Determinant of Adoption of Improved Varieties of Sorghum in Center-north and Boucle du Mouhoun in Burkina Faso

Sorghum is the first food cereal in Burkina Faso. Its average yield is of the order of 850 kg/ha (Barro-Kondombo, 2010). Sorghum is listed as a priority plant in research and food security strategies. Local varieties remain dominant in the traditional farming system with a predominance of the botanical race Guinea (Barro-Kondombo et al., 2008). This study was conducted on 300 farmers in center-north and the Boucle du Mouhoun in Burkina Faso. The objective of this paper is to determine the socioeconomic factors influencing the adoption of improved varieties of sorghum in the aforementioned regions. An econometric model Probit was used. The results show that the area, training on the improved varieties, membership of a farmer organization, land area planted with sorghum, access to credit and the availability of improved varieties positively affect the probability of adopting improved varieties of sorghum. In contrast, the age of the producer negatively affects the probability of adoption of improved varieties of sorghum.

used an econometric modulation with Probit model as analysis tool because it has been judged as appropriate for the certification of relations between a probability of adoption and its determinants (Ntsama et Kamgnia, 2012;Barry, 2016). In fact, it is assumed that the adoption of an improved variety of sorghum by randomly selected farmers' follows a normal distribution.

Methodology
The methodology comprises the study area, data collection and analysis.

Study area
The study area concerns the regions of center-north and Boucle du Mouhoun in Burkina Faso. These regions was selected because they present two contrasting situations as regard to their climate. These two regions are located in two distinct agro-ecologies mainly differentiated by their rainfall. In fact, the region of center-north is located in the Sahelian eco-climate (less than 600 mm rainfall per year) and the region of Boucle du Mouhoun is in the Soudano-sahelian zone between the isohyets 600-900 mm rainfall per year. It is important to point out that these two regions had density of 13 and 21 inhabitant/km 2 in 2011 for center-north and Boucle du Mouhoun respectively (INSD, 2011). This is lower than the national average which is 38 inhabitants/km 2

Econometric modeling
The economic theory underlying the econometric model on the adoption is the maximization of utility. The theory of the maximization of utility is often used to explain a producer's response to a new technology. The literature review of adoption of innovations allowed to distinguish three types of models frequently used to analyze the decision of adoption new agricultural technology: models of linear probability, Logit and Probit. Logit and Probit models are often used in most adoption studies. In this study we used an econometric modulation with Probit model as analysis tool because it has been judged as appropriate for the certification of relations between a probability of adoption and its determinants (Ntsama et Kamgnia, 2012;Barry, 2016). In fact, it is assumed that the adoption of an improved variety of sorghum by randomly selected farmers' follows a normal distribution. Usually, to determine households' determinants of choice for adopting agricultural innovations, the economic theory uses adoption models. The goal of an adoption model is to explain a variability of an Y vector, representing the adoption of good practices, by X variable vector that represents explicative variables (such as the characteristics of the adopter, the innovation and the institutional environment). Formally, the model can therefore be written: = + where β is a vector of coefficients to be determined, ε an unknown vector which effect is imputable to variables that are not observed. For a given producer i we have: = 1 if producer adoptes improve varieties of sorghum = 0 if not The decision of this producer i to adopt seeds of an improved variety of sorghum is assumed to be based on the maximization of the utility of these improved varieties. Several models allow to perform regression of dependent variables for dichotomy variables (1, 0) the most expended being the Logit and Probit models. The difference between these two models lies in the specification of the distribution function.
The probability P such as: = ! ( =1) = ! (" 1>" 0) = ! ( 1+ 1> 0+ 0) =#; when the agricultural innovation is adopted is therefore function of independent variables and error term. In this equation, F represents the distribution function. If the distribution function is logistic, it refers to Logit model. If it is normal centered reduced, it is a Probit model (Bourbonnais, 2015). The Probit model fits to our dependent variable which is dichotomous. In addition, in the case of probit model, the dependent variable is qualitative. The distribution function will be therefore written: &67. In this expression, Xi is a characteristic vector of the adopter i, the seeds of the improved sorghum and the institutional environment.

Data collection and analysis
Three hundred farmers were interviewed by two investigators; 150 farmers being interviewed by each investigator per region. Surveyed districts are: Pissila in center-north and Dedougou and Bourasso in the Boucle du Mouhoum.
Overall six villages were surveyed in the three districts. The villages from Pissila are: Nioko, Noaka and Poulallé. The villages in Bourasso are: Barakuy and Lekuy and the village from the district of Dedougou is Wétina. Fifty producers were interviewed per village making a total of 300 producers for the six villages. A reasoned choice of producers was made emphasizing on sorghum producers. Data were typed in SPSS software and analyzed with SPSS and STATA.

Results and discussion
Results of socio-economic characteristics and econometric modulation 4.1. Socio-economic characteristics of the interviewed farmers Table 2 shows the socio-economic characteristics of the interviewed farmers. The results show that the variables agro-ecological zone, membership of farmers' organization, training, access to credit and availability of improved varieties statistical influence the adoption of improved varieties at 1% significance level. Education level influences the adoption of improved varieties at 5% significance level. Average age of the interviewed farmers is 42.99 year while the average age of those who do not adopt is 41.11 and that of those who are adopting is 43.897 year. The age influences the adoption of improved varieties of sorghum at 10% significance level (Table 2).

Results of the econometric modeling
The results of the econometric modeling show that some variables positively influence adoption of improved varieties while some negatively influence it. Some variables are not significant meaning that they do not influence the adoption of improved varieties of sorghum in both regions. Statistically significant variables are: the agroecological zone, the age of the farmer, the training received on improved varieties, membership to farmers' organization, land area planted with sorghum, access to credit, availability of improved varieties. All these variables positively influence except farmer's age which negatively influence the adoption of improved varieties of sorghum. Agro-ecological zone: the agro-ecological zone in center north positively and significantly influence the adoption of improved varieties of sorghum at 1%. This means that being in an area with low rainfall leads producers to protect themselves from the effects of climate change by adopting more . It also appears that the region of Boucle du Mouhoun, the second study area, benefited from favorable natural resources in the large majority of its territory such as its climate (rainfall and temperature), the presence of the river Mouhoun and rich enough soils for agricultural production. These favorable natural factors in terms of physical geography have contributed to make this region the food store in Burkina Faso in terms of cereals production (sorghum, pearl millet, maize). The importance of cereal crops in the study area prevailed in selecting it to conduct research on improved varieties of sorghum and pearl millet for several years. In fact, the second study area is not out of uncertainty in terms of irregularity in agricultural production caused by climate change. The geographic area comprising the two study zones, the region of Boucle du Mouhoun, mainly the provinces of Mouhoun, Kossi and Banwa et the province of Sanmatenga in the region of center north, represents in fact two agro-ecological zones differentiated by their rainfall and their production system, whereby their selection and implication in projects of participative selection and/or production and diffusion of improved varieties of sorghum for several years (Trouche G. et al., 2016). Membership of farmers' organization: this variable positively and significantly influences the adoption of improved varieties of sorghum at 10%. This implies that a producer, member of farmer organization is apt to adopt improved varieties of sorghum than the one who does not belong to an organization. Marginal effects show that the fact of being member of a farmer organization increases the probability of adoption of improved varieties of sorghum for 0.16 points, all other things being equal. These results are in agreement with those found by other authors (Abdoulaye et al., 2014;Ndjadi. et al., 2017).
Two farmers' organizations UGCPA and AMSK are involved in a relatively long space-time innovation started in the 2002's in participative breeding, production and diffusion projects of improved seeds of sorghum (Trouche G. et al., 2016). From this long experience of partnership, these two organizations, mainly producers who are affiliated to them have benefited from trainings in different fields on improved varieties of sorghum and pearl millet in many locations in the regions of Boucle du Mouhoun and center north. Capacity building of producers, specialization of some of them in improved seed production, training in commercialization and marketing, construction of seed storage facilities, dissemination of seeds and improvement of the local seed system were some factors among others that contributed to make improved sorghum varieties available in the studied areas. Packaging seeds in mini-bags of 100 g to 1 kg, available in seed shops tested by the project HOPE, renders improved sorghum seeds accessible to small producers (men, women and young). Age of producer: the age of the producer significantly and negatively influences the adoption of improved varieties of sorghum. This means that young producers adopt more improved sorghum varieties than old producers.  Prob > chi2 = 0.0000 Pseudo R2 = 0.3865 Log pseudo likelihood = -61.204977 *: significant at 10%, **: significant at 5%, ***: significant at 1%

Conclusion and policy implication
The objective of the present study was to determine and analyze the socio-economic factors that influence the adoption of improved varieties of sorghum in the regions of center north and Boucle du Mouhoun in Burkina Faso. To achieve these results, statistical and econometric methods were used to identify and analyze the factors influencing the adoption of improved varieties of sorghum. Results of the study showed that seven variables influence the adoption of improved varieties of sorghum in these regions. These variables are: agro-ecological zone, age of producer, training received on improved varieties, membership of farmers' organization, land area planted with sorghum, access to credit and finally the availability of improved varieties. All these variables positively influence the adoption of improved varieties of sorghum except the age of the producer that negatively influences it. These results suggest that training in favor of producers should be intensified and their access to credits should be facilitated as well. Improved varieties seeds should be made available to them. A combination of these actions will undoubtedly increase agricultural productivity and thus help to achieve food security.