The proportion of women (PW) across occupations shows considerable variation. Here we hypothesize that occupational segregation could be moderated by the effect of testosterone (T), leading individuals to gender-typical choice of occupation. To test this, we examined the relationship between PW across 22 occupations and three putative correlates of T (the 2nd to 4th digit ratio [2D:4D], a supposed correlate of prenatal T [PT]; body height, a possible correlate of adult T [AT]; and a systemizing–empathizing score [SQ–EQ], a putative behavioural correlate of PT and AT) in a large internet survey. PW varied from 17% (Engineering/R&D) to 94% (Homemaker) per occupation. Compared to participants in female-typical jobs, participants in male-typical jobs tended to have low right hand 2D:4D and low right–left hand 2D:4D [Dr-l] (higher PT, women only), were taller (higher AT, men and women), and had higher SQ–EQ scores (higher PT and AT, men and women). With regard to women, the relationships for Dr-l and SQ–EQ (but not body height) remained significant when Whites only were considered. We conclude that in women Dr-l, and SQ–EQ are related to occupational segregation, suggesting that high PT and AT are found in women who are in male-typical occupations.