Findings, implications and lessons learned

• We can also apply the second formula to Scenario 1, (equal probability of sampling) where M_i can be assumed to be the value of any ancillary variable associated with a dip-tank area.

  This could be from census information such as total number of homesteads (including those without livestock), dip-tank area size in hectares or human population, each of which might correlate with T_i.

Thus, if the number of homesteads with livestock in the sampled areas is unknown then one should look for other national census statistics that might be available.