How the self-concept structures social role learning: insights from computational models

Learning about the social expectations tied to upcoming social roles is crucial to promoting adaptation. However, such learning can prompt a strong need for personal change, undermining the stability of individuals' self-concept. Here, we provide a mechanistic account of how individuals at the onset of significant life transitions utilize their self-concept to modulate self-role dissonances during social role learning. Participants engaged in a learning task where they first provided self-ratings for different traits and then estimated how these traits would apply to an individual well-adapted to their forthcoming social role and received trial-by-trial feedback from reference groups. We hypothesized that individuals would employ strategies to minimize dissonances between role expectations and their current self-concept during the learning process. Our computational models included strategies that straightforwardly integrate role expectations to more complex strategies that involve leveraging the self-concept against the pure incorporation of role-related information. The best-performing model demonstrated that the self-concept functions as a modulatory mechanism, guiding the integration of role information to avoid self-role dissonances. Notably, this strategy was strongly accentuated in individuals learning about their upcoming contexts. Our work offers a mechanistic perspective on role learning that may inform interventions to support those facing significant life transitions.