Abstract
The study on social relationships that influence health, as well as the development of reliable measures to assess this construct has been highlighted in the academic literature. The aim of this study was to estimate new evidence of validity based on the internal structure and reliability of the MOS-SSS, as well as the parameters of items and participants by Item response theory. The sample consisted of 998 people (age: M = 27.18, SD = 9.90, 65.1% women) from different sampling strata. Confirmatory factor analysis (CFA) revealed better goodness of fit of the four-factor model when compared to factor structures shown in other Brazilian studies. The multigroup CFA demonstrated invariance of the factor model when comparing the different sampling strata. The partial credit model indicated items with mean difficulty and appropriate adjustments indices (infit/outfit) and desirable reliability for the factors. The analysis of the maps indicated the tool's strengths and limitations to assess the construct.
Keywords
psychological assessment; item response theory; social interaction; health
