Project
Meta-Analytic Structural Equation Modeling with Group Data

Meta-analytic structural equation modeling (MASEM) is a method to systematically synthesize results from primary studies. The technique is increasingly popular in various research fields as it allows a researcher to simultaneously examine multiple relations among variables by fitting a structural equation model to the pooled correlations from a meta-analytic dataset. In MASEM, one can include the effect sizes of primary studies even when they do not include all variables of interest. Still, with current MASEM methods, it is not evident how one can include natural dichotomous and artificially dichotomized variables. A dichotomous variable indicates, for instance, whether participants are in the experimental or control group. A dichotomized variable represents, for example, whether a person scored above or below a cut-off score on a continuous variable.

In this PhD project, I evaluate MASEM methods with the aim of advising researchers on how to answer research questions involving group data. Specifically, one obstacle would be that MASEM requires correlation (or covariance) matrices as input, whereas the summary statistics reported in studies with group data are typically converted to standardized mean differences (e.g., Cohen’s d or Hedges’ g). To use MASEM, we need to convert standardized mean differences to a correlation coefficient. I will consider and evaluate methods to transform effect sizes obtained from group data with natural dichotomous and artificially dichotomized variables into effect sizes appropriate for MASEM. To assist meta-analysts, I plan to implement the relevant conversion formulas in a user-friendly online application.


Supervisors
dr. S. (Suzanne) Jak
dr. K.J. (Kees-Jan) Kan
Prof. dr. F. J. (Frans) Oort

Financed by
NWO-Vidi Suzanne Jak (“No data left behind”)

Period
01 November 2021 – 1 November 2024