Date: Fri, 5 Dec 1997 02:20:43 -0500
Reply-To: Zihua lin <linzh@UCLINK4.BERKELEY.EDU>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: Zihua lin <linzh@UCLINK4.BERKELEY.EDU>
Subject: 95%CI for two-part models
Content-Type: text/plain; charset="us-ascii"
We need to calculate 95%CI for the expected value from the two-part models.
Any information and advice is highly appreciated.
The typical two-part models are as follows:
Logit p = a + BX + u (p=0 or 1)
Log Y = c + ZY + w (Y is continuos variable)
The expected value = p * Y
The question is how construct the 95%CI for p*Y.