| Sampling |
The sampling frame for this survey is closely aligned with other CCAFS research activities. CCAFS is supporting research and interventions that address the increasing challenge of global warming and declining food security in five regions – East Africa, West Africa, South Asia, Latin America and Southeast Asia. To better coordinate and learn from the various activities undertaken as part of CCAFS, many research activities are concentrated in selected sites within these five focal regions. Sites were selected through a consultative process between the CCAFS management team, CGIAR researchers already engaged in climate change research, and experts based within each region. Sites were selected based on a set of criteria developed through these consultations, including biophysical and agro-ecological conditions, socio-economic and demographic factors, anticipated temperature and precipitation change, anticipated land use pressure, agricultural potential, and governance and institutional capacity among other factors. The final sites are each 10x10km blocks and are not representative of national or sub-national administrative boundaries. (For details on the site selection process, see Kristjanson et al 2010 ). Within these sites, different research activities were carried out including a CCAFS baseline survey and an IMPACT Lite survey utilizing different sampling approaches. To minimize the burden on respondents and to link the survey with existing data on agricultural production, the survey team for the CCAFS Gender and Climate Change Survey chose to sample the same set of households surveyed by ILRI as part of the IMPACT Lite survey conducted in all CCAFS baseline sites. IMPACT Lite teams used secondary data on each 10x10km site (including satellite images, maps, CCAFS consultations, etc.) to develop a typology of agricultural production systems. Villages were then stratified and selected so that equal numbers of villages would be selected from each type of production system identified within the CCAFS site. The final sample contains 200 households per CCAFS baseline site and 10 households per village. Within each selected village, households were randomly selected from a list of agricultural households (defined by whether they were cultivating land or keeping livestock or aquaculture) developed by the survey teams. For more details on the IMPACT Lite survey see Rufino et al. 2012. IMPACT Lite collected a wide range of data on household production practices, inputs, outputs, labor, and food consumption, at a plot and subplot level. Specifically, the IMPACT Lite survey contained modules on: 1) household composition and livelihood activities, 2) sketch of the farm and seasonality of crops, 3) inventory of livelihood sources, 4) land allocation, 5) farm production activities and inputs, 6) production of main crops and aquaculture 7) residue production from crops and aquaculture, 8) non-ruminant species, 9) ruminant livestock inventory, 10) livestock activities and inputs, 11) livestock feeding, 12) livestock products, 13) other income, 14) other expenses, 15) household consumption of on-farm products, 16) household consumption of off-farm products, and 17) farm and domestic assets. Therefore, rather than collect these household-level modules on production, assets, consumption etc. again, the gender survey was intended to complement and extend the IMPACT Lite database through the addition of the gender-disaggregated modules listed above. Given that the CCAFS Gender and Climate Change Survey re-surveyed households already selected for the IMPACT Lite survey, the sampling strategy depended on the IMPACT Lite sampling strategy. IMPACT Lite sites were chosen to represent major farming systems and agro-ecological zones in the various regions, but the purposeful selection means that the sites cannot be considered nationally or regionally representative. The IMPACT Lite survey aimed for 200 households in each site, stratified to reflect the different farm production systems that exist. Only households engaged in livestock and agricultural activities were interviewed. While IMPACT Lite covered all CCAFS sites. The CCAFS Gender and Climate Change Survey focused on selected CCAFS/IMPACT Lite sites within East and West Africa. In East Africa the sites were Nyando and Wote in Kenya and Rakai in Uganda while in West Africa the site chosen was Kaffrine in Senegal. These sites were selected based on the availability of local partners to conduct the surveys. The survey was implemented by the World Agroforestry Center (ICRAF) in Nyando and Wote; the International Institute of Tropical Agriculture (IITA) implemented the survey in Rakai and the Senegalese Institute for Agricultural Research (Institut Senegalais du Recherches Agricoles, ISRA) implemented the survey in Kaffrine. To better understand gender dynamics, household bargaining, and gender-differentiated adaptation strategies and preferences, the enumerators interviewed multiple household members, usually the principal male and female adult decision-makers (See discussion in Doss 2013). For polygamous households, the survey teams identified additional wives and conducted the interview with them. IFPRI staff field tested and trained enumerators from each country in conducting the survey and entering data |
| Unit of analysis |
Household
Individual
Plot
|
| Universe |
Households engaged in agricultural and livestock activities
|
