Problem with SPSS Complex Sample GLM - Unweighted Sample Size

Problem

I am using the SPSS Complex Samples Module to fit a general linear model (CSGLM) to a subpopulation in my data set. There seems to be an error in the unweighted valid cases reported in the 'Sample Design Information' table of the CSGLM output. The subpopulation of interest is defined as those cases with a value of 1 on the variable SUBPOP. The CSGLM subcommand /DOMAIN was used to focus analysis on these cases in the following syntax command. CSGLM SelfConcept BY income Age urban WITH st2 Educ /PLAN FILE='C:\spsssupp\edstudy.csaplan' /DOMAIN VARIABLE=subpop(1) /MODEL income Age urban st2 Educ /INTERCEPT INCLUDE=YES SHOW=YES /STATISTICS PARAMETER SE CINTERVAL TTEST DEFF /PRINT SUMMARY VARIABLEINFO SAMPLEINFO /TEST TYPE=F PADJUST=LSD /SAVE PRED RESID /MISSING CLASSMISSING=EXCLUDE /CRITERIA CILEVEL=95. SPSS is reporting n = 1000 valid unweighted cases in the 'Sample Design Information' table. When Descriptives is run on these variables without a filter in place, then the Listwise N (the number of cases with valid values on all of the variables analyzed) also equals 1000. However, when the Descriptives procedure is run with this subpopulation selected by a filter, there are only 250 valid cases with complete data for this model in the subpopulation of interest. If I also run the CSGLM with the filter applied, then the Unweighted Valid N reported in the Sample Design Information table is the expected 250. Is the /DOMAIN subcommand insufficient to focus the analysis on the subpopulation of interest?