Lack of independence: An insidious and subtle pitfall in model fitting

Summary: The somatic growth of individuals is crucial in the life history of a species and for assessing populations. This study evaluates seven techniques for estimating growth parameters using mark-recapture tagging data and finds that Bayesian implementations of Fabens and Francis methods are the most reliable in simulated data with errors. When applied to observed data of 14 elasmobranch stocks, only BFa gives biologically plausible results. BFa is a reliable alternative to conventional length-at-age methods, especially in data-limited situations for elasmobranchs.