Use of Factor Scores in Multiple Regression Analysis for Estimation of Body Weight by Several Body Measurements in Brown Trouts (Salmo trutta fario)

This study was conducted to estimate body weight from several biometrical traits viz. total and fork lengths, body height, head length, adipose fin width and adipose fin length of Salmo trutta fario trouts. A sample of 140 Salmo trutta fario trouts (70 male & 70 female) was used in this investigation. First, only multiple regression analyses were applied to the data on each gender that caused to multicollinearity problem. In order to eliminate multicollinearity problems, multiple regression analysis after factor analysis was used for each fish gender data. Thus the problems were removed by using factor scores in multiple regression analyses. Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy and Bartlettʹs test of sphericity were used to prove whether factor analysis was appropriate for analysis of both genders. According to these two criteria, the data were fit for application of factor analysis. Three factors with eigenvalues greater than one were selected as independent variables for multiple regression analysis. It was concluded that fish body weight increased when total length, fork length, body height, head length, adipose fin width and length increased. Use of biometrical measurements such as total and fork lengths body height, head length, adipose fin width and length for breeding purposes might provide valuable information on improvement of body weight. © 2010 Friends Science Publishers