Illusory gender-equality paradox, math self-concept, and frame-of-reference effects: New integrative explanations for...

Gender-equality paradoxes (GEPs) posit that gender gaps in math self-concepts (MSCs) are larger-not smaller-in countries with greater gender equality. These paradoxical results suggest that efforts to improve gender equality might be counterproductive. However, we show that this currently popular explanation of gender differences is an illusory, epi-phenomenon (485,490 students, 18,292 schools, 68 countries/regions). Between-country (absolute) measures of gender equality are confounded with achievement and socioeconomic-status; tiny GEPs disappear when controlling achievement and socioeconomic-status. Critically, even without controls GEPs are not supported when using true gender-gap measures-within-country (relative) female-male differences, that hold many confounds constant. This absolute/relative-gap distinction is more important than the composite/domain-specific distinction for understanding why even tiny GEPs are illusory. Recent developments in academic self-concept theory are relevant to GEPs and gender differences, but also explain other, related paradoxes. The big-fish little pond effect posits that attending schools with high school-average math achievements leads to lower MSCs. Extending this theoretical model to the country-level, we show that countries with high country-average math achievements also have lower MSCs. Dimensional comparison theory predicts that MSCs are positively predicted by math achievements but negatively predicted by verbal achievements. Extending this theoretical model, we show that girls' low MSCs are due more to girls' high verbal achievements that detract from their MSCs than to their low math achievements. In support of the pan-human wide generalizability of our findings, our cross-national results generalize over 68 country/regions as well as multiple math self-belief constructs (self-efficacy, anxiety, interest, utility, future plans) and multiple gender-equality measures. (PsycInfo Database Record (c) 2021 APA, all rights reserved).