Combined household and GIS analysis of farmer strategies: an application to feeding practices on smallholder Kenyan dairy farms
MetadataShow full item record
Staal, S.; Wolff, T. de; Baltenweck, I.; Romney, D.; Waithaka, M.; Njoroge, L.; Kruska, R.; Wokabi, A.; Njubi, D.; Thorpe, W. 2000. Combined household and GIS analysis of farmer strategies: an application to feeding practices on smallholder Kenyan dairy farms. Paper presented at the Fifth Seminar on GIS in Developing Countries (GISDECO 2000) - “GIS Tools for Rural Development”, 2-3 November 2000, IRRI, Los Banos, Laguna, Philippines. Nairobi (Kenya): ILRI
Permanent link to cite or share this item: https://hdl.handle.net/10568/2055
Traditional studies of agricultural technology adoption have long been constrained by a limited ability to include spatially-differentiated data. Typically, crude proxies or location dummy variables are used to approximate spatial effects. GIS tools, however, now allow spatially explicit data to be included in household econometric models of technology adoption. This paper describes a study that combined GIS and survey variables to examine the cattle feeding strategies on farms in highland Kenya. Data from a large geo-referenced household survey were combined with GIS-derived variables to comprehensively evaluate the spatial, agro-ecological, market and farm resource factors that determine variability of feeding strategies on smallholder dairy farms. Roads, urban populations, milk collection and processing facilities were digitised, and integrated with spatial coverages of agro-ecology. These were then combined, using econometric methods, to quantify the main spatial and local determinants of the probability of adoption of: a) stall feeding or zero-grazing, and b) planted fodder in the form of Napier grass. The results show the influence not only of agro-ecology, but also of market infrastructure and support services on the adoption of improved feeding strategies. A comparison of predicted uptake using GIS and household variables shows that after first calibrating GIS-derived variables through a household survey, broad but reliable predictions of technology uptake in other areas may be possible.