Analysis of innovation networks

In recent years Mexico has generated findings regarding the use of network analysis in the rural sector. These results come from work and research seeking alternatives interaction with farmers to generate competitiveness through improved processes extensionism and rural innovation (Muñoz et al., 2004; 2007; ^{Zarazúa et al., 2011}; ^{Díaz et al., 2013}; ^{Sánchez et al., 2013}).

Network analysis innovation is a tool to analyze a territory to a network of actors and public relations develop in their social, technical and commercial environment for the purpose of promoting intervention strategies in the rural sector around a production system. Network analysis supports innovation bases in the theory of Social Network Analysis (ARS). This analysis tool allows us to find new ways of intervention to join efforts and existing extensionism ATyC in the rural sector; It has been used for locating key stakeholders to further the process technology transfer (^{Muñoz et al., 2004}; ²⁰⁰⁷; ^{Zarazúa et al., 2011}), adoption of innovations (^{Muñoz et al., 2004}; ^{Díaz et al., 2013}; ^{Sánchez et al., 2013}), as well as the design, monitoring and evaluation of intervention strategies change agents.

The aim of this research was to analyze the coverage of services ATyC five cases of agents of change, using social network analysis, seeking to improve the efficiency of public and private investment and access to knowledge in rural areas.

Identification of key players

In rural areas it is possible to identify different types of actors according to some features such as relationships and whether these have indicated or are identified by others as a source of information. Identifies 3 types of actors: source actor, collector actor and an actor articulator. When one of these actors is of some importance by the number of relationships or their position in the network can be called key player. The first two types of actors were included in the study to calculate coverage. The Table 2 defines three types of actors, including its graphical representation where each actor is represented by a node and the relationship is directed, represented by an arrow.

Table 2. Types of actors can be identified in an innovation network. 

To determine which producers are key source and which as collectors in an innovation network is used scope concept developed by ^{Borgatti (2006)}. For actors sources node degree is considered based on the input area; for collectors’ nodal degree considering the proximity output (^{Borgatti, 2006}).

Coverage estimator

The concept of coverage, this study used is based on the scope defined by indicator ^{Borgatti (2006)} shown in the following equation.

K= group of key players or dominant set j is an actor of the network that does not belong to K; d_Kj= length or minimum distance from any member of the group K to a node j, measured in steps and N= j set of actors; ^DR = ratio of all nodes reached by the set K, where nodes are weighted by the inverse of the minimum distance.

A range of 100% of the network will be obtained when the set K reaches all actors set N and a range of 0% is obtained when the set K is completely detached from the set N. The distance is defined by Everett and Borgatti (1988) as the number of links or ties between an actor and another, it is possible to calculate minimum and maximum distances.

In the application, the algorithm selects a group of key players K with which a change agent can intervene directly to these through their relationships (steps) with other actors are capable of transmitting knowledge previously received , meaning that they can optimize the dissemination of information. Therefore the coverage, scope is interpreted as two steps an agent of change achieved as a result of attending directly to a particular group of key stakeholders that are part of a network. The research was conducted in two stages, the first to contrast the coverage taking into account as key players only to producers, which was named "selection of actors", while the second contrasts coverage taking into account as key players to producers and other actors so it was called "diversity of actors."

For the two stages of the joint research network resulting from the merger of social, technical and commercial networks each observation was built. To calculate the scope defined by ^{Borgatti (2006)} Ucinet software version 6.288 for Windows (^{Borgatti et al., 2002}) was used. This software was possible to plot the networks of the five observations; subsequently, they were transferred in implementing Ucinet: KeyPlayer 2; with the two "harvest" indicators (input range) for source identification actors and "diffuse" (reach out) for the identification of collectors actors were obtained.

Selection of actors

In the step of selecting actors a comparison was performed between the coverage obtained by the potential change agent and coverage, both producers. The obtained coverage represents the percentage of the network to which the producers served directly by the change agent had access to total network. On the other hand, the potential coverage is estimated as a result of choosing the same number of players but considering criteria for identifying networks of actors sources and collectors coverage. For the calculation of coverage obtained they are entered KeyPlayer two actors who had direct care as required actors and with the same calculation process the harvest and diffuse indicators, which were identified source application and what collectors. To calculate coverage potential key producers were identified in the joint network requesting half of source actors and half of collectors actors for each case, with the same number as the direct care at each observation, using only producers actors.

Diversity of actors

In the diversity of actors stage a comparison between the coverage obtained by the agent of change and potential coverage range of stakeholders was made. For this stage only the potential diversity of actors coverage was calculated as the coverage calculated obtained in the previous step was used. To calculate the potential coverage range of stakeholders: stakeholders were identified in the joint network, so that a combination of source and collectors actors in the same ratio was obtained, these may be different producers, with the same number the direct care cared for each observation. Finally indicators of these key actors were calculated and summed to obtain the potential coverage range of stakeholders.

Selection of actors

At the stage of selection of players, the calculation of the indicators allowed to choose closer and better connected to the network producers, where even producers who were treated were selected directly, and other producers which coincided with those who were selected by change agents. In Figure 1 the coverage achieved with elected by the change agents conventionally producers and chosen based on criteria producers network analysis are compared; i.e. shows the comparison between the potential coverage and coverage obtained, the coverage obtained in all cases is less than the potential difference with an average of 22.4%.

Figure 1. Coverage potential and obtained in each network with producers. 

Change agents may have increased their coverage in 22.4% have chosen his actors taking into account the structure of the network and selecting producers network criteria, at least the identification of sources and collectors actors. The maximum increase (32.4% of the network) occurred with the change agent 2.

Diversity of actors

Regarding the stage of diversity of actors, Figure 2 shows the comparison of coverage between a) the coverage obtained with producers direct care b) the potential coverage by selecting all kinds of actors, the latter is the one that could have been achieved considering criteria of connectivity and closeness, but this time with different actors such as government institutions, educational institutions and research organizations, traders, agro-industries, suppliers and consultants.

Figure 2. Coverage of each network by type of actor. 

The average increase of coverage that can be achieved by ATyC services with diverse stakeholders regarding selected only UPR traditionally is 32.2%. The inclusion of diverse actors in an ATyC intervention strategy by the agents of change can reach higher levels of coverage, with spans ranging from 68% to 93% of the actors in the joint network. The smallest increase was in observation 3 with 9.6% and the largest increase was in observation 2 with 50% of the network. Derived from the first two stages of research a new way to calculate the coverage where the inclusion of diversity of actors considered and the selection of key stakeholders that enable knowledge transfer at least two steps is proposed. The Figure 3 shows the agent of change and different combinations of actors who can access directly or indirectly. The calculated so far by SAGARPA coverage takes into account only the coverage achieved directly with actors A (UPR receiving services ATyC), excluding coverage a step with actors B (UPR not receiving services ATyC), and C (other actors) and cover two steps in any combination. That is, only that in calculating the percentage of the UPR receiving services ATyC omitted the rest of the target population has access to knowledge as a result of the intervention of the change agent.

Figure 3. Coverage of a change agent. 

The Figure 3 shows the concept of coverage ATyC services that this research suggests. Directly where the change agent can intervene with actors A, B and C. While indirectly communicating A, B and C with A and B grow other options that they can perform activities as a way: replica, testing, verification, reaffirmation and dissemination of knowledge, information or further collaboration technologies based on the combination of actors, through which these actors access to knowledge.

This proposed coverage calculation takes into account access to knowledge as a result of an intervention by an agent of change. So then direct and indirect intervention of an agent of change is defined. direct intervention is considered when there is a direct relationship (which may be of assistance, training, bonding, collaboration among others) between the change agent and an actor who can be UPR or other actors (government agencies, educational institutions and research organizations, traders, agro-industries, suppliers and consultants), representing intervention coverage agent to a step change. Moreover, intervention is indirect when an actor who has received a direct intervention transmits information or knowledge to a third actor, this intervention represents the coverage change agent two steps. Indirect intervention is important in the participation of actor who has received a direct intervention, as it serves as articulator, "filter" or "information bridge".