Doubts about Mark

[Natalia]

We are starting with Mark and have some doubts that you may find to be too basic but, after reading the manual we were not able to find out how to proceed so we decided to ask somebody more experienced with this software (may be you can point some paper or another version of the manual...).

We are interested in estimations of nest survival and associated habitat features for 12 grassland birds in Argentinean Pampas, one of our main doubts are:
1- Can we model the entire set of species together? For doing so, do we have to define “species” as a factor? How do we do so? How can we find differences in nest survival among the species?
2- We have data for three reproductive seasons that have defined as three groups (we used Run, Pre-defined model(s), and selected (g)). We are confused about the interpretation of this result because we obtained just one AICc, how can we find differences in nest survival between years? What does this AICc mean? Which is the reference to contrats it against?
3- We supose that there are differences in nest survival along the reproductive season, how can we test this hypothesis? We have already defined B1 date, but when comparing with the B0 model (B0 + B1 vs B0) B0 appears associated with a better fit (less AICc). Does this result mean that the program can not detect differences along the season?

We know these are many doubts but will be very grateful for any help in order to proceed with our analysis
Thank you very much,

Natalia

[jlaake]

Q1- Can we model the entire set of species together? For doing so, do we have to define “species” as a factor? How do we do so? How can we find differences in nest survival among the species?

Answer: Yes you can model all species together. Any factor variable like species is best handled as a group variable in MARK. Differences between species can be measured by differences in the betas or the real parameters.

2- We have data for three reproductive seasons that have defined as three groups (we used Run, Pre-defined model(s), and selected (g)). We are confused about the interpretation of this result because we obtained just one AICc, how can we find differences in nest survival between years? What does this AICc mean? Which is the reference to contrats it against?

Answer: AICc is a model selection criterion for choosing between alternative models. This is a different statistical paradigm from hypothesis testing. You should read the book by Burnham and Anderson or another on model selection. As said above differences in nest survival between species or years/seasons is done with the parameter estimates and not the model selection criterion.

3- We supose that there are differences in nest survival along the reproductive season, how can we test this hypothesis? We have already defined B1 date, but when comparing with the B0 model (B0 + B1 vs B0) B0 appears associated with a better fit (less AICc). Does this result mean that the program can not detect differences along the season?

Answer: I don't understand what you mean by B1 date and B0 versus B0+B1. Are these parameters?


You can use 2 or more varibables to define groups. If you have 3 seasons and 12 species then you'll have 36 groups if each species is represented in each season. When you start to get that many groups MARK can get rather clumsy to use. Some alternatives are a standard statistical package like SAS or R. You have a lot of reading to do before you are ready to analyze your data and MARK may not be the best program for you in this case. You can analyze nest survival data with other programs. As a start, you should look at the following articles:

ROTELLA, J. J., S. J. DINSMORE and T. L. SHAFFER. 2004. Modeling nest–survival data: a comparison of recently developed methods that can be implemented in MARK and SAS. Animal Biodiversity and Conservation 27: 187-205.

HEISEY, D. M., T. L. SHAFFER and G. C. WHITE. 2007. The ABCs of nest survival: theory and application from a biostatistical perspective. Studies in Avian Biology No. 34: 13-33.