Date:         Mon, 28 Jan 2008 06:21:50 -0500
Reply-To:     Peter Flom <peterflomconsulting@mindspring.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Flom <peterflomconsulting@MINDSPRING.COM>
Subject:      Re: Analyzing Bimodal ata
Comments: To: Ron Do <rondo@HOTMAIL.COM>
Content-Type: text/plain; charset=UTF-8

Ron Do <rondo@HOTMAIL.COM> wrote
>
>Thanks for this.  Unfortunately I am not familiar with R at all, although I
>am thinking of picking it up since there are some programs that are useful
>for me that are only in R.
>

It takes some getting used to, but it can be very useful


>But I do have some experience using the proc mixed procedure in SAS.  If I
>categorize values around the minor mode as "0" and then the values around
>the major mode as "1" (new variable: modes1) and then fit a mixed model with
>mode1 as a random effect, would this work?

Nope.  Mixture models have nothing in common with mixed models.  A mixed model (as you know) combines fixed and random effects to allow for nonindependence.  A mixture model is a model that assumes the sample (and the population) comes from two (or more) different distributions.


HTH

Peter