Re: A regression question

Article View: bit.listserv.stat-l

Article #1200

Back to bit.listserv.stat-l Previous Next

#1200

From: Dennis Roberts
Date: Mon, 26 Jul 1993 14:16

8 lines
1291 bytes

if the difference between the two subgroups is a matter of intercept but slopes
 are similar ... then using the coding variable as a second independent variabl
e and then a multiple regression ... would tend to produce an overall scatter p
lot that would not show as good of a trend as for each subgroup separately ...
but the coding variable should be a significant weight in the model. Seems like
 if you have evidence that shows that the general areas of the scatterplots are
 different for the two subgroups ... you would be better off treating each
separately

Technical Details

Message-ID: <STAT-L%93072610160847@VM1.MCGILL.CA>

Path: arc1-ca.pugleaf.net!gmd.de!xlink.net!howland.reston.ans.net!paladin.american.edu!auvm!PSUVM.BITNET!DMR

Show Full Headers

Complete Article Headers

X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: f5efb,99c3f6d29272e528
X-Google-Attributes: gidf5efb,public
X-Google-ArrivalTime: 1993-07-26 07:23:44 PST
Path: arc1-ca.pugleaf.net!gmd.de!xlink.net!howland.reston.ans.net!paladin.american.edu!auvm!PSUVM.BITNET!DMR
Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
Newsgroups: bit.listserv.stat-l
Return-Path: <@AUVM.AMERICAN.EDU,@VTBIT.BITNET:owner-stat-l@VM1.MCGILL.CA>
Message-ID: <STAT-L%93072610160847@VM1.MCGILL.CA>
Date: Mon, 26 Jul 1993 10:16:00 EDT
Sender: Stat-l Discussion List <STAT-L@MCGILL1.BITNET>
From: Dennis Roberts <DMR@PSUVM.BITNET>
Subject: Re: A regression question
In-Reply-To: prgst2 AT VMS.CIS.PITT.EDU -- Sun, 25 Jul 1993 23:12:13 GMT
Lines: 7