Analyzing latent state-trait and multiple-indicator latent growth curve models as multilevel structural equation models

Abstract

Latent state-trait (LST) and latent growth curve (LGC) models are frequently used in the analysis of longitudinal data. Although it is well-known that standard single-indicator LGC models can be analyzed within either the structural equation modeling (SEM) or multilevel (ML; hierarchical linear modeling) frameworks, few researchers realize that LST and multivariate LGC models, which use multiple indicators at each time point, can also be specified as ML models. In the present paper, we demonstrate that using the ML-SEM rather than the SL-SEM framework to estimate the parameters of these models can be practical when the study involves (1) a large number of time points, (2) individually-varying times of observation, (3) unequally spaced time intervals, and/or (4) incomplete data. Despite the practical advantages of the ML-SEM approach under these circumstances, there are also some limitations that researchers should consider. We present an application to an ecological momentary assessment study (N = 158 youths with an average of 23.49 observations of positive mood per person) using the software Mplus (Muthén and Muthén, 1998–2012) and discuss advantages and disadvantages of using the ML-SEM approach to estimate the parameters of LST and multiple-indicator LGC models.

Consistency of health-related quality of life among people living with HIV: Latent state-trait analysis

“…Thus, the domains seem to be differently prone to occasion-specific influences. Also, we observed that T1 and T3 values across domains are affected by situation or person and situation interactions (as they are indistinguishable in the LTS models [ 58 ]), whereas T2 scores are more strongly linked to stable disposition (52–81%), as well as method (12–43%). Taken together, this may suggest problematic homogeneity of domain indicators.…”

“…As the strong MI allows for mean comparison, for this purpose, the mean of the first latent state factor was fixed to be zero, whereas the means of the two remaining latent state factors were freed [ 58 ]. Consequently, the intercepts of those factors can be then interpreted as a difference relative to the first factor.…”

“…To verify the research hypotheses, a latent state-trait (LST) analysis was performed [ 57 , 58 ]. The LST models have been used increasingly in analyses of longitudinal data to capture the within-time consistency vs. situation variability of individual differences of a particular variable over time.…”

“…However, as our participants were not newly diagnosed patients, but had been infected and under treatment for some time already, we did not expect any significant systematic changes within HRQoL. Therefore, LST models were used instead of LGC models [ 58 ].…”

The micro-social risk environment for injection drug use: An event specific analysis of dyadic, situational, and network predictors of injection risk behavior

“…2 Unfortunately, there is currently no method for estimating a multilevel random slope model with categorical indicators that includes global measures of fit such as Chi-square, RMSEA, CFI, or TLI (Geiser, Bishop, Lockhart, Shiffman, & Grenard, 2013). Although global fit indices are not available when using MLR with random slopes, an iterative approach is possible to examine if nested models significantly improve fit using the Wald test of nested models.…”