Testing differences between two samples of continuous data

Abstract

Summary: In this article the circumstances and techniques used to test a hypothesis by comparing information from two random samples to identify possible or existing difference amongst the target population are presented. This is based on continuous data collected from two samples after which a comparison can be made that may then be generalised to the source population. The terminology of the method is briefly explained including basic concepts such as survey errors and probablistic chance. Hypothesis test methods are described for two types of samples: independent and dependent, and examples of some of the most commonly use test in health research are given for parametric and non-parametric source distributions. Testing difference between two samples of continuous data is an important and frequently applied process of making a decision about the main differences in a population of interest in health research.