about group size assumptions

[antom...@gmail.com]

Hi distancers!, reviewing chapter 1 of Buckland et al (2015) in regard to the assumption that in for clustered populations, the size of the detected clusters is recorded without error, and then the book states that: "Using the default method in Distance for estimating mean cluster size, we can relax this assumption so that for clusters on or close to the line or point, cluster size estimates are assumed to be unbiased".  

I would like to be sure that this also applies to the distance package in R.  

Regards!

Antonela

[Eric Rexstad]

Antonela

The phrase you quote from the first chapter of Buckland et al. (2015) is referring to the issue of size bias, whereby the average size of groups detected in a distance sampling survey is a biased estimate of the average size of groups in the population, because small groups at large distances are not detected in the field.  This is explained in our online workshop slides:

https://workshops.distancesampling.org/online-course/lecturepdfs/Ch5/L5-3%20Cluster%20Size.pdf

A regression approach is used in the Distance for Windows software to remove the size bias problem. (See the above PDF slides 7-8).

The ds() function in the R package Distance takes a different approach to this problem.  When group size is included as a covariate in the detection function, the effect of group size is incorporated into the estimation process.  A Horvitz-Thompson estimator is used to estimate both the number of groups and the number of individuals.  The ratio of those two estimates is a derived estimate of average group size in the population.

Have a look at this case study on our Examples page for a more detailed discussion.

http://examples.distancesampling.org/Distance-groupsize/Remedy-size-bias-for-dolphin-surveys.html

Hope that's useful.

--
You received this message because you are subscribed to the Google Groups "distance-sampling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to distance-sampl...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/distance-sampling/a64db2ef-9dca-4545-84f4-412861a9a300n%40googlegroups.com.

-- 
Eric Rexstad
Centre for Ecological and Environmental Modelling
University of St Andrews
St Andrews is a charity registered in Scotland SC013532

[Antonela Marozzi]

Thanks!