dataset minimum size

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

dataset minimum size

Postby Claudio » Wed Jan 07, 2015 4:11 pm

Hi all,
my question his: there is any rule of thumb to understand if a dataset is so small that try to analyze it is only a waste of time?
E.g., I have a mark-recapture dataset from 44 individuals sampled in 8 years with secondary occasions. I'd like to fit CRD models, but I suspect to run the risk of overparametrization. The same, I got estimates of phi with quite narrow 95% C.I. (less than 0.17). Likely, data structure has its weight on the analysis output.
However, I think it will be useful to have some general, and even coarse, advice on how to assess if a dataset is too small (numeber of individuals / sampling occasions) to be analised.
Thanks in advance.
Claudio
Claudio
 
Posts: 11
Joined: Wed Jan 23, 2008 6:28 pm
Location: Sezze (Italia)

Re: dataset minimum size

Postby claudiapenaloza » Mon Mar 09, 2015 9:58 am

I'm no expert on this, but no one else has stuck his/her head out yet so, here goes...

I vaguely remember reading about a 5-year minimum for multistate models (Schaub et al. 2004, maybe?), so you'd have that covered (a CRD being a special case of an MS model)... and other than having relatively "few" individuals overall (though I guess this would ultimately depend on the species you're studying), I would suggest you determine how many recaptures you have.

P. Doherty once told me something like, ">10% go for it, <1% increase your recaptures or you'll be stuck with a constant CJS at best".

So if you have a lot of recaptures, model on!

Hope this helps,
cp
claudiapenaloza
 
Posts: 63
Joined: Tue Sep 23, 2008 9:48 pm
Location: York, SC


Return to analysis & design questions

Who is online

Users browsing this forum: No registered users and 14 guests