Study design

The second stage involved purposive sampling in order to choose herds that satisfied the following criteria: a) herds within a 7km radius for a bicycle journey from the centre where the recording assistant operated, b) cattle owners who agreed to have cows ear-tagged and the milk of individual cows recorded.

This type of approach is typical of surveys of this nature and the researcher has to decide on the likely nature of any bias in the selection of herds on the final results. It is important to report the selection criteria precisely when presenting the findings so that a reader can judge for himself/herself on how representative the results are likely to be.

Note that only two herds were monitored from two of the areas where thickets had not been cleared. This might have been a reflection of the nature of the vegetation and the difficulties of raising cattle for milk production or difficulties in accessing different herds. Some of the pastoralists were nomadic and this could also have restricted the numbers of herds that were possible to select within these areas.

Milk offtakes were recorded once a day in the morning each fortnight. A total of 606 cows were recorded during the study, of which 467 were observed from the beginning to the end of their lactation periods. Of the 467 cattle, 388 were in their second or higher lactations and 79 were first parity cows.