s-news
[Top] [All Lists]

Survival analysis - survfit

To: s-news@wubios.wustl.edu
Subject: Survival analysis - survfit
From: Manoel Pacheco <mpacheco@glec.com>
Date: Thu, 10 Jan 2008 13:59:43 -0500
Organization: Great Lakes Environmental Center
User-agent: Thunderbird 2.0.0.9 (Windows/20071031)
I am trying to estimate survival of fishes exposed to multiple concentrations of a contaminant. Only fishes exposed to the highest concentration experienced high mortality. Consequently, most of the records in my file are censored data at the last day of the experiment.

The analysis,

se3.surv <- survfit(Surv(time, cens) ~ treat, data = se3t)

results in an error message: "Problem: No. of rows in Surv(time, cens) (variable 1) is 798 != length of others (364)"

Records on the last day start on row 365. I get no error message when data from the last day is excluded, but then most n.risk values, and survival estimates, are wrong.

I would appreciate suggestions,

thanks

S-Plus 6.2 for Windows (Professional Edition)

<Prev in Thread] Current Thread [Next in Thread>