Performance of non-binary athletes in mass-participation running events

Objectives: To test the hypothesis that, controlling for age, natal-sex differences in running performance are lower among non-binary athletes than in the rest of the population. To test the hypothesis that natal-male non-binary athletes outperform natal-female non-binary athletes. Methods: A secondary analysis of 166 race times achieved by non-binary athletes within a data set of 85 173 race times derived from races with a non-binary category in the New York Road Runners database. The natal sex of non-binary athletes was modelled probabilistically using US Social Security Administration data when it could not be derived from previous races. Race times were used as the outcome variable in linear models with explanatory variables derived from natal sex, gender identity, age and the event being raced. Statistical significance was estimated using Monte Carlo methods as the model was not Gaussian. Results: There was no evidence that controlling for age, natal-sex differences in running performance are lower among non-binary athletes. Natal-male non-binary athletes outperform natal-female non-binary athletes at a confidence level of p=0.1%. Conclusions: Both natal sex and gender identity may be useful explanatory variables for the performance of athletes in mass-participation races. It is, therefore, valuable to include both variables in data collection.