Generalized multilevel structural equation modeling

By Sophia Rabe-Hesketh, A. Skrondal, Andrew Pickles. 2004

861 citations recorded by the source corpus.

A unifying framework for generalized multilevel structural equation modeling is introduced. The models in the framework, called generalized linear latent and mixed models (GLLAMM), combine features of generalized linear mixed models (GLMM) and structural equation models (SEM) and consist of a response model and a structural model for the latent variables. The response model generalizes GLMMs to incorporate factor structures in addition to random intercepts and coefficients. As in GLMMs, the data can have an arbitrary number of levels and can be highly unbalanced with different numbers of lower-level units in the higher-level units and missing data. A wide range of response processes can be modeled including ordered and unordered categorical responses, counts, and responses of mixed types. The structural model is similar to the structural part of a SEM except that it may include latent and observed variables varying at different levels. For example, unit-level latent variables (factors or random coefficients) can be regressed on cluster-level latent variables. Special cases of this framework are explored and data from the British Social Attitudes Survey are used for illustration. Maximum likelihood estimation and empirical Bayes latent score prediction within the GLLAMM framework can be performed using adaptive quadrature in gllamm, a freely available program running in Stata.

Subfields: Experimental and Cognitive Psychology, Health, Management, Monitoring, Policy and Law

View Paper: https://www.semanticscholar.org/paper/6277b4c89a95acf69bf3311c0f5107a4c5b802f0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Experimental and Cognitive Psychology

In Health

In Management, Monitoring, Policy and Law

Citations in the corpus

Multilevel modelling of complex survey data (2006, methodology): “Robins et al. (1995)).”
Maximum likelihood estimation of limited and discrete dependent variable models with nested random effects (2005, methodology): “The adaptive quadrature method can also be used for the more general class of multilevel factor and structural equation models (Rabe-Hesketh et al., 2004) since they have the same conditional independence structure as random coefficient models: variables (at level 1) are conditionally independent…”
Latent Variable Modelling: A Survey * (2007, methodology): “Recognizing the mathematical similarity of a wide range of latent variable models, Muthén (2002), Rabe-Hesketh et al. (2004), Skrondal & Rabe-Hesketh (2004), among others, have developed very general frameworks for latent variable modelling which accommodate and extend the models mentioned in…”
Discovering Unobserved Heterogeneity in Structural Equation Models to Avert Validity Threats (2013, background): “…unobserved heterogeneity in the factor means and covariance structure (i.e., structural parameters, measurement error variance, and factor covariance), thereby generalizing multilevel SEM models (Muthén 1994; Rabe-Hesketh et al. 2004) that only account for heterogeneity in the mean structure.”
Maximum Likelihood Estimation of Endogenous Switching and Sample Selection Models for Binary, Ordinal, and Count Variables (2006, background): “(see, for example, Rabe-Hesketh et al. 2004a; Skrondal and Rabe-Hesketh”
A Bayesian Approach to Multilevel Structural Equation Modeling With Continuous and Dichotomous Outcomes (2015, methodology): “Similar models have been presented by Lüdtke et al. (2011), Marsh et al. (2009), and Rabe-Hesketh et al. (2004).”
Maximum Likelihood Estimation of Generalized Linear Models with Covariate Measurement Error (2003, methodology): “When the outcome has been measured several times, it can also be useful to construct models for a latent true outcome, see Rabe-Hesketh, Skrondal, and Pickles (2004), Skrondal and Rabe-Hesketh (2004), and Rabe-Hesketh, Pickles, and Skrondal (2004).”
The psychometric and empirical properties of measures of risk preferences (2017, positive): “Also for that reason, we do not include a constant in the vector of covariates x. 13We used the Stata program GLLAMM (Rabe-Hesketh et al. 2004) to estimate all models by maximum likelihood, using spherical quadrature with nine integration points.”
Assessing the impact of learning environments: How to use student ratings of classroom or school characteristics in multilevel modeling (2009, methodology): “, 2006) When multilevel exploratory factor analysis (see Reise, Ventura, Nuechterlein, & Kim, 2005) and multilevel confirmatory factor analysis (Mehta & Neale, 2005; Rabe-Hesketh et al., 2004) are used, it is important to determine the extent of”
Do I Follow My Friends or the Crowd? Information Cascades in Online Movie Ratings (2015, positive): “Since causal processes operate at the individual level and not the aggregate level of user ratings, it follows that investigation of causality requires individual-specific effects (Rabe-Hesketh et al. 2004a).”
Reliable Estimation of Generalized Linear Mixed Models using Adaptive Quadrature (2002, methodology): “…discussed adaptive quadrature in the context of estimating generalized linear mixed models, gllamm allows this method to be used for estimat-S ing general multilevel latent variable models, including multilevel factor models and multilevel structural equation models; see Rabe-Hesketh et al. (2002).”