Doubly latent multi-level modelling

Value adding in diverse educational contexts: substantive-methodological synergies that address complex issues with sophisticated methodology – Professor Herb Marsh

Professor Herb Marsh was awarded an ESRC Professorial Fellowship 2008-2012. The fellowship is extremely prestigious, with only 3-5 awarded each year, to allow leading researchers in social science “to create and evolve their own creative and ground-breaking research agendas”. The fellowship was based on the premise that complex substantive issues require sophisticated methodologies. Bringing the two together creates a powerful synergy that contributes to understanding critical educational issues. Marsh and Hau (2007) drew attention to the dearth of quantitative skills in educational research and the need of substantive-methodological synergy; first to integrate important complex issues and sophisticated, methodology appropriate to the task; second to provide models of good research, set standards, and motivate researchers to develop the requisite skills to understand advanced quantitative methods.

Particularly in education and the social sciences more generally, Professor Marsh has been a leading contributor to Structural Equation Modelling (SEM) and Multilevel Modelling (MLM) methodologies. Multilevel modelling was an overarching theme of the fellowship. Educational research is inherently multilevel; students are nested in classrooms, classrooms and teachers in schools, etc. Traditional single level models are inherently biased, ignoring contextual effects of group-level attributes (e.g., classroom motivational climate). One key element of the fellowship was on doubly latent contextual effects models that integrate SEM and MLM. These models simultaneously control measurement error through the use of multiple indicators (as in SEMs), and control of sampling error at the group level with a latent aggregation of latent individual variables. Although these models have wide applicability across the social sciences, he has applied these to school effectiveness studies, university rankings and self-concept research. His latest methodological work has been on Exploratory Structural Equation Modelling (ESEM) an integration of the best features of exploratory and confirmatory approaches to factor analysis with broad relevance across the social sciences.

These studies not only make important advances in substantive aspects of educational research, but also provide a model for UK educational researchers of appropriate substantive-methodological synergies. Hence, the overarching agenda was to use this research program to advocate substantive-methodological synergies; to demonstrate their power to address diverse and complex issues of practical significance, and to provide “best practice” models of this approach to assure researchers, particularly doctoral students and early career researchers, of the need to develop appropriate methodological skills. This ongoing commitment to capacity building is demonstrated through the successful ERSC Research Development Initiative that funded Prof. Marsh’s workshops on new approaches to meta-analysis conducted around UK, sessions at the 2008 & 2010 ESRC Research Methodology Festivals, and the founding of a quantitative special interest group at Oxford.