The usefulness of examining the structural validity of scores on multidimensional measures using nested hierarchical model comparisons was evaluated in two studies using the Social Problem-Solving Inventory (SPSI). In Study 1, the authors systematically explored the factor structure of the measure using hierarchical confirmatory factor analysis, contrasting the results with past research using more traditional approaches. In Study 2, the authors developed and validated a shorter version of the SPSI based on the most promising items identified in the first study. Results from Study 2, using an independent sample, confirmed the stability of the structure and offered construct-related validity evidence for the new scales. Taken together, the results suggest that the modest fit of the SPSI measurement model reported in past research is principally the result of the inclusion of psychometrically poor items rather than a problem with the conceptual model itself. Implications are discussed as related to the organization of social problem-solving dimensions and strategies for evaluating structural validity.
|Journal||Educational and Psychological Measurement|
|State||Published - 1996|