The robustness of the win-win effect

We demonstrate that positive relationships between measures of national gender equality and Olympic medal wins are robust across a variety of appropriate statistical approaches to analyzing cross-national data. First demonstrated by Berdahl, Uhlmann, and Bai (2015), who controlled for GDP, population, latitude, and income inequality, we show that relationships between gender equality and medal wins remain positive when controlling for GDP per capita, consistently log-transforming positively skewed variables, and fully analyzing all four gender gap subindexes. The Win-Win effect is most robust for gender equality in education and earnings. Controlling for arbitrarily-defined world regions ("Anglo-Saxon countries" vs. "Africa") is inappropriate, as such groupings are based on folk stereotypes, not objective scientific criteria, and risks masking meaningful differences between countries. There is, however, often more than one right way to analyze a dataset; we discuss how this can be addressed by crowdsourcing the analysis of complex datasets prior to publication.