Validation of fish ageing methods should involve bias estimation rather than hypothesis testing: a proposed approach for bomb radiocarbon validations

• Published in: Canadian Journal of Fisheries and Aquatic Sciences
• Main Authors: Francis, R. I.C. Chris, Campana, Steven E., Neil, Helen L.
• Other Authors: Gillanders, Bronwyn
• Format: Article in Journal/Newspaper
• Language: English
• Published: Canadian Science Publishing 2010
• Subjects: Aquatic Science; Ecology, Evolution, Behavior and Systematics; Antarc*; Antarctica
• Online Access: http://dx.doi.org/10.1139/f10-068; http://www.nrcresearchpress.com/doi/full-xml/10.1139/F10-068; http://www.nrcresearchpress.com/doi/pdf/10.1139/F10-068

The need to validate methods of ageing fish is widely accepted and several approaches to validation have been used. Most validations are essentially informal tests, using graphical methods, of the null hypothesis of zero bias in the age estimates. It is argued that it would be more useful to estimate a confidence interval for this bias. This would provide both a quantitative measure of the strength of the validation and a means of formalising the hypothesis test. A method of estimating this confidence interval is proposed for validations based on bomb radiocarbon, and this is illustrated using data for bluenose ( Hyperoglyphe antarctica ) and haddock ( Melanogrammus aeglefinus ).