The objective of this study was to estimate cow variability that can be used to determine the optimal sample size for digestibility trials using lactating dairy cows. Experimental design was randomized complete block design having three blocks and three dietary treatments. Three similarly managed nearby intensive farms were considered as blocks, and three diets were formulated to have 0.7, 1.0, and 1.3 neutral detergent fiber (NDF): starch ratio. In each farm, 18 cows were assigned for each dietary treatment and five sample sizes per each treatment group were simulated by simple random sampling of data from 18, 15, 12, 9 and 6 cows respectively. Intake was not affected by diet or sample size (p > 0.05). Estimated cow variability (as standard deviation) for digestibility of dry matter, NDF and starch were 3.8 g/kg, 5.1 g/kg and 3.3 g/kg, respectively. A major implication of this study is that cow variability is greatest for NDF digestibility and the use of a minimum of 12 cows per dietary treatment is adequate to reliably detect treatment effects on the digestibility of NDF, starch and dry matter using cows fed in groups with randomized block design under these experimental conditions.
Keywords: dairy cow; sample size; statistical power; variation.