Date:         Thu, 24 Feb 2005 18:25:38 +0100
Reply-To:     Pere-Joan Ventura <pjventura@gmail.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Pere-Joan Ventura <pjventura@gmail.com>
Subject:      Re: Fwd: Prediction confidence intervals after logistic regression
In-Reply-To:  <1142816741.1109244379@SSW0243.ssw.buffalo.edu>
Content-Type: text/plain; charset=US-ASCII

Thanks to all,

But I'm asking for the SE of Z_i, which it is the linear combination
of the Betas estimated by means of a logistic regression

In my last message I wrote the following formulas:

>Logistic regression model:

>log(p/1-p)_i = bo + b1*x_i

>then I've calulated the index as the linear combination of the betas

>Z_i = bo + b1*x_i

Then I can calculate the probability of the event as:

P(y = 1)_i = exp(b0+b1*x_i) / (1 + exp(b0+b1*x_i)) or

P(y = 1)_i = exp(Z_i) / (1 + exp(Z_i))

Regards

--
Pere-Joan Ventura