# st: RE: concitional means in regression

 From "Maarten Buis" To Subject st: RE: concitional means in regression Date Fri, 27 Oct 2006 13:18:48 +0200

```Tomas:
The real reason why the regression line is not equal to the conditional means is that
with regression we usually assume that we that these conditional means lie on a
straight line. In reality this constraint does not have to be true. Also, we can relax
this assumption by adding the variable as a series of dummies. See the example
below:

*----------------begin example--------------------
sysuse auto, clear
recode rep78 1/2=3
tabstat price, stat(mean) by(rep78)
reg price rep78
xi: reg price i.rep78
*------------------end example-------

HTH,
Maarten

-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

http://home.fsw.vu.nl/m.buis/
-----------------------------------------

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of Blau Blau
Sent: vrijdag 27 oktober 2006 13:01
To: statalist@hsphsun2.harvard.edu
Subject: st: concitional means in regression

Hallo,

by definition the regression line should go through the conditional means of Y-variable in each X-variable. But when I compute conditional means of Y in each category of X the regression line does not go through these conditional means.

In statistical literature concerning regression I have found this explanation: we  work with sample and therefore we do not have all observations of Y variable in each X-variable, and this is the reason why we do not observe the correspondence among measured conditional means of Y variable and predicted conditional means given by regression line.

Is this explanation correct and just one possible?

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```