As part of a longitudinal research project, an instrument for assessing a multifaceted self-concept of social competence (contact, empathy, emotion regulation) of German elementary school children was investigated concerning its psychometric properties. Moreover, reciprocal effects between social preference and self-representation of one’s own social skills were analyzed. From the spring of 2007 to the summer of 2010, 26 schools with 54 classes took part in the study that comprised four time points of measurement (T1–T4). A social self-concept questionnaire (SKSozKomp) with 15 items was administered as well as additional questions to assess the peer status. 1Confirmatory factor analysis revealed unsatisfying results for the assumed three-dimensional structure of the SKSozKomp at T1. Including only contact and empathy items led to acceptable model f t indices at T1 to T4. Measurement in-variance tested stepwise for a two-factor latent state (LS) model showed good model f t even for a model with strong factorial invariance though Chi-square difference testing argued for configural invariance. A cross-lagged panel model with 2nd order autoregressive paths revealed small but significant paths from social preference to self-concept (T1 to T2 and T3 to T4) but no significant paths from social self-concept to social preference. Results and limitations of the study are discussed.

Keywords
Social competence; Social self-concept; SEM; Elementary school