Date: Thu, 10 Feb 2000 15:45:35 +0100
Reply-To: Markus Quandt <Markus.Quandt@UNI-KOELN.DE>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Markus Quandt <Markus.Quandt@UNI-KOELN.DE>
Organization: Universitaet zu Koeln
Subject: Tables problem
Content-Type: text/plain; charset=us-ascii
Dear all,
I have the following problem. I have a rather large dataset (N>20.000)
for which I want to produce sets of separate tables. Each set is
supposed to show aggregate statistics for a different *subgroup* (# of
subgroups: K>400) of cases in the dataset. No problem so far: Just use
TABLES or even DESCRIPTIVES, and run it behind a SPLIT FILE BY groupvar.
BUT I want the tables to show the statistics for each subgroup (mean,
min, max, N, stddev) AND as a standard of comparison the mean and maybe
the stddev over all groups. How do I get both the "grand mean" and the
"group mean + other statistics" into one single table?
The ideal output would conceptually look like this:
*********
Subgroup 1:
Var. Mean min max Std.Dev. Grd Mean Grd Std.Dev.
v21 2.2 1 5 .68 2.1 .70
v22 1.8 1 4 .53 2.0 .54
....more variables
...more separate tables like the one above with different variables...
Subgroup 2:
Var. Mean min max Std.Dev. Grd Mean Grd Std.Dev.
v21 2.4 1 5 .78 2.1 .70
v22 2.6 1 5 .62 2.0 .54
....more variables
...more separate tables like the one above with different variables...
Subgroup 3:
... the same again ~ 400 times ...
*********
I've been fiddling around with TABLES for some time now, and I can't get
it to do what I need. Is it possible? How? Right now I'm thinking of
using AGGREGATE and LIST/WRITE to produce ASCII output of the values,
import it into a database like ACCESS and format/print the stuff there.
Not very elegant.
Many thanks for your time and advice, MQ
--
_______________________________________________________________________
Markus Quandt Markus.Quandt@Uni-Koeln.DE
Institut f. Angewandte Sozialforschung der Universitaet zu Koeln
Greinstr. 2 50939 Koeln
Tel. +49-221/470-4232 Fax. +49-221/470-5169
Find my PGP public key and other information here:
http://www.uni-koeln.de/wiso-fak/ifas/html/personal/homepage/quandt-markus/
