GOF Multistrata: Fixed parameters, Median c-hat and U-care.

[rems]

Hi everybody,

I have a series of models which assess if there are differences in mortality in a waterbug species that has the characteristic that males care the eggs carrying them in their back until eclosion. So I want to know if there are differences in mortality between females, parental males and non parental males.
I have a multistrata model with 28 intervals, 2 sexes, and 2 states. The states specify if the waterbugs I study are doing paternal care (1=non paternal care, 2= paternal care). Since females do not care, I constrained parameters: psif phi f2, p f2, and psif 1:2, psif 2:1 to 0 in the model that does not account time variation.

So I have

Real Function Parameters of {phig. sg. psig.}
95% Confidence Interval
Parameter Estimate Standard Error Lower Upper
------------------------- -------------- -------------- -------------- --------------
1:S 1:np 0.8947976 0.0077874 0.8785286 0.9091131
2:S 1:np 0.8958762 0.0135382 0.8661993 0.9195819
3:S 2:p 0.0000000 0.0000000 0.0000000 0.0000000 Fixed
4:S 2:p 0.8906154 0.0202817 0.8440916 0.9244981
5:p 1:np 0.3449712 0.0137864 0.3184727 0.3724696
6:p 1:np 0.2505643 0.0215968 0.2106479 0.2952158
7:p 2:p 0.0000000 0.0000000 0.0000000 0.0000000 Fixed
8:p 2:p 0.2729467 0.0382571 0.2046343 0.3539163
9:Psi 1 to 2 0.0000000 0.0000000 0.0000000 0.0000000 Fixed
10:Psi 1 to 2 0.0365401 0.0096013 0.0217391 0.0607919
11:Psi 2 to 1 0.0000000 0.0000000 0.0000000 0.0000000 Fixed
12:Psi 2 to 1 0.0885077 0.0212866 0.0547249 0.1400555

The problem comes when I try to asess overdispersion parameters , U-care suffers a problem of allocation memory (Windows XP, 2.0 GHz, 1 G RAM). When I try to use median c-hat approach mark crashes.

I ran a multistrata model with only males data, and I did not have problems estimating overdispersion via median c-hat, this model did not have fixed parameters. Is there a bug that does not allow estimation of median c-hat with fixed parameters? Am I doing something wrong?

Thank you very much for your help

Roberto Munguia Steyer

Instituto de Ecologia, A.C
Departamento de Biol. Evol.
Xalapa, Veracruz.
MEXICO

robermunguia@gmail.com
munguiar@posgrado.ecologia.edu.mx

[rems]

I forgot to say that when trying to estimate overdispersion c via median c-hat method, Mark crash report an Error Code 33 [Data Type Error]

Roberto Munguia Steyer

Instituto de Ecologia, A.C
Departamento de Biol. Evol.
Xalapa, Veracruz.
MEXICO