Path: news.mathworks.com!not-for-mail
From: "Daniel " <dan.merfeld@gmail.com>
Newsgroups: comp.soft-sys.matlab
Subject: Re: glm with probit model
Date: Mon, 10 Nov 2008 16:43:02 +0000 (UTC)
Organization: Harvard University
Lines: 7
Message-ID: <gf9oam$jvi$1@fred.mathworks.com>
References: <geapba$62l$1@fred.mathworks.com> <gecis6$hje$1@fred.mathworks.com> <gf6c79$87j$1@fred.mathworks.com> <gf9k8o$e3s$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 1226335382 20466 172.30.248.35 (10 Nov 2008 16:43:02 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Mon, 10 Nov 2008 16:43:02 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 45502
Xref: news.mathworks.com comp.soft-sys.matlab:500004


Thanks very much.  We will try your suggestions and let you know. 

FYI, the 2 comes from the fact that people can lapse on both ends, so we want the cdf to range from lambda on the low end to (1-lambda) on the high end.  For example, if lambda equals 0.01, then the lowest probability for very small values of the independent variable would be 0.01 and the highest probability for very large values of the independent variable would be 0.99.

Hope that makes sense.  Will let you know if this works.