Sample size calculation

Let us now return to the different options proposed earlier.

Option 4. Stay with the altered treatment design and include four goats fed C, B and A+B. This option would allow the experiment to continue as planned, but with diet A+D rather than A+B.

Comment: This would give the opportunity to determine the increase in milk yield expected from feeding A+B compared with B and C, i.e. giving extra feed during the first 30 days. The researcher explained that this would not really be a useful or interesting result, but we include it here for illustration purposes. To determine whether the sample size of 10, 10, 10 will be adequate we would need to apply the formula introduced earlier, namely n=2s²t²/()²