Print

Print


Dear Sivaram, list,

This is a recurrent question posted to the list. The fundamental point 
is to understand why is it that the HN model produces a significant GOF 
statistic. I would say, based on what you are describing, that you have 
many small distances (maybe even too many 0 distances) that lead to a 
spiked detection function. What is funny is that in such cases quite 
often the AIC favours strongly the NE, which does not seem to be the 
case for your data.

I would say that blindly using AIC or GOF measures is never the best 
approach. One needs to understand the data, and model accordingly. If 
say you have a measurement error problem, with too many distances 
recorded as 0, a model like the HN will likely provide you better 
density estimates than the more than likely overestimation given the NE 
is used.

Of course, even better for next time, would be to identify what might be 
the cause of the problem and solve it by changing survey/field methods.

All the best,

Tiago

On 28/05/2012 07:42, sivaram, kfri wrote:
> I have two competing detection functions in estimating density estimates
> from line transect data with same truncation level and class intervals.
> While both of them provide very similar AIC values, chi-sqaure value is
> significant for one of the models. The model with non-significant
> chi-square value is Negative exponential plus simple polynomial. However,
> this gives unacceptably high density estimates. The model with significant
> chi-square value (ie. Halfnormal with Simple Polynomial) gives reasonable
> estimate. Should I go strictly with model with lowest AIC and
> non-significant chi-sqaure value?. Any one can advice me.
>
> Sivaram
>
> --
> distance-sampling list home page:
>    http://www.jiscmail.ac.uk/lists/distance-sampling.html
> To unsubscribe, go to the home page and click "Join or leave the list",
>    or email [log in to unmask] with the following message body:
>    leave distance-sampling
> Before posting to the list, read the list guidelines at
>    http://www.jiscmail.ac.uk/files/distance-sampling/
>      distance-sampling.welcome
> Any questions, email the list owner at
>    [log in to unmask]
> --
>
>

-- 
____________________________________________________________________________________________________ 

Tiago André Marques
Research Unit for Wildlife Population Assessment
Centre for Research into Ecological and Environmental Modelling
University of St Andrews
The Observatory
Buchanan Gardens
St Andrews
Fife KY16 9LZ
Scotland

Tel: 00441334461842 Fax: 00441334461800 (Scotland office)
Tel: 00351210198736 (Portugal home)

http://www-maths.mcs.st-andrews.ac.uk/homepages/tam2.html
http://www.creem.st-and.ac.uk/tiago/

The University of St Andrews is a charity registered in Scotland : No 
SC013532

--
distance-sampling list home page:
  http://www.jiscmail.ac.uk/lists/distance-sampling.html
To unsubscribe, go to the home page and click "Join or leave the list",
  or email [log in to unmask] with the following message body:
  leave distance-sampling
Before posting to the list, read the list guidelines at
  http://www.jiscmail.ac.uk/files/distance-sampling/
    distance-sampling.welcome
Any questions, email the list owner at
  [log in to unmask]
--