Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Graphical and Statistical Methods for Determining the Consistency of Age Determinations
Ist Teil von
Transactions of the American Fisheries Society (1900), 1995-01, Vol.124 (1), p.131-138
Ort / Verlag
Taylor & Francis Group
Erscheinungsjahr
1995
Link zum Volltext
Quelle
Taylor & Francis Journals Auto-Holdings Collection
Beschreibungen/Notizen
Many laboratories rely on periodic rereading of reference collections of scales or otoliths to ensure that their age readers remain consistent in their age interpretations, both through time and with other age readers. Measures of both systematic difference (bias) and precision are required for this purpose, because measures of bias are not suitable as measures of precision, and vice versa. Using data from an age comparison study of haddock Melanogrammus aeglefinus for demonstration purposes, we evaluated a variety of graphical and statistical approaches for making paired age comparisons from the standpoint of both detecting age differences and assessing precision. Parametric and nonparametric matched‐pair tests, regression analysis, analysis of variance, and age difference plots were all capable of detecting systematic over‐ or underaging. However, only the age bias plot was sensitive to both linear and nonlinear biases. The coefficient of variation (CV = SD/mean) was a robust measure of precision, whereas the widely used percent agreement statistic was not. The combination of an age bias plot, an age frequency table, and a CV provides a powerful and easily prepared comparison of matched pairs of age determinations.