Comparación entre tamaños de muestras de medias pareadas con análisis de poder usando paquetes del software R y el software G*Power.

Abstract

Hypothesis verification requires a prior approach of factors such as the sample size and its reliability of statistical tests to address experimental studies, since the presence of an effect derived from a treatment can be rejected, when in reality there is not enough statistical power to arrive at that conclusion. The objective of this article is to define what is the power of a statistical test, explain its calculation and determine the level of approximation of the sample sizes of the statistical test of paired means generated by three R software packages, and identify if there are similarities in the generation of results with the G * power software; giving a degree of security and confidence for statistical tests in power analysis.