In this section we generalize what we have seen so far concerning the expected value of a normal population. We consider the following problems:
- Estimating and testing the difference in the mean of two populations, which is used to assess if there are significant differences between them
- Hypothesis testing for variances of normal populations
- Testing Bernoulli populations, i.e., estimating and testing proportions
- Estimating covariances and related issues
- Estimating and testing a coefficient of correlation
- Estimating skewness and kurtosis
As you can see, this is a rather long list of topics. Because of space limitations, we will take a somewhat “cookbook” approach, cutting some corners to keep the treatment to a reasonable size. Nevertheless, we think that the previous discussions have provided the reader with a sensitivity for the pitfalls of inferential statistics techniques. Our aim is to illustrate the tools provided by statistical software packages and their rationale. We refer readers to the references for a more thorough and solid treatment.
Leave a Reply