508 proc mianalyze parms=mixparms edf=66;
509 by type sex;
510 class group day;
511 modeleffects group day day*group;
ERROR: Within-imputation Estimate missing for effect group*day in _Imputation_= 1 in the input PARMS= data set.
NOTE: The above message was for the following BY group:
at which point MIANALYZE stops.
So, I look at the dataset mixparms (created by ODS from PROC MIXED as in the documentation Example 55.4), and it looks exactly like the printout of the dataset in the documentation, except for the usual zeroes that one gets out of the solution for the fixed class effects.
In an effort to keep my evergrowing bald spot from getting even larger through headscratching, I appeal to the good readers--What obvious thing am I missing here?