Computational obstacles remain a great barrier to progress in a range of microdata research areas, including the analysis of semi-parametric structural models, research involving very large datasets, the analysis of models with dynamics and complexity, and the analysis of networks, interactions and equilibrium. One result is that researchers resort to restrictive and misspecified parametric models to get around computational difficulties. While this produces valuable information in many cases, it severely limits the information that can be extracted from data about economic and social processes and often prohibits consideration of important phenomena. Analysis of models involving essential non-linearities and rich forms of heterogeneity requires advances in computational methods. This strand of the Centre's research focusses on advancing computational social science. Computational tools will be developed for estimation and testing, for understanding longitudinal data and dynamic models, and for analysing models of networks, of interactions and of equilibrium. Applications will use the most sophisticated computational tools available to expand the frontier of applied microdata research.