Social experiments are characterized by their high cost. A tempting alternative to the establishment of a contemporaneous statistical control group is pre-experiment observation of the treatment group. In this paper we analyze the trade-off between these two types of experimental "control" as a function oftheir relative cost and information content in the context of a multi-period error components framework, where the allocation of observations across the two groups is always done in an optimal manner. Solutions for the optimal proportion of the sample to be devoted to a contemporaneous control group are presented and their behavior as a function of relevant parameters is studied.