Date: Thu, 26 Jun 2003 10:41:20 +1000
Reply-To: paulandpen@optusnet.com.au
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: Paul Dickson <paulandpen@optusnet.com.au>
Subject: Re: regression questions
Content-Type: text/plain
Hi Paul,
Are you looking to enter this covariate to remove its impact on the model you are
using. Have you considered the nature of the impact of the covariate before
partitioning out its variance by examining whether is has moderator or mediator
effects on your variables. There are ways you can model this type of analyses in
linear regression using SPSS.
Cheers Paul
> Mark Davenport <madavenp@OFFICE.UNCG.EDU> wrote:
>
> Enter the covariate first to get model fit and parameter estimates
> for
> the covariate alone. Then, in future steps, enter the other
> variables.
> However, the question about how I would choose to enter/remove
> any/all
> subsequent variables in subsequent blocks would depend on the
> hypotheses
> I was testing. I have had hypotheses wherein I entered all variables
> in
> one fell swoop. However, I generally don't like 'shotgun' methods of
> hypothesis testing.
>
> Mark
>
>
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%
> Mark A. Davenport Ph.D.
> Asst to the Vice Chancellor for Student Affairs/Research and
> Evaluation
> The University of North Carolina at Greensboro
> 149 Mossman Bldg.
> Greensboro, NC 27402-6170
> 336.334.5099
> madavenp@office.uncg.edu
>
> 'An approximate answer to the right problem is worth a good deal more
> than an exact
> answer to an approximate problem' -- J. W. Tukey
>
> >>> Paul Mcgeoghan <Mcgeoghan@CARDIFF.AC.UK> 6/25/2003 11:26:03
AM
> >>>
> Mark,
>
> Can I clarify the following:
> Are you saying use the ENTER method and enter all the independents at
> once, but
> entering the control variable first.
>
> This is then controlling for that independent variable in the
> equation.
>
> Paul
>
>
> Interms of using an IV as a control, simply enter it into the
> equation
> first and interpret the statistics for IV entered after.
>
> Mark
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%
> Mark A. Davenport Ph.D.
> Asst to the Vice Chancellor for Student Affairs/Research and
> Evaluation
> The University of North Carolina at Greensboro
> 149 Mossman Bldg.
> Greensboro, NC 27402-6170
> 336.334.5099
> madavenp@office.uncg.edu
>
> 'An approximate answer to the right problem is worth a good deal more
> than an exact
> answer to an approximate problem' -- J. W. Tukey
>
> >>> Paul Mcgeoghan <Mcgeoghan@CARDIFF.AC.UK> 6/25/2003 8:50:06
AM >>>
> Hi,
>
> I have 2 questions:
>
> 1. What is the opinion on including ordinal predictors in linear
> regression?
> Should one create dummy variables if the independent variable is
> ordinal or
> should one look at using One Way Anova.
>
> 2. Also, if you need to control for a single independent variable in
> a
> multiple
> linear regression analysis, what is the best way of doing it
> (controlling for a
> single continuous variable or an ordinal variable).
>
> It is possible using Partial Correlations to examine the relationship
> between 2
> continuous variables adjusting for a control variable. But can you do
> such a
> thing in multilple linear regression.
>
> Hopefully I am not being too vague, I am a little out of practice in
> statistical methodology.
>
> Thanks in advance.
> Paul
| 
