Date:         Tue, 23 Oct 2001 09:21:17 -0400
Reply-To:     "Powhatan J. Wooldridge, Ph.D." <pjw@ACSU.BUFFALO.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Powhatan J. Wooldridge, Ph.D." <pjw@ACSU.BUFFALO.EDU>
Subject:      Re: Betas greater than 1
Comments: To: Hector Maletta <hmaletta@FIBERTEL.COM.AR>
In-Reply-To:  <3BD49414.754CAC8B@fibertel.com.ar>
Content-Type: TEXT/PLAIN; charset=US-ASCII

Hector--

The situation is a bit more confused than your memo suggests. When there
are no control variables involved, beta is numerically identical to the
Pearsonian correlation coefficient, and cannot exceed 1, any more than r
can. It is only PARTIAL betas that can exceed 1, whereas partial
correlation coefficients cannot, of course, exceed 1.

   ---   Pow
***************************************************************************
Powhatan J. Wooldridge, Assoc. Professor, Nursing, State Univ. NY at Buffalo


On Mon, 22 Oct 2001, Hector Maletta wrote:

> Carla,
> betas, the standardized regression coefficients, are regression
> coefficients of an equation dealing with standardized variables, i.e.
> variables with a zero mean and unit standard deviation (SD). A beta=1
> would mean that an increase of one SD in the independent variable
> concerned would be associated with a one SD increase in the dependent
> variable. There is no reason for betas to be necessarily above or below
> one: they can be anywhere.
>
> What caused your confusion, I suspect, is the fact that linear
> correlation coefficients vary between -1 and +1. There is a relationship
> between correlation and regression coefficients, but the latter can
> exceed +/- 1, unlike the former.
>
> Hector Maletta
> Universidad del Salvador
> Buenos Aires, Argentina
>
> Carla Grayson wrote:
> >
> > Sorry for this beginner stats question, but what does it mean if your
> > standardized regression coefficient (Beta) is greater than 1? (I got
> > this in the context of a multiple regression, 2 main effets and their
> > interaction.) I always thought Betas had an absolute maximum of 1, like
> > the Pearson correlation coefficient.
>
>