Monday 20 May 2013

May 24: Shirley Pledger - Using Mixture Models to Model Abundance from Capture-Recapture Studies

Our next speaker is Shirley Pledger, an expert in capture-recapture analysis. She has suggested the following article explaining the use of mixture models to model abundance in the face of differing capture/re-capture probabilities:

http://www.onlinelibrary.wiley.com/doi/10.1002/bimj.200810446/abstract

Feel free to post any questions or comments in advance of the discussion group meeting at 2-3 on the 24th of May.

8 comments:

  1. There seems to be an implicit assumption that not accounting for heterogeneity in capture probability would tend to yield underestimates in population size. The results presented in Table 3 show that the models that account for heterogeneity have higher estimates of population size. The text states that this therefore has better bias correction than the model that includes only a time component. Is it possible that the model accounting for heterogeneity could overestimate the population size to such a degree that it is highly biased in the other direction, and in fact more biased?

    ReplyDelete
  2. Large group of people today, which is great to see!

    ReplyDelete
  3. The discussion starts with the study design for capture-recapture experiments.

    ReplyDelete
  4. Talking about how the heterogeneity is incorporated by the different classes: a class is a latent variable in the mixture model, etc.

    ReplyDelete
  5. Discussing Ian's question on heterogeneity, (see above). Also, talking about CR studies where the population size is known in advance, that way we know the true population and can see the actual bias from our fitted models, etc.

    ReplyDelete
  6. Now discussing Section 4, using similar CR mixture models in the SDM context. Good discussion, some agree, some disagree with this link... Moved on to discussing the style of the paper.

    ReplyDelete
  7. Some brief discussions on Figure 1, AIC, interpretation of classes, ...

    ReplyDelete
  8. Anyone interested in mixture models more generally in ecology, and particularly their use in the analysis of multi-species datasets, might be interested in reading the following article: http://www.esajournals.org/doi/abs/10.1890/12-1322.1

    ReplyDelete