Premium
An Alternative Approach to Detection of Length‐Related Biases in Standard Weight Equations
Author(s) -
Gerow Kenneth G.,
Hubert Wayne A.,
AndersonSprecher Richard C.
Publication year - 2004
Publication title -
north american journal of fisheries management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.587
H-Index - 72
eISSN - 1548-8675
pISSN - 0275-5947
DOI - 10.1577/m03-035.1
Subject(s) - percentile , curvilinear coordinates , statistics , regression , mathematics , line (geometry) , computer science , regression analysis , algorithm , geometry
We propose a new method for assessing length‐related biases in standard weight ( W s ) equations computed by the regression‐line−percentile method. We evaluated the performance of the new method relative to two previous methods for assessing length‐related biases using 15 data sets from which W s equations have been computed. The new method detected potentially serious length‐related biases in 10 W s equations, whereas one of the previously used methods failed to detect any biologically significant biases and the other method detected biases in only one equation. The new method can detect curvilinear relationships between W s and length, so it provides insight that is not available from previous methods.