Hypothesis-testing
Chapter 10
10.65 The criteria used for selecting a hypothesis-testing procedure are as follows.
- The first criterion is based on the number of groups being tested. The hypothesis test to be used for two samples is different from that used for three samples. Thus, it is crucial to determine the number of samples being compared before choosing a hypothesis-testing procedure.
- Secondly, the type of data should be considered while choosing a testing procedure. Different procedures are used for discrete data and different for continuous data.
- Lastly, the selection of the hypothesis testing procedure should be guided by the type of population parameter being compared. There is a difference between the methods used to compare means, medians, and standard deviations.
10.68 Two independent populations refer to populations that don’t have common characteristics. For each population, the means and standard deviations are different. On the other side, related populations refer to populations that have a common origin, and their characteristics are connected. They could share a similar mean or standard deviation.
10.73 While using the ANOVA for analysis, the following assumptions are considered:
- Samples come from populations exhibiting normal distribution. Shapiro Wilks Test is used to test for the normality of samples. If the Shapiro Wilks test results in a non-significant p-value, then the sample is normally distributed.
- The populations being studied have a common variance.
- The process used to assign subjects to samples was random. The selection of subjects in one sample is independent of the selection in another sample.
- Samples are drawn independently.