An increasing number of psychological studies are devoted to the analysis of g-factor structures. One key purpose of applying g-factor models is to identify predictors or potential causes of the general and specific effects. Typically, researchers relate predictor variables directly to the general and specific factors using a classical mimic approach. However, this procedure bears some methodological challenges, which often lead to model misspecification and biased parameter estimates. We propose 2 possible modeling strategies to circumvent these problems: the multiconstruct bifactor and the residual approach. We illustrate both modeling approaches for the application of g-factor models to longitudinal and multitrait-multimethod data. Practical guidelines are provided for choosing an appropriate method in empirical applications, and the implications of this investigation for multimethod and longitudinal research are discussed. (PsycINFO Database Record

Symmetrical bifactor models are frequently applied to diverse symptoms of psychopathology to identify a general P factor. This factor is assumed to mark shared liability across psychopathology dimensions and mental disorders. Despite their popularity, however, symmetrical bifactor models often yield anomalous results, including but not limited to non-significant or negative specific factor variances and non-significant or negative factor loadings. To date, these anomalies have often been treated as nuisances to be explained away. In this paper, we demonstrate why these anomalies alter the substantive meaning of P such that it (1) does not reflect general liability to psychopathology and (2) differs in meaning across studies. We then describe an alternative modeling framework, the bifactor-(S − 1) approach. This approach avoids anomalous results, provides a framework for explaining unexpected findings in published symmetrical bifactor studies, and yields a general factor with well-defined meaning across studies. We present an empirical example to illustrate these points and provide concrete recommendations to help researchers decide for or against a specific variant of bifactor structures. In summary, bifactor-(S − 1) models provide an approach to answer questions posed in symmetrical bifactor models in a more comparable and replicable manner.

No abstract

Symmetrical bifactor models are frequently applied to diverse symptoms of psychopathology to identify a general P factor. This factor is assumed to mark shared liability across all psychopathology dimensions and mental disorders. Despite their popularity, however, symmetrical bifactor models of P often yield anomalous results, including but not limited to nonsignificant or negative specific factor variances and nonsignificant or negative factor loadings. To date, these anomalies have often been treated as nuisances to be explained away. In this article, we demonstrate why these anomalies alter the substantive meaning of P such that it (a) does not reflect general liability to psychopathology and (b) differs in meaning across studies. We then describe an alternative modeling framework, the bifactor-( S−1) approach. This method avoids anomalous results, provides a framework for explaining unexpected findings in published symmetrical bifactor studies, and yields a well-defined general factor that can be compared across studies when researchers hypothesize what construct they consider “transdiagnostically meaningful” and measure it directly. We present an empirical example to illustrate these points and provide concrete recommendations to help researchers decide for or against specific variants of bifactor structure.

BackgroundStandardized and individualized Internet‐based interventions (IBI) for depression yield significant symptom improvements. However, change patterns during standardized or individualized IBI are unknown. Identifying subgroups that experience different symptom courses during IBI and their characteristics is vital for improving response.MethodsMildly to moderately depressed individuals according to self‐report (N = 1,089) were randomized to receive module‐wise feedback that was either standardized or individualized by a counselor within an otherwise identical cognitive‐behavioral IBI for depression (seven modules over six weeks). Depressive symptoms were assessed at baseline and before each module (Patient Health Questionnaire; PHQ‐9). Other individual characteristics (self‐report) and the presence of an affective disorder (structured clinical interview) were assessed at baseline. Growth mixture modeling was used to identify and compare subgroups with discernable change patterns and associated client variables across conditions.ResultsModel comparisons suggest equal change patterns in both conditions. Across conditions, a group of immediate (62.5%) and a group of delayed improvers (37.5%) were identified. Immediate improvers decreased their PHQ‐9 score by 5.5 points from pre to post, with 33% of improvement occurring before treatment commenced. Delayed improvers were characterized by stable symptom severity during the first two modules and smaller overall symptom decrease (3.4 points). Higher treatment expectations, a current major depressive disorder (interview), and lower social support were associated with delayed improvement.ConclusionInternet‐based interventions for depression with individualized and with standardized feedback lead to comparable patterns of change. Expectation management and bolstering of social support are promising strategies for individuals that are at risk for delayed improvement.