Maize legume relay intercrops in Malawi - Microbial Ecology in Sustainable Agroecosystems - page 240

9.5.9 Socioeconomic data collection and analysis
Socioeconomic data were collected through farmer interviews and focus groups. For
detailed methods, see the work of Sirrine, Shennan, and Sirrine (2010). Briefly, semifor-
mal farmer interviews were carried out in 2001 to obtain demographic data and farmers'
experience with the legumes and in 2003 to obtain farmers' assessments of the legume sys-
tems. For the latter, we then solicited information on farmers' preferred legume system(s)
and farmers' perceptions of the legume systems' labor requirements, secondary benefits,
impacts on food security, biophysical performance, and variability. Husbands and wives
were interviewed separately for both interviews. From the 34 farm households still partici-
pating, 47 individual farmers were interviewed in 2001 and 51 in 2003, which represented
87% and 94% of the potential study population, respectively.
Adoption surveys were also performed in 2001 and 2003 to collect information on
legume system adoption in farmers' fields. We carried out 34 surveys in 2001, interview-
ing husbands and wives together since they generally farm the same plots of land. In
2003, the adoption survey was combined with the preference survey; thus, we inter-
viewed 51 farmers. Although heads of households were interviewed separately, data
were only reported for each household. The first adoption survey also requested farm-
ers to recall on-farm presence of any of the three legume species prior to the inception
of the project's experimental phase in 1995. We analyzed legume system adoption and
intensity of cropping chronologically using farmer recall for 1994 and actual adoption
results for 2001 and 2003.
In 2004, we held four focus groups that included five to nine farmers each with the
dual purpose of informing farmers of the research results and to obtain further insights
into cropping system impacts on livelihoods. The composition of the focus groups was
reflective of the overall population of farmers participating in the project and included
men, women, FHHs, and a range of poor to wealthier farmers. Following the focus groups,
we also held a training session focused on legume management.
9.6 Case study results and discussion
When averaged across all landscapes for D1, the highest yields were obtained in all treat-
ments receiving inorganic fertilizer (F) ( Table 9.1 ) , with a consistent, but nonsignificant,
trend toward higher average yields when legumes were present, especially S. sesban . In
general, yields were lowest on the hillside and highest in the dambo margin, with simi-
lar patterns of response to fertilizer and legume treatments. Legume treatment differ-
ences were not significant, however, on the hillside, whereas they were in the dambo and
dambo margin ( Figure 9.2 ). In the last cases, the legumes with fertilizer treatment yields
were all significantly higher than the unfertilized control, with the maize with fertilizer
and legume-only treatment yields between the unfertilized control and the fertilized
legumes. D1 and D2 fertilized maize control yields ( Table 9.1 ) were substantially higher
than those found by Harawa et al. (2006), Kamanaga, Waddington, et al. (2010), and Snapp
et al. (2010) from other on-farm research trials in Malawi; but unfertilized maize yields
from D2 were similar. Maize yields were found to increase by between 0.2 and 4 ton ha -1
when legumes were introduced in a similar study (Kamanga, Waddington, et al., 2010),
which is in line with the increases observed in this study. Snapp et al. (2010) also observed
improved fertilizer response with long-lived legumes, such as pigeon pea and Tephrosia,
as was seen in D1 here.