LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (February 2003)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Wed, 19 Feb 2003 11:49:56 +0100
Reply-To:     fvillamayor-research@ferrergrupo.com
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Frederic Villamayor <fvillamayor-research@ferrergrupo.com>
Subject:      Re: ONEWAY witn MATRIX DATA and SPLIT FILE: a bug?
Comments: To: Asesoría Bioestadística
          <bioestadistica@eresmas.net>
Content-Type: text/plain; charset="iso-8859-1"

Hi,

This happens with version 8.01, too. I've tried to explicitly demand SPLIT FILE LAYERED BY testnum. before running ONEWAY, and the result is the same. If I try SPLIT FILE SEPARATE BY testnum.

only the results for the seconf split group are reported.

Frederic Villamayor Departamento de Toxicología C.I.D.F. GRUPO FERRER

Asesoría Bioestadística <bioestadistica@eresmas.net> Enviado por: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU> 19.02.03 10:28 Por favor, responda a Asesoría Bioestadística

Para: SPSSX-L@LISTSERV.UGA.EDU cc: Asunto: ONEWAY witn MATRIX DATA and SPLIT FILE: a bug?

Hi everybody:

I am working on a big set of summary data (mean, sd & n) for several groups and several variables. The idea is to perform a oneway ANOVA for each variable (labeled as "testnum") in the dataset. I have used MATRIX DATA to create a matrix suitable for ONEWAY... /MATRIX=IN(*). This a sample of the dataset (just two variables):

MATRIX DATA VARIABLES=testnum rowtype_ group depvar /SPLIT=testnum /FACTORS=group. BEGIN DATA 1 MEAN 1 62.80 1 MEAN 2 70.20 1 MEAN 3 79.40 1 MEAN 4 80.20 1 STDDEV 1 6.97 1 STDDEV 2 6.21 1 STDDEV 3 6.24 1 STDDEV 4 6.11 1 N 1 10 1 N 2 10 1 N 3 10 1 N 4 10 2 MEAN 1 72.80 2 MEAN 2 80.20 2 MEAN 3 89.40 2 MEAN 4 90.20 2 STDDEV 1 7.97 2 STDDEV 2 7.21 2 STDDEV 3 7.24 2 STDDEV 4 7.11 2 N 1 15 2 N 2 15 2 N 3 15 2 N 4 15 END DATA.

SPSS syntax guide says that ONEWAY should recognize split groups specified in matrix data and perform a separate ANOVA for each level of the splitting variable:

ONEWAY depvar BY group /STATISTICS=DESCRIPTIVES /POSTHOC=TUKEY /MATRIX=IN(*).

This syntax should give 2 descriptives, 2 ANOVA tables, 2 multiple comparisons tables and 2 homogeneous subsets tables. I get the last (2 homogeneous subsets tables, one for each "testnum" level), but only one descriptive, ANOVA & multiple comparison (for "testnum"=2).

I have tried this syntax with SPSS 9 and SPSS 11. I get the same results for both versions of the program.

Is this a bug or should MATRIX DATA be specified in some other way? (I followed the instructions given in the manual, and, in fact, it looks like the splitting is recognized, at least partially).

Of course, I know if can do something like:

temporary. select if (testnum=1). ONEWAY depvar BY group /STATISTICS=DESCRIPTIVES /POSTHOC=TUKEY /MATRIX=IN(*).

temporary. select if (testnum=2). ONEWAY depvar BY group /STATISTICS=DESCRIPTIVES /POSTHOC=TUKEY /MATRIX=IN(*).

And so on.

Or design a "looping macro" to do the job, but curiosity is one of my defects - I'm a scientist - ;)

Regards

Marta Garcia-Granero.


Back to: Top of message | Previous page | Main SPSSX-L page