Comparing models

questions concerning analysis/theory using program MARK

Comparing models

Postby Mardagosa » Thu Sep 23, 2021 11:20 am

Context: I have a database that includes the variables color, sex and body size. As I construct my models it seems easy to build every possible combination (e.g: sex*color*bodysize; sex*color+bodysize; sex:color:bodysize) so I can, at the end, have the best combination for my data. I understand how some of them may overlap in several parameters, but again, it seems an easy way to be sure I am not missing the best combination of variables. Any thoughts on why you would or would not do this approach?
Mardagosa
 
Posts: 10
Joined: Wed Sep 08, 2021 10:28 am

Re: Comparing models

Postby jhines » Fri Sep 24, 2021 11:42 am

If you try enough covariates, even random ones, one or more of them will show a significant effect due to random variation. This is why it is recommended that you only try covariates which correspond to a hypothesis you have. The conclusions will be stronger if the result confirms a pre-conceived hypothesis than just a significant result.

I'm not saying you shouldn't do all combinations it might show effects you did not predict, but the combinations you didn't have a hypothesis for might be more appropriate for the "future research" section.
jhines
 
Posts: 599
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Comparing models

Postby sbonner » Fri Sep 24, 2021 11:45 am

sbonner
 
Posts: 71
Joined: Mon Aug 15, 2005 7:38 pm
Location: University of Western Ontario, Canada

Re: Comparing models

Postby ehileman » Fri Sep 24, 2021 11:46 am

It's not clear to me what is your objective. Are you interested in model prediction or only parameter estimation? The answer to this question will likely dictate the approach you use. Here are a few papers you might find useful related to strategies using multimodel inference.

Arnold, T. W. 2010. Uninformative Parameters and Model Selection Using Akaike's Information Criterion. Journal of Wildlife Management 74:1175-1178.

Doherty, P. F., G. C. White, and K. P. Burnham. 2012. Comparison of model building and selection strategies. Journal of Ornithology 152:S317-S323.

Morin, D. J., C. B. Yackulic, J. E. Diffendorfer, D. B. Lesmeister, C. K. Nielsen, J. Reid, and E. M. Schauber. 2020. Is your ad hoc model selection strategy affecting your multimodel inference? Ecosphere 11:e02997.

Cheers,

Eric
ehileman
 
Posts: 51
Joined: Sat Nov 26, 2011 6:40 pm
Location: West Virginia University


Return to analysis help

Who is online

Users browsing this forum: Google [Bot] and 13 guests