Date:         Sat, 11 Jan 2003 02:16:33 GMT
Reply-To:     Jonathan Greenberg <greenberg@UCDAVIS.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Jonathan Greenberg <greenberg@UCDAVIS.EDU>
Organization: EarthLink Inc. -- http://www.EarthLink.net
Subject:      Floating point zero divide error in PROC LIFEREG
Content-Type: text/plain; charset="US-ASCII"

I'm getting a floating point zero divide error using PROC LIFEREG, and it is
being caused by a class variable that has two potential values, 1 or 2.  I
know for a fact that both of them have > 1 observations, so why would this
be choking on that class variable?  If I run the PROC LIFEREG and don't
include the class variable, it runs fine (which is why I know the class
variable is causing the error).  Anyone have a guess?


--j