In Journal of adolescence
INTRODUCTION : We review the longitudinal evidence documenting that middle and high school students with school-focused possible future identities subsequently attain better school outcomes. Consistent results across operationalizations of possible identities and academic outcomes imply that results are robust. However, variability in study designs means that the existing literature cannot explain the process from possible identity to academic outcomes. We draw on identity-based motivation theory to address this gap. We predict that imagining a possible school-focused future drives school engagement to the extent that students repeatedly experience their school-focused future identities as apt (relevant) and actionable (linked to strategies they can use now).
METHODS : We operationalize aptness as having pairs of positive and negative school-focused possible identities (balance) and actionability as having a roadmap of concrete, linked strategies for school-focused possible selves (plausibility). We use machine learning to capture features of possible identities that predict academic outcomes and network analyses to examine these features (training sample USA 47% female, Mage = 14, N1 = 602, N2 = 540. Test sample USA 55% female, Mage = 13, N = 247).
RESULTS : We report regression analyses showing that balance, plausibility, and our machine algorithm predict better end-of-school-year grades (grade point average). We use network analysis to show that our machine algorithm is associated with structural features of possible identities and balance and plausibility scores.
CONCLUSIONS : Our results support the inference that student academic outcomes are improved when students experience their school-focused possible identities as apt and actionable.
O’Donnell S Casey, Oyserman Daphna
2022-Dec-08
academic expectations, academic outcomes, identity-based motivation, machine learning, natural language processing, possible selves
