Evaluation of dairy milk yield in cows through repeated measures analysis

Abstract

The objective of this study was to analyze the milk yield data at 14-day intervals of seven Holstein cows submitted to three treatments with alfalfa [T1: control (confinement in a rest area with corn silage +11.0 kg concentrate); T2: restricted grazing of alfalfa + corn silage+11.0 kg concentrate; T 3 : grazing of alfalfa ad libitum + corn silage + 8.0 kg concentrate] in randomized experiment. Data were analyzed as repeated measures through MIXED procedure, where each cow was a subject. In order to model the error variance-covariance structure of the evaluations within the subjects, the Huynh-Feldt matrix was selected, meeting the sphericity condition, indicating that for analysis of the milk yield data in this experiment, besides the mixed linear model using maximum likelihood resources, the ordinary least squares model can also be used, considering (co)variances in uniform time. Through the use of repeated measures analysis, it was possible to verify significant reduction (p ≤ 0.01) in the milk yield in the three treatments; however, there was no global significant difference (p > 0.05) between treatments or in treatments within each control (p > 0.05).