BackHome
Refresh: Off
Paper Summary
Share...

Direct link:

Dynamic Self-Concept and Emotional Responses: Insights from a Situated Laboratory Study in Mathematics

Sun, April 12, 9:45 to 11:15am PDT (9:45 to 11:15am PDT), Westin Bonaventure, Floor: Level 3, Avalon

Abstract

Objectives

We examined the association between state self-concept and state emotions during a laboratory math task, and whether trait mathematics self-concept moderates this within-person relationship.

Perspective(s) or theoretical framework

The academic self-concept has been traditionally viewed as a relatively stable trait. Recent research has begun exploring it as a dynamic state (e.g. Hausen et al., 2022; Niepel et al., 2022, 2025). Despite the current research interest, the relationship between state self-concept, state emotions, and the influence of trait self-concept remains poorly understood. Building on self-concept research (Marsh et al., 1992) and CVT (Pekrun, 2024), we tested the within-person association between state self-concept and state emotions. Further, we explored if trait self-concept moderates this association, hypothesizing that lower trait self-concept increases vulnerability to self-worth threats (Covington, 1984) in moments of low perceived competence, strengthening the link between state self-concept and state emotions.

Methods

Sample and Procedure: N=70 elementary-school teacher students reported their mathematics trait self-concept (Marsh et al., 1992). In a follow-up laboratory session, they completed five blocks of numerical-reasoning tasks. After each block, they rated their state self-concept (“How would you judge your performance in this task?”) on a five-point Likert scale (very bad to very good), and reported their current emotions (joy, pride, anger, fear, frustration, stress) via single items.
Analyses: We used linear multilevel models in R with lmerTest (Kuznetsova et al., 2017) a) to examine within-person associations between self-concept and emotions, including random intercepts and first fixed, then random slopes, and b) to test whether trait self-concept moderated these associations.

Results

a) Within-person associations: We observed significant fixed effects of state self-concept predicting emotions in the expected directions across all achievement emotions (joy: .45, p <.001; pride: .48, p <.001; anger: −.43, p <.001; anxiety: −.09, p <.05; stress: −.12, p <.05; frustration: −.67, p <.001). Model fit improved significantly (p <.001) when including random slopes for joy, pride, anger, and frustration, but not for anxiety (p = .243) or stress (p = .067). Notably, histograms showed high variability of the slopes.
b) Moderation effects: Trait mathematics self-concept positively predicted the slopes for pride (.08, p <.001), anger (.07, p <.05), and frustration (.06, p <.001), while the slope for joy was negatively predicted (-.08, p <.01).

Scientific significance

We adopted an idiographic, within-person approach, providing novel evidence for the link between state self-concept and emotions. Firstly, when participants felt more competent during the task, they experienced more positive and fewer negative emotions. Secondly, substantial variability in the slopes indicated that this within-person relationship varies considerably across individuals. Trait self-concept moderated these associations, qualified in that participants with low self-concepts tended to experience particularly elevated anger and less joy in blocks where they felt they were doing poorly, whereas participants with high self-concepts were more susceptible to experiencing pride in response to higher in-situ competence experiences. We aim to replicate these findings using two additional datasets: a laboratory dataset similar in design but using a domain-unspecific working memory task, and an ESM dataset to enhance ecological validity.

Authors