More knowledge with the same amount of data: advantage of accounting for parameter correlations in hierarchical meta-analyses

Abstract

MEPS Marine Ecology Progress Series Contact the journal Facebook Twitter RSS Mailing List Subscribe to our mailing list via Mailchimp HomeLatest VolumeAbout the JournalEditorsTheme Sections MEPS 443:29-37 (2011) - DOI: https://doi.org/10.3354/meps09368 More knowledge with the same amount of data: advantage of accounting for parameter correlations in hierarchical meta-analyses Henni Pulkkinen1,3,*, Samu Mäntyniemi1, Sakari Kuikka1, Polina Levontin2 1Fisheries and Environmental Management Group, Department of Environmental Sciences, 00014 University of Helsinki, Finland 2Imperial College London, South Kensington Campus, London SW7 2AZ, UK 3Present address: Finnish Game and Fisheries Research Institute, Rakentajantie 3, 90570 Oulu, Finland *Email: henni.pulkkinen@helsinki.fi ABSTRACT: In fisheries stock assessment, the amount of information available from a specific stock is often limited, and the resources to gather new data can be insufficient. This is especially the case when management advice is required for by-catch species which are not always well monitored. However, information is often available from other stocks of the same or closely related species. Also, potential correlations between the life history parameters may contain information which is not usually taken into account in stock assessments. Utilizing all available information can be a cost-efficient way to diminish the amount of uncertainty about key parameters for a case with limited data or when constructing an informative prior for a new case study. In hierarchical modeling, different hierarchical levels can be used to connect closely related and more distant stocks. For example, a lower level of hierarchy can link the stocks within the same species, and a higher level, stocks of related species. We use length-weight and length-fecundity datasets from the FishBase database. Using these datasets, we demonstrate how Bayesian hierarchical correlation analysis can improve understanding of fecundity and formalize available knowledge about length-weight and length-fecundity relationships in terms of informative priors for future analysis. KEY WORDS: Bayesian hierarchical models · Correlation · Reducing uncertainty · Informative priors Full text in pdf format Supplementary material PreviousNextCite this article as: Pulkkinen H, Mäntyniemi S, Kuikka S, Levontin P (2011) More knowledge with the same amount of data: advantage of accounting for parameter correlations in hierarchical meta-analyses. Mar Ecol Prog Ser 443:29-37. https://doi.org/10.3354/meps09368 Export citation RSS - Facebook - Tweet - linkedIn Cited by Published in MEPS Vol. 443. Online publication date: December 20, 2011 Print ISSN: 0171-8630; Online ISSN: 1616-1599 Copyright © 2011 Inter-Research.