Problem working nest survival example in Gentle Intro

[zugunruhe1]

Hi everyone! I ran into a problem working the example with mallard data in the nest survival chapter, I keep going back and trying to figure out where I went wrong but I'm stumped. I get all the numbers I should be getting up until the example of fitting Model 3 (DSR varying with Robel score). I get:

LOGIT Link Function Parameters of {Robel}
95% Confidence Interval
Parameter Beta Standard Error Lower Upper
------------------------- -------------- -------------- -------------- --------------
1: 2.9088395 0.1744304 2.5669560 3.2507231
2: 0.0272700 0.0466151 -0.0640957 0.1186357

when the manual indicates the beta values should be 3.012 and 0.108, respectively. I do get the same AIC and deviance values as the example, however. (And the option to "standardize individual covariates" is unchecked, and I use mean individual covariate values for the real parameter estimates).

Thank you for any help.

[Rotella]

There are 2 issues here.

First, the estimated beta's presented on page 17-13 of the book chapter are for a model that relates daily survival to proportion grass (Model 4) not to Robel readings (Model 3). You can see this if you look at the model name that's presented a couple of lines above the estimated beta's you are referring to.

Second, you only get the estimates that are presented on page 17-13 for Model 4 if you use standardized values for proportion grass. The use of standardized values is a hold-over from an earlier version of the chapter; sorry for the confusion. I will try to replace those with estimated beta's that are produced when you use un-standardized covariate values in a future version of the chapter. If you do not standardize the values, you get the following beta estimates for Model 4:
estimate se lcl ucl
S:(Intercept) 2.7517290 0.1402710 2.4767979 3.0266601
S:PpnGrass 0.4097794 0.2109604 -0.0037031 0.8232618

The estimates you obtained are the correct ones for Model 3 for unstandardized values of Robel readings.

[zugunruhe1]

Wow- DUH. I guess I got such tunnel vision from thinking I'd done something wrong I overlooked that, er, minor detail. Thanks a lot!