A Latent Transition Mixture Model Using the Three-Step Specification

Abstract

The 3-step method for estimating the effects of auxiliary variables (i.e., covariates and distal outcome) in mixture modeling provides a useful way to specify complex mixture models. One of the benefits of this method is that the measurement parameters of the mixture model are not influenced by the auxiliary variable(s). In addition, it allows for models that involve multiple latent class variables to be specified without each part of the model influencing the others. This article describes a unique latent transition analysis model where the measurement models are a latent class analysis model and a growth mixture model. We highlight the application of this model to study kindergarten readiness profiles and link it to elementary students' reading trajectories. Mplus syntax for the 3-step specification is provided.

A Comprehensive Examination of Reading Heterogeneity in Students with High Functioning Autism: Distinct Reading Profiles and Their Relation to Autism Symptom Severity

“…This was accomplished by estimating profilespecific means. This process has been shown to result in a shift in the latent profiles, thereby altering the substantive interpretation of them (Asparouhov and Muthén 2014a;Nylund-Gibson et al 2014). Therefore, we implemented the BCH approach (Asparouhov and Muthén 2014b;Bakk et al 2013;Bolck et al 2004;Vermunt 2010) in order to account for classification error and avoid profile shifts.…”

“…The final profile with a solid line with circular markers accounted for about 32% of the sample and was delineated by scores in the average range on all language and reading variables; this subgroup is called Average Readers. Did not converge We also examined potential differences in age and gender among the emergent profiles using the three-step method (Asparouhov and Muthén 2014a;Nylund-Gibson et al 2014). There were no effects of either age or gender.…”

A Review of Graphical Approaches to Common Statistical Analyses. The Omnipresence of Latent Variables in Statistics

“…), have pointed to the ‘doubly latent’ nature of the classification categorical latent variable C that represents the classes (class 1, 2, etc.). For example, in GMM models, the ‘measurement model’ component is meant to extract/uncover the classes using some indicators of the class latent variable (similar to indicators of a common factor in factor analysis [51]), while the ‘predicting the class’ part of the model allows for regressing this categorical C classification unto chosen predictors of class membership, yet a multi-class part of the model allows for ‘effects’ of the class variable unto distal outcomes [52]. Class variables however are estimated imperfectly, i.e.…”

Trajectories of Psychological Adaptation to Marital Breakup after a Long-Term Marriage

“…The model was estimated following the recommendations of Nylund-Gibson et al [34] and Asparouhov and Muthén [35]. The profile probabilities for the most likely profile membership extracted from the LCA, run separately for each wave, were used to calculate the classification uncertainty rate, which is the average probability that members of each profile could also be classified into the other profiles.…”