Why model: Our goal of using model is to describe data, to make inferences about a population or to make predictions about the future.
Review of linear regression: We have known that a model is linear if it is linear in the parameters which is quite important when we would like to identify which model is linear.
Linear regression assumptions: “L” refers that our model should be linear; “I” refers that the errors or observations are independent; “N” refers that the errors are normally distributed with expected value zero; “E” refers to equal or constant variance which is also called homoscedasticity.
ANOVA: Analysis of Variance (ANOVA) is a statistical method used to test differences between two or more means and it is quite useful to compare more than two groups.
ANOVA assumptions: Errors or observations are independent; Errors are normally distributed with expected value zero; Equal or constant variance (homoscedasticity).
A note on Normality: The normality assumption is most important when the size n is small, highly non-normal and small effect size.
