ABSTRACT
Weight-length relationships are most often modeled with simple linear or dummy variable regressions. The data required to examine fish weight-length relationships are measurements of the length and weight of individual fish. Length measurements may be total, fork, or standard lengths and weights may be wet or dry, whole or dressed. There are two reasons why it is inappropriate to use linear regression to model the raw weight-length data for most species. First, the relationship between fish weight and length is often nonlinear because most fish add a linear amount of length but a three-dimensional volume of mass. Second, the weights of shorter fish are often less variable than the weights of larger fish. Simple linear regression is used to model the linear relationship between a single quantitative response variable and a single quantitative explanatory variable. A fisheries scientist may use the fitted model to predict the weight of a fish given its length.
