The United States legislation prohibiting sex discrimination in federally funded education programmes was enacted in 1972. Its most visible consequences have been in school and college athletics, though its text does not mention sport.
Its record is genuinely instructive, in both directions, and it is worth examining carefully because it is one of the few large natural experiments available in this area.
The participation effect
The change in participation is dramatic and is among the largest documented effects of any equality legislation.
Before the legislation, girls' participation in high school sport in the United States was a small fraction of boys'. It has since risen by more than an order of magnitude, and the ratio has moved from something close to one in twenty to something approaching parity in participation numbers.
College participation shows a similar pattern.
The counterfactual question — how much of this would have happened anyway, given broader social change — is genuine. Studies exploiting variation in the timing and intensity of enforcement have generally found effects attributable to the legislation itself, and the timing of the acceleration corresponds to enforcement rather than to general trends.
The downstream effects
This is where the research becomes most interesting, because it addresses whether sport participation causes anything.
Studies exploiting state-level variation in the speed of implementation have found associations between increased girls' sport participation and later outcomes including educational attainment, labour force participation and employment in previously male-dominated occupations.
These studies use the legislation as an instrument, which addresses the selection problem that undermines most research on sport participation — the difficulty of separating the effect of playing sport from the characteristics of people who choose to.
The findings should still be held with some caution. Instrumental variable designs rest on assumptions that can be questioned, and effect sizes vary between studies. But this is among the better-identified evidence available that sport participation has effects beyond sport.
What did not change
The record on leadership and employment within sport is close to the opposite of the participation record.
The proportion of women's college teams coached by women has fallen substantially since the legislation. Before it, the large majority of women's teams were coached by women; the figure is now well under half.
The mechanism is reasonably well understood. As women's programmes gained resources, coaching positions became better paid and more attractive, and hiring for them moved into athletic departments whose leadership was predominantly male. Women's teams became jobs worth having, and the people who got them changed.
This is the same dynamic described in occupational research more generally: as a role gains status and pay, its demographic composition shifts.
Athletic director positions and senior administrative roles show similar patterns, with women substantially underrepresented.
The resource question
Participation parity has not produced resource parity.
Analyses of institutional spending data continue to find substantial differences in expenditure per athlete, recruiting budgets, coaching salaries and facilities between men's and women's programmes at many institutions.
Compliance is assessed against several tests, and institutions have considerable latitude in how they demonstrate it. Participation-based compliance can be achieved without expenditure equality.
High-profile disputes over facility differences at major tournaments have periodically demonstrated the gap in ways that institutional reporting does not.
What generalises
Three lessons that apply beyond the specific legislation and beyond the country.
Quantitative requirements produce quantitative results. The participation effect is large because the compliance standard was participation, which is measurable and enforceable. Standards defined in terms of opportunity or culture do not produce comparable movement.
What is not measured does not change. Coaching composition was not a compliance metric and it went backwards. This is the single most instructive detail in the whole record.
Enforcement is the variable. The legislation was enacted in 1972 and the largest changes followed periods of active enforcement rather than the enactment itself.
The current arguments
Contemporary disputes concern the application of the legislation to sexual harassment and assault procedures in institutions, and questions about eligibility and participation categories.
These are substantial and contested debates that this article does not attempt to resolve. What is worth noting is that the framework has proved adaptable to questions its drafters did not anticipate, which is unusual for legislation of that age.
And the core lesson stands regardless of position on the current questions: the parts that were measured changed, and the parts that were not measured did not.