Data management

Before deciding on how the data should be processed we first need to consider the way that the data are structured. A number of measurements (up to 6 for measurements such as tsetse density, trypanosome prevalence, body weight, etc, or at most one for such events as calving, death etc.) are collected on each animal over a 6-month period. In order to compare means during and before the intervention we need to assemble the results into 6-monthly mean values.

It was decided at the outset that, because of the annual cyclic nature of crop/livestock production systems, a 6-month interval, defined according to seasonal rainfall patterns, was the smallest observational unit that could be used for assessing impact of tsetse control. Thus, the raw data first needed to be reduced to mean values for each animal and then these means summarised over 6-month periods.