General research concerns multivariate techniques for studying the simultaneous behavior and effects of numerous variables and large samples of complex structure in a subject area related context. We study distributional properties, estimation and testing methods in connection with multivariate linear or nonlinear multilevel mixed models with latent structures and endogenous factors. Such techniques are applied to panel data analysis of rural development of the agricultural structural transformation in Africa. Another area is within studies of diversification of income sources among farm households in Vietnam, to study the impact on agricultural transformation, economic growth and development in time.

Other branches concern the statistical analysis of data with certain natural restrictions. This includes efficient methods to combine and smooth diverging results from many election polls carried out by different organizations. Other problems concern modelling of clusters in directional data or orientation data with restrictions due to spatial boundaries. 

My main areas of teaching are multivariate statistics and spatial statistics often in a subject area related context. I also teach, at different levels, on the foundations of statistical theory.