Factor stationarity and invariance of the POMS in cocaine patients

The Profile of Mood States (POMS) factor scores are commonly used to evaluate the clinical pharmacologic effects of a drug. Two assumptions made when using the POMS in this manner are that its factor structure remains constant across time (factor stationarity) and that the factor structure is identical across experimental groups (factor invariance). If these assumptions are not met, then interpreting change in POMS factor scores becomes problematic. Structural equation modeling programs (e.g., LISREL) may be used to explicitly test assumptions of factor stationarity and invariance. As part of a National Institute on Drug Abuse grant-funded, double-blind, placebo-controlled trial to assess effectiveness of treating cocaine-abusing patients with fluoxetine, POMS data were collected at baseline and at 6 weeks (n = 90). Applying a longitudinal factor analysis strategy, factor stationarity and invariance of two POMS factors were assessed in this patient sample.