182
number of years of residence in the village meant that a farmer who stayed in
the village for more than 1 year was 7.19% more likely to adopt new seed varieties. This could be attributed to strong social networks along with a greater number of years’ farming experience in the village. Simtowe et al. (2012) also
reported that farmers who’d lived in their village for a longer time were more
likely to be exposed to the availability of improved pigeon pea varieties, unlike
their counterparts, because of the social capital in information sharing. Asset
index was used as a proxy for estimating the wealth of the farmers. Farmers
with more assets are likely to have more money, equipment and materials that
will aid easy access to new technologies. The results in Table 15.2 show that a
1% increase in the asset index increases the probability of adopting new varieties by 19.55%. This is in line with Tesfaye et al. (2016), where the authors
reported that asset ownership was positively correlated with the adoption of
improved wheat varieties in rural Ethiopia. Lastly, the significance of the variable ‘noticed change in climate change’ indicates that farmers who had noticed
change were 10.98% more likely to adopt improved varieties. We can argue that
such farmers know about the negative impacts of climate change and would,
therefore, prefer to adopt technologies that will increase production and make
them food secure, unlike their counterparts. Asayehegn et al. (2017) similarly
argue that farmers who were aware of climate change were more willing to
implement climate adaptation measures to mitigate themselves from the
dangers.
15.3.2 Estimating the Impact of Improved Varieties Adoption
Decision
In the second step of PSM, we applied three different matching algorithms: nearest neighbour matching, kernel matching, and radius matching. The PSM model
was used to determine the impact of the different CSA technologies on household
welfare. After matching, ATE was computed. The propensity scores for both
adopters and non-adopters ranged from 0 to 1. The reduced magnitude of
Pseudo-R
2
as well as the statistical insignificance of the p-values associated with
the likelihood test, justified the choice of PSM model for our data. In addition, as
shown in Table 15.3, there was a substantial reduction in bias after matching
which is important in examining balancing powers of estimation. The reduction in
the value and the insignificance of Pseudo-R
2
after matching indicated that there
were no significant differences in the values of the independent variables for the
adopters and non- adopters of stress-tolerant varieties after matching. Likewise,
the p-values of the likelihood ratio test were insignificant after matching. Lastly,
the mean and median bias were all below 20% justifying the choice of PSM model
in this study.
C. M. Mwungu et al.
number of years of residence in the village meant that a farmer who stayed in
the village for more than 1 year was 7.19% more likely to adopt new seed varieties. This could be attributed to strong social networks along with a greater number of years’ farming experience in the village. Simtowe et al. (2012) also
reported that farmers who’d lived in their village for a longer time were more
likely to be exposed to the availability of improved pigeon pea varieties, unlike
their counterparts, because of the social capital in information sharing. Asset
index was used as a proxy for estimating the wealth of the farmers. Farmers
with more assets are likely to have more money, equipment and materials that
will aid easy access to new technologies. The results in Table 15.2 show that a
1% increase in the asset index increases the probability of adopting new varieties by 19.55%. This is in line with Tesfaye et al. (2016), where the authors
reported that asset ownership was positively correlated with the adoption of
improved wheat varieties in rural Ethiopia. Lastly, the significance of the variable ‘noticed change in climate change’ indicates that farmers who had noticed
change were 10.98% more likely to adopt improved varieties. We can argue that
such farmers know about the negative impacts of climate change and would,
therefore, prefer to adopt technologies that will increase production and make
them food secure, unlike their counterparts. Asayehegn et al. (2017) similarly
argue that farmers who were aware of climate change were more willing to
implement climate adaptation measures to mitigate themselves from the
dangers.
15.3.2 Estimating the Impact of Improved Varieties Adoption
Decision
In the second step of PSM, we applied three different matching algorithms: nearest neighbour matching, kernel matching, and radius matching. The PSM model
was used to determine the impact of the different CSA technologies on household
welfare. After matching, ATE was computed. The propensity scores for both
adopters and non-adopters ranged from 0 to 1. The reduced magnitude of
Pseudo-R
2
as well as the statistical insignificance of the p-values associated with
the likelihood test, justified the choice of PSM model for our data. In addition, as
shown in Table 15.3, there was a substantial reduction in bias after matching
which is important in examining balancing powers of estimation. The reduction in
the value and the insignificance of Pseudo-R
2
after matching indicated that there
were no significant differences in the values of the independent variables for the
adopters and non- adopters of stress-tolerant varieties after matching. Likewise,
the p-values of the likelihood ratio test were insignificant after matching. Lastly,
the mean and median bias were all below 20% justifying the choice of PSM model
in this study.
C. M. Mwungu et al.