Weighted data - Different totals between MEANS procedure and UNIANOVA

[Steve Barron-2]

Hi there

If i run the following code on weighted data i get different numbers in
the 4 cells of my 2x2 factorial design - the total number of cases
according to the MEANS procedure is 2,405; the total number of cases
accoring to UNIANOVA is 2,764.

USE ALL.
GRAPH
  /HISTOGRAM(NORMAL)= depvar.
MEANS
  TABLES=depvar  BY indvar1  BY indvar2
  /CELLS MEAN COUNT STDDEV  .


UNIANOVA
  depvar  BY indvar1 indvar2
  /METHOD = SSTYPE(3)
  /INTERCEPT = EXCLUDE
  /PRINT = DESCRIPTIVE HOMOGENEITY
  /CRITERIA = ALPHA(.01)
  /DESIGN = indvar1 indvar2 indvar1*indvar2.

When i turn off the weighting the numbers are the same according to both
procedures - 2,405 which is the correct number of cases in the data file.
Does anybody know why this is happening?

Many thanks
Steve Barron