The gender gap in STEM fields — science, technology, engineering, and mathematics — has been studied so extensively that the volume of research now exceeds the clarity of its conclusions. What the evidence does establish with reasonable confidence is that the gap is not primarily explained by innate differences in aptitude. Girls and boys perform comparably on mathematical assessments through early adolescence in most developed countries, and in some countries girls outperform boys. The gap that opens in later education and career choices is a social and structural phenomenon, not a biological one.

The mechanisms are more contested than the existence of the gap. The stereotype threat literature — showing that women perform worse on math tests when reminded beforehand that women are expected to perform worse — has faced replication challenges in recent years, though the broader finding that social expectations influence performance remains robust. The more durable explanations involve the cumulative effects of discouragement: the teacher who calls on boys more frequently in mathematics classes, the parent who expresses more concern about a daughter's social life than her science grade, the career counselor who steers girls toward people-oriented careers.

The cross-national comparisons are instructive. Countries with lower overall gender equality also tend to show larger STEM gender gaps, which suggests that the gap is downstream of broader gender norms rather than specific to scientific culture. Countries that have made the most progress in closing the gap have typically done so through changes in early education and through deliberate exposure of girls to female scientists and engineers at ages when identity formation is most malleable.

The corporate diversity initiatives that have attracted attention in recent years address a different part of the problem — the pipeline after women have already arrived at STEM careers — without necessarily affecting the cultural influences that reduce the number of women who enter those pipelines in the first place. Both interventions matter; neither is sufficient alone.