How to defend the sine link and why use it?

Forum for discussion of general questions related to study design and/or analysis of existing data - software neutral.

How to defend the sine link and why use it?

Postby tgar3881 » Thu Jul 02, 2015 3:58 pm

Forum,

I am currently writing up some of my analyses for a thesis project and I need some good citations or reasoning behind using the sine link.

I used it for known-fate models. I had no covariates and model building was completely binomial. I understand that most survival data comes from a non-normal distribution. Therefore, I would think that an appropriate statement in my methods section would be: "I used the sine-link transformation for models because survival data mostly comes from a non-normal distribution (Need citation here)".

Any thoughts?

Taylor
tgar3881
 
Posts: 9
Joined: Tue Apr 28, 2015 4:37 pm

Re: How to defend the sine link and why use it?

Postby jlaake » Thu Jul 02, 2015 4:09 pm

The link function doesn't have anything to do with the data distribution. It is simply a convenient way to constrain the estimates of survival to be probabilities in this case. The particular shape of the link function can matter if you use numeric covariates but you didn't. The sin link is a good choice when you don't have covariates and aren't using an additive model because with the sin link MARK doesn't typically confuse parameters at boundaries and ones that are not-identifiable so parameter counting is more reliable.
jlaake
 
Posts: 1031
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA


Return to analysis & design questions

Who is online

Users browsing this forum: No registered users and 1 guest

cron