Optimal design of observational studies

Motivated by the universal need for cost-efficient data collection, we propose a unifying framework for optimal design of observational studies. In this framework, the design is defined as a probability measure in the space of observational processes that determine whether the value of a variable is observed for a specific unit at the given time. The framework allows one to optimize decisions on three design dimensions: unit, variable and time. The optimal observational design maximizes the utility defined as a function of model parameters, data and policy decisions. We discuss computational methods that can be used to find optimal or approximately optimal designs. We review a wide variety of design problems that can be cast in the proposed framework, including sample size determination, subsample selection, selection for re-measurements, choice of measurement times, and optimization of spatial measurement networks.