Survival Analysis help

[Judy Harmon]

Hello,
A quick question.  What's the smallest sample one can use for survival analysis?  I have a small dataset of only 25 individuals that I'd like to run a survival analysis on- those who have completed a counseling program, and using a subsequent arrest following their completion.  Would this be too small dataset to use for this analysis? Thanks!
Judy Harmon

[Poes, Matthew Joseph-2]

http://www.stattools.net/SSizSurvival_Tab.php  does this help by chance?

 

In general I would say, yes its too small given certain assumptions, but depending on the nature of the data and the hazard ratio you must find, maybe not.

 

Matthew J Poes

Research Data Specialist

Center for Prevention Research and Development

University of Illinois

510 Devonshire Dr.

Champaign, IL 61820

Phone: [phone]

email: [hidden email]

 

 

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Judy Harmon
Sent: Friday, March 30, 2012 11:59 AM
To: [hidden email]
Subject: Survival Analysis help

 

Hello,
A quick question.  What's the smallest sample one can use for survival analysis?  I have a small dataset of only 25 individuals that I'd like to run a survival analysis on- those who have completed a counseling program, and using a subsequent arrest following their completion.  Would this be too small dataset to use for this analysis? Thanks!
Judy Harmon

[Rich Ulrich]

What do you mean by "too small?"


With large samples, a survival test will have a bit more
power than a simple 2x2 test, since it uses more information.
If everyone has an "arrest" (in your sample), you have the
information about the time-to-arrest -- And a 2x2 table for
"ever-arrested" would show no difference.   However, if there
few arrests in all, that number of arrests puts the limit on
the expected power to be about the same as for the 2x2
table. 

It is possible to find a difference in two groups with a simple
2x2 test when the N is 10 or less. But you won't have much
power unless you expect an outcome like (0,5), (5,0).

--
Rich Ulrich.

---

Date: Fri, 30 Mar 2012 12:58:47 -0400
From: [hidden email]
Subject: Survival Analysis help
To: [hidden email]

Hello,
A quick question.  What's the smallest sample one can use for survival analysis?  I have a small dataset of only 25 individuals that I'd like to run a survival analysis on- those who have completed a counseling program, and using a subsequent arrest following their completion.  Would this be too small dataset to use for this analysis? Thanks!
Judy Harmon