Date:         Fri, 28 Apr 2006 21:15:03 -0400
Reply-To:     Richard Ristow <wrristow@mindspring.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Richard Ristow <wrristow@mindspring.com>
Subject:      Re: R in Linear Regression
Comments: To: "Marc Halbrügge" <marc.halbruegge@GMX.DE>
In-Reply-To:  <307.1146263046@www083.gmx.net>
Content-Type: text/plain; charset=iso-8859-1; format=flowed;
              x-avg-checked=avg-ok-5816356E

At 06:24 PM 4/28/2006, Marc Halbrügge wrote:

>>Also, note that R (usually noted by an
>>uppercase r) is the correlation  coefficient,
>>has a range of possible values from -1 to +1
>>(indicating from > perfect inverse correlation
>>to perfect direct correlation).
>
>In linear regression, the correlation between
>observed and predicted values cannot become
>negative. This follows from minimizing the
>squared differences between observed and predicted values.

You're right; it can't. But the correlation
between observed, or predicted, values with
predictOR variables (IVs) can very readily be
negative; and that's what's being discussed.