Hi all,
I have built the following linear mixed models to understand the association of biological aging (grimAge) with pulmonary function assessed annually. Instead of using time as the time-scale, we used chronological age incremented by 1 year at each annual visit.
mixed pulmonary c.chronologicalage##c.chronologicalage##c.grimage i.sexe || id: chronologicalage, residuals(ar 1, t(chronologicalage))
My supervisor is asking me about assessment of model assumptions and diagnostics. I am familiar with what needs to be checked for linear regression : Visual inspection of scatterplots with fitted regression lines, residual-versus-fitted plots to check heteroscedasticity, kernel density and Q–Q plots for normality of residuals, eventually variance inflation factors for multicollinearity but I am struggling with finding what really needs to be checked for linear mixed models.
Could you please help me with this? I do not need anything super fancy here but just to make sure my model is OK to answer my supervisor.
Thank you so much for your help,
Best regards, Pierre
I have built the following linear mixed models to understand the association of biological aging (grimAge) with pulmonary function assessed annually. Instead of using time as the time-scale, we used chronological age incremented by 1 year at each annual visit.
mixed pulmonary c.chronologicalage##c.chronologicalage##c.grimage i.sexe || id: chronologicalage, residuals(ar 1, t(chronologicalage))
My supervisor is asking me about assessment of model assumptions and diagnostics. I am familiar with what needs to be checked for linear regression : Visual inspection of scatterplots with fitted regression lines, residual-versus-fitted plots to check heteroscedasticity, kernel density and Q–Q plots for normality of residuals, eventually variance inflation factors for multicollinearity but I am struggling with finding what really needs to be checked for linear mixed models.
Could you please help me with this? I do not need anything super fancy here but just to make sure my model is OK to answer my supervisor.
Thank you so much for your help,
Best regards, Pierre
