plot of residuals vs predicted Hi all,
I have a linear mixed random intercept model in r. I have applied this model to longitudinal/panel data of customers of a time period. For each customer, i have a different intercept. As i want to obtain their general effect over time to make conclusion i have chosen for a linear random intercept model.
For this model, as for each linear model, the error term has to be homoskedastic.
I want to test this assumption statiscally, however a bptest can't be applied because this assumes I.D.D. So, next to my plot which i attached. Do you guys know any method to test this statiscally?
Thank you in advancw