Simulation question

[F. Gabarrot]

Hello,

I am trying, for fun, to make a simulation of the consequences of dichotomizing a continous predictor. I made a syntax that asks SPSS to perform 2 regression analyses (one with the continuous variable treated as continous, and one with the continuous variable treated as a categorical variable, dichotomized using a median-split). My syntax creates 1000 samples of 80 people each, and thus performs 2x 1000 regression analysis.

My question is : Can I ask SPSS not to report every regression tables (Variables, ANOVA and coefficient table) in the output file? I know there is such a command in SAS (the "noprint" command), and I wondered if there was an equivalent one in SPSS.

Thanks for your answers.

Fabrice.

[Peck, Jon]

The regression command syntax gives you some control over the output -  see the choices on the STATISTICS subcommand, but you can suppress any item using OMS.  Here is an example suppressing several regression tables.

OMS /if subtypes=['coefficients ' 'ANOVA' 'Variables Entered Removed'] /destination viewer=no.
REGRESSION
  /STATISTICS COEFF OUTS R ANOVA
  /DEPENDENT salary
  /METHOD=ENTER educ  .
OMSEND.
You can find the subtype names for tables from the Utitiles/OMS Identifiers menu or by right clicking on the table in the outline.

If you are going to run a lot of replications, you may want to put your loop outside the OMS/Regression/OMSEND sequence in order to avoid tying up a lot of memory.  Until OMSEND, the output can't be discarded.

You can also use OMS to direct tabular output to a dataset, so that may be useful in capturing the results for analysis.

HTH,
Jon Peck
SPSS

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of F. Gabarrot
Sent: Monday, April 23, 2007 1:15 AM
To: [hidden email]
Subject: [SPSSX-L] Simulation question

Hello,

I am trying, for fun, to make a simulation of the consequences of
dichotomizing a continous predictor. I made a syntax that asks SPSS to
perform 2 regression analyses (one with the continuous variable treated as
continous, and one with the continuous variable treated as a categorical
variable, dichotomized using a median-split). My syntax creates 1000 samples
of 80 people each, and thus performs 2x 1000 regression analysis.

My question is : Can I ask SPSS not to report every regression tables
(Variables, ANOVA and coefficient table) in the output file? I know there is
such a command in SAS (the "noprint" command), and I wondered if there was
an equivalent one in SPSS.

Thanks for your answers.

Fabrice.
--
View this message in context: http://www.nabble.com/Simulation-question-tf3629407.html#a10134421
Sent from the SPSSX Discussion mailing list archive at Nabble.com.