Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: OLS parameters differences Stata vs SPSS

From   "Maarten Buis" <>
To   <>
Subject   st: RE: OLS parameters differences Stata vs SPSS
Date   Thu, 29 Jun 2006 14:42:47 +0200

Statalist is a multidisciplinary list, so though it may 
be obvious what you mean with the Mankiw-Romer model in 
your discipline, it is not so outside. So you have 
given (at least) me very little to go on. One thing I 
can tell you is that estimation of OLS parameters is 
well established, so I would find it inconceivable that 
a difference in the estimation algorithm would be the 
cause of the difference. A much much more likely cause 
is that you specified a different model in Stata than 
in SPSS. For instance, facilities for dealing 
categorical values for SPSS (by calling a independent 
variable a factor) and Stata (by using either -xi-, 
-xi3-, or -desmat-), could easily use different 
defaults which would result in mathematically 
equivalent models which would still result in parameter 
estimates that look very different.

Could you tell us what the Mankiw-Romer model is, and 
the exact SPSS and Stata commands you used? (For SPSS 
you can press paste instead of OK in the menu to get 
the text file with the command in it)


Maarten L. Buis
Department of Social Research Methodology 
Vrije Universiteit Amsterdam 
Boelelaan 1081 
1081 HV Amsterdam 
The Netherlands

visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z214 

+31 20 5986715

-----Original Message-----
From: []On Behalf Of Gerardo MagaƱa
Sent: donderdag 29 juni 2006 13:46
Subject: st: OLS parameters differences Stata vs SPSS


Sorry for this silly question......I'm running an OLS regression  both in
Stata and SPSS, to replicate the Mankiw-Romer model. Nevertheless, the
output parameters from both programs are substantial different. The problem
is that the parameters from SPSS resembles very well the original results of
Does anybody knows why the difference in the parameters estimation? Is it
possible to alter the way stata runs the OLS procedure ?
Thanks for your kindly help.

*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index