Measures of gender identity have almost exclusively relied on positive aspects of masculinity and femininity, although conceptually the self‐concept is not limited to positive attributes. A theoretical argument is made for considering negative attributes of gender identity, followed by five studies developing the Positive–Negative Sex‐Role Inventory (PN‐SRI) as a new measure of gender identity. Study 1 demonstrated that many of the attributes of a German version of the Bem Sex‐Role Inventory are no longer considered to differ in desirability for men and women. For the PN‐SRI, Study 2 elicited attributes characterizing men and women in today's society, for which ratings of typicality and desirability as well as self‐ratings by men and women were obtained in Study 3. Study 4 examined the reliability and factorial structure of the four subscales of positive and negative masculinity and femininity and demonstrated the construct and discriminant validity of the PN‐SRI by showing that the negative masculinity and femininity scales were unique predictors of select validation constructs. Study 5 showed that the new instrument explained variance in the validation constructs beyond earlier measures of gender identity. Key message: Even in the construction of negative aspects of gender identity, individuals prefer gender‐congruent attributes. Negative masculinity and femininity make a unique contribution to understanding gender‐related differences in psychological outcome variables. Copyright © 2013 John Wiley & Sons, Ltd.