Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: RE: Transforming regressors prior to estimation in count models

From   [email protected]
To   [email protected]
Subject   Re: st: RE: Transforming regressors prior to estimation in count models
Date   Tue, 7 Jun 2005 17:18:54 EDT


I reestimated their models using  the same variables with the same 
transformed variables (some logs, some other  transforms), but with a slightly expanded 
set of observations in a different  time period. I find no evidence of an 
effect of their key regressor. I then  thought I should do what you suggest, i.e., 
"start over" without assuming the  regressors need to be or should be 
transformed (at least initially).

I  wasn't sure if there was a "rule of thumb" regarding whether one should  
transform prior to estimation or whether one should wait until an assessment of 
 model fit suggests a need to do so. I also wasn't sure if there was a 
difference  between count models and OLS regarding assumptions of the RHS  variables.

But thank you (and the others) for the responses so far. They  are helpful.


In a message dated 6/7/2005 4:54:16 P.M.  Eastern Standard Time, 
[email protected] writes:
At 04:33 PM  6/7/2005 -0400, [email protected] wrote:
>Carter and  Richard,
>Let me clarify that they were estimating count models  (not OLS), and that a
>rationale for why they started with transformed  variables was not given
>(theoretical or  otherwise).

Given that it is a replication, you  probably have to do the same things 
they did - but that doesn't mean that  you can't eventually argue that a 
non-transformation or an alternative  transformation is better.  It may also 
become obvious early on why they  did what they did.

It seems odd to me that they offer no rationale for  the transformations, 
but maybe the substance of the problem is such that it  is just taken for 
granted that things should be done this  way.

Richard Williams,  Notre Dame Dept of Sociology
OFFICE: (574)631-6668,  (574)631-6463
FAX:    (574)288-4373
HOME:    (574)289-5227
EMAIL:  [email protected]
WWW  (personal):
WWW  (department):

*    For searches and help try:

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index