Statistical hypothesis testing: motivation and methods for the comparison of means between 2 patient samples.

摘要

The logic behind statistical reasoning may appear to proceed in the opposite direction from scientific reasoning. Thus, understanding and interpreting statistical tests and their conclusions may be counterintuitive for many researchers and WOC nurses. For example, scientists often reason deductively and aggressively: "I hypothesize that there is a difference between treatment A and treatment B, and I will collect data to prove it." Statisticians think inductively and defensively: "The data show some difference between treatment A and treatment B, but is it enough evidence to conclude that a no-difference hypothesis is wrong?" A statistical hypothesis test is a method for quantifying how much evidence constitutes "enough." In this article, we will outline the motivation for statistical hypothesis testing, and discuss specific examples for 2-sample testing in the context of WOC nursing. This is the first in a series of 6 articles; subsequent articles will cover other types of statistical tests. To set the stage for 2-sample testing (eg, the comparison of 2 treatment groups), consider these examples: