Exploration and description / Summaries

Please note that these are preliminary estimates. At this early stage, however, we can say that chickens occur, as one might expect, in the greatest numbers, followed by cattle and goats. Other species exist in much smaller numbers.

We have, however, made various assumptions.

For example, we have assumed for the purpose of this case study that dip-tank areas were selected completely at random and that there was no stratification. Correct population estimates will need to be obtained taking into consideration the precise structure of the sampling design.

There are also a couple of somewhat large standard errors appearing for chickens (shaded yellow on previous page). The source of this problem needs to be investigated by studying extreme values in the raw data as we did for sheep.

The standard errors for cattle and goats are 5-7% of the mean. This is reasonable. However, for sheep, pigs and donkeys, they are much higher. This is understandable because of the large numbers of homesteads without these species.

Finally, note how the results in the second table have been presented. Numbers are reported with 3 significant figures. These make the results very much easier to follow than those presented in the first table. This is an important point for reporting.