CHAPTER 8
TESTS OF HYPOTHESES ON
POPULATION MEANS
In Chapter 7 we presented the methods used in constructing confidence intervals
for the population mean, a widely reported parameter. Although we also consider
population means in this chapter, here we show how to test hypotheses concerning
them. The data being analyzed here are interval or ratio data that are also described
as continuous data. Hypothesis testing is a commonly used method for assessing
whether or not sample data are consistent with statements made about the population.
In addition to this chapter, hypothesis testing is discussed in Chapters 9-14.
In Section 8.1, hypothesis testing is introduced by presenting an example for a
single mean. Note that this test is used less than the common test for two means.
Tests for two means using data from independent populations are given in Section 8.2,
and tests for paired data are given in Section 8.3. In the first two sections, we first
present how to perform the tests when 0 is assumed to be known. This was done since
the tests are simpler to understand when the normal distribution can be assumed. It is
followed by tests that are generally used when 0 is not known. The general concepts
used in testing hypotheses are listed in Section 8.4. The sample size needed for testing
for two means is described in Section 8.5. In Sections 8.6-8.8 we discuss comparisons
between the use of confidence intervals and hypothesis testing, correction for multiple
testing, and reporting the results.Basic Statistics: A Primer for the Biomedical Sciences, Fourth Edition.
By Olive Jean Dunn and Virginia A. Clark
Copyright @ 2009 John Wiley & Sons, Inc.95