A Conceptual Framework for the Evaluation of Social Agriculture: An Application to a Project Aimed at the Employability of Young People NEET

Agriculture can be a possible provider of social services of relevant importance for the whole society. In order to generate a valuable social service, a multi-actor approach is often applied, based on an active collaboration among public institutions, non-profit organizations, and private firms, and capable of generating multiple positive impacts. This new approach may both favorite agricultural diversification and enhance the quality of life of rural communities. However, in order to enable policymakers in motivating the public support to these types of initiatives, an evaluation method capable of disentangling the multiple benefits generated by social agricultural projects is required. In this paper, we adapted the evaluation method previously developed by the SIMRA consortium for Social Innovation initiatives, to a project aimed at the employability of NEETs in the south of Italy. A selection grid, framed by cross-referencing the national policy objectives of social agriculture and the criteria of eligibility adopted in public calls is proposed, to choose the suitable indicators for the evaluation. The evaluation experience allowed the measurement of 34 indicators of performance. The results prove that 12 indicators are positive, while 12 are moderate, and 10 are low and are mainly related to the enhancement of social inclusion. The evaluation exercise may be useful to disentangling the multiple outcomes generated by initiatives based on social innovation, which are highly based on intangible assets, and exert a positive effect on the internal cohesion and the engagement of the civil society.


Introduction
In the last decades, the agricultural sector has been attracting growing attention in the EU as a possible provider of social services of relevant importance for the whole society [1,2]. Beyond the multifunctional role of agriculture in the provision of ecosystem services, the domain of social services is comprehensively described in Di Iacovo and O'Connor [3]. More recently, several EU member states have enacted specific laws and regulations, to provide a clear definition of social agriculture and to activate some policy intervention to its promotion and deployment. In this context, Italy enacted the 141/2015 law, which defines social agriculture as an aspect of the multifunctionality of agricultural enterprises, aimed at the development of social and socio-sanitary services, educational, and sociooccupational placement. The aim is to facilitate plain and proper access to basic services to disadvantaged individuals, families, and local communities in all of the national territory, particularly in rural or lagging behind regions [4][5][6][7][8][9]. In order to generate a valuable social service, a multi-actor approach is required, where local authorities may establish a clear and flexible regulatory framework promoting an active collaboration among public institutions, non-profit organizations, and private firms. In this regard, social innovation (SI) may represent a promising approach, as the key players may experiment with novel interaction patterns and new informal institutions, based on co-creation, capable of valorizing local resources, and pursuing the co-production of multiple goods and services [10]. This new approach may be promising in terms of agricultural diversification and in terms of the viability of rural communities, due to the provision of social services, which are essential to ensure an adequate quality of life.
However, there is a high degree of complexity in promoting endogenous models of development, as there is no clear and straightforward relationship between the resources needed for the promotion of social innovation (e.g., voluntary workers, natural and financial resources) and the desired outcomes (e.g., services provided to marginalized individuals, marketed goods to sustain the project). In other words, policy support will be aimed at funding initiatives for the creations of immaterial capital, which will be capable of activating the generation of both social services for the targeted individuals, and other goods and services which may contribute to the sustainability of the SI initiative (e.g., sells of agricultural products). Therefore, policymakers should be able to evaluate whether the resources allocated for supporting IS initiative are effective in achieving the objectives of social agriculture. The evaluation process should be capable of disentangling the multiple benefits generated by SI initiatives, and isolating the specific outcomes. Definitely, this is primarily important for public institutions (i.e., public managers, policymakers), in order to perform a costbenefit analysis and to motivate the public support to social agriculture, from a political perspective. In addition, the evaluation is also important for the actors involved (i.e., innovators, consultants) in a specific initiative, in order to achieve an adequate level of self-awareness of the efforts and achievements during the evolution of the initiative [11,12].
In this paper, we present a case study that referred to a project for the employability of NEETs. In fact, in the context of Southern Italy, it is perceived that agriculture is the most relevant sector with the highest employability, due to the fact that farmers face enormous difficulties in finding young and professional workers for specialized jobs. In specific terms, organic farming is particularly emerging in this area and, therefore, the project was aimed at bridging the gap between NEETs and farms. It is an example of initiatives based on the collaboration of multiple heterogeneous actors and aimed at generating multiple relevant impacts for the advancement of the local agro-food system. However, we envisage the problem of identifying and discriminating the impacts of this sort of SI initiative, which we try to overcome by proposing a selection grid formed by cross-referencing the policy objectives pursued by means of social agriculture, and the selection criteria applied for funding a specific category of SI initiatives. The question we try to answer is the following: "which indicator best measures the impact of the SI initiative in terms of specific social agriculture objectives?".
In our contribution, we borrowed the indicators which had already been developed in previous research, and that have already been tested in the evaluation of several types of SI initiatives, within the EU Horizon project "Social Innovation in Marginalised Rural Areas" (SIMRA) [13]. During four years of activity (from 2016 to 2020), the project contributed to the advancement and the understanding of SI in the agricultural sector, with special attention to forestry and rural domains (http://www.simra-h2020.eu/, accessed on 20 June 2021).
As a study case, we performed an evaluation exercise related to the promotion of a social agricultural project aimed at young people who are not in employment, education, or training (NEETs), in a rural area located in Southern Italy. In particular, we selected a project which is still at the early stage of advancement, whose promoters are keen to perform an evaluation exercise, to understand how to steer the activities to improve project effectiveness and sustainability, in the long run. This paper faces two main issues. The first relates to the need to understand whether a specific project can effectively generate valuable social services for a rural community, which are consistent with the objectives of social agricultural policy. The second relates to the need to identify a set of pre-defined indicators, which are suitable to check some project critical aspects, and support the project managers in the adoption of some specific improvements.
In order to achieve these goals, we propose a methodology to adopt the evaluation indicators which are either consistent with the "social innovation" theoretical framework and are also to measure the outcomes of "social agricultural" policies. The evaluation exercise has been conducted by using a set of indicators, already tested in several case studies within the SIMRA project, and that are specifically addressed at SI projects or initiatives [14,15].
The further development and fine-tuning of the methodology proposed in this paper may be addressed at different types of audience: (i) actors who are involved in policy evaluation, such as policymakers and analysts, as well as evaluation agencies, (ii) project designers and managers, (iii) experts of communication activities, to disseminate the achievements of specific initiatives and facilitate the public engagement and involvement.
The structure of the paper is the following. In the next section, we present an introduction to the evaluation of social initiatives. In the third section, we describe the adaptation of the SI evaluation methodology developed by the SIMRA consortium, in order to perform an early-stage evaluation of an initiative aimed at improving the NEETs employability. The study case is reported in the fourth section, while the conclusions and implications for project managers and policymakers are drawn in the last section.

Evaluation of SI Initiatives
There are already several methodologies for the evaluation of third sector organizations [16][17][18][19][20][21], which may be adapted to evaluate the structural part of social innovation (i.e., resources involved, activated processes, etc.). However, to the best of our knowledge, there is still a gap in evaluation tools suitable for the analysis of early-stage initiatives stemming from multi-objective policies (e.g., creation of job opportunities for the youths, favoring women entrepreneurship) and based on multi-stakeholders interplay. Here, the perspective is that multiple policy measures are synergically designed to boost cooperation and collaboration between diversified groups of private and public institutions, and non-profit organizations in order to achieve multiple social goals, through social innovation. Consequently, the difficulty in performing the evaluation relies on the comparison between the resources mobilised (most of which are immaterial), and the bundle of benefits generated by the initiative. In addition, considering the fact that the arrangement of new relationships, or the re-definition of already existing ones, will likely generate some impacts only in the long terms, some indications are needed, particularly at the early stages of the initiative, where some relevant adjustments and improvements may be crucial, for the sustainability of the project.
In this context, SI is conceived as a novel approach to stimulate the synergy between different actors, aiming at satisfying common social needs. To this regard, the EU promotes SI for the development of the marginalised rural areas across Europe, considering that, as claimed by [22], along with the progressive reduction of public and private sources allocated for social interventions in the third sector, the politics of resources allocation has been changing, requiring objectivity, transparency, and rationality.
In this regard, one of the main achievements of the SIMRA project has been the creation of new knowledge to foster the scientific progress regarding the evaluation tools of SI initiatives in the agricultural sector, with special attention to those realized in marginalized rural areas (http://www.simra-h2020.eu/, accessed on 20 June 2021). The research consortium identified and analysed 54 examples of SI across Europe and Mediterranean Areas, which contributed to forming a robust theoretical and methodological framework, enriched with a wide empirical investigation of the SI. One of the main project's outputs is represented by a comprehensive evaluation manual, embracing all the relevant aspects of the SI [13]. The manual covers all the essential aspects of the evaluation, starting from the definition of SI, which was previously provided by [23] and is stated as follows: "the reconfiguring of social practices, in response to societal challenges, which seeks to enhance outcomes on societal well-being and necessarily includes the engagement of civil society actors" (p.12). Next, it describes the comprehensive evaluation framework, providing specific concepts and mechanisms useful to explore the most important aspects of SI, which are subsequently covered by the evaluation criteria and the specification of indicators. In addition, instructions for the preparation of the evaluation, facilitating the application of the methodology to practical cases are provided. Another relevant contribution of the manual is the collection of tools, a set of 11 questionnaires addressed at the different types of actors of SI (e.g. people who conceived the idea and started the process, individuals who joined afterwards, beneficiaries of the activities, etc.). Within each tool, an ordered sequence of questions is listed, which are finalised for data collection needed to calculate the indicators. More than one hundred indicators have been elaborated, in order to capture the multifaceted aspects of SI, and they are grouped in 34 composite indicators, corresponding to aggregated concepts.
Certainly, the all-encompassing nature of the indicators, requires an accurate selection of the suitable indicators needed for the specific evaluation purpose (i.e., policy analysis), consistently with the local specificity of the study case which, in the case of SI, do play a relevant role. The selection of indicators presented in this paper has been performed by a selection grid, resulting from the cross-referencing of the different policy objectives of social agriculture and the selection criteria adopted by diverse administrative levels (e.g. central government, regional authorities, local agencies) in public calls.

Methodology
The starting point to explain the evaluation approach adopted in this paper, is represented by the introduction of the legal framework defining social agriculture in Italy, which involves at least the national and the regional administration government levels. This multilevel approach augments the complexity of the analysis, due to the interaction of different public bodies, which are responsible for the design and selection of public calls, through which local groups of actors are empowered and financially supported, to activate the specific SI initiative aimed at NEETs social inclusion and their employability. In this context, it is important to verify whether the objectives defined by the different policy measures are actually and adequately achieved by the SI initiative.
The methodology applied in this work can be summarized as follows: step 1: identification of criteria for the selection of indicators; step 2: selection of indicators for the evaluation of SI; step 3: data collection, by using the tools defined by the manual of evaluation; step 4: data elaboration.
Step 1. The Italian legal framework defining social agriculture is defined by the national law 141/2015. The regional governments are supposed to enact specific regulations, providing the rules for the elaboration of specific policy measures (e.g. calls for financial support, tenders for public procurement). Among the regional calls specifically addressed at social agriculture, three of them are relevant, as they provide financial support on an annual basis: PuLSA (Apulia Region), "Welfare, che Impresa!" (promoted by various Italian public fundations), and "Coltiviamo Agricoltura Sociale" (promoted by national agricultural associations). Besides some specificity, the calls recently opened have a similar structure, regarding the procedure for the empowerment and funding of local multi-agents groups, the eligibility criteria for the project application, and the selection criteria, according to the relevance and the target groups. The eligibility of projects depends on the consistency of the selection criteria of local calls with the national policy.
Concerning the criteria for project selection provided by the calls, the most important are the following: -activities and strategies to strengthen social farming practices: definition of precise strategies, activities and concrete commitment in the generation and reinforcement of local actors' networks, both formal (i.e., contracts, agreements), and informal (i.e., practices, routines) relationships, for the achievement of the project objectives and the strengthening of the local social agriculture practices; -involvement of traditional farms: strong involvement of the traditional agricultural enterprise; -new networks: potential to create connections between the different actors at bonding (i.e., horizontal links between practitioners or beneficiaries), bridging (i.e., links between firms of beneficiaries with research and professionals), and linking (i.e., links between local actors and institutions) levels. Within these criteria, a specific focus is put on the social involvement of under 35s and women, which can be formal (i.e., hired labor) or informal (i.e., temporary collaboration), and should be directed at favouring the creative contribution to the project implementation; -innovation potential: expected innovative impact of the project in the specific context it will be implemented; -replicability of the project: formulation of the project based on standardised protocols, practices, and experiences, which represents the pre-conditions for possible replication by other subjects and in other contexts; -link with the context of action and with third parties: capacity to match local needs and a good potential to involve local actors; -presence of the technological component to facilitate scalability: partnership with technological partners and technical elements, favouring the scalability of the project, its best practices, and/or its format and protocols; -organization capacity and management skills of the project team: project teams endowed with better planning skills, strategic abilities, and competencies; -facilitation of the emergence of new projects: potential to foster co-design experiences (participation process) and possess concrete tools to implement social agriculture. -Concerning the second dimension of the evaluation framework (i.e., the relevance and target groups), we consider the national framework objectives, which can be summarised as follows: -social inclusion: integration of disabled workers and marginalised or disadvantaged peoples within the local agricultural sector; -creating conditions for social inclusion: provision of social services to local communities to promote, accompany and carry out actions aimed at developing skills and capacity, social and labour integration, and inclusion of recreational and services for daily life; -goods and services to improve social conditions: the creation of new health services supporting medical, psychological, and rehabilitation therapies aimed at improving the health conditions and social, emotional, and cognitive functions of the person, also through the help of farmed animals and the cultivation of plants; -education in nutrition and environmental awareness: capacity of the project to promote environmental and food education, and biodiversity safeguard. It is a broad objective aiming at the spread of local knowledge, by the organization of social and educational farms, such as initiatives for the reception and care of preschool children and marginalised people.
The crosscheck of the consistency of the regional criteria and the national objectives forms the basis for the selection of the indicators actually consistent with the nature of the specific project. Table 1 reports regional selection criteria (in rows) and national objectives (in columns). The labels reported within some cells identify the composite indicators elaborated in the SIMRA evaluation manual [13].
Step 2. After the identification of the composite indicators, the following step consisted in the selection of indicators reported in the evaluation manual produced by the SIMRA project [13], which are consistent with social agricultural projects aimed at NEETs employability. Each indicator listed in the manual is well-rooted on the theoretical framework of the SI, and can be interpreted accordingly. Table 2 reports, for each indicator, the description, the question posed to the respondent, the calculation method, and the tool required for the data collection, as extensively explained in the SIMRA manual.      The indicator measures whether the SI idea was amongst the motivations that drove the transformers to get involved in the initiative. The higher the value, the more likely the transformers considered the SI initiative interesting and valid. This is a proxy of SI capability to attract others Dichotomous choice among following options: they liked the idea and it made sense, they wanted to serve a good cause, they liked the leadership and charisma of the innovator(s) and followers, they wanted to share their expertise for the project, they wanted to feel personally fulfilled, they wanted to receive economic benefits, they was part of the duties of their job, for previous relationships they had with people involved, other motivations % "yes" in options options (they liked the idea, and it made sense) and (they wanted to serve a good cause) Range of values: [0-100]; Acquisition data Tool: 4 Ca2. Innovativeness of the SI idea in the territory Q: "To what extent do the actors in the Social Innovation process perceive the idea as innovative in their territory?" The indicator measures to what extent the actors in the whole SI network perceive their SI idea to be innovative in their territory. The higher the indicator value, the greater the perceived innovativeness of the SI idea in the territory The perception is expressed on a Likert Scale from 1 (=not at all) to 10 (=to a great extent The indicator measures if innovators, followers and transformers have a specific relationship with the territory where the SI has been implemented. The higher the indicator value, the greater the role of newcomers (i.e., someone who comes from other provinces) in the development of the SI process It specifies the percentage of actors in the core group (Innovator(s) and Follower(s)) and in the network (Transformer(s)) that are newcomers for the territory. Respondents have 1 option for which to choose for describing their own relationship to the local territory: (i) I have always lived here (endogenous); (ii) I have lived here, but I studied or worked away (neo-endogenous); (iii) I come from outside, but I have been living here for a while (neo-endogenous); (iv) I come from outside and I consider myself a newcomer (exogenous).        Source: Description of the indicators, labels and page references are referred to the SIMRA evaluation manual [13].
Step 3. The data collection requires particular attention to the correct specification of the individuals to be interviewed, as specified by [13], who are classified according to their roles within the SI initiative.
The first group of agents is represented by the key leaders and initial drivers of innovation. These are labelled as innovators and are the individuals who first had the idea and elaborated on it. They form the first nucleus of the SI initiative.
The second group of agents joining the innovators, are called followers. They provide the initial support as they believe in the initiative and firstly adopt it. Together with the innovators, they form the core group of the initiative.
As the core group starts to implement the idea, new agents join the initiative. These early adopters are referred to as transformers, who support the core group by adopting the idea, and further by spreading it to other people.
Innovators, followers, and transformers together form the SI network. While the SI network relates only to the agents directly involved in the stages of creation, implementation, and development of the SI initiative, other individuals are involved in the later stages of the initiative, as they are affected by its outcomes and impacts. These agents are identified as beneficiaries.
Step 4. Most of the collected information are the result of a verbal judgement and, therefore, are qualitative data, which can be elaborated in terms of counts and scores on the likert scale. Subsequently, each indicator has been elaborated according to the formulae specified in [13]. In order to achieve an overall evaluation, results of the elaborations have been coded according to tertiles, expressing a low, medium, and high performance. Finally, we obtained a synthetic evaluation of the overall SI initiative, based on the distribution of the measure of performance indicators, according to three tertiles.

Description of the Case Study
The evaluation exercise refers to an overall project named "SeminaMenti-L'Orto dei Principi", which literally means "Mind Seeding-The Princes' Vegetable Garden", whose vision consists in educating consumers to high quality vegetables, locally produced by farmers who provide some working positions reserved to include socially marginalised people. It is a relatively small project, promoted by a partnership of three actors, a social cooperative, a professional training agency, and a charity organization, and the endorsement of the municipal government of San Severo town. In this context, the SI initiative object of evaluation is aimed at the improvement of the employability of 10 young people NEET who will likely be hired by the organic farms involved in the whole SI initiative, for producing, processing, distributing, and selling agricultural products to the local community. The municipal government of San Severo town (Apulia region, Italy) allocated about 30,000 EUR budget to incentivize the participation of the youngsters, while the Apulia region supported the cost for the training activities. The core of the initiative is represented by the training activities of NEETs, whose purpose is twofold: to provide and reinforce professional and soft skills, and to favorite the creation of relationships among all people involved in the project, who should boost the potential development of organic farming and the provision of high-quality food for the local community.
In order to apply the SIMRA evaluation method, the basic question to be answered is whether this project can be considered as a "social innovation". In order to perform this first check, we followed the SIMRA approach to verify the presence of the fundamental prerequisite of the SI: i) the trigger, or the prime-mover of the SI, that is a happening or condition no longer acceptable by the community (e.g., marginalisation of social groups), or that represents unexpected chances for the development of the area (e.g. normative change turning the attractiveness of the area). It can be a single event (e.g. an environmental disaster) or the accumulation of happenings (e.g. consistent migration flows due to unemployment); ii) the unmet social needs, which are challenges referred to the societal, economic, environmental, and/or institutional domains; iii) the perceived context, that is the conditions that influence the actors' behaviour; iv) the agency, that is a group of actors (with their ideas, values, willingness, and capacity) preparing and implementing the transformation; v) the preparatory actions, including objects, activities, discourses, and narratives of change used by the agency for preparing and starting the reconfiguration of social practices and networks since the early stage of the process. Table 3 reports the arguments supporting the hypothesis that the case study characteristics comply with the fundamental requirements for SI definition.
The project activities started in October 2019, while the data collection for the evaluation occurred in March 2020, that is, the mid-term of the one-year full duration of the training of NEETs. The group of people involved in the project management was aware of the importance of getting an intermediate evaluation, to get some advice and recommendations, and, therefore, exhibited full cooperation to conduct the direct interviews. Table 3. Characterization of the SI initiative, within the "SeminaMenti" project.

Trigger
The municipality of S. Severo needed to provide social support to NEET, who are not currently targeted by any specific social policy; the presence of social cooperatives able to provide social services, with a new "inclusive" vision; the availability of land to be used as experimental fields for organic farming

Social needs
High unemployment rate; young people with poor education and skills, unable to catch regular job opportunities and exposed to engagement to illegal activities; agricultural products are not valorised; local community is not educated in food quality and seasonality Perceived context Agriculture is the economic sector with highest employability; local consumers are willing to buy high quality products Agency Institutions: municipality of S. Severo, Social Cooperative "Attivamente", education and training agency "SMILE", non-profit charity organisation "Caritas"; about 10 people belonging to the institutional agents, who are directly involved in the initiative, one external consultant (agronomist) Preparatory actions Identification of external actors to activate possible collaboration for further development of the SI; market observation of organic food Source: own elaboration, based on interviews of actors involved in the project.

Data Collection and Elaboration
A preliminary introductory focus group was conducted in February 2020, with all representatives of institutional actors promoting the "SeminaMenti" project (i.e., the social cooperative, the professional training agency, the charity organization, and the municipal councilor for social affairs). Subsequently, due to the containment measures of the Covid-19 adopted in Italy (Decree of the Council of Ministers DPCM 26 April 2020, starting 4 May, and up to 17 May 2020), we conducted a series of in-depth interviews via online web conferencing platforms. We adopted the same interview protocols used in the SIMRA case study analysis, namely Tools 3, 4, 5, and 7.
After the survey, we calculated the indicators for the evaluation of SI selected for this specific initiative (Table 2), by applying the formulae specified in the evaluation manual [13]. The results of the data analysis are reported in Table 4.   The first evidence that emerged from results, is that 34 indicators have been evaluated and the majority of them are satisfactory (i.e., 70% belong to the high and medium category), while only one 30% are poorly performant. A critical issue of this evaluation is to understand the mechanism underlying the performance of the different criteria. We will proceed with the interpretation of results for each composite indicator, explained in terms of the basic elements characterizing the SI according to the theoretical framework, as described in Table 4.

Indicators and Sub-Indicators
X4-Engagement of civil society: all the indicators are very poor, probably due to the fact that the project is still at the early stage of development, and the main role is still played by some local organizations (i.e., the non-profit organization, the training agency, and the municipality), while the local community has not been extensively involved; X2-Response to societal challenges: the initiative has been oriented at the training of NEETs towards organic agriculture and, therefore, emphasizing the relevance of environmental sustainability; the score is quite satisfying, as the design of the project is adequate to respond to some relevant social challenges.
X6.2-Perceived Opportunities and Threats: the scores of the three indicators are heterogeneous, due to contrasting forces. On the one hand, the strong cohesion among the institutional actors, endorsing and funding the initiative, provides a positive perception towards the opportunities of developing an innovative type of agriculture, capable of redeeming NEETs and promoting organic farming. On the other hand, there is some sort of discomfort regarding the financial sustainability of the SI process, in the mid and long terms, especially in case other actors (e.g., farms willing to recruit the trained NEETs) may not collaborate; X7.1-Social Innovation Idea: the indicators of this group reveal that the idea behind the project is original and well suited to the perceived context, as it is able to valorise some local strengths (i.e., local consumers are willing to buy high-quality products), exploiting the capacity of the primary sector to provide high employability.
X7.4-Capabilities: all actors forming the agency already have some experience in activities which were closely related to the provision of social services; however, the current experience was the first multi-actor initiative they conceived and planned together; X7.5-Endogenous versus exogenous drivers of the SI process: the performance of these indicators is moderate, demonstrating that, despite the need to valorise the role of external members, it may contrast with the internal cohesion of the core group (i.e., agency); X8.1-New networks: the performance in this group of indicators is polarized as low values in some indicators alternate with high performance in other indicators. This is probably due to the fact that the preparatory actions of this project have been effectively targeted at the involvement of young people, who are the direct beneficiaries of the intervention. However, less attention has been paid towards the balance of the partnership and the female inclusion, as the action mainly engaged male farmers.
X8.2-New attitude: the initiative seems effective in favoring the collaboration of actors actively involved in the training of NEETs (the so-called transformers), who exhibited a pro-active attitude and foresee some further development of the SI process; X7.2-Leadership: the value of the indicator is moderate, which demonstrates that there is not a prevalence of any actor on others. This can be positive, as all actors may act responsibly and aware of their role within the project, and will exhibit a more pro-active role; X10.1-Feedback loops and multiplier effects: the project is able of providing high symbolic value as it addresses one of the most relevant social challenges, that is to provide job opportunities for NEETs, and this reflects on high performance for this composite indicator which assesses the replication potential of the idea.
X11.1-Relevance of the SI process: the performance of these indicators are crucial, as they demonstrate the adequacy of the SI in meeting the social needs, which is one of the main pre-requisite for the activation of a SI process, and its future development; X11.3-Relevance of the Social Innovation Initiative: the good performance of these indicators are consistent with the previous one. In this case, the evaluation refers more specifically to the NEETs training initiative.
These figures reveal the need for some improvement, which could be achieved if more efforts are focused on about one-third of aspects affecting the indicators with lower scores. In order to identify possible corrective actions, the following strategy may be considered: (i) to assign the higher priority to actions addressed at critical issues, identified by poor indicator performance; (ii) consolidate and reinforce the activities underlying the indicators falling in the medium category; (iii) keeping a close monitoring of the well performing indicators, by considering them as strengths of the SI initiative.
Certainly, most of the actions must be planned and implemented by the agents directly involved in the SI. However, external partners and policymakers may also play an important role, in the endorsement of the SI initiative and the activation of parallel initiatives exerting a synergistic effect.
In order to meet the evaluation needs of policymakers, we rearrange the indicators, according to the cross-reference between the composite indicators and the policy objectives presented in Table 1. Accordingly, we are now able to evaluate the SI initiative aimed at NEETs employability, in terms of the objectives of social agriculture (Table 5). The distribution of the indicators shows that the SI initiative was mainly evaluated with regards to its contribution to the two first objectives of social agriculture, such as social inclusion and the enabling conditions to favor social inclusion. On the contrary, only a few indicators were suitable to evaluate the contribution to the other two objectives, regarding the creation of goods and services and the education of society in nutrition and environment. This distribution can be explained by the fact that the SeminaMenti project is still at its early stage of development, which is mainly focused on the training of NEETs in organic farming; afterwards, they are expected to be recruited by organic farms and continue to collaborate with the SI agency to promote organic food to the local community (e.g. schools, open market, specialized shops).
Regarding the performance of the first objectives, despite the overall performance seems quite satisfactory, some weaknesses emerged regarding the gender balance (NEETs were only male), and the active involvement of local people in the initiative. This can be explained by the specific cultural context of Southern Italy, where agriculture has always been considered a male domain while, on the contrary, women have played a main role in food preparation, and in restaurants. Probably, the SI agency has not been sufficiently effective in advertising that the training activities for NEETs were based on the integration of cultivation of horticultural crops with post-harvest operations and marketing; if this were the case, some women would have been interested in the project.

Conclusions
This paper focuses on the importance of an evaluation methodology aimed at initiatives for the promotion of social agriculture, based on SI. The main feature of SI is that the most relevant outcomes are represented by intangible assets, such as the strengthening of social capital, the reinforcement of the sense of belonging to the local community, and a more pro-active attitude towards new opportunities offered by the market (e.g., provision of organic products to the local market), and the social services provided by the agricultural sector (e.g., the social inclusion of NEETs and their employability). The complexity involved in SI initiatives, arising from the multi-agent collaboration (e.g., non-profit organizations, municipalities, private firms, training institutions), and the multi-faceted nature of benefits which may be created by these activities (most of which are immaterial), poses a relevant challenge for the evaluation of the outcomes and, consequently, for grasping the whole relevance in terms of impacts for the whole society. It is worth mentioning that SI initiatives are highly dependent on intangible assets, typically included in the concept of social capital, as shared values, perceptions, expectations, trust, reciprocity [24][25][26][27], which deserves some form of representation (e.g., by means of evaluation), in order to exert a positive effect on the internal cohesion and external engagement (i.e., civil society).
Despite the difficulty of performing an evaluation exercise at the early stages of development of a SI initiative, we claim that even by collecting some (unavoidably) approximate qualitative data, it is possible to identify some critical aspects of the project, which may provide some valuable insights to project managers, policymakers and the local community, either for the steering and the improvement of the project, but also for the strengthening of mutual relationships among the participants and the civil society.
The difficulty of disentangling the multiple benefits of SI initiatives in social agriculture is particularly relevant for policymakers, who are responsible for policy measures that must comply with the "best value for money" principle. In this paper, we proposed a methodology based on an evaluation grid, through that we selected a set of indicators which have already been applied for the evaluation of SI initiatives. In this way, we have been able to perform a consistent evaluation exercise referring to a SI aimed at training and favoring the employability of NEETs, within the frame of an agricultural policy oriented at supporting social agriculture.
We have been able to demonstrate that the SeminaMenti project performed quite well, mainly in terms of two objectives of the national policy of social agriculture, such as social inclusion and the social inclusion and the enhancement of enabling conditions for social inclusion. On the contrary, only a few indicators for SI evaluation were suitable for the evaluation of the other two objectives, such as the provision of goods and services and the improvements in education in nutrition and environmental awareness. This was mainly due to the fact that the SeminaMenti project is still at an early stage of development, and it will take a long time in order to produce some relevant results in these domains.
The evaluation experience presented in this contribution demonstrates that, despite the valuable efforts made by civil society in promoting some virtuous initiatives of social inclusion (e.g., the enhancement of the employability of NEETs), the role of public funding is indispensable, in order to support these initiatives in the short and long rum, and to replicate them. For this reason, an evaluation exercise is needed in order to provide a robust analysis of benefits arising from a specific initiative. This is a new domain of investigation, as the available evaluation frameworks are suitable for single project assessment and for the single institution (private, public, or non-profit), while only a few approaches and experiences are available to evaluate multi-actor and multi-goal interventions (or process), typical of holistic and bottom-up frameworks.
The adaptation of the methodological framework developed by the SIMRA consortium, to a social agricultural project aimed at fostering the employability of NEETs, needs some further efforts, especially with regards to the robustness of the selection grid, formed by the cross-reference between national policy and criteria of relevance used project selection in public calls. Though we could have applied some consolidated methodologies to obtain a robust selection grid of indicators, we were not able to interact with policymakers and experts who are fully familiar with concepts such as social agriculture, social innovation, selection criteria of public calls, as they refer to a domain which is still at a niche level. Therefore, the limitation of this study is that it is still valid as a conceptual framework at an academic level, while some time will still be required until its application at an operational level. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy commitment with interviewed persons.