A meta‐analytic investigation of the factor structure of the PANAS

The 20-item Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988) is a self-report measure to assess two global measures of psychological well-being, namely positive affect (PA) and negative affect (NA). Its brevity and repeated evidence of sufficient levels of reliability and validity has contributed to a frequent use in all areas of psychology. Moreover, the PANAS has been translated into various languages and is administered all over the world. Despite its widespread use, there is still an ongoing discussion with regard to the internal structure of the PANAS. Though originally developed to provide distinct and independent measures of PA and NA, empirical studies identified different factor structures including two- and three-factor models, second order models, and bifactor models. Additionally, there is few information on the robustness of the internal structure of the PANAS across, for example, different application contexts and questionnaire characteristics. In light of the ongoing discussion, the present study evaluates the nature and the generalizability of the PANAS factor structure by means of a meta-analytic structural equation modeling approach (MASEM; Cheung & Chan, 2005). In a first step, inter-item correlation matrices from 76 independent samples (total N = 54,976) were pooled. Then, popular factor models for the PANAS were compared using confirmatory factor analysis. Overall, the originally proposed orthogonal two-factor model exhibited a rather inferior fit (CFI = .884, TLI = .871, RMSEA = .052). In contrast, a bifactor model was the most appropriate representation of the PANAS (CFI = .930, TLI = .912, RMSEA = .043). This model included two specific factors for PA and NA as well as a general factor that represents a fundamental approach or withdrawal tendency (i.e., affective polarity). Moderator analysis revealed profound differences in the internal structure of the PANAS between the original English version and translated versions, leaving some doubts on the appropriateness of the application of the PANAS in cross-cultural research. References: Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. Cheung, M. W.-L., & Chan, W. (2005). Meta-analytic structural equation modeling: A two-stage approach. Psychological Methods, 10(1), 40–64. https://doi.org/10.1037/1082-989X.10.1.40