Checking model assumptions

To use linear regression, your data must satisfy the following four core assumptions:

• Linearity

• Independence

• Normality

• Equal variance

It may be helpful to think of these assumptions by their first letters. You can remember that LINE is an important aspect of linear regression. Next, you will learn about each of the assumptions as well as tests that you can perform in R to check whether the data satisfies them.

Note

Learn more: Checking the assumptions of a statistical model is important. The power and accuracy of any model comes from its adherence to the assumptions. David Robinson (2015) has written a blog called VARIANCE EXPLAINED that describes this topic in an enjoyable way: http://varianceexplained.org/r/kmeans-free-lunch/

Linearity

Note

Linearity assumption: The relationship between the predictor and response variables is linear.

In an SLR situation, a quick way to determine linearity is to plot the variables with a scatterplot. Earlier, you saw a strong correlation...