Statistical modelling

To enable us to explore this problem further the following model:

y_ijk = µ+ b_i+d_j+ ßx_ijk+ e_ijk

was fitted to average milk yield over 1 – 30 days and 1 – 60 days, respectively, stored in CS8Data2, where y_ijk is the average milk yield for goat ijk, µ is the grand mean, bi is the effect of batch i (i = 1,2), d_j is the effect of diet j (j = C, A, B for the first 30 days; j = C, A + B, B for the 60 days), ß is the regression coefficient on pre-treatment milk yield xijk and e_ijk is the residual term.

An analysis of covariance on the average milk yield for the 18 animals from the first two batches can be obtained using the dialogue box below produced via Stats → Analysis of Variance… and changing Design from General Analysis of Variance to One-way ANOVA (in Randomized Blocks).