From: Bruce Southey <southey_at_UIUC.EDU>

Date: Thu, 13 Sep 2007 08:55:23 -0500

Date: Thu, 13 Sep 2007 08:55:23 -0500

Hi,

Typically survival is model in terms of mortality so that death is coded as 1 rather than 0.

*>Is the model right for this analysis? which is right for survival rate
*

*>logit or probit link function?
*

The two links that are most commonly used in survival literature are either logit (odds scale) or complementary log log (hazards scale) and other links for binary data are extremely rare. Provided the probability of mortality is low (less than 0.5) these link functions should be the same. The only reason for one over the other is the whether you want the odds scale or the hazard scale. The variances of the link function to estimate heritability are different but this is irrelevant because the genetic variances should be the same (allowing for numerical error).

Bruce

---- Original message ----

*>Date: Thu, 13 Sep 2007 16:27:00 +0800
*

*>From: luansheng <luansheng_at_GMAIL.COM>
*

*>Subject: Is my model right for survival rate?
*

*>To: ASREML-L_at_AGRIC.NSW.GOV.AU
*

*>
*

*>Dear friends:
*

*>I am the beginner of asreml. I am estimating the variance of survival
*

*>rate which is from the challenge test data of turbot. Turbot is one of
*

*>marine fishies.
*

*>this is my model:
*

*>survival !bin !probit !wt=0 ~ mu !r animal fullsib 0.2
*

*>Is the model right for this analysis? which is right for survival rate
*

*>logit or probit link function?
*

*>The format of data file is just like this:
*

*>animal Sire Dam fullsib survival (binary,0 for death, 1 for survival)
*

*>20070028 2006041 20060159 3 0
*

*>20070029 2006041 20060159 3 0
*

*>20070030 2006041 20060159 3 0
*

*>20070031 2006041 20060159 3 0
*

*>20070032 2006041 20060159 3 0
*

*>20070033 2006041 20060159 3 0
*

*>20070034 2006041 20060159 3 0
*

*>20070035 2006041 20060159 3 1
*

*>20070036 2006041 20060159 3 1
*

*>20070037 2006041 20060159 3 1
*

*>20070038 2006041 20060159 3 1
*

*>20070039 2006041 20060159 3 1
*

*>20070040 2006046 20060165 4 0
*

*>20070041 2006046 20060165 4 0
*

*>20070042 2006046 20060165 4 0
*

*>
*

*>
*

*>
*

*>--
*

*> Yours sincerely,
*

*>
*

*> luan sheng
*

*> ----------------------------------------
*

*> Yellow Sea Fisheries Research Institute
*

*> Chinese Academy of Fisheries Sciences
*

*> Nanjing Road 106
*

*> Qingdao 266071
*

*> China
*

*> luansheng_at_gmail.com
*

*> luansheng_at_ysfri.ac.cn
*

*> ----------------------------------------
*

Received on Mon Sep 13 2007 - 08:55:23 EST

This webpage is part of the ASReml-l discussion list archives 2004-2010. More information on ASReml can be found at the VSN website. This discussion list is now deprecated - please use the VSN forum for discussion on ASReml. (These online archives were generated using the hypermail package.)