7.6.2 Explaining Knowledge Diffusion and Adoption
As shown in Sect. 7.2 a wide range of SCT are available, entailing varying resource
requirements, initial investments, maintenance costs and benefits. In Chap. 5 it was
shown that many of these techniques are being practiced in Yen Chau district, albeit
by relatively few farmers. Due to the diversity of available SCT, the factors
influencing their adoption also varied widely, making an aggregate investigation
inappropriate. We therefore focused our analysis on agroforestry,
7 which is one of
the most widely known SCT in the study area and is also perceived to be one of the
most effective (cf. Table 5.9).
An investigation into the adoption determinants of a technology in a population
where knowledge diffusion is incomplete may lead to biased estimates (Diagne and
Demont 2007). Selection bias arises when farmers who are aware of a technology
differ in their propensity to adopt it from those who are unaware, which may be the
case for at least two reasons. First, knowledge acquisition is part of a farmer’s
adoption decision and, therefore, endogenous, and second, for efficiency reasons
agricultural extension may especially target farmers or communities with a high
innovative capacity. We found that only 43 % of our sample farmers were aware of
agroforestry as an SCT (cf. Table 5.9), necessitating the use of a regression model
that corrects for potential selection (exposure) bias (see Technical Note 1 at the end
of this chapter).
We used a Heckman full-maximum likelihood procedure to jointly estimate the
probability of knowing and adopting a technology, while controlling for selection
bias (Heckman 1979). The model predicts a household’s probability to adopt and
maintain agroforestry techniques on at least one of its plots, conditional on a set of
explanatory variables X 1i and on awareness of agroforestry as an SCT. The latter is
itself a function of a set of explanatory variables X 2i . Households’ observed
knowledge and adoption status (i.e., the dependent variables) is reflected by binary
variables that take on the value of 1 in the case of ‘yes’ and the value of 0 in the case
of ‘no’. Table 7.7 summarizes the explanatory variables contained in X 1i and X 2i .
Following the literature on knowledge acquisition and learning (Feder and Slade
1984; Foster and Rosenzweig 1995; Conley and Udry 2001), we expected information access to be closely linked to education and social capital levels, the possession
of communication assets, access to agricultural extension services and income. The
social capital variable measured how well each household was connected to mass
7 “Agroforestry is a collective name for land-use systems in which woody perennials are deliberately grown on the same piece of land as agricultural crops and/or animals” Lundgren (1982). By
agroforestry, we refer to a cultivation technique consisting of planting trees and/or shrubs on
cultivated land, so as to limit soil erosion and improve soil fertility. The plants mostly used in the
study area are wild tamarind (Leucaena leucocephala), teak (Tectona grandis) trees and pine
(Pinus spp.) trees.
7 Soil Conservation on Sloping Land: Technical Options and Adoption Constraints
261
As shown in Sect. 7.2 a wide range of SCT are available, entailing varying resource
requirements, initial investments, maintenance costs and benefits. In Chap. 5 it was
shown that many of these techniques are being practiced in Yen Chau district, albeit
by relatively few farmers. Due to the diversity of available SCT, the factors
influencing their adoption also varied widely, making an aggregate investigation
inappropriate. We therefore focused our analysis on agroforestry,
7 which is one of
the most widely known SCT in the study area and is also perceived to be one of the
most effective (cf. Table 5.9).
An investigation into the adoption determinants of a technology in a population
where knowledge diffusion is incomplete may lead to biased estimates (Diagne and
Demont 2007). Selection bias arises when farmers who are aware of a technology
differ in their propensity to adopt it from those who are unaware, which may be the
case for at least two reasons. First, knowledge acquisition is part of a farmer’s
adoption decision and, therefore, endogenous, and second, for efficiency reasons
agricultural extension may especially target farmers or communities with a high
innovative capacity. We found that only 43 % of our sample farmers were aware of
agroforestry as an SCT (cf. Table 5.9), necessitating the use of a regression model
that corrects for potential selection (exposure) bias (see Technical Note 1 at the end
of this chapter).
We used a Heckman full-maximum likelihood procedure to jointly estimate the
probability of knowing and adopting a technology, while controlling for selection
bias (Heckman 1979). The model predicts a household’s probability to adopt and
maintain agroforestry techniques on at least one of its plots, conditional on a set of
explanatory variables X 1i and on awareness of agroforestry as an SCT. The latter is
itself a function of a set of explanatory variables X 2i . Households’ observed
knowledge and adoption status (i.e., the dependent variables) is reflected by binary
variables that take on the value of 1 in the case of ‘yes’ and the value of 0 in the case
of ‘no’. Table 7.7 summarizes the explanatory variables contained in X 1i and X 2i .
Following the literature on knowledge acquisition and learning (Feder and Slade
1984; Foster and Rosenzweig 1995; Conley and Udry 2001), we expected information access to be closely linked to education and social capital levels, the possession
of communication assets, access to agricultural extension services and income. The
social capital variable measured how well each household was connected to mass
7 “Agroforestry is a collective name for land-use systems in which woody perennials are deliberately grown on the same piece of land as agricultural crops and/or animals” Lundgren (1982). By
agroforestry, we refer to a cultivation technique consisting of planting trees and/or shrubs on
cultivated land, so as to limit soil erosion and improve soil fertility. The plants mostly used in the
study area are wild tamarind (Leucaena leucocephala), teak (Tectona grandis) trees and pine
(Pinus spp.) trees.
7 Soil Conservation on Sloping Land: Technical Options and Adoption Constraints
261