Comparison of length-weight relationships in fishes using dummy variable(s)

Abstract

At present length-weight relationships in fishes of the form w = alb are fitted by logarithmic transformation of the non-linear relationships into linear form and then using the method of least squares to estimate the constants. These relationships are tested for their homogeneity by analysis of covariance technique. In the present paper, regression analysis is used to fit multiple linear regression equation of log weight on log length by including sex as dummy variable. This equation is tested for its significance by analysis of variance technique, various coefficients are tested by t-test, separate relationships are derived and growth conditions tested for isometric growth. The method is illustrated for 2 sexes and the results are compared with the results of analysis of covariance. The technique of deriving multiple linear regression equations for more than 2 groups is also explained