A study on the statistical comparison methods for engineering applications

Abstract

In this paper, we introduce a data comparison method, Matching Rule (MR). By setting up a relation between MR and F/T, an empirical criterion of MR for comparing two groups of data is defined. It is based on a study on the classical statistic method “Hypothesis test - F/T”. The F/T test is widely used to compare variations & means of two normal populations. However, the empirical criteria of MR do not consider type I error. To make MR truly useful, we develop a program to simulate the sample size needed for the two groups at different levels of type I error. Then MR criteria and the minimal sample size can be determined based on the required type I error. Our simulation results show that the type I error of MR can approach to the traditional F/T test method when sample size is close to 30. With a large sampling size, MR tool is more useful for engineering application than statistical comparison test [1].