RD w Markovian emigration. Most S and gammas not estimable?

Forum for discussion of general questions related to study design and/or analysis of existing data - software neutral.

RD w Markovian emigration. Most S and gammas not estimable?

Postby Dani_Silva » Sat Feb 27, 2016 10:58 am

Hi
I ran a RD under different emigration models and the Markovian model with S(t), gamma"(t) and gamma'(t) (with the constrains g"k = g"k-1 and g'k = g'k-1) was the best fitted model. However, most S and gammas parameters either have very high estimates and low SE or very low estimates and low SE. I have 8 primary periods with 27 secondary periods and low recaptures (1 or 2) in 6 of those primary periods. I wonder if the low recaptures are causing this parameters to not be estimated properly?
I cannot figure out why that is or how to interpret those. :?
Any insight is greatly appreciated!

Thanks in advance

Dani

Real Function Parameters of {Markovian S(t), g"(t), g'(t), c=p(t)}
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 0.5926217 0.5800465 0.0129408 0.9938429
4:S 0.7642203 0.7451460 0.9776270E-003 0.9999069
5:S 0.9999998 0.1100125E-003 0.6089753E-297 1.0000000
6:S 0.9999994 0.2055220E-003 0.3053419E-287 1.0000000
7:S 0.6754896 0.5168165 0.0200755 0.9952941
8:Gamma'' 0.1401739E-011 0.0000000 0.1401739E-011 0.1401739E-011
9:Gamma'' 1.0000000 0.0000000 1.0000000 1.0000000
10:Gamma'' 1.0000000 0.0000000 1.0000000 1.0000000
11:Gamma'' 1.0000000 0.0000000 1.0000000 1.0000000
12:Gamma'' 0.4545057 0.2038633 0.1425692 0.8067687
13:Gamma'' 0.8028686 0.1239202 0.4674737 0.9497377
14:Gamma' 0.5746924E-006 0.0000000 0.5746924E-006 0.5746924E-006
15:Gamma' 1.0000000 0.1758454E-005 0.9999966 1.0000034
16:Gamma' 0.3016692 0.1927802 0.0670467 0.7219674
17:Gamma' 0.6415582E-010 0.0000000 0.6415582E-010 0.6415582E-010
18:Gamma' 0.4646890E-004 0.0147712 0.1192507E-274 1.0000000
Dani_Silva
 
Posts: 2
Joined: Fri Jan 16, 2015 11:47 am

Re: RD w Markovian emigration. Most S and gammas not estimab

Postby cooch » Fri Apr 27, 2018 10:33 pm

Its undoubtedly 'poor data' causing your problem -- in this instance, either low p, or the likelihood that there is little to no temporary emigration. Try fitting a p(dot) model, and perhaps a model with 'no movement' ' (see Chapter 15).

If you have sparse data, fitting a Markovian movel is always going to be problematic, unless you have informative covariates. But, low p is what it is -- a violation of the 'big law' (being, do whatever you can with your sampling to increase p).

Good luck.
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University


Return to analysis & design questions

Who is online

Users browsing this forum: No registered users and 14 guests