Study design

The experiment was planned for execution in three batches to accommodate goats that kidded at different times. Each batch on its own can be considered as a completely randomised design. Together they form a randomised block design with batch taking the role of block. The planned structure for the analysis of variance takes the following form:

Source of variation	d.f
Batch	2
Diet	2
Covariate	1
Residual	24
Total	29

After the completion of the first two batches with 12 goats in the first batch and 6 in the second it was discovered that the animal technician had unfortunately misunderstood the instructions. The six animals that had received 200g Calliandra + 100g maize per day from days 1 to 30 (diet A) were put on diet B, not D, by mistake for the subsequent 30 days.

This meant that these goats had been provided with extra supplementation of Calliandra over the last 30 days than had been planned. Advice needed to be sought on what to do with the next batch of 12 animals in order to obtain the desired information on the three alternative feeding regimes that had been planned.