Different coding systems for the modeling of lactation milk yields of Awassi sheep

Abstract

This study evaluated the feasibility of using different coding systems for categorical variables when using continuous and categorical variables together for the modeling of the lactation milk yield of Awassi sheep. In the study, when all variables were included in the model, and Dummy Coding and Effect Coding methods were used for age, the effects of lactation duration, average daily milk yield, type of birth, and age 5 group were found to be statistically significant in addition to the constant term. When the Deviation Coding method was used for age, the effects of lactation duration and average daily milk yield were found to be statistically significant in addition to the constant term. On the other hand, when Forward and Backward Coding methods were used, the effect of the age 5 group was found to be statistically significant, along with the effects of lactation duration and average daily milk yield. The results of the study indicated that different results can be obtained depending on the various coding systems used. The results also indicated that the choice of coding system affected the interpretation of the obtained coefficients. Therefore, it can be stated that the aims of the researcher in the study should be defined clearly and the proper codding system should be selected according to the variables to be included in the model.