Date:         Thu, 10 Oct 2002 20:51:26 +0200
Reply-To:     anne5432@uni.de
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Annette <anne5432@uni.de>
Organization: Annette
Subject:      Very tricky filter-Problem
Content-Type: text/plain; charset="us-ascii"

Hi, I have a very tricky filter-problem. Please do help! It's very-very
urgent!
Background: I have to analyze missing data of a repeated measurements
analysis. The analysis data set contained three variables VARTIME 1 to 3
and the var COVARIABLE. I counted the data by "count DATABASE= vartime1
vartime2 vartime3 COVARIABLE (missing sysmis)." DATABASE=0 gives me the
correct number of nonmissing variables of the analysis performed before
(including the COVARIABLE, very important!). DATABASE=1 to 4 0 give me
the numbers of missing variables. However, missings values of the
COVARIABLE are counted too. However, I need DATABASE 1 to 3 categories
that containing the number of missings of VARTIMEs without counting the
missings of COVARIABLE too. In short, I need a group variable whose
values (categories) fulfill the following conditions: 0=number of
nonmissing variables including COVARIABLE (count-sample above produces
correct value). 1=number of one missing VARTIMEx ("x" indicates it
doesnt't matter which) without numbers of missing COVARIABLE. 2=number
of two missing VARTIMEx. ("x" indicates it doesnt't matter which)
without numbers of missing COVARIABLE. 3=number of all VARTIMEs missing
without numbers of missing COVARIABLE. Thank you so much in advance for
your precious time and effort! Annette