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

st: RE: ...stepwise regression...

From   "Nick Cox" <[email protected]>
To   <[email protected]>
Subject   st: RE: ...stepwise regression...
Date   Tue, 13 Apr 2004 11:03:58 +0100

I suspect that -sw- does not offer all the procedures
which have been dubbed "stepwise". In addition there
is a terminology problem. As far as I understand the matter, 

stepwise sensu lato (broad sense) includes "forwards 
only" and "backwards only" methods as special cases; 

but stepwise sensu stricto (narrow sense) means a procedure
in which you can proceed backwards _and_ forwards (which
is _not_ available in -sw-, I believe). 

I also guess, wildly, that -sw- might be regarded 
with perfect hindsight as a mistake. Put it this way: 
if -sw- did not exist in Stata, I doubt that StataCorp 
would implement it now. The collective statistical attitude 
seems to have shifted towards regarding such methods as 
snake oil. 

It is difficult, indeed impossible, to say what you should do 
without knowing any of the scientific or practical background to your 
project. It's your decision to make. 

[email protected] 

[email protected]
> Nick, many thanks for your help.
> I have read the FAQ file on Stata Site. I share their (Harrel 
> & Conroy)
> view on this point, but I have to set a interpretation model 
> of a dipendent
> variable and I have more than 500 indipendent variables. I 
> have select the
> more correlate variables but I need a procedure to select the 
> most explanatory
> variables. 
> If I don't apply the sw linear regression I don't know how 
> select the indipendent
> variables. I have to do.
> But the procedure in Stata doesn't correspond to the model I 
> studied in
> many manuals. In particular, the sw procedure should go 
> forward and back
> but in Stata I don't see the different between the sw and the 
> forward or
> backward simple selections.
> If I don't use the sw procedure, how select the variables??

*   For searches and help try:

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