Misspecification in Latent Change Score Models: Consequences for Parameter Estimation, Model Evaluation, and Predicting Change

Latent change scores models for applied research: A practical guide using Mplus

The present article provides a practical guide for modelling and interpreting several basic applications of the latent change scores (LCS) model, a useful and flexible approach for the analysis of change. The article is addressed to students, researchers and practitioners who are familiar with structural equation modelling but new to LCS. We first provided a gentle introduction to the LCS model using non-technical language and minimal mathematical formalism. We illustrated the basic ideas behind this approach, introducing LCS in its simplest form. We show how this model can be straightforwardly extended to more complex applications, including the dual change score (DCS) model and some of its variants (i.e., the proportional change and the constant change models). We illustrated how the univariate LCS model can be used to determine the growth trajectory of a variable across multiple waves of assessment. Next, we focused on the bivariate case, which allows for the modelling of the dynamic relations between two variables. For each model, we provided easy-to-follow examples of applications based on Schwartz's theory of basic personal values. The examples are accompanied by annotated syntax and output showing how they can be implemented with the Mplus software and how results can be interpreted.

Decline in positive future orientations among adolescents during covid-19: The role of socioeconomic status, parental support, and sense of control

INTRODUCTION

Before coronavirus disease (covid-19), adolescents from a lower socioeconomic status (SES) background tend to have less positive future orientations, receive less parental support, and have a weaker sense of control than adolescents from a higher SES background. The covid-19 pandemic has potentially increased the socioeconomic gaps in positive future orientations, parental support, and sense of control among adolescents who are currently in vocational education. As societies are aiming to return back to precovid norms, certain groups of adolescents might require more attention for ensuring a stable future than others.

METHODS

Two-wave questionnaire data of 689 Dutch adolescents (Mage  = 17.8; 56% female) from the Youth Got Talent project was analyzed. Latent Change Score models are a relatively novel approach that allows two-wave data to estimate associations between precovid predictor variables and changes in outcome variables from before to during covid-19 (e.g., SES, positive future orientations, parental support, and sense of control). Analyses were preregistered.

RESULTS

The precovid socioeconomic differences in adolescent's positive future orientations and sense of control remained stable during covid-19, whereas the socioeconomic difference in parental support decreased during covid-19. A decline in parental support, an increase in sense of control, and more covid-19 hardships were associated with an increase in future orientations.

CONCLUSION

The covid-19 situation has not substantially increased socioeconomic differences in positive future orientations and sense of control, but did decrease socioeconomic differences in parental support among adolescents. Short-term policies should aim to facilitate parental support and positive future orientations to all adolescents who experienced a decline, while also long-term focusing on the more consistent socioeconomic difference in sense of control among adolescents.

Dynamic modelling of chronotype and hypo/manic and depressive symptoms in young people with emerging mental disorders

There is significant interest in the possible influence of chronotype on clinical states in young people with emerging mental disorders. We apply a dynamic approach (bivariate latent change score modelling) to examine the possible prospective influence of chronotype on depressive and hypo/manic symptoms in a youth cohort with predominantly depressive, bipolar, and psychotic disorders (N = 118; 14-30-years), who completed a baseline and follow-up assessment of these constructs (mean interval = 1.8-years). Our primary hypotheses were that greater baseline eveningness would predict increases in depressive but not hypo/manic symptoms. We found moderate to strong autoregressive effects for chronotype (β = -0.447 to -0.448, p < 0.001), depressive (β = -0.650, p < 0.001) and hypo/manic symptoms (β = -0.819, p < 0.001). Against our predictions, baseline chronotypes did not predict change in depressive (β = -0.016, p = 0.810) or hypo/manic symptoms (β = 0.077, p = 0.104). Similarly, the change in chronotype did not correlate with the change in depressive symptoms (β = -0.096, p = 0.295) nor did the change in chronotype and the change in hypo/manic symptoms (β = -0.166, p = 0.070). These data suggest that chronotypes may have low utility for predicting future hypo/manic and depressive symptoms in the short term, or that more frequent assessments over longer periods are needed to observe these associations. Future studies should test whether other circadian phenotypes (e.g. sleep-wake variability) are better indicators of illness course.

lcsm: An R package and tutorial on latent change score modelling

Latent change score models (LCSMs) are used across disciplines in behavioural sciences to study how constructs change over time. LCSMs can be used to estimate the trajectory of one construct (univariate) and allow the investigation of how changes between two constructs (bivariate) are associated with each other over time. This paper introduces the R package lcsm, a tool that aims to help users understand, analyse, and visualise different latent change score models. The lcsm package provides functions to generate model syntax for basic univariate and bivariate latent change score models with different model specifications. It is also possible to visualise different model specifications in simplified path diagrams. An interactive application illustrates the main functions of the package and demonstrates how the model syntax and path diagrams change based on different model specifications. This R package aims to increase the transparency of reporting analyses and to provide an additional resource to learn latent change score modelling.

Study Length, Change Process Separability, Parameter Estimation, and Model Evaluation in Hybrid Autoregressive-Latent Growth Structural Equation Models for Longitudinal Data

Hybrid autoregressive-latent growth structural equation models for longitudinal data represent a synthesis of the autoregressive and latent growth modeling frameworks. Although these models are conceptually powerful, in practice they may struggle to separate autoregressive and growth related processes during estimation. This confounding of change processes may, in turn, increase the risk of the models producing deceptively compelling results (i.e., models that fit excellently by conventional standards despite highly biased parameter estimates). Including additional time points provides models with more raw information about change, which could help improve process separability and the accuracy of parameter estimates to a degree. This study thus used Monte Carlo simulation methods to examine associations between change process separability, the number of time points in a model, and the consequences of misspecification, across three prominent hybrid autoregressive-latent growth models: the Latent Change Score model (LCS; McArdle, 2001), the Autoregressive Latent Trajectory Model (ALT; Bollen & Curran, 2006), and the Latent Growth Model with Structured Residuals (LGM-SR; Curran et al., 2014). Results showed that including more time points increased process separability and robustness to misspecification in the LCS and ALT, but typically not at a rate that would be practically feasible for most developmental researchers. Alternatively, regardless of how many time points were in the model process separability was high in the LGM-SR, as was robustness to misspecification. Overall, results suggest that the LGM-SR is the most effective of the three hybrid autoregressive-latent growth models considered here.

Dynamical Properties and Conceptual Interpretation of Latent Change Score Models

Latent Change Score models (LCS) are a popular tool for the study of dynamics in longitudinal research. They represent processes in which the short-term dynamics have direct and indirect consequences on the long-term behavior of the system. However, this dual interpretation of the model parameters is usually overlooked in the literature, and researchers often find it difficult to see the connection between parameters and specific patterns of change. The goal of this paper is to provide a comprehensive examination of the meaning and interpretation of the parameters in LCS models. Importantly, we focus on their relation to the shape of the trajectories and explain how different specifications of the LCS model involve particular assumptions about the mechanisms of change. On a supplementary website, we present an interactive Shiny App that allows users to explore different sets of parameter values and examine their effects on the predicted trajectories. We also include fully explained code to estimate some of the most relevant specifications of the LCS model with the R-packages lavaan and OpenMx.

Longitudinal associations between borderline personality disorder and five-factor model traits over 24 years

Are five-factor traits and borderline personality symptoms the same features with different names? The existing literature offers reasons to think they are the same and reasons to think they are different. We examined longitudinal associations between these variables in a sample of patients assessed 12 times over 24 years using latent curve models with structured residuals. Mean trajectories for all variables were in the direction of symptom reduction/personality maturation and could be parsed into an initial, rapid improvement phase and a subsequent, gradual improvement phase. We found robust between-person associations among intercepts and long-term slopes of traits and symptoms. Specifically, higher levels of neuroticism as well as lower levels of extraversion, agreeableness, and conscientiousness were associated with higher levels of borderline personality symptoms, and changes in these traits were correlated with reduction in symptoms over time. Associations among time-structured residuals allowed for examinations of within-person deflections from these general trends at briefer (two year) intervals. All variables exhibited robust within-person carry-over effects. Other within-person effects were more specific to certain traits. These results suggest that, despite their distinct theoretical and methodological bases, normal trait and psychiatric diagnostic approaches largely converged on a similar conception of borderline personality.

Latent Change Score Models for the Study of Development and Dynamics in Organizational Research

The empirical study of change has proven to be one of the most vexing challenges in organizational science. Fortunately, contemporary methodologies originating from developmental psychology may provide a potential solution and are consequently working their way into the literature. In particular, organizational researchers are increasingly employing variations of latent change score (LCS) models to address questions regarding change, development, and dynamics. Although these models may indeed be used to reliably study change, development, and dynamics, many studies utilizing these models—and published in premier outlets—are characterized by questionable methodological choices, improper modeling procedures, and suboptimal research designs. Thus, the purpose of the present article is to (a) provide a critical review of LCS models, (b) outline appropriate modeling procedures (with corresponding Mplus and R syntax), (c) compare and contrast LCS modeling with other analytical techniques, and (d) delineate best practices. Ultimately, we endorse the use of LCS models by organizational researchers interested in studying longitudinal phenomena. However, we also heed researchers to do so judiciously because their misuse may lead to their unwarranted rejection by the field.

Testing prospective effects in longitudinal research: Comparing seven competing cross-lagged models

In virtually all areas of psychology, the question of whether a particular construct has a prospective effect on another is of fundamental importance. For decades, the cross-lagged panel model (CLPM) has been the model of choice for addressing this question. However, CLPMs have recently been critiqued, and numerous alternative models have been proposed. Using the association between low self-esteem and depression as a case study, we examined the behavior of seven competing longitudinal models in 10 samples, each with at least four waves of data and sample sizes ranging from 326 to 8,259. The models were compared in terms of convergence, fit statistics, and consistency of parameter estimates. The traditional CLPM and the random intercepts cross-lagged panel model (RI-CLPM) converged in every sample, whereas the other models frequently failed to converge or did not converge properly. The RI-CLPM exhibited better model fit than the CLPM, whereas the CLPM produced more consistent cross-lagged effects (both across and within samples) than the RI-CLPM. We discuss the models from a conceptual perspective, emphasizing that the models test conceptually distinct psychological and developmental processes, and we address the implications of the empirical findings with regard to model selection. Moreover, we provide practical recommendations for researchers interested in testing prospective associations between constructs and suggest using the CLPM when focused on between-person effects and the RI-CLPM when focused on within-person effects. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Personality trait change at work: Associations with organizational socialization and identification

OBJECTIVE

This study investigates associations between Big Five personality trait change, organizational socialization, and organizational identification during a 3-year police officer training program (N = 416 police officer cadets).

METHOD

Participants completed a questionnaire measuring the Big Five personality traits when they entered the training academy, and then, completed the same personality questionnaire, along with measures of organizational socialization and identification, during their 2nd (n = 360) and 3rd (n = 397) year of training.

RESULTS

Results corroborated the hypotheses that (a) the Big Five traits can show systematic changes even across a relatively short time period and (b) this change is functional, given that the latent difference scores of all Big Five traits significantly predicted increases in organizational socialization and identification.

CONCLUSION

The Big five personality traits showed significant mean level changes across the 3-year training program. Although these changes were not fully consistent with theoretical expectations, they did predict two aspects of organizational adjustment (socialization and identification). The theoretical and practical implications of these findings were discussed.

Mapping differential responses to cognitive training using machine learning

We used two simple unsupervised machine learning techniques to identify differential trajectories of change in children who undergo intensive working memory (WM) training. We used self‐organizing maps (SOMs)—a type of simple artificial neural network—to represent multivariate cognitive training data, and then tested whether the way tasks are represented changed as a result of training. The patterns of change we observed in the SOM weight matrices implied that the processes drawn upon to perform WM tasks changed following training. This was then combined with K‐means clustering to identify distinct groups of children who respond to the training in different ways. Firstly, the K‐means clustering was applied to an independent large sample (N = 616, M_age = 9.16 years, range = 5.16–17.91 years) to identify subgroups. We then allocated children who had been through cognitive training (N = 179, M_age = 9.00 years, range = 7.08–11.50 years) to these same four subgroups, both before and after their training. In doing so, we were able to map their improvement trajectories. Scores on a separate measure of fluid intelligence were predictive of a child's improvement trajectory. This paper provides an alternative approach to analysing cognitive training data that go beyond considering changes in individual tasks. This proof‐of‐principle demonstrates a potentially powerful way of distinguishing task‐specific from domain‐general changes following training and of establishing different profiles of response to training.

Modeling Latent Change Score Analysis and Extensions in Mplus: A Practical Guide for Researchers

Many developmental and life course researchers are interested in modeling dynamic developmental processes. Latent change score (LCS) modeling is a potentially powerful modeling technique that can be used to assess complex life course processes, as well as the direction of longitudinal bivariate associations. Advances in modeling software, like Mplus, as well as widening adoption of software by researchers has made LCS modeling simpler. Thus, in the present paper, we provide 1) a theoretical overview of LCS analysis, 2) information on the interpretation of these models, 3) a practical guid7e for estimating these models in Mplus (including example syntax), 4) illustrative examples of LCS analysis, and 5) potential caveats for researchers.