The present study compared the sex-typed preferences for playmates and play styles in children referred for concerns about their gender identity development (199 boys, 43 girls) with that of controls (96 boys, 38 girls). Each child was administered the Playmate and Play Style Preferences Structured Interview (PPPSI) developed by Alexander and Hines (Alexander, G. M., & Hines, M. (1994). Child Development, 65, 869-879). In the two single dimension conditions (playmates and play styles), the controls significantly preferred same-sex playmates and same-sex play styles whereas the gender-referred children significantly preferred cross-sex playmates and cross-sex play styles. Effect sizes ranged from 1.56-2.78. In the conflict condition (which required a choice between same-sex playmates and cross-sex play styles vs. cross-sex playmates and same-sex play styles), there was a general indication of a hierarchical preference for the preferred play style in the single dimension condition as opposed to the preferred playmate except for the gender-referred boys, who showed an inverted pattern. For the gender-referred group, the PPPSI data were significantly correlated with other measures of sex-typed behavior, providing evidence of predictive validity. The PPPSI also discriminated between probands threshold and subthreshold for the diagnosis of gender identity disorder. The results were discussed in relation to both basic and applied issues in the assessment of sex-typed behavior in children.