Assessing Length‐Related Bias and the Need for Data Standardization in the Development of Standard Weight Equations

Abstract

AbstractThe recently developed empirical percentile (EmP) method, a technique for deriving standard weight (Ws) equations, putatively reduces the length‐related biases that often plague such equations. To determine whether the EmP method is superior to the regression line–percentile (RLP) method in reducing length‐related biases, we developed new Ws equations by applying both methods to two morphologically distinct species, walleye Sander vitreus and black crappie Pomoxis nigromaculatus. We also investigated diagnostic approaches to provide quality control for weight–length data. We evaluated the new Ws equations with filtered independent data to determine which equation reduced length bias the most. We suggest a protocol for evaluating length‐related bias using an independent data set. Our results showed that for randomly selected walleye populations, the RLP method did not have any length‐related biases when relative weight (Wr) was plotted as a function of length. However, the Wr values calculated from the EmP Ws equations were length biased when the latter were applied to those same populations. Both methods generated Ws equations that were length biased when Wr was plotted as a function of length for black crappies. Further, the absolute difference in Wr between the RLP and EmP methods indicates that there is little difference between the methods as far as their relevance to management is concerned. Based on these results, we believe that revising existing Ws equations using the EmP method is unnecessary and that the RLP technique should remain the standard for developing Ws equations pending the development of an approach that clearly eliminates methodological length bias.