Using Logistic Regression to Characterise Communal Cattle Farmers in Botswana

Abstract

This paper aims to characterise and profile subsistence cattle farmers according to selected attributes to establish their influence on the type of cattle breed kept by farmers in Botswana. The development of communal cattle production can be a sustainable way to improve the livelihoods of the rural population in Botswana. However, there needs to be more information or research conducted to characterise and profile communal cattle farmers with a precondition that the farmers can keep any of the three breeds (Tswana, Cross, or Exotic). A logistic regression model was fitted to determine the influence of 11 predictor variables on the type of cattle breed kept by the farmers. Results revealed that female-headed households were 50% more likely to have the Tswana breed of cattle than male-headed households. In contrast, female-headed households were 30% less likely to have cross-breed or exotic-breed cattle than male-headed households. Results further show that resource-poor farmers tend to keep Tswana breed cattle. These are holdings with no farm labour, no other economic activities, female-headed households, and their primary source of income specified as “other”.