Rural smallholder agricultural field management surveys across Mozambique

All data were quality assured by the lead author. When in the field, all interviewers were trained before conducting interviews with farmers to ensure consistent robust interview techniques and recording of data. Data were continually checked for errors as it was collected. All interviewers were able to translate directly from the local language to Portuguese, and all data were recorded in Portuguese. All translations from Portuguese to English were done by the lead author. Wherever possible, pre-set answers or binary responses to questions were used to systematically record qualitative observations and avoid ambiguous results.All quantitative estimates of crop yields must be used with caution, as data are based on recall data and unit conversions are uncertain and should be considered as rough estimates of yields. Unit conversions for yield data were based on local measurements of harvest units and weights where possible, or from government data. Other estimates are based on best available literature or rough estimates of weights and volume conversions. In cases where the volume or size of the local unit of harvest was not specified (e.g. sack or bucket), the most common volume or size of the unit reported in household surveys in each village were used.© University of Edinburgh