Hi Pat,
[It's better to start a new thread with a new subject, when you send a
new question.]
Pairwise exclusion might be used sometimes, but it should never be used.
There is a Sage little green book, by Paul Allison, called Missing
Data, which explains why pairwise is never better, and often worse,
than listwise, deletion.
Jeremy
On 29/05/07, Pat Burton <[log in to unmask]> wrote:
>
> Could anyone advise me when if ever pairwise exclusion rather than list wise
> is ever used in multiple regression analysis?
>
> Pat
>
>
>
> -----Original Message-----
> From: Stephen Caswell
> To: [log in to unmask]
> Sent: Sun, 27 May 2007 10.57am
> Subject: SPSS error bars
>
>
>
> Dear All,
> after hours of getting no where I thought I'd ask for advice.
> I am trying to display error bars for results of an intervention.
> Variables are usual care (UC1) and intervention (I1) at baseline and
> follow up (UC2 I2). I can only get them to be displayed in the form UC1
> UC2 I1 I2 but would prefer UC1 I1 then UC2 I2
> i.e baseline results for both conditions followed by follow up for both.
>
> Can anyone please advise?
>
> Many Thanks
> Stephen
>
> Stephen Caswell
> Centre for Public Health Nutrition Research
> Department of Medicine and Therapeutics
> Ninewells Hospital and Medical School
> Dundee
> Tel: 01382 496448
> Fax: 01382 496452
>
>
> ________________________________
> Get a FREE AOL Email account with 2GB of storage. Plus, share and store
> photos and experience exclusively recorded live music Sessions from your
> favourite artists. Click Here for more information.
>
--
Jeremy Miles
Learning statistics blog: www.jeremymiles.co.uk/learningstats
Psychology Research Methods Wiki: www.researchmethodsinpsychology.com
|