Michael Browne's research has been primarily concerned with statistical modeling of multivariate psychological data. He has made contributions on asymptotically distribution free estimation in the analysis of moment structures, asymptotic robustness of multivariate normal theory against violation of assumptions, multiplicative models for multitrait-multimethod data, circumplex models for investigating the personality circle and rotation methodology in factor analysis. His main current interest is in the modeling of longitudinal data. He is carrying out further research on nonlinear latent curve models, first considered in Browne & du Toit (1991) and Browne (1993) and in time series models (du Toit & Browne, 2001).