Infuence diagnostics are not estiamted (Update failed)

[marina_unn]

Dear SAS users,

I have a short question about SAS PROC MIXED influence diagnostics. I run a mixed model where each patient is a random effect, I wanted to check influence diagnostics, by adding the option INFLUENCE with 5 iterations. I have 658 patinets in total, and I see in the Results that for many patients the influence diagnostics are not estimated (the Notes columns says "Update failed").

model ecs=
  age_start 
  time_num 
  time_num2
 age_start*time_num/  INFLUENCE(iter=5 effect=id) residual;

My questions is about this failed update: does it happen because I have too many patients or because I need to include more iterations for the influence diagnostics?

If someone knows the nature of this error, please, let me know. Even if I cannot fix it, I'd like to know the nature of the problem.

Many thanks in advance!

Marina

[jiltao]

Here is the related information from the MIXED documentation.

EFFECT=effect

specifies an effect according to which observations are grouped. Observations sharing the same level of the effect are removed from the analysis as a group. The effect must contain only class variables, but need not be contained in the model.

Removing observations can change the rank of the ( X' V-1 X)- matrix. This is particularly likely to happen when multiple observations are eliminated from the analysis. If the rank of the estimated variance-covariance matrix changes or its singularity pattern is altered, you would receive that message in the Log, and no influence diagnostics are computed. I suspect that is the case here.

Thanks,

Jill

[marina_unn]

Thank you, @jiltao , I will check the log when I rerun the models this week.

I guess there is no solution to this problem otherwise? 

[jiltao · Accepted solution]

Not that I know of.