Tips for troublesome sample-size calculation

Abstract

Properly set sample size is one of the important factors for scientific and persuasive research. The sample size that can guarantee both clinically significant differences and adequate power in the phenomena of interest to the investigator, without causing excessive financial or medical considerations, will always be the object of concern. In this paper, we reviewed the essential factors for sample size calculation. We described the primary endpoints that are the main concern of the study and the basis for calculating sample size, the statistics used to analyze the primary endpoints, type I error and power, the effect size and the rationale. It also included a method of calculating the adjusted sample size considering the dropout rate inevitably occurring during the research. Finally, examples regarding sample size calculation that are appropriately and incorrectly described in the published papers are presented with explanations.

Keywords: Biostatistics; Effect size; Independent t-test; P value; Power; Sample size.

Fig. 1.

Description and relevance of dropout rate. Appropriate: Appropriate calculations and clear descriptions of the sample size derivation process, Inappropriate: Inappropriately calculated sample size, Miscellaneous errors include unclear primary outcome, absence of value in spite of mention of dropout rate, absence of data about calculation of sample size, and miscalculation other than dropout rate. KJA: Korean Journal of Anesthesiology, APM: Anesthesia and Pain Medicine.