Reconstruing past selves following threats to self-esteem and effects on construal level

Abstract

Previous research on Temporal Self-Appraisal Theory demonstrated that people make downward comparisons to their past selves. Researchers have discussed this tendency to denigrate past selves as a self-esteem maintenance strategy; however, little research has been done on how people recall their past selves following active threats to their self-esteem. Reconstruing one’s past self for self-enhancement may lead to changes in construal level. I conducted three studies in which participants were randomly assigned to either an intelligence self-esteem threat or control condition and then tasked to recall an autobiographical memory, rate attributes of their recalled past self, and complete a measure of construal level. In the pilot study (N = 113), participants were free to recall any memory of their choosing. In Studies 1 and 2, participants recalled and rated two memories from early high-school: pre- and post-threat manipulation. Participants in Study 1 (N = 240) recalled their academic experience, while participants in Study 2 (N = 243) recalled their interpersonal relationships. A pattern emerged across studies suggesting that when people recall autobiographical memories related to the domain in which their self-esteem has been threatened (an academic memory and intelligence threat), threatened participants are more likely to denigrate their past selves (lower endorsement of positive self-attributes post-manipulation than pre-test) than controls in threat-relevant traits (competence, knowledge). This pattern is accompanied by an increased likelihood to recall positive transformations and periods of growth, as opposed to stability. However, a relationship with construal level was not observed. Further, when people recall autobiographical memories unrelated to the domain in which their self-esteem has been threatened (an interpersonal memory and intelligence threat), threatened participants are more likely to idealize their past selves (higher endorsement of positive self-attributes post-manipulation than pre-test) than controls in threat-irrelevant traits (likeable, attractive). However, this pattern was only observed for those who demonstrated fixed mindsets, emphasizing stability, and was not associated with an impact on construal level.