ABSTRACT
Most studies that are of consequence are expensive undertakings, and experience accumulated over time is conclusive about the value of designing them—investing in their preparation and planning by reviewing the know-how and experience from past studies with similar agenda and context. The result of planning a study, that is, designing it, is a protocol. In a general example of such a study, the data have been collected in the past for another purpose, and so they are now available at a nominal cost. There is a vast literature on sample size calculation for estimation and hypothesis testing. It is a key aspect of clinical trials and other studies in which established ethical standards have to be met. A study should have its inferential goals clearly formulated at the design stage and should not be conducted with the sole purpose of searching for some finding that would have been remarkable and worth reporting had it been declared at the design stage.
