Re: Multiple Regression with Continuous and Categorical

[Martin P. Holt-2]

On 24th July, SR Millis gave an list of excellent reasons for not doing stepwise procedures. OK, so what does one do ?

Best Wishes,

Martin Holt

====================To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[SR Millis-3]

Martin,

Among the alternatives to stepwise variable selection:

--Theoretically-driven variable selection.

--Bayesian model averaging.

--Least angle regression (if the response variable is continuous).

--the stepdown method of Lawless & Singhal.



Scott R Millis, PhD, MEd, ABPP (CN,CL,RP), CStat
Professor & Director of Research
Dept of Physical Medicine & Rehabilitation
Wayne State University School of Medicine
261 Mack Blvd
Detroit, MI 48201
Email:  [hidden email]
Tel: [phone]
Fax: [phone]


--- On Tue, 7/29/08, Martin Holt <[hidden email]> wrote:

> From: Martin Holt <[hidden email]>
> Subject: Re: Multiple Regression with Continuous and Categorical
> To: [hidden email]
> Date: Tuesday, July 29, 2008, 1:29 PM
> On 24th July, SR Millis gave an list of excellent reasons
> for not doing stepwise procedures. OK, so what does one do
> ?
>
> Best Wishes,
>
> Martin Holt
>
> ====================To manage your subscription to SPSSX-L,
> send a message to
> [hidden email] (not to SPSSX-L), with no body
> text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the
> command
> INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Art Kendall]

In many situations the best best approach is to use hierarchical or
"stepped" regression.  In this approach, variables are entered in a
predetermined order to see if there is satistcally significant and
meaningful improvement in the fit of the equation to the data.

A lot depends on the nature of your inquiry, what data you can  gather,
and what rhetorical points you are seeking to establish.

Art Kendall
Social Research Consultants


Martin Holt wrote:

> On 24th July, SR Millis gave an list of excellent reasons for not doing stepwise procedures. OK, so what does one do ?
>
> Best Wishes,
>
> Martin Holt
>
> ===================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
>
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD