The Design of Time-Series Comparisons Under Resource Constraints

Abstract

We compare two methods for dividing an interrupted time-series study between baseline and experimental phases when study resources are limited. In fixed designs the baseline duration is chosen before the baseline begins. In flexible designs the baseline duration is not predetermined but made contingent on remaining resources and the degree to which the results match the prior expectations of the evaluator. Flexible designs are much pre ferred when the evaluator's priors are of good quality.