Each observers watched same 6 cattle in Uganda

For the first data set, explain that if there is no baseline treatment, there are usually too many pairs of treatments to compare them all pairwise.

If differences between the animals (blocks) are ignored, the anova table finds no evidence of differences between the observers.

Select Correct analysis using blocks from the pop-up menu. We do not explain how the p-value is obtained here — say that computer software will find it for you. With the correct analysis, there is strong evidence of differences between the observers.

The other examples are similar but in the second, there are 3 replicates for each treatment in each block. In the last example, the blocks were intentionally chosen to be different.

Comparing observers who watched grazing cattle in Uganda

Five observers simultaneously watched the same group of 10 cattle for 88 minutes in one afternoon. Each person reported the number of minutes that each animal was observed grazing.

The animals can be treated as blocks — we are mostly interested in differences between the observers in this study.

Amino acid uptake by fish (unavoidable variability between experimental units)

Amino acid uptake varies considerably between fish. The intestines of each fish only give six perparations on which to do an experiment about how NaCN affects uptake of amino acid.

Penicillin production (intentional variability in experimental units)

Four variants of the basic process for manufacture of penicillin (A, B, C and D).

Corn steep liquor, an important raw material for the process, is quite variable. A single blend could have been used, but five blends (blocks) were used to ensure that the results were not specific to one particular blend.