The present study assesses the factor structure, psychometric adequacy, and invariance across sex and age of the Relational Provisions Loneliness Questionnaire (RPLQ). Discriminant validity with an external criterion was also tested. In a sample of Portuguese adolescents, from 7th to 9th grade (N = 817), Confirmatory Factor Analyses (CFA) were conducted to test a four-factor model of loneliness (lack of integration and/or intimacy in peer group and/or family). Results evidenced empirical support for the structure of the RPLQ loneliness scale, which fitted very well the proposed model, and provided adequate fit to the Portuguese data. There was substantial support for the construct validity (factorial, convergent, and discriminant) and reliability of the RPLQ. Measurement invariance (configural, metric, and scalar) was established across sex and age. Finally, it was assured discriminant validity, provided by the contrast with the social functioning dimensions in peer group. Overall, our findings support the conceptualization of loneliness in adolescence by peer- and family-related loneliness through lack of integration and intimacy. In a single instrument, the RPLQ loneliness scale combines measures of four important aspects of adolescents' social life. This seems to be an adequate instrument to be used in the study of adolescents' loneliness, in its different forms and across relational contexts.