Dear list

When running a two-level logistic model in spss (21)  using "genlinmixed" the following message appears:

--------
     glmm: An internal error happened during execution.
     Execution of this command stops.
-------

The data are handed over in the format  "nr. of events, nr. of trials".
The nr. of events is the total turnout on a given election date.
The nr. of trials is the total nr. of eligible voters.

The null-model without any predictor variables is estimated withouit problem.
The command below, using  'rain' as a fixed effect predictor variable, causes the problem mentioned.

GENLINMIXED
  /FIELDS TARGET=turnout  TRIALS=field(kiesgerechtigd) OFFSET=NONE
  /TARGET_OPTIONS DISTRIBUTION=binomial LINK=logit
  /FIXED  EFFECTS=rain
   USE_INTERCEPT=TRUE
  /RANDOM USE_INTERCEPT=true SUBJECTS=date
   COVARIANCE_TYPE=identity
  /BUILD_OPTIONS TARGET_CATEGORY_ORDER=descending
   INPUTS_CATEGORY_ORDER=ASCENDING MAX_ITERATIONS=100
   CONFIDENCE_LEVEL=95 DF_METHOD=RESIDUAL COVB=MODEL.

Thanks