Path: news.mathworks.com!not-for-mail
From: "Daniel " <dan.merfeld@gmail.com>
Newsgroups: comp.soft-sys.matlab
Subject: glm with probit model
Date: Wed, 29 Oct 2008 22:50:18 +0000 (UTC)
Organization: Harvard University
Lines: 7
Message-ID: <geapba$62l$1@fred.mathworks.com>
Reply-To: "Daniel " <dan.merfeld@gmail.com>
NNTP-Posting-Host: webapp-05-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1225320618 6229 172.30.248.35 (29 Oct 2008 22:50:18 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Wed, 29 Oct 2008 22:50:18 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 45502
Xref: news.mathworks.com comp.soft-sys.matlab:497901


I have successfully implemented a glmfit with binary response and a probit model.  My model has the form of p(t)=probit(-x^2).  This seems to work fairly well until a lapse occurs and subjects miss a level that they "should" have detected.  For example, maybe they fall asleep for a few moments.  I understand that this can be modeled by including a lapse rate lambda (usually assume to be less than 0.05 or so) yielding a model of the form:

lambda + (1 - lambda)*probit (-x^2)

I haven't been able to figure out how to fit this with glmfit.  Am I missing something?  Ideas?