While a substantial literature in economics and finance has concluded that ‘women are more risk averse than men’, this conclusion merits investigation. After briefly clarifying the difference between making generalizations about groups, on the one hand, and making valid inferences from samples, on the other, this essay suggests improvements to how economists communicate our research results. Supplementing findings of statistical significance with quantitative measures of both substantive difference (Cohen's d, a measure in common use in non‐Economics literatures) and of substantive overlap (the Index of Similarity, newly proposed here) adds important nuance to the discussion of sex differences. These measures are computed from the data on men, women and risk used in 35 scholarly works from economics, finance and decision science. The results are considerably more mixed and overlapping than would commonly be inferred from the broad claims made in the literature, with standardized differences in means mostly amounting to considerably less than one standard deviation, and the degree of overlap between male and female distributions generally exceeding 80%. In addition, studies that look at contextual influences suggest that these contribute importantly to observations of differences both between and within the sexes.