Further Reflections on Disentangling Shape and Level Effects in Person-Centered Analyses: An Illustration Exploring the Dimensionality of Psychological Health

Abstract

Morin and Marsh (2015) proposed a methodological framework to disentangle shape and level effects in latent profile analyses. We discuss limitations of this framework (based on a logic similar to that of higher-order measurement models), and suggest that these limitations are easily solved by a more thorough examination of the variable-centered measurement models underlying profile indicators. This study presents complementary variable- and person-centered approaches aiming to assess the dimensionality of psychometric constructs. Psychometric measures often assess separate conceptually-related facets of global overarching constructs, based on the assumption that these overarching constructs exist as global entities including specificities mapped by the facets. The framework proposed here explicitly models this dimensionality in both variable- and person-centered analyses. To illustrate this revised psychometric framework, we use ratings of psychological health collected from 1,232 teachers, and show how this revised framework provides a clearer picture of teachers’ profiles of psychological health.

Keywords:

• bifactor
• dimensionality
• ESEM
• latent profiles
• person-centered
• variable-centered
• well-being

FUNDING

Preparation of this article was supported by a research grant from the Australian Research Council (LP140100100) awarded to Alexandre J. S. Morin and Herbert W. Marsh, and by an international research collaboration support grant from the University of Montreal–Direction des Relations Internationales awarded to the Jean-Sébastien Boudrias and Pascale Desrumaux.

SUPPLEMENTAL MATERIAL

Supplemental data for this article can be accessed at www.tandfonline.com/hsem.

Notes

¹ Alternative bifactor-CFA and bifactor-ESEM models including two G-factors representing psychological well-being and psychological distress were also estimated, but these alternative models failed to result in a meaningful improvement in the fit of the model, and resulted in highly correlated G-factors (r values close to .80) so that these models were not retained.

² The reason why both first-order and bifactor measurement models were retained at this stage is that both are needed for the estimation of the person-centered models described in Figure 1 (Models 1 and 3 are based on factor scores saved from first-order models, whereas Model 2R relies on factor scores from a bifactor model).

³ It is interesting to note that the unexpected pattern of loadings and cross-loadings observed for these items makes sense substantively. For instance, Item W12 assess whether participants “are curious and interested in all sorts of things,” which can as easily reflect a sense of harmony as a sense of involvement. Item W20 cross-loads negatively on the Distance factor (Distress), which is the logical opposite of the Involvement factor (Well-being). Finally, Items D17 (I generally lack initiative and drive) and D18 (I feel useless), although they are designed to assess Distance, clearly also tap into known manifestations of depression.

Additional information

Funding

Preparation of this article was supported by a research grant from the Australian Research Council (LP140100100) awarded to Alexandre J. S. Morin and Herbert W. Marsh, and by an international research collaboration support grant from the University of Montreal–Direction des Relations Internationales awarded to the Jean-Sébastien Boudrias and Pascale Desrumaux.