Sample size calculation

By following the same procedure as before this results in the analysis of variance structure:

Source of variation	d.f
Batch	2
Diet	2	(C and B pooled)
Covariate	1
Residual	25
Total	29

The 95% t-value with 25 degrees of freedom = 2.06. Repeating the calculation with t = 2.06 gives n = 9.9. This confirms that 10 goats will be necessary for group A and 10 each for the other two groups. This was the number planned at the outset. Thus, if the researcher were to decide to continue the experiment along the lines of Option 4, the prediction is that the sample might be just large enough to determine a statistically significant effect (P<0.05). 

But as already indicated this would not provide information of biological interest to the researcher. Other options are unlikely also to yield data that would give results of statistical significance. Thus, there might be little value in continuing the experiment.