I'm trying to make sure I understand how to best think about the residual and scale location plot as it relates to heteroscedasticity.

A couple plots below. My initial thought was that that since the residuals are not spread 'evenly' across the x-axis, it breaks the 'constant variance' assumption. However, it does look random in the sense that it doesn't look like there is a pattern to it.

Looking at the scale location plot, the plots look the same, but the redline is decently flat, which tells me the average magnitude of the residuals is flat, so does that mean there is constant variance?

I do get confused about what to consider when looking at these things (shape, vs "randomness" vs spread, vs the red line)

Residual vs fitted

scale location


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.