The extent of sex differences in psychological traits is vigorously debated. We show that the overall sex difference in the pattern of adolescents' achievement and academic attitudes is relatively large and similar across countries. We used a binomial regression modeling approach to predict the sex of 15 and 16 year olds based on sets of academic ability and attitude variables in three cycles of the Programme for International Student Assessment (PISA) data (N = 969,673 across 55 to 71 countries and regions). We found that the sex of students in any country can be reliably predicted based on regression models created from the data of all other countries, indicating a common (universal) sex-specific component. Averaged over three different PISA cycles (2009, 2012, 2015), the sex of 69% of students can be correctly classified using this approach, corresponding to a large effect. Moreover, the universal component of these sex differences is stronger in countries with relative income equality and women's participation in the labor force and politics. We conclude that patterns in academic sex differences are larger than hitherto thought and appear to become stronger when societies have more socioeconomic equality. We explore reasons why this may be the case and possible implications.