CHAPTER 10 Nonnormality and Heteroskedasticity

This is part 2 of the diagnostic checking part.

Recall the assumption about the random error term: \varepsilon_i\sim Normal(0,\sigma^2) To test assumptions about the error term, we explore the characteristic of the observed values of this random variable, which are the residuals e_i=Y_i-\hat{Y_i}.

In this section, we want to know if the residuals are sampled from a normally distributed population and has the same variance regardless of the valuue of Y_i.

10.1 Effects and Usual Causes of Non-Normality and Heteroskedasticity

10.2 Testing for Nonnormality

10.3 Testing for Heteroskedasticity

10.4 Remedial Measures

10.5 Exercise