About nest age model in nest survival Chapter18

[Ray]

Dear all,
I have a question about nest age model, in the book of introductory user's guide to mark Chapter18,Page 18-15 to 16, when I built the model of DSR varies with nest age, why the result is DSR varies with the nesting season (for me ,is 45days), not the the nesting days(Kentish Ploveqr is 25days), So how could I get the date of DSR changing wiht nest age, additonal, how to make the graph on the page 18-17?

Thank you very much!

Ray

[Rotella]

Dear Ray,
Each nest in the sample has a different age on each day of the nesting season. The model described in chapter 17 (pages 15-16) of the current version of the book uses the "add" function in MARK to calculate the age of each nest on each day of the nesting season. Those calculated ages are then used as covariates in the model, and the estimated coefficient relating age to DSR is for age not the day of the season.

Perhaps if you contrast this age model (designated as model 6) with the date model (designated as model 5 on the previous page), you can see the difference more clearly. In both models, the design matrix has as many rows as you have dates in your nesting data. But, in model 5, you simply put a continuous string of numbers in the column (from 1 to the number of days in the season). In contrast, for model 6 (age model) you use code to calculate the age of each nest on each date.

I hope this helps.
Jay

[Ray]

Dear Ray,
Each nest in the sample has a different age on each day of the nesting season. The model described in chapter 17 (pages 15-16) of the current version of the book uses the "add" function in MARK to calculate the age of each nest on each day of the nesting season. Those calculated ages are then used as covariates in the model, and the estimated coefficient relating age to DSR is for age not the day of the season.

Perhaps if you contrast this age model (designated as model 6) with the date model (designated as model 5 on the previous page), you can see the difference more clearly. In both models, the design matrix has as many rows as you have dates in your nesting data. But, in model 5, you simply put a continuous string of numbers in the column (from 1 to the number of days in the season). In contrast, for model 6 (age model) you use code to calculate the age of each nest on each date.

I hope this helps.
Jay

Thanks so much, Rotella, I am clear now, the result is the DSR varies according wiht nest age, not just the DSR of the nesting days, I got it~
But in the graph of Chapter page17, how does the dat of DSR age1 、15、30 come from？ I try in individual convariate plot，estimates put into excel ，but I am not sure the result is right. Sorry , this is my first time to use mark, And english is a block for me to understand it. Thanks again~

[Rotella]

Dear Ray,
To obtain the graph on page 17 of the nest survival chapter, you need to (1) record the estimated beta's, (2) decide what the relevant covariate values are for your problem and model, and (4) work outside of MARK with the estimated beta's and covariate values to calculate and then plot the estimates of DSR for various covariate combinations.

For the example on page 17, the 3 beta's are estimated to be: 2.4255 (intercept), 0.0187 (age coefficient), and 0.3689 (proportion grass coefficient). If we want to estimate DSR for a nest that is age 15 days on a site with half of its cover in grass, we would use the following equation to calculate the log-odds of survival: 2.4255 + 0.0187*15 + 0.3689*0.5.

DSR = exp(log-odds of DSR)/(1+exp(log-odds of DSR)), where we are working with the natural logarithm.

In this case the estimated log-odds of DSR = 2.89045
and
estimated DSR = 0.9473723

I often use R to calculate and plot such things. There is an example in the RMark help (look under 'Mallard' in the RMark help file) of how to do these things in R in conjunction with the RMark package.

All the best with your work.
Jay

[Ray]

Dear Rotella,
Thank you very much. You are really patient, through your guidance, I have made out this graph.Your guide is so helpful for those of us beginners. Hope I have more opportunity to get help from you in the future.

Best wish!