Comparing Two Normal Populations

Abstract

In Chapter 3 we considered problems of making inferences about the parameters of a single normal distribution. In this chapter, we will deal with various problems involving parameters of two normal populations. Statistical inferences concerning the parameters of two normal distributions arise in the comparison of two treatments, brands, market segments, services, etc. Experimenters in a variety of fields including biomedical research, agricultural experiments, quality engineering, and socioeconomic investigations often need to perform tests and construct confidence intervals for the difference in two normal means or variances. For example, this is the case when an experimenter compares the yield of a new hybrid corn with a standard kind.