Use of factor scores in multiple regression analysis for estimation of body weight by certain body measurements in Romanov Lambs

Abstract

The study investigates the solution of the multicollinearity between certain body measurements of Romanov lambs and prediction of the body weight of Romanov lambs using the thus calculated factor analysis scores and a multiple regression model. For this purpose, the body measurements (wither height (WH), croup height (CH), body length (BL), chest depth (CD), chest circumference (CC), chest width behind shoulders (CWS) and head length (HL)) and body weight (BW) of 6-month-old 50 Romanov lambs born in 2015 were used. The factor analysis scores were used to obtain the prediction equation for the relationship between the investigated traits. The analysis results showed that there was a multicollinearity between the wither and croup height traits used in the prediction equation. Moreover, the results revealed that the variables for the body measurements can be represented by two factors. These factors explained 50.89% and 22.86% of the total variance, respectively. The multicollinearity between the independent variables was eliminated with the use of the factor scores obtained with the factor analysis in the multiple regression model, and thus it was observed that better results can be obtained by using the factor analysis scores in the prediction of the body weight of 6-month-old Romanov lambs.