HELP! - Quadratic term for Nest Age

[MartinColombo]

Hello everyone!

I've been following manual's instruction to run nest survival models, and I've been successful except for a model that incorporates Nest Age and quadratic nest age.
As the book says, I added a column in the design matirx with "add(AgeDay1,0) and so on... and a third column "power(add(AgeDay1,0),2)". The nesting season for my study is 130 days.
The model runs, but it had an extreme AIC value, all the other models 17 points below,which made me suspect something.
I looked the real estimates in notepad and I got this:

95% Confidence Interval
Parameter Estimate Standard Error Lower Upper
-------------------------- -------------- -------------- -------------- --------------
1:S 1.0000000 0.0000000 1.0000000 1.0000000
2:S 1.0000000 0.0000000 1.0000000 1.0000000
3:S 1.0000000 0.0000000 1.0000000 1.0000000
4:S 1.0000000 0.0000000 1.0000000 1.0000000
5:S 1.0000000 0.0000000 1.0000000 1.0000000
6:S 1.0000000 0.0000000 1.0000000 1.0000000

Then there are some lower values in the middle (from 29:S to 95:S) and then repeats the same.
PIM chart is set as "all different", "standradized covariate values are unchecked" and so far I followed all the instructions as said in the book.

My concern is that I'm missing something in the formula or in the data input (it worked well with only linear effects of nest age nonetheless).

If anyone is familiar with the quadratic trend and knows if that is normal, I will really welcome the help!

Thank you in advance!

[Rotella]

The real parameter estimates will be for a specific set of covariate values. If you chose to produce them for the 1st nest in your input file when you setup the model run, they will be just for that nest, and that nest was likely only being studied on a subset of all the days in your study. I would instead focus on the estimated beta's for the model to see if those appear to have been estimated well or not.