The present study examined measurement invariance of the Center for Epidemiologic Studies Depression Scale (CES-D) in community groups of Australian heterosexual men (N = 1106), heterosexual women (N = 2111), gay men (N = 527), and lesbians (N = 712). Confirmatory factor analysis of CES-D item scores supported the theorized oblique 4-factor model. There was support for full measurement invariance across the 4 groups, based on differences in approximate fit indices. In contrast there was support for only partial invariance when the chi-square difference test was applied. Lack of invariance was mostly for depressed affect and somatic symptom items, with noninvariant somatic symptom items showing consistently high factor loadings and thresholds among lesbians compared with the other groups. The findings are discussed in relation to the use of the CES-D, the relevance of different depression symptoms to how depressions is experienced by the different gender and sexual orientation groups, and gender role socialization and minority sexual orientation theories.