Background: Gender differences in educational outcomes encompass many different areas. For example, in some educational settings, boys lag behind girls on indicators of educational success, such as leaving certificates and type of school attended. In studies testing performance, boys typically show lower competence in reading compared with girls, yet tend to show higher competence in school subjects related to mathematics. While such differences in competence between the genders can be relatively small, they coincide with much greater differences in motivation-related variables emerging during the school years, and thus seem to channel students into lifelong gendered pathways via gendered educational and occupational preferences. Purpose: From a psychological perspective, we propose the "Interests as Identity Regulation Model" (IIRM) as a useful tool for understanding many of the gender differences in educational outcomes. Specifically, the focus is on two areas of research: girls' and women's under-representation in subjects such as maths and science; and boys' lower engagement at school in general. Sources of evidence: Findings from recent research, mostly from a psychological perspective using quantitative measures and empirical studies testing the IIRM, are reported to illustrate different aspects of the interplay between students' gender identity and gendered social meanings of academic domains (such as maths), as well as academic engagement in general. Main argument: IIRM suggests that the perceived fit between students' gender identity and the gendered social meanings associated with different possible behaviours at school (e.g. choosing a subject, investing effort or not) is a relevant heuristic for students' directing of their learning activities. The male stereotyping of maths and science implies a greater misfit between girls' gender identity and engagement in these domains. The perception that displaying effort and engagement at school is feminine leads to a misfit between boys' gender identity and academic engagement in general. Conclusions: Attempts to alleviate gender differences in educational outcomes that channel students into lifelong gendered pathways with regard to qualifications and occupations will benefit from an understanding of how closely these academic choices are related to students' gender identity. Interventions should aim at enhancing the individually perceived fit between a student's gender identity and engagement in specific subjects or learning activities. The nature of such interventions will be an important topic of future research.