This study investigated the psychometric properties of the Loneliness and Social Dissatisfaction Scale (LSDS) in a sample of African-American and Hispanic-American children. Participants were a non-clinical sample (N = 186) of children ages 11 to 13 in the fifth and sixth grades in a school in the Metropolitan New York area. Confirmatory factor analyses revealed a two-factor model fits the sample data better than a one-factor model. Internal consistencies were acceptable across the two factors, and convergent validity of the LSDS was supported by a moderately positive relation with a self-report measure of depressive symptomatology. In a structural equation model, ethnicity, grade, and gender predicted little variance in each LSDS factor, suggesting little measure bias. These preliminary findings suggest that the LSDS is a psychometrically sound instrument for African-American and Hispanic-American children, yet future inspection of its factor structure in more diverse samples of children is warranted.