Smallholder Rice Farmers in Ghana’s Kpong and Weta Irrigation Projects

The sample design for the surveys conducted with primary cultivators of the 2013 Smallholder Rice Farmers in Ghana's Kpong and Weta Irrigation Project involved the following four main steps:

1. The Ghana Irrigation Development Authority (GIDA) provided a list of plots within the Kpong and Weta Irrigation Projects. This list included 3,065 plots in Kpong and 1,087 plots in Weta, as well as the name of the leaseholder for each plot.

2. The study met with the Kpong Irrigation Project (KIP) and Weta Irrigation Project (WIP) block leaders. With the help of the block leaders, plots that were vacant, waterlogged, had no owner, or were owned by a private firm or institutions were removed from the list. Plots that were de facto subdivided for use by multiple cultivators were counted as multiple plots. And names of the primary cultivators (PC)s for each plot were secured. After this process the resulting number of plots and PCs in the KIP was 2,606 plots with 1,875 PCs and in WIP was 1,093 plots with 993 PCs. In total, 31% of PCs were female; 78% were plot owners; 16% were renters; and 6% were both owners and renters.

3. The study decided to enroll in the sample 1,600 of these PCs. All female PCs were drawn into the sample, to achieve as close to a balanced sample as possible by gender. And a random subset of male cultivators were drawn into the sample, stratifying on number of plots (one plot or more than one plot), tenure type (owner, renter, or owner and renter), and irrigation site (Kpong or Weta). For Kpong only, the study also stratified on plot section.

4. The selected individuals were screened for their willingness to participate in the study and interest in participating in the Global Agricultural Development Company's (GADCO) Copa Connect Smallholder Program (CCSP). The study sample was then formed only of those individuals who were interested in participating in CCSP and who completed the baseline survey. Those who did not meet these criteria were replaced, with replacements randomly drawn from within the same strata as the replaced farmer, where possible. For cases where the sample frame contained no more cultivators within that strata (primarily in the case of female cultivators), a cultivator was randomly drawn from all remaining primary cultivators. The target number was exceeded by 16, such that a total of 1,616 PCs were enrolled in the study sample.

Two additional survey modules were conducted:

A) For the Input Seller Survey, all individual persons or shop establishments that sell chemical inputs within the study communities, including herbicides, insecticides, fungicides and fertilizer for rice production, were included.

B) For the Labor Wage Survey 14 focus groups were created: one in each of the 14 community groupings. Each focus group was comprised of between 7 and 10 individuals. Focus groups were formed from a listing of agricultural laborers, which gave particular attention to gender, community of residence, the agricultural tasks the laborer specialized in, and whether the laborer has been involved in this activity since the 2012 major season, to ensure a fair representation of gender, labor, task and communities. The listing was done with the assistance of a popular person, block leader, or family head who directed the survey enumerators to places where laborers could be found.