Psychometric properties of the MHC‐SF: An integration of the existing measurement approaches

Abstract

The structure of the MHC-SF is best represented as combination of the bifactor and ESEM model.

Well‐being in Patients with Affective Disorders Compared to Nonclinical Participants: A Multi‐Model Evaluation of the Mental Health Continuum‐Short Form

“…However, not surprisingly, SWB had the lowest internal consistency and composite reliability of the three subscales across samples, in line with most nonclinical (e.g. Hides et al, ; Keyes et al, ; Petrillo et al, ; Rogoza et al, ) and recent clinical findings (Franken et al, ; van Erp Taalman Kip & Hutschemaekers, ) and probably due to the aforementioned heterogeneity of the SWB subscale.…”

“…Considering well‐being under a broader perspective, it is noteworthy that only few nonclinical studies (Joshanloo et al, ; Keyes et al, ; Lamers et al, ; Lupano Perugini et al, ; Petrillo et al, ; Rogoza et al, ) have specifically reported MHC‐SF subscale means, thereby permitting assessment of their relative sizes with respect to cultural, societal, political and economic context. Interestingly, SWB consistently had the lowest means of the three subscales in all nonclinical samples, including the present one.…”

“…The recently developed bifactor ESEM model addresses both limitations inherent in traditional CFA as it combines both a general factor and specific factors orthogonal to each other, and it also allows items to load on target factors but also cross‐load on nontarget factors (Morin et al, ). Bifactor ESEM has most recently been studied with MHC‐SF (Lamborn et al, ; Longo et al, ; Rogoza et al, ; Schutte & Wissing, ) and it was found to produce a better fit than both bifactor CFA and three‐factor ESEM. In the current study, bifactor ESEM attained the best fit than all models tested, replicating previous findings.…”

“…">We examined the construct validity, assessed with CFA and ESEM techniques. Based on previous research, the following six models were evaluated in all three samples: (a) a single factor CFA model with all 14 items; (b) a two correlated factors CFA model with EWB in one factor and both SWB and PWB in a second factor; (c) a three correlated factors CFA model with EWB, SWB, and PWB, as proposed by Keyes (), (); (d) a bifactor CFA model (Schmid & Leiman, ), where each item loads on a specific factor (i.e., EWB, SWB, or PWB) and simultaneously on a general factor (GF), all four factors being orthogonal to each other (de Bruin & du Plessis, ; Hides et al, ; Jovanovic, ); (e) a three‐factor ESEM model (Asparouhov & Muthén, ; Marsh et al, ), where all 14 items are loading on all three correlated factors (i.e., EWB, SWB, and PWB) simultaneously (Joshanloo, ; Joshanloo & Jovanovic, ; Joshanloo & Lamers, ; Joshanloo, Jose, et al, ); (f) a bifactor ESEM model (Morin et al, ), where all 14 items are simultaneously loading on EWB, SWB, PWB, and GF, all four factors being orthogonal to each other (Lamborn, Cramer, & Riberdy, ; Longo, Jovanovic, Sampaio de Carvalho, & Karas, ; Rogoza et al, ; Schutte & Wissing, ). We did not test a higher‐order model (Dore et al, ; Lupano Perugini et al, ; Petrillo et al, ) because for a three first‐order factorial structure, like MHC‐SF, the second‐order is just identified.…”

“…Despite the well‐documented association between psychopathology, especially affective disorders, and well‐being, as assessed with the MHC‐SF, its psychometric properties have been primarily evaluated in general population samples or national epidemiological surveys. In particular, the MHC‐SF has been translated and validated in various countries and cultural contexts (Zemojtel‐Piotrowska et al, ), including Argentina (Lupano Perugini, de la Iglesia, Castro Solano, & Keyes, ), Australia (Hides et al, ), Brazil (Machado & Bandeira, ), Canada (Dore, O'Loughlin, Sabiston, & Fournier, ; Orpana, Vachon, Dykxhoorn, & Jayaraman, ), Chile (Echeverria et al, ), China (Guo et al, ), France (Ismaïl & Salama‐Younes, ), Iran (Joshanloo, ; Joshanloo, Wissing, Khumalo, & Lamers, ), Italy (Joshanloo, Capone, Petrillo, & Caso, ; Petrillo, Capone, Caso, & Keyes, ), the Netherlands (Joshanloo et al, ; Lamers, Westerhof, Bohlmeijer, ten Klooster, & Keyes, ), New Zealand (Joshanloo, Jose, & Kielpikowski, ), Poland (Karas, Cieciuch, & Keyes, ), Portugal (de Carvalho, Pereira, Pinto, & Marôco, ), Serbia (Joshanloo & Jovanovic, ; Jovanovic, ), South Africa (Joshanloo et al, ; Keyes et al, ), South Korea (Lim, ), United States (Joshanloo, ) and Vietnam (Rogoza, Truong Thi, Rozycka‐Tran, Piotrowski, & Zemojtel‐Piotrowska, ), providing an ample body of evidence in support of Keyes's model of well‐being. Although all of these studies were performed in nonclinical samples, they often included several mental illness measures to investigate the discriminant validity of the MHC‐SF.…”