• No results found

2. CHAPTER 2: LITERATURE REVIEW

4.8 Data used in the study

Using a pre-tested structured questionnaire (Appendix A5.5, Section 5), this study collected data on a set of variables, among others, as presented in the next sub-section.

4.8.1 Description of variables used in the probit model Dependent variable

In this study, the dependent variables are 1 and 0 dummy variables which indicates whether or not a household sourced seed from certified channels respectively. In this regard, the probability of a household sourcing seed from certified channels is explained and estimated by: the sign, the statistical significance, and the magnitude of the parameter of estimates in the probit adoption model.

Independent variables

The household’s decision whether or not to source cassava seed from certified sources is hypothesized to be associated with several independent variables. Accordingly, the study classifies the independent variables into seven (7) categories: (a) decision-making dynamics, (b) farmer perceptions, (c) institutional factors which include information, seed, extension, and credit access; vulnerabilities and shocks; group membership, participation in AIS initiatives, farmer experience and transaction cost variables; (d) household demographics and social networks; (e) land access; (f) wealth status and; (g) regional dynamics. A description of these independent variables is presented in the next section.

Decision-making: Joint decision-making involving more than one household

member (especially inter-spousal decision-making) increases chances of making informed decisions which would meet objectives of all household members.

Farmer perceptions: In respect to trustworthiness of cassava seed source

and satisfaction with seed inspection and certification services, it can be hypothesized that higher levels of trust reduce perception of risk and hence transaction costs in an exchange relationship (Woldie and Nuppenau, 2009). Trust and perception variables are expected to positively influence farmers’ channel choice and are included as binary dummy variables to reflect presence or not of: (a) trust in cassava seed source and (b) satisfaction with cassava seed source inspection and certification services.

Institutional factors: As suggested in Belay et al. (2017), these include:

information access, extension services access, and credit access; vulnerabilities and shocks; group membership, participation in AIS initiatives, farmer experience and transaction cost variables. With regard to household membership to an association, farmer group or Agricultural Innovation Platform (AIP), it may be hypothesized that this increases access to information critical to production and marketing decisions (Olwande and Mathenge, 2012). Since government and donor support targets farmer groups as opposed to individual farmers (usually as a way of increasing their bargaining power at time of output marketing, accessing extension advice and input procurement), membership to a farmer association increases chances of a farmer

accessing seed from government and donors. This is modeled as a binary response variable with 1 if the household is a group or AIP member and 0 if otherwise.

Access to extension services: Mmbando et al. (2014) explain that agricultural extension services are expected to increase access to production and marketing information and technical skills of farmers. Extension services are also expected to facilitate smallholder linkages with input and output markets (Gebremedhin et al., 2009). Therefore, access to extension services may lead to a farmer using clean certified seed usually from government agencies than elsewhere. This is modeled as a dummy with 1 if farmer accessed extension services and 0 if otherwise. Information access is further increased by ownership of ICT equipment such as mobile phones, radio and TVs. Informed farmers are more likely to use certified channels because of the advantages that come with them.

Access to Credit: Modeled as a dummy with 1 if farmer accessed credit and 0 if otherwise; access to credit is understood to increase the farmers’ purchasing power and therefore financial ability to access and use technological innovation.

Transaction costs variables (vehicle or motorcycle ownership and road conditions: Omamo (1998) suggests that farmers will choose closer sources to avoid transportation costs. This is especially so if they lack means of transport and if the road infrastructural conditions are poor.

With regard to farming experience, it is hypothesized that older farmers are more likely to use own saved seed or seed from neighbor (uncertified channels) because they have lost the energy and vibrancy to move extensive distances in search of certified seed (NARO, 2014). Farming experience may positively or negatively influence technology adoption in a sense that risk averse farmers may want to stick to their old proven ways of doing things as opposed to trying out new innovations.

Vulnerabilities and shocks: High input price shocks were included as dummy variables that take on the value of 1 if a household experienced the shock and 0 if otherwise. Shocks may significantly influence a household’s ability to adopt a new agricultural technology especially if they are related to input access.

Household demographics and social networks: Besides its influence on

productivity of family labor and therefore the ease with which improved agricultural practices are adopted. Strauss et al. (1991) assert that the level of formal education attained is used as a proxy for the farmers’ ability to acquire, process and effectively use information gathered from different sources. Thus, the household head’s years of formal education is expected to increase the likelihood of accessing seed from sources that have some inspection and certification services embedded (certified channels). It can also be argued that farmers without formal education are able to innovate if provided with appropriate extension and innovation support through various means including adult education. Furthermore, both family size and dependence ratio have a direct bearing on family labor availability and therefore adoption of technological innovations (Belay et al., 2017). Bigger family sizes and higher dependence ratios may on the contrary constrain resource availability required to access innovations resulting into non adoption and therefore use of uncertified seed sources. Social networks too may negatively or positively influence adoption since they could result into farmers using free uncertified seed from neighbors or certified seed from innovative friendships. Household head’s gender is included as a dummy categorical variable that takes on the value of 1 if household head is female and 0 if male. This household head gender variable is included to control for and explain the cultural institutional limitations imposed on women with regard to free association and technology adoption decision-making (Mishra et al., 2015; Forsythe, 2017).

Wealth status and land access: Farmers with relatively higher wealth will

have a lower degree of risk aversion and will thus more easily adopt new innovations that are more efficient (Alemu et al., 2012).

Regional dynamics: captured as dummy variables taking on the value of 1 if

a household is domiciled in the Mid-western or Northern regions and 0 for eastern region. Regional dynamics are included to assess the influence of geographical location on household’s seed sourcing decision-making process.

In the next sub-section, the study presents descriptive statistics results and discussion of socio-economic and demographic characteristics followed with empirical results and discussion of adoption determinants using the probit model.

4.9 Results and discussions