Dynamic Structural Equation Models

Asparouhov, Tihomir; Hamaker, Ellen L.; Muthén, Bengt

doi:10.1080/10705511.2017.1406803

Cited by 573 publications

(715 citation statements)

References 24 publications

Supporting

Mentioning

709

Contrasting

Unclassified

Order By: Relevance

“…Thus, it does not require large sample sizes to properly estimate standard errors. Moreover, Bayesian estimation can be less computationally demanding than maximum‐likelihood estimation while providing very similar results to maximum‐likelihood estimation (Asparouhov, Hamaker, & Muthén, ). We specified 50 000 iterations and a thinning of 10 iterations, and two chains with different starting seeds.…”

Section: Methodsmentioning

confidence: 96%

Analysis of dynamic, bidirectional associations in older adult physical activity and sleep quality

Best

Falck

Landry

et al. 2018

Journal of Sleep Research

View full text Add to dashboard Cite

Sleep quality and physical activity (PA) appear to be interrelated; thus, by promoting one behaviour, it may be possible to improve the other in older adults. Examination of the within-person day-to-day variation in PA and sleep quality could potentially elucidate the directionality of the association of these behaviours. We measured sleep quality (i.e. fragmentation, efficiency, duration and latency) and moderate-to-vigorous PA using the MotionWatch8© over 14 consecutive days and nights in community-dwelling adults (n = 152; age range 53-101 years). Multilevel modelling estimated within-subject autoregressive and cross-lagged effects and between-subject associations between PA and sleep quality. On days when individuals engaged in a high amount of PA on one day (relative to their averages), they were more likely to engage in a high amount of PA on the next day (estimate, 0.19; 95% CI, 0.14, 0.24). Nights in which individuals had a long sleep latency were followed by nights in which they also had a long sleep latency (estimate, 0.09; 95% CI, 0.03, 0.14). In contrast, nights in which individuals slept for a long period of time were followed by nights in which they slept relatively less than their averages (estimate, -0.09; 95% CI, -0.13, -0.04). When individuals engaged in a large amount of PA during the day, they tended to sleep longer that following night (estimate, 0.01; 95% CI, 0.001, 0.02). All other associations between PA and sleep quality were not significant. Increasing PA therefore might increase sleep duration in older adults.

show abstract

Section: Methodsmentioning

confidence: 96%

Analysis of dynamic, bidirectional associations in older adult physical activity and sleep quality

Best

Falck

Landry

et al. 2018

Journal of Sleep Research

View full text Add to dashboard Cite

show abstract

“…The indices of centrality were calculated using the centrality Plot and centrality Table functions in q graph (Epskamp et al, ). Missing data were estimated using the Kalman filter as part of the estimation procedure (Haan‐Rietdijk et al, ; Asparouhov, Hamaker, and Muther, ). Furthermore, over all we had a 73% compliance rate for EMA data, which is as high or higher than most EMA studies in eating disorders.…”

Section: Methodsmentioning

confidence: 99%

“…In words, outStrength represents that a symptom provides input to many other symptoms. For individual networks, centrality was calculated only for contemporaneous networks (Asparouhov, Hamaker and Muthén, 2018). The indices of centrality were calculated using the centrality Plot and centrality (Epskamp et al, 2012).…”

Section: Data Analytic Procedure: Network Analysismentioning

confidence: 99%

Longitudinal and personalized networks of eating disorder cognitions and behaviors: Targets for precision intervention a proof of concept study

Levinson

Vanzhula

Brosof

2018

Intl J Eating Disorders

View full text Add to dashboard Cite

Introduction Despite the high mortality and significant societal and personal costs associated with eating disorders (EDs) there are few evidence‐based treatments. Part of the difficulty developing and implementing evidence‐based treatments in EDs is due to the extremely high heterogeneity (e.g., variability in treatment outcome, symptom presentation etc) present. Methods To begin to identify specific symptom heterogeneity within persons, the current study used novel within and between group and intra‐individual network analyses to create longitudinal and within‐person networks of ED cognitions and behaviors (N = 66 individuals diagnosed with an ED). This article provides a proof of concept study for how to use between and within‐person network analyses both for the EDs and other forms of psychopathology. Results We found that cognitions focused on desiring thinness played a likely maintaining role in ED pathology, across network type and across time. Furthermore, we showed that three individuals with the same diagnosis (anorexia nervosa) differed in which symptoms maintained the disorder. We use these participants to exemplify how to use intra‐individual network analysis to personalize treatment focused on the primary maintaining symptoms. Finally, we found that amount of time (e.g., 4 hr vs. simultaneously) impacts how symptoms maintain each other. Conclusions These findings have implications for the development of novel personalized evidence‐based treatments for EDs, as well as implications for how the field understands how psychopathology maintains itself. These data represent a first‐step towards using intra‐individual network analyses in the ED field, as well as for hypotheses generation in future research.

show abstract

“…This technique of obtaining an SVAR model leads to multiple solutions (Beltz & Molenaar, 2016). Step-wise model selection can also be used to estimate sparse SVAR, for example by using model selection in (unified) SEM (Gates et al, 2010) or Bayesian dynamical SEM models (Muthén & Muthén, 2017;Asparouhov et al, 2016).…”

Section: Temporally Ordered Data Of a Single Subjectmentioning

confidence: 99%

The Gaussian Graphical Model in Cross-Sectional and Time-Series Data

Epskamp

Waldorp

Mõttus

et al. 2018

Multivariate Behavioral Research

659

876

View full text Add to dashboard Cite

We discuss the Gaussian graphical model (GGM; an undirected network of partial correlation coefficients) and detail its utility as an exploratory data analysis tool. The GGM shows which variables predict one-another, allows for sparse modeling of covariance structures, and may highlight potential causal relationships between observed variables. We describe the utility in three kinds of psychological data sets: data sets in which consecutive cases are assumed independent (e.g., cross-sectional data), temporally ordered data sets (e.g., n = 1 time series), and a mixture of the 2 (e.g., n > 1 time series). In time-series analysis, the GGM can be used to model the residual structure of a vector-autoregression analysis (VAR), also termed graphical VAR. Two network models can then be obtained: a temporal network and a contemporaneous network. When analyzing data from multiple subjects, a GGM can also be formed on the covariance structure of stationary means-the between-subjects network. We discuss the interpretation of these models and propose estimation methods to obtain these networks, which we implement in the R packages graphicalVAR and mlVAR. The methods are showcased in two empirical examples, and simulation studies on these methods are included in the supplementary materials.

show abstract

Dynamic Structural Equation Models

Cited by 573 publications

References 24 publications

Analysis of dynamic, bidirectional associations in older adult physical activity and sleep quality

Analysis of dynamic, bidirectional associations in older adult physical activity and sleep quality

Longitudinal and personalized networks of eating disorder cognitions and behaviors: Targets for precision intervention a proof of concept study

The Gaussian Graphical Model in Cross-Sectional and Time-Series Data

Contact Info

Product

Resources

About