From micro to macro: an ecological model of test anxiety in a large-scale assessment

Objective The present study examines the extent to which demographic backgrounds (H1), self-belief (H2) and school climate (H3) predict test anxiety among adolescents in the United States by applying Bronfenbrenner’s ecological theory on human development as a theoretical framework to synthesize previously established predictors within a unified analytical model.Method We conducted a cross-sectional, three-block hierarchical regression analysis using PISA 2015 data from 5,712 U.S. adolescents. After listwise deletion, analytic samples ranged from 5,385 in Block 1 to 5,023 in the final model. Predictors were entered as demographics, self-beliefs and school climate. Sampling and replication weights were applied to account for the complex survey design.Results Demographic factors accounted for 8% of the variance in test anxiety. Adding self-belief factors increased explained variance to 15% with all self-belief predictors emerging as significant unique predictors. School climate contributed only a negligible increment and did not provide clear support for H3. The strongest unique predictors were sex and subjective well-being, followed by science self-efficacy.Conclusions Findings are consistent with the view that test anxiety is associated primarily with demographic background and psychological resources, while the available school-climate indicators added little explanatory value beyond these factors.