On Tue, 10 Oct 2000, =?iso-8859-1?B?wMzB2L/1?= (June Rhee, apparently)
wrote:
> Hi, Donald.
> Thank you for your kind respond.
> My refined question is as follows:
>
> > Ambiguous question. By "beta" do you mean (as some would) a standardized
> > regression coefficient? Or do you mean (as some would, perhaps
> > especially in the context of testing hypotheses) the population value of
> > a raw regression coefficient?
>
> I mean the 'beta the standardized reg coefficient.
>
> > Further, you specify "multiple regression equations", not "simple
> > regression", which implies that each equation has several betas.
> > Did you wish to compare only one of them, between the two equations,
> > or several of them, or the whole vector of betas?
>
> That's right, I am intrested in only one pair of betas; for example
>
> (Eq.1 from the male sample) y = beta(1) * x(1) + beta(2) * x(2) + ....
> (Eq.1 from the female sample) y = beta(1) * x(1) + beta(2) * x(2) + ....
>
> I just want to compare the betas(1) from the equations.
Since the value of beta(1) in each case depends in part on the other
variables in the regression equation (and the structural relationships
between those variables and x(1)), and this dependency may be affected by
the distinction between the two samples, I am not convinced that this is
an appropriate or wise undertaking.
> > Do you really mean "a formula", or are you asking for a procedure that
> > mith be implemented in a statistical computing package?
>
> I will need a formular.
Sorry, can't help you there, without making what appear to me
unreasonable assumptions regarding the other variables.
> But a SPSS syntax for getting the significance level for this
> comparison, if any, will also be greatly appreciated.
Alexander Tsyplakov's advice would be appropriate in generating such
syntax. Essentially, combine the two data sets, use an indicator
variable to distinguish between the two (0/1 for M/F, e.g.), and include
the product of that indicator variable with x(1). The significance level
you seek would be that associated with the regression coefficient for
this product variable.
> Thank you
>
> June Rhee
----------------------------------------------------------------------
Donald F. Burrill [EMAIL PROTECTED]
348 Hyde Hall, Plymouth State College, [EMAIL PROTECTED]
MSC #29, Plymouth, NH 03264 (603) 535-2597
Department of Mathematics, Boston University [EMAIL PROTECTED]
111 Cummington Street, room 261, Boston, MA 02215 (617) 353-5288
184 Nashua Road, Bedford, NH 03110 (603) 471-7128
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
