The survey targeted two maize-legume based farming systems in the eastern and northern zones of Tanzania. In the eastern zone the survey target two districts; Kilosa and Mvomero. While in the northern zone the study focused on Mbulu and Karatu districts. A combination of purposive and stratified sampling methods were used to select the four districts (Karatu, Mbulu, Mvomero and Kilosa) included in this survey. As the project focus is on maize based farming system, maize production used as an important criterion to select districts, wards and villages. A total of 39 wards were sampled. A multi-stage random sampling procedure was used to arrive at a total sample of 65 villages. In each district, wards selected with probability to proportional size. The number of villages selected with probability to proportional size was 60. In every selected village, probability to proportional size sampling was used to identify the households that were interviewed. A total of 701 households were interviewed.
