Abstract
Keyes’s theory-driven model of mental health uses a diagnosis which leads to three levels of positive feelings and positive functioning: languishing, moderately mentally healthy and flourishing (Keyes 2002). Although these three-level categories may be justified for a unidimensional factor structure, or a factorial structure using summated scales, the recent works supporting a multidimensional structure of the Mental Health Continuum Short Form (MHC-SF) suggest the adoption of shape-based rather than level-based categories of mental health (Morin et al. 2017). This research aims at testing the empirical validity of Keyes’s taxonomy and its relationship to psychosocial functioning. We first adopted a variable approach by selecting the optimal factor structure for the MHC-SF: the bifactor exploratory structural equation modeling (Bi-ESEM). Following with a person-centered approach, we used a latent profile analysis on the factor scores of the Bi-ESEM. Psychosocial risks indicators were used as outcomes for testing the criterion validity of the models on a sample of 1065 French workers. Results show that the mental health subgroups are more intricate than the three-levels categories originally theorized by Keyes (2002, 2005). While the general Bi-ESEM factor warrants three levels akin to those of Keyes, accounting for the specific factors reveals two profiles of languishing: the hedonic languishers characterized by a low level of emotional wellbeing, and the eudaimonic languishers characterized by a low level of psychological wellbeing. Both variable and person-centered approach confirm Keyes’s initial statements (Keyes 2007) that those who are languishing exhibit the highest levels of psychosocial risks while those flourishing display the lowest. A rule-based diagnosis derived from a decision tree method is suggested as an alternative to a level-based taxonomy.