Date:         Thu, 6 Dec 2001 20:54:25 -0300
Reply-To:     hmaletta@fibertel.com.ar
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Hector Maletta <hmaletta@FIBERTEL.COM.AR>
Subject:      Re: log-linear regression?
Comments: To: Seumas Rogan <seumasr@MARTYNIA.HRP.ARIZONA.EDU>
Content-Type: text/plain; charset=us-ascii

Log-linear models are used to analyze contingency tables, i.e. the
relationship between categorical variables. The phrase "log linear
regression" is ambiguous. Of course, if you take the logs of all your
(interval-level) variables you may have a linear regression on the log
variables, and you may call that "log linear regression", but that would
even more ambiguous, I'm afraid.

Hector Maletta
Universidad del Salvador
Buenos Aires, Argentina

Seumas Rogan wrote:
>
> Hi everyone,
>
> I am new to statistics. I think I understand the following:
>
> Multiple Linear Regression:
> Continuous Normal Dependent, Continuous & Discrete Independents
>
> Logistic Regression:
> Dichotomous Dependent, Continuous & Discrete Independents
>
> Where, or when would you use Log-linear regression? What is it?
>
> Thanks for any help!
>
> SR
>
> seumasr@hrp.arizona.edu