I run a ols regression and want now check the linearity assumption. I found out that i have to plot the residuals vs the fitted values and if there is no non linear pattern the linearity assumption is fullfiled.

Here is the plot:

enter image description here

On my opinion i would say there is no real non linear pattern so i would say the linearity asumption is fullfiled.

Am i right?

  • $\begingroup$ Those residuals look like they're fanning. Constant variance will be violated. Also it might be non linear something's causing those clear boundaries in your residuals. Without knowing more about the data it is hard to know why but definitely assumptions are violated currently. $\endgroup$ – Huy Pham Apr 23 at 2:27
  • $\begingroup$ But i thought if the red line is near 0 the linearity assumption is fullfiled. For example like its explained here sthda.com/english/articles/39-regression-model-diagnostics/… $\endgroup$ – MasterStudent1992 Apr 24 at 13:04
  • $\begingroup$ I'm not positive if it's non linear, but that boundary on the bottom left looks like an area that your predictions won't go beyond. stats.stackexchange.com/questions/197694/… Maybe you have discrete integers as your DV? Either way, the other assumptions will not hold from that scatterplot. $\endgroup$ – Huy Pham Apr 24 at 13:27
  • $\begingroup$ My depended variable is log(wage). I'am only interessted in the linearity assumption. So on your opinion the linearity assumption is fullfiled but other assumptions not ? $\endgroup$ – MasterStudent1992 Apr 24 at 14:19
  • $\begingroup$ Looks like that boundary is happening because the model won't log 0. Yeah looks linear enough but constant variance is violated. $\endgroup$ – Huy Pham Apr 24 at 15:41

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.