|Date: ||Thu, 11 Apr 2002 09:24:17 +0200|
|Sender: ||"SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>|
|From: ||ABIDI <hassanab@LYON-SUD.UNIV-LYON1.FR>|
|Organization: ||Hospices Civils de Lyon|
|Subject: ||Re: Gender codes|
|Content-Type: ||text/plain; charset=iso-8859-1|
Jeremy Miles a écrit :
> At 15:41 10/04/02, Beckman, Anthony wrote:
> >Greetings All
> >I am looking at some code that was given to me for help in learning SPSS. I
> >see within it a few lines of syntax that transforms gender values from 1=
> >female and 2 = male to -1 = female and 1 = male.
> >Any specific reasons why someone might do this? Does it help in running
> Yes, using -1 and +1 is a type of coding for using categorical variables in
> regression analysis - this is usually referred to as effect coding. Using
> this method, the intercept becomes the mean of the DV.
But if the two groups have the same size !
> Have a look at a book on regression for more info (or do a web search, that
> would probably do the trick).
> Jeremy Miles
> mailto:firstname.lastname@example.org http://www-users.york.ac.uk/~jnvm1/
> Department of Health Sciences, University of York, Heslington, York, YO10 5DD
> *** New phone number: 01904 321375, *** Mobile: 07941 228018
| Hassane ABIDI (PhD) |
| Unite d'Epidemiologie; Centre Hospitalier Lyon-Sud |
| Pavillon 1.M, 69495 Pierre Benite Cedex, France |
| Tel: (33) 04 78 86 56 87 ; Fax: (33) 04 78 86 33 31 |
| E. mail: Hassanab@lyon-sud.univ-lyon1.fr |