There are a million ways to be a woman and a million ways to be a man: Gender differences across personality nuances and nations

• Items designed to assess the same traits often vary in their gender differences. • Gender can be very accurately predicted from (multivariate) item sets. • But (univariate) trait scores cannot be accurately predicted from gender. • Cultures vary in traits’ gender differences. • Each gender can correspond to countless trait configurations. Gender differences in personality are typically summarised using broad trait domains. To acknowledge personality traits’ multidimensionality, we studied gender differences on different levels of the personality hierarchy. Using data spanning 74 countries and four different inventories, we compared the multivariate and univariate associations of gender with aggregate trait domains and facets, and unaggregated items as markers of trait nuances. Many gender differences were specific to narrow traits and geographical contexts. While gender is accurately predictable by aggregating small gender differences in multiple traits (multivariate gender difference), individual traits cannot be accurately predicted from gender (univariate gender differences), making many simplistic statements about gender differences inconsistent with data.