LSE individuals reported similar initial experiences in their events. In study 2, however, global self-esteem was positively correlated with state-self-esteem ratings for all events selected for recall and negatively correlated with negative emotion ratings for most events. (These differences were controlled for in further analyses.)
Relative accuracy: participants’ recall of state self-esteem was strongly related to their diary ratings of state self-esteem across both studies.
Bias: Global self-esteem was uniquely related to bias in recall of state self-esteem for six of the eight events tested. As global self-esteem increased, so did their tendency to recall feeling better about themselves following pleasant events. Global self-esteem also predicted bias for unpleasant events, but to a lesser extent.
Relative accuracy: recall of positive emotion was strongly related to diary ratings of positive emotion.
Bias: Global self-esteem predicted bias, but only in study 1 at the second recall session. Higher self-esteem was associated with less underestimation of positive emotion and lower self-esteem was associated with greater underestimation of positive emotion. Global self-esteem did not predict emotion in study 2 and the authors speculate that this could be due to the administration of the mood measure prior to the recall sessions.
Relative accuracy: recall of negative emotion was strongly related to diary ratings of negative emotion in both studies.
Bias: Global self-esteem did not predict bias in remembered negative emotion.
Relative accuracy: Recall of valence was related to diary ratings for all events except unpleasant ones in study 1.
Bias: There was a trend for HSE participants to recall unpleasant events at recall 1 as more pleasant than LSE participants (study 1) and for HSE participants to recall pleasant events as more pleasant at recall 1 (study 2).
The authors are fairly confident that these effects are not driven solely by mood at the time of recall.
Discussion
The results of these two studies suggest that global self-esteem is related to memory bias for different aspects of autobiographical memories, including state self-esteem, positive emotion, and event valence. HSE participants recalled their daily events more positively and LSE participants recalled their daily events more negatively. Bias was strongest for feelings about the self and weakest for memory of event evaluation.
Interestingly, the bias effects diminished after the delay, and one explanation offered for this effect is that bias occurs up to the point that the event is still relevant to the self and then diminishes as the event becomes less salient.