LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (July 2006)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
=========================================================================  
Date:   Sun, 16 Jul 2006 15:44:29 -0400
Reply-To:   "Dogan, Enis" <edogan@air.org>
Sender:   "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:   "Dogan, Enis" <edogan@air.org>
Subject:   Re: regression -- predicted values
Content-Type:   text/plain; charset="us-ascii"

Thank you for the answers One more related question: I wanted adjust my DV for a covariate and plot the means of three different groups on Y after this adjustment. I ran a simple regression with the covariate as the only predictor. I saved the adjusted predicted values (and I understand this adjustment is not the same adjustment I am after) and plotted the mean 'adjusted predicted values' for three groups. Is this an OK procedure? My hesitation is that these predicted values are "predicted" after all and the accuracy of that prediction depends on the nature of the relationship between the covariate and the DV; I don't think one can pretend as of these are observed scores and compare the groups according to their mean adj predicted score. I know I should be checking group by covariate interaction (and I did actually) but I am not sure that addresses my issue.\

Best,

Enis

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] On Behalf Of Hector Maletta Sent: Friday, July 14, 2006 10:42 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: Re: regression -- predicted values

Agree with Stephen. Moreover, even the adjusted predicted values are practically the same as the unadjusted, except in the case of outliers with a disproportionate influence on the estimated coefficients. In many datasets there are no cases with enough influence as to cause the correlation to be less than (nearly) perfect. Hector

-----Mensaje original----- De: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU] En nombre de Statisticsdoc Enviado el: Friday, July 14, 2006 11:12 PM Para: SPSSX-L@LISTSERV.UGA.EDU Asunto: Re: regression -- predicted values

Stephen Brand www.statisticsdoc.com

Enis,

The Unstandardized Predicted Value and the Standardized Predicted Value have a perfect correlation because they are simple linear transformations of one another. Illustrattively, the standardized predicted value involves the subtraction of a constant (the mean predicted value) from each predicted value, and division by a constant (the standard deviation of the predicted values).

The adjusted predicted value is somewhat more complicated. This is the predicted value for a case when it is excluded from the computation of the regression coefficients.

HTH,

Stephen Brand

For personalized and professional consultation in statistics and research design, visit www.statisticsdoc.com

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of Dogan, Enis Sent: Friday, July 14, 2006 3:52 PM To: SPSSX-L@LISTSERV.UGA.EDU Subject: regression -- predicted values

Dear list

I hope this email finds you all well on this late Friday afternoon. Quick question:

I ran a simple regression and saved Unstandardized Predicted Value; Adjusted Predicted Value; and Standardized Predicted Value

What is the difference between Unstandardized Predicted Value and Adjusted Predicted Value?

I get almost identical values on these variables, correlation = 1.00 almost however they are not necessarily 100% equal.

Any ideas?


Back to: Top of message | Previous page | Main SPSSX-L page