Estimation of Parametric Single Index Ordered Logit Model on Milk Yields

Abstract

This article aims to determine some important factors affecting the milk yield of Holstein Friesian cows and introduce the use of single index ordered logit model in milk yield studies. Considering the three-level ordered structure of the dependent variable of milk yield, a single index ordered logit model is applied to the data set. The data set used in this study is consisted of the 305-day milk records of the Holstein Friesian cows that calved between 2001-2011 years. 14487 records, obtained from 1840 herds belonging to the members of Cattle Breeders Association in Isparta province, Turkey, are analyzed. The direct effects of the variables: parity, lactation length, first year of calving and calving season on milk yield are investigated and comprehensive interpretations are presented. Results show that the cows having longer lactation lengths produce more milk than the cows having middle lactation lengths. The highest amount of milk yield is produced by the cows calving for the first time in their middle ages and in autumn season. The amount of milk obtained after a birth occurring in spring season is higher than the one reached after a birth in summer. Cow being on its 1st parity decreases the milk yield compared to the 6th parity. This result is consistent with the finding in literature in that cows reach their highest amount of milk on their 4th parities.