In the mid 90s, Nike ran a commercial extolling the virtues of letting high school girls play sports. “If you let me play, I will . . .” Many studies have backed up Nike’s commercial, showing that those who play sports make more money, receive more schooling, are more self confident, etc. However, these studies may be biased by a selection effect—people who choose to play sports may simply be different than those who do not. For example, motivated people that naturally do better in the workplace may also choose to play sports at a younger age. Since the people who play sports are not a random sample of the population, it is hard to determine whether playing sports benefits the people who play them. There has never been a “natural” experiment where people have been randomly assigned to play sports that would allow us to determine whether playing sports has positive effects.
However, Title IX and the sudden increase in girls playing sports in the 1970s provide a very good substitute for such a natural experiment. Title IX was passed in 1972 and banned gender discrimination in schools. It wasn’t immediately clear how or whether this would apply to sports, but it subsequently became clear that schools would have to make efforts to increase girls’ participation rates. Regardless, once Title IX was passed, schools got the hint immediately and started providing opportunities for girls to play sports. The increase was drastic with girls’ participation rates increasing from almost zero in 1972 to over 25% in 1978, when schools were required to be in compliance.
Professor Betsey Stevenson in a recent paper examined the increase in girls’ sports participation rates in different states to determine the effects of playing sports. Before Title IX, boys’ participation rates varied throughout the United States, anywhere from 25% to near full participation. With the passage of Title IX, schools had to attempt to raise girls’ participation rates to match the boys. This eventually led to girls’ participation rates that were very well correlated with boys’ participation rates. On average, the girls’ participation rate became about half of the boys’ participation rate in each state. Using this, Stevenson could get around the selection effect problem by comparing the changes in educational and vocational outcomes across cohorts for different states and comparing that with the change in girls’ participation as predicted by the pre-Title IX boys’ participation rates for the older cohort, while controlling for other variables such as regional differences and economic status. Since there is no a priori reason why the change in educational and vocation outcomes should be correlated with the change in girls’ sports participation as predicted by the pre-Title IX rates of boys’ participation, the relationship between the two can be interpreted as causal.