Data weighting in statistical fisheries stock assessment models

Author:
Publisher website
Google Scholar
Abstract
The conclusions drawn from fisheries stock assessment models can depend strongly on the relative weights assigned to different data sets. However, there is no consensus amongst practitioners as to the best approach to data weighting. From a discussion of some key questions concerning data weighting in stock assessment models, I draw three guiding principles: (i) do not let other data stop the model from fitting abundance data well; (ii) when weighting age or length composition data, allow for correlations; and (iii) do not down-weight abundance data because they may be unrepresentative. I propose an approach to data weighting based on these principles. Two factors that complicate this approach are that some decisions are inevitably subjective (which underlines the need for expert knowledge in stock assessment), and some technical problems are unresolved.

This publication has 31 references indexed in Scilit: