Summary

This case study uses data on the numbers of livestock species in homesteads, collected in a livestock breed survey implemented in Swaziland, to estimate population numbers and their standard errors. Whilst the survey covered the whole of Swaziland, the results reported here are just for the Hhohho region used in Case Study 11.

Sampling was done in two stages. Firstly, dip-tank areas were selected for sampling in each sub-region, taking into account such factors as livestock density, proximity to commercial farms and ecological zone; secondly, homesteads were randomly sampled within each sampled dip-tank area.

This case study shows how estimates of the numbers of livestock in the Hhohho region (and their standard errors) can be calculated. This has been done by using different formulae for calculating standard errors, assuming that the selection of dip-tank areas was random.

The two contrasting situations, whereby the numbers of homesteads in each dip-tank area throughout the region (whether sampled or not) are known or not known, are also analysed. This is followed by a discussion on the effect of different methods of sampling, not only on the methods of calculation but also on the reliability of the population estimates that are derived.

Chickens, followed by cattle and goats, in that order, were kept in the largest numbers by smallholder farmers.