Multilevel Modeling: A Review of Methodological Issues and Applications

Abstract

This study analyzed the reporting of multilevel modeling applications of a sample of 99 articles from 13 peer-reviewed journals in education and the social sciences. A checklist, derived from the methodological literature on multilevel modeling and focusing on the issues of model development and specification, data considerations, estimation, and inference, was used to analyze the articles. The most common applications were two-level models where individuals were nested within contexts. Most studies were nonexperimental and used nonprobability samples. The amount of data at each level varied widely across studies, as did the number of models examined. Analyses of reporting practices indicated some clear problems, with many articles not reporting enough information for a reader to critique the reported analyses. For example, in many articles, one could not determine how many models were estimated, what covariance structure was assumed, what type of centering if any was used, whether the data were consistent with assumptions, whether outliers were present, or how the models were estimated. Guidelines for researchers reporting multilevel analyses are provided.

Explaining Student Achievement: the Influence of Teachers’ Pedagogical Content Knowledge in Statistics

“…In line with recommendations from Dedrick et al (2009) underlying assumptions of the model, including the absence of an autoregressive structure, were assessed and found to be satisfactory. As is seen in Table 3, the null model predicted that students' participation in the study for one year was associated with an increase in SLK of 0.14 logits.…”

What drives retirement income worries in Europe? A multilevel analysis

“…One advantage of multilevel analysis is that it allows one to take into account the dependency of observations between respondents from the same country. The practical benefit of multilevel modeling is that mean scores and standard errors of country-level variables can be estimated in an unbiased fashion (Dedrick et al 2009). Furthermore, multilevel modeling allowed us to estimate the extent to which dependent measures vary across countries, and the degree to which variance on each criterion can be explained by individual-level (i.e., micro) and countrylevel (macro) effects.…”

Impact du soutien affectif et du contrôle abusif parental sur l'ajustement psychologique des adolescents de 14 à 18 ans /

“…La flexibilité qu'offre cette technique permet également de tenir compte des données manquantes dans l'estimation des paramètres (Dupéré et ai., 2007). Ainsi, SAS PROC MIXED estime des paramètres de façon à maximiser la vraisemblance (Dedricks, et al, 2009) (Akaike, 1974;Dupéré, et al, 2007;Singer, 1998) et BIC (Dupéré, et al, 2007;Schwarz, 1978) sont basés sur la vraisemblance et encourent une correction pour le nombre de paramètres à inclure dans le modèle final. En effet, parmi les modèles obtenus à l'aide des analyses, on retient le modèle qui minimise ces critères de sélection.…”

“…En effet, parmi les modèles obtenus à l'aide des analyses, on retient le modèle qui minimise ces critères de sélection. Toutes les variables sont centrées autour de leur moyenne, ce qui facilite l'interprétation des résultats (Dedricks, et al, 2009). …”