Thailand Social Capital Evaluation:
A Mixed Methods Assessment of
the Social Investment Fund’s
Impact on Village Social Capital
July 14, 2017
The World Bank
EASES
Acknowledgements
The initial impetus to conduct this research came from Khun Paiboon Wattanasiritham,
Director of the Thailand Social Investment Fund, who was interested in gathering the
strongest possible evidence of the social capital impact of the operation to promote global
learning. While the study moved forward independently, it benefited from his intellectual
support. The research was guided by an Advisory Committee representing several
government stakeholders including Dr. Maitree Wasuntiwongse, Dr. Priyanut
Piboolsravut, NESDB, Khun Vichol Maneetiersiri, Ministry of Interior - Community
Development, Prof. Anuchart Poungsamalee, Mahidol University, Khun Jirawan
Boonpem, National Statistics Office.
The study benefited from close and effective collaboration among Thai researchers and the
World Bank team. The Core Thai Research Team was led by Assoc. Prof. Dr. Napaporn
Havanon and Principal Investigator, Prof. Dr. Maniemai Thongyou; Dr. Numchai
Suparerkchaisakul conducted much of the initial data analysis, with Mr. Apichart
Thongyou advising the work as rural and social development expert. The field researchers
for the team, offering extensive experience in mixed methods research and operations,
included Mr. Paisal Chuangcham, Mr. Monchai Pongsiri, Mr. Kovit Kulsuwan, Ms.
Benjalak Taraporn, Mr. Adirek Rengmanawong, Ms. Supawadee Boonjuer, Mr. Voragarn
Tirasirachot, Mr. Pornvit Taraporn, Ms. Sasikarn Chomchuen, Ms. Banjong Siri, Mr.
Pornamarin Promkerd, Mr. Adisorn Pakpong, and Dr. Montana Pipatpen.
The World Bank team benefited from the support and advice of Country Director Ian
Porter, Country Program Coordinator Lynne Sherburne-Benz, and East Asia Social
Development Manager Cyprian Fisiy. Gillian Brown and Rob Chase, co-task-team
leaders of the research project, enjoyed the able support of task team members Rikke
Nording and Pamornrat Tangsanguanwong. Lawrence Salmen provided expert advice on
the conception of the work. Peer Reviewers Ana Revenga, Lant Pritchett, and Anna
Wetterberg provided valuable comments throughout the project.
The research was supported financially from the Trust Fund for Environmentally and
Socially Sustainable Development (TF051253) and East Asia and Pacific Results
Secretariat. Rob Chase drafted the concept note and final report and Rikke Nording
conducted the empirical analysis for the propensity score matching and produced the final
empirical tables.
Tables and figures
Executive Summary
Figure I. Framework for Social Capital Dimensions
Table I. Social Capital Variables Significantly Different Between SIF and Comparison
Villages
Main Document
Figure 1. Model of Social Capital Structural Relationships
Figure 2. Social Capital Indicators and Sub-indicators for SIF Impact Evaluation
Figure 3. Pre-match Kernel Densities of participation propensity
Figure 4. Post-match Kernel Densities of participation propensity (6 nearest neighbor
within provinces)
Figure 5. Social Capital Structural Relationships for SIF villages
Figure 6. Social Capital Structural Relationships for non SIF villages
Figure 7. Effects of Training and Network support in SIF villages
Table 1. Number of SIF Projects and Amount of SIF Funding Support by Menu
Table 2. Probit regression results of participation in SIF
Table 3. Number of Sample SIF Villages by Sub-regions
Table 4. Mean scores in SIF and Non-SIF Villages
Table 5. Regression results
Appendix A: Semi-Structured interview guide
Appendix B: Interviewer’s rating form
Appendix C: Village Resource Profile
Appendix D: Additional analysis tables
Table 1. Means and standard deviations of variables used in propensity analysis for
treatment and matched villages and total villages.
Table 2. Comparison of differences in means across SIF and Non-SIF Villages
Table 3. Comparison of percentages across SIF and Non-SIF Villages
Table 4. Types of Support Received from SIF by SIF Village
Table 5. Regional differences across regions (comparing one region with the rest of
Thailand)
Thailand Social Capital Evaluation:
A Mixed Methods Assessment of the Social Investment Fund’s
Impact on Village Social Capital
Executive Summary
Social capital is a vital yet underappreciated development asset. Broadly defined
by the World Bank as the “norms and networks that enable collective action,” social
capital refers to a class of assets that reside in social relationships, which makes those with
access to it more effective, and that can be enhanced for lasting effects like other forms of
capital. Evidence from many different contexts suggests communities and individuals
endowed with more social capital enjoy better services, more effective governance, and
improved welfare. Given this conceptual and empirical appeal, the government of
Thailand has underscored its importance, with the Prime Minister explicitly calling for
efforts to enhance village social capital.
Assessing the impact of the Thai Social Investment Fund (SIF), this study
strengthens our understanding of social capital in Thailand and presents evidence about
how community driven development (CDD) approaches can enhance it. The study
presents a set of indicators for several dimensions of social capital that reflect the Thai
context, though they have broader regional applicability. Further, it uses an innovative
and pragmatic mixed-methods evaluative approach to collect and analyze evidence about
social capital characteristics in treatment villages that participated in SIF and matched
comparison villages that did not. Finally, from that evidence it concludes that SIF acted
both as a mechanism to select villages with pre-existing cooperative norms and as an
effective instrument to enhance leadership, networks, and villagers’ capability to exercise
voice to formal authorities.
Framework for Understanding Social Capital Dimensions
As a development asset social capital has many aspects. To clarify work on social
capital, one should separate this broad category into component parts, adapt a social
capital conceptual framework to local context, and identify appropriate proxies and
indicators that capture those definitions.
In this study, it worked well to disaggregate social capital into component parts.
Anchored by accomplished qualitative researchers well-versed in Thai village dynamics,
the team identified dimensions of social capital relevant to the Thai context. As
summarized in Figure 1, those consisted of stock dimensions, that are existing
characteristics of village institutions and norms; channel dimensions, that are means
through which social capital operates to make village actions more effective; and outcome
dimensions, which refer to the ends to which social capital assets are applied. Within
these three categories, the research investigated the stock variables of solidarity and trust,
groups and organizations, and networks and linkages. In the channel dimensions it looked
at cooperation, collective action, information sharing and communication. Finally, it
considered outcome variables of social cohesion and empowerment. To capture these
variables, the research team identified a set of indicators valuable for understanding Thai
social capital.
Figure I. Framework for Social Capital Dimensions
STOCK
CHANNEL
Solidarity
and Trust
OUTCOME
Social Cohesion
Cooperation and
Collective Action
.
Group and
Organization
.
Information
Sharing and
Communication
.
Network and
Linkages
Empowerment
Based on the Thai SIF example, this study shows how CDD operations work with
existing social capital characteristics in communities. But many of these operations also
have as development objectives to enhance social capital characteristics. Considering
separate social capital indicators, it is important to discern selection effects from impact
effects in CDD operations. For example, evidence from this study shows that CDD
operations can act as selection mechanisms among communities, promoting those wellendowed with particular social capital characteristics to receive program funding. Further,
CDD operations can have impact on village dynamics. Presenting several dimensions of
social capital and separating selection from impact effects, this study offers an example of
how CDD operations work with village social capital and whether they can enhance it.
Innovative Evaluation Methodology
Ideally, those seeking to learn from program experience would collect baseline
social capital information to know beforehand how treatment villages that participated in a
program differed from control villages that did not. However, as with many development
operations, particularly those CDD operations aiming to change social capital, the Thai
SIF did not collect baseline information about disaggregated social capital indicators for
treatment and control villages. To address this common challenge, the research team
innovated with a rigorous, pragmatic approach to understanding the interaction between a
CDD program and village social capital.
The research methodology combined quantitative matching techniques with
qualitative field research to identify how and why villages that participated in the Thai SIF
differed from others that did not. It used existing household survey data to match each of
72 sample villages that participated in SIF to six potential comparison villages. Field
teams of highly qualified qualitative researchers1 then consulted with local authorities and
collected additional information to pair each treatment village with its most similar match.
After teams of three field researchers spent three days in each of the 144 sampled villages,
they scored each on a one to five scale for each social capital variable. This scoring
summarized qualitative information with a series of quantitative scores. To establish
statistically significant differences between treatment and comparison villages, the team
analyzed differences in means and regressions on each social capital indicator. Finding
several significant differences, the team then asked field researchers to put those findings
in context, judging whether the observed differences were due to SIF selecting villages
well endowed beforehand with certain social capital characteristics, or whether SIF
activities had impact on those characteristics. With a nuanced discussion of how SIF
processes interacted with village social capital characteristics through selection and impact
effects, the study gives more operational guidance than do purely quantitative impact
evaluations.
Table I. Social capital variables differing
significantly between SIF & comparison villages
Thai SIF’s Selection and
Impact Effects on Social
Capital
The research produced several
interesting and relevant findings on
how Thai SIF worked with and
enhanced village social capital. After
correcting for village socio-economic
traits through extensive matching and
regression analysis, the research
showed that SIF villages still differed
from matched comparison villages on
several social capital indicators.
Adopting conservative statistical
standards2, Table I summarizes the
social capital variables that differed
between treatment and comparison
communities.
Some observed differences
resulted from SIF processes selecting
villages better endowed with certain
1
Social capital indicators
1.1.3 Solidarity: Self sacrifice for common benefits
2.2.2 Strength of Leadership: Diversified Capability
2.4 Organizational capacity
2.4.1 Organizational capacity: Effectiveness
2.4.3 Organizational capacity: Learning Ability
3. Network and Linkages
3.2 Strength of Horizontal Linkages of Groups and Communities
3.2.2 Horizontal Linkages: Multi-dimensionality
3.3 Strength of Vertical Linkages
3.3.1 Vertical Linkages: Breadth
3.3.2 Vertical Linkages: Multi-dimensionality
3.3.3 Vertical Linkages: Benefits
3.3.4 Vertical Linkages: Accessibility
5.2 Information sharing outside community
5.2.6 Villager’s Voices Responded by Government Agencies
7.1 Empowerment: Capacity building
7.1.3 Empowerment: Strength of Multi-party Mechanism
7.3 Empowerment: Ability to Sustain Development
The principal investigators responsible for the field work are leading Thai researchers on village social
characteristics. Much of the field work was conducted using a team from the University of Khon Kaen,
which Professor Robert Chambers noted “did most to establish [participatory rural appraisal’s] credibility,
emphasizing the management of multidisciplinary teams and the techniques and value of semi-structured
interviewing” (PRA Note 24, IDS, Sussex).
2
The indicators summarized in this table include those found to differ between SIF treatment villages and
matched comparison villages to a statistically significant degree using 95% confidence intervals in both the
difference in means and regression findings presented below.
social capital characteristics. Specifically, the analysis shows that SIF villages began with
more trust among neighbors and stronger traditions of cooperation and collective action.
This is not surprising: to participate in SIF, community members must work together and
commit to implementing proposals that entail self sacrifice for common benefit.
Given that it collected extensive social capital data, the study also sheds light on
village characteristics generally associated with social capital. For example, regardless of
their participation in SIF, villages were better endowed with social capital where more
people work in agriculture and where fewer own their farms. This evidence suggests that
cooperative norms emerge when farmers work together to cultivate crops. And, while
spending levels made little difference, villages with more unequal expenditures generally
had more social capital, suggesting an insurance motive in that cooperative norm. Finally,
since more education was associated with less social capital, there is evidence that these
cooperative norms and networks are maintained and valued by the less well educated.
The study shows that SIF had direct impact to enhance several social capital
variables. While SIF selected villages that already had stronger norms of self-sacrifice,
the research found that SIF activities also enhanced that norm. The project demonstrated
the value of sacrificing to improve village welfare. SIF activities helped build leadership
of local organizations, both through its support for networks and training and through a
procedure that encouraged leaders to get things done outside of the formal government
system. The intervention itself strengthened networks and linkages within and among
villages and improved information flows between villagers and government officials.
These impacts result from SIF’s concerted efforts to create horizontal links between
organizations and vertical links to formal authorities. Further, the research found that SIF
activities weakened social cohesion: its focus on reaching clear objectives tended to
exclude those people viewed as least effective. Finally, through its focus on networking
and leadership strengthening, SIF changed relations with government officials,
empowering villagers to support multi-party mechanisms, express their voice to local
officials, and sustain development activities through their own initiative.
Through invitation of the Thai government, this research established differentiated
social capital indicators for Thailand and the region. It developed a pragmatic
methodology for understanding the relationship between village social capital and
community driven development operations. And it identified several ways that the Thai
SIF selected villages with strong social capital and had direct impact on several social
capital dimensions. Through these successes, it opens up opportunities for a broader and
deeper understanding of social capital and CDD in the region.
I. Introduction
Social capital is a vital yet underappreciated development asset. Defined by the
World Bank as “the norms and networks that enable collective action,” social capital
refers to a class of assets that inhere in social relationships, makes those with access to it
more effective, and can be enhanced for lasting effects. Further, evidence from many
different contexts suggests communities and individuals better endowed with social
capital enjoy better services, more effective governance and improved welfare. Given
this intuitive appeal and burgeoning evidence base, the Thai government has underscored
its importance: the Prime Minister explicitly called for efforts to enhance village social
capital in Thailand.
Social capital has many different aspects as a development asset. To understand
and work with the concept, it is necessary to disaggregate it into component parts and
consider how it manifests itself in different contexts. This report presents a careful
decomposition of social capital in Thailand, breaking it into several dimensions according
to three categories. Stock dimensions are contextual characteristics of village institutions
and norms. Channel dimensions are the means through which social capital operates to
make the village more effective. And outcome dimensions refer to the ends to which
social capital assets are applied. Articulating a framework for understanding Thai social
capital, the study provides the opportunity to apply the concept in other initiatives in
Thailand and the region.
Through understanding of social capital dimensions in context, development
actors need to develop ways to support and enhance different social capital aspects.
Based on experience in the region and elsewhere, the World Bank’s community-driven
development (CDD) operations are a promising approach to support local social capital.
As an example of community driven development, the Thailand Social Investment Fund
(SIF) made strengthening village social capital one of its prime objectives. This
assessment takes the Thailand SIF as a case study to investigate whether and how CDD
operations enhance social capital.
CDD operations provide opportunities for communities to apply and compete for
resources. Some apply for and receive funding. Others apply and receive nothing. Still
others do not apply, lacking information or relevant skills. Some dimensions of social
capital, such as willingness for self-sacrifice or links to formal local authorities, may
increase the chance that a village will prepare a successful CDD financing proposal and
thus become a treatment community. Given this self-selection, CDD operations likely
have two effects that together imply that, viewed after the CDD operation has been
completed, communities that participated will have different social capital characteristics
than non-participants:

Selection Effect. Villages that participate in the CDD operation already had
different social capital than non-participating communities, for the CDD approach
selects villages with high capacity to work collectively.
1

Impact Effect. The CDD operation implements activities that directly impact the
social capital characteristics of participating villages.
These selection and impact effects are common elements of any effort to understand and
assess program results. To separate these effects with the highest standards of rigor,
those seeking to learn from program experience need baseline information on the social
capital characteristics of treatment and non-treatment communities.
This operational research develops and applies an innovative methodology to
understand the social capital effects of Thailand’s SIF. As a project that focused on
learning by doing, it only became clear as SIF was being implemented what dimensions
of village social capital were most relevant to project operations. Accordingly, the
operation collected no baseline social capital information because it had not defined
social capital characteristics and how they would likely inteact with the CDD operation.
Unfortunately, most operations of any type lack baseline data adequate to understand
project impact. Missing baseline data but purporting to have changed social capital
characteristics of participant villages, the Thailand SIF is representative of many CDD
operations.
This report tackles this methodological challenge with a creative and practical
approach to evaluation. The study combines ex-post quantitative and qualitative
assessment approaches. Taking advantage of existing quantitative data sources from
before the project began, it matches 72 treatment and 72 comparison communities, based
on observed characteristics. Leading Thai academics and practitioners developed a range
of social capital indicators. Field teams of three qualitative researchers spent three days
in each of these 144 matched villages. They used a methodology that would allow them
to transform qualitative observations into quantitative scores. Separating social capital
into separate dimensions appropriate for the Thai context and distinguishing selection
from impact effects, the research reaches operational conclusions, presenting how CDD
operations can enhance social capital. Given that baseline data is often missing, this
mixed method assessment approach will be useful for other operations, particularly CDD
operations that purport to have enhanced village social capital.
The paper is organized as follows. Section 2 presents the detailed decomposition
of social capital into components appropriate for the Thai context and for other countries
in the region. Section 3 summarizes the operations of the Thailand Social Investment
Fund as an example of a Community Driven Development operation. Section 4 explains
the mixed methods evaluation. Section 5 presents the findings of the observed social
capital differences between SIF treatment and matched comparison villages, discussing
how much of these observed differences can be attributed to a selection effect and how
much to an impact effect. Section 6 concludes with implications of this research.
2
II. Social Capital in Thailand
Defined broadly by the World Bank as “the norms and networks that enable
collective action”, social capital is a concept with broad intuitive and operational appeal.
As a rubric encompassing many institutional characteristics important for development
efforts success or failure, social capital represents an important asset for practitioners to
understand and enhance. Taking a practical approach, this evaluation developed social
capital indicators based on inputs from several sources, including recommendations from
the research steering committee,3 staff involved in the Thailand Social Investment Fund,
and an extensive review of the literature. As a starting point for this study, social capital
refers to the internal social and cultural coherence of society, the norms and values that
govern interactions among people and the institutions in which they are embedded. It is
the glue that holds societies together and without which there can be no economic growth
and human well-being (Grootaert and van Bastelaer, 2001).
Social Capital from the Thai Perspective
Thai scholars have long realized the distinguishing characteristics of Thai rural
villages and considered them to be positive social assets. The Thai ‘community culture’
school of thought, composed of prominent academics and NGO activists, have produced
significant research, articles, books, and other literary works since the early 1980s (for
example, Chatthip, 1984; Apichart, 1983; Boonthien, 1984) which constitute a counternarrative to the dominant modernization discourse. They argued that rural village
communities had ideologies, worldviews, social relationship systems and values that
differed from the capitalist culture. This perspective noted with alarm the decline of the
culture-based rural economy resulting from commodification and differentiation.
Napaporn (2003) suggests that the strength of rural communities lies in their
knowledge, social and spiritual capital. Social capital gained its strength from kinship ties
and social networks within and across communities. In addition, Maniemai (2003)
proposes that various mechanisms have maintained traditional networks within and
across villages. Culturally, religious practices help maintain people’s connection.
Ceremonies, festivities and even life-course rituals rely on material and human resources
from within and across villages. Economically, the exchange system - a form of survival
strategy of rural people in the marginalized economy - maintains linkages among villages
depending on who has or lacks resources. In time of crisis such as droughts or floods,
villages elevate the exchange to a moral principle that allows those in trouble to gain
much more from the exchange than they give. In the past, migration for better lands
enlarged people’s connections; today migration for work expands their networks to urban
areas.
Anan (1998) suggested that social capital in Thai society was governed by the
principles of reciprocity and communality. Labor exchange in farming as well as labor
3
This research initiative was requested by the implementing agency for the Thailand Social Investment
Fund. To maintain the independence of the research effort, the World Bank convened a Steering
Committee for the research with representatives from National Economic and Social Development Board,
National Statistics Office, Ministry of Labor, Ministry of Interior – Community Development Department,
and Mahidon University.
3
contribution in village public works are examples of reciprocity based on equality.
Communality principles can be observed in village common property, which all members
have access to and benefit from, following specified rules.
Akin (1997) argued that Thai kinship relations were based on the obligatory
principle. Children are under an obligation to their parents for giving them birth, while
kin are under obligation to each other for the assistance, goodwill or mercy rendered
among them. In many cases, ‘equal’ reciprocity is out of the question, because some
mercy is beyond reciprocity or even beyond measurement. This gives rise to a patronclient relationship system as the primary social institution. Amara (2003) emphasized that
the study of social capital in the Thai context should recognize the importance of the
patron-client relationship.
Social Capital Conceptual Framework
Building on this body of Thai specific literature, the research involved significant
effort from leading Thai academics and practitioners to develop an analytic framework
for understanding social capital. Further, it used a methodology to collect qualitative data
systematically and transform that qualitative data into a reliable set of scores on several
social capital dimensions. An advisory group of leading Thai government officials,
academics, and civil society practitioners provided valuable guidance and advice.
Through extensive consultations, the team reached consensus that the analytic framework
and resulting indicators reflected social capital in Thai society comprehensively and
appropriately.
The analytic framework organizes social capital into three categories: stocks,
channels, and outcomes. Stocks or assets are characteristics that establish an
environment for social relations. The channels or flow of benefits are the primary means
by which social capital stocks operate within communities to facilitate or enhance the
productivity of activities requiring collective efforts. Outcomes are major areas to which
social capital is applied within communities. For the purposes of this research into Thai
social capital, we have established a structural relationship model based on the
assumption that stocks determine channels and channels determine outcomes. Figure 1
illustrates this structural relationships model.
The stock of social capital includes solidarity and trust, groups and organizations,
and networks and linkages. Compared to a village characterized by distrust and selfinterest, a village with the cognitive traits of trust, shared values, compassion and mutual
assistance would more likely embrace cooperation and collective action, and exhibit
better communications and information sharing. Similarly, groups and organizations are
structural features of village social capital through which members cooperate. Similarly,
though somewhat less formally, existing networks and linkages are relational assets that
allow communities to share information, build trust and work together.
Flowing from these social capital stocks, there are important channels through
which social capital operates. In the Thai context, those channels are cooperation and
collective action as well and information sharing and communication. For example, if a
4
community were to tap into bonds of solidarity and trust, it would do so through efforts to
get community members to work together toward some common goal. Further, through
the presence of existing networks or organizations, it could share information that would
make it easier for the community to achieve whatever goals it may have set out for itself.
Figure 1. Model of Social Capital Structural Relationships
STOCK
CHANNEL
OUTCOME
Solidarity
and Trust
Social Cohesion
Cooperation and
Collective Action
.
Group and
Organization
.
Information
Sharing and
Communication
.
Network and
Linkages
Empowerment
The final dimension of social capital is the outcomes arising from the use of social
capital. In the Thai context, these were identified as social cohesion and empowerment.
Information sharing and cooperation could lead to increased social cohesion, for a society
with greater tolerance will more likely include socially marginalized groups and promote
a greater sense of safety, stability and hope for the future. In addition, the social capital
assets and streams of benefits should empower individuals and the community to develop
management capability and the ability to sustain development, to increase political
participation and action, and to influence and control government.
Having decomposed social capital into these eight dimensions, the team then
developed specific social capital indicators to gather evidence about each social capital
dimension. These 34 aggregated indicators and 71 sub-indicators provide the specific
characteristics field researchers investigated to understand social capital characteristics in
Thai villages.
Stocks: Solidarity and Trust
Solidarity and trust are essential elements of social capital. Generally, it takes a
long time for solidarity and trust to develop in a society, though they could deteriorate
more rapidly. These characteristics make up a reservoir of social capital, from which
other elements, such as information sharing or cooperation and collective action, could
flow to benefit society.
5
Figure 2. Social Capital Indicators and Sub-indicators for SIF Impact Evaluation
Dimension
of Social
Capital
Stock
Indicators
1. Solidarity and
Trust
2. Groups and
Organizations
3. Network and
linkages
Channels
(Streams of
Benefits)
4. Cooperation and
Collective Action
Sub-indicators
1.1.1 Collective Actions when Community Members have Problems
1.1.2 Voluntary Cooperation for Common Benefits
1.1.3 Self Sacrifice for Common Benefits
1.1.4 Overall Community Solidarity
1.2.1 Trust within Kin Group
1.2.2 Trust among Close Neighbor
1.2.3 Trust in Community Leaders
1.2.4 Trust in Community Groups and Organizations
2.1 Strength of Membership
2.1.1 Inclusion of Diversified Groups
2.1.2 Voluntary Contribution of Members
2.2 Strength of Leadership
2.2.1 Availability
2.2.2 Diversified Capability
2.2.3 Honesty
2.2.4 Voluntarism and Sacrifice
2.3 Level of Participation
2.3.1 Decision Making Process
2.3.2 Consultation and Debate
2.3.3 Inclusiveness of Diversified Groups
2.4 Organizational Capacity
2.4.1 Effectiveness
2.4.2 Adaptability
2.4.3 Learning Ability
2.4.4 Sustainability
2.4.5 Transparency
2.5 Level of Benefits
2.5.1 Responsive to Needs
2.5.2 Benefit Sharing
3.1 Strength of Horizontal Linkages of Individuals and Households
3.1.1 Breadth
3.1.2 Multi-dimensionality
3.1.3 Benefits
3.1.4 Accessibility
3.2 Strength of Horizontal Linkages of Groups and Community
3.2.1 Breadth
3.2.2 Multi-dimensionality
3.2.3 Benefits
3.2.4 Accessibility
3.3 Strength of Vertical Linkages
3.3.1 Breadth
3.3.2 Multi-dimensionality
3.3.3 Benefits
3.3.4 Accessibility
4.1 Size of People Involved
4.2 Degree of Cooperation
4.2.1 Scale of Cooperation
4.2.2 Diversity of Types of Cooperation
4.2.3 Common-benefit Motivation
4.2.4 Level of Contribution
4.2.5 Outside Resource Tapped
4.3 Inclusiveness and Diversified Groups
4.4 Effectiveness
4.5 Equal Benefit Sharing
6
5. Information
Sharing and
Communication
Outcomes
(Major Areas
of Application)
6. Social Cohesion
7. Empowerment
5.1 Within Community
5.1.1 Between Leaders and Villagers
5.1.2 Among Villagers
5.1.3 Between Groups and Organizations
5.1.4 Equal Access Information
5.2 With Outside Community
5.2.1 Villagers' Access to Production and Marketing Information
5.2.2 Villagers' Access to Development Information
5.2.3 Leaders' Access to Production and Marketing Information
5.2.4 Leaders' Access to Development Information
5.2.5 Voicing of Problems and Needs to Government Agencies
5.2.6 Villagers' Voices Responded by Government Agencies
5.2.7 Adequacy and Timeliness of Information Received
6.1 Tolerance of Differences
6.2 Social Inclusion and Marginalized Groups
6.3 Conflict Management Ability
6.3.1 Personal Conflicts
6.3.2 Communal or Public Conflicts
6.4 Sociability
6.5 Sense of Safety and stability
6.6 Hope for Better Future of the Community
7.1 Capacity Building
7.1.1 Planning
7.1.2 Monitoring and Evaluation
7.1.3 Strength of Multi-party Mechanism
7.2 Ability to Influence and Control Government
7.2.1 To be More Responsive to People's Need
7.2.2 To be More Accountable to People
7.3 Ability to Sustain Development
7.4 Political Participation and Action
7.4.1 Participation in Local and Nation Election
7.4.2 Join or Support Political Parties
7.4.3 Voice Problems to Government, Mass Media or Public for Changes
Stocks: Groups and Organization
Like solidarity and trust, groups and organizations are considered a stock of social
capital essential to the development process, especially the development of marginalized
rural communities. Through groups and organizations, people can mobilize their
diversified strengths, cooperate and exchange information, learn from each other, and
utilize greater bargaining power. The strength of groups and organizations depend on
various factors, many of which are interrelated: strength of membership, strength of
leadership, level of participation, organizational capacity, and the level of benefits gained
by group or community members.
Stocks: Network and Linkages
Networks and linkages can be classified into horizontal and vertical dimensions.
Horizontal networks and linkages of rural people are those that exist among people of
similar social class or social strata, for example fellow villagers in other communities,
both nearby and far away. Conversely, vertical linkages refer to connections with those of
different social classes, for example between villagers and traders, government officers,
academics, and NGOs.
This study considered several aspects of networks and linkages. The first aspect
was the breadth of network. For horizontal networks this was indicated by the extent of
geographical reach (e.g., within a small locality or extending to provincial, regional,
7
national sphere and beyond). For vertical networks, the breadth of network took into
account the key sectors that might be mobilized for people’s well-being and village
development, such as government agencies, non-governmental agencies, academia and
the private sector. The second aspect of network strength was multi-dimensionality,
which involved the extent to which the network responded to the various dimensions of
people’s lives, including their economic life (e.g., production, consumption, marketing),
and their social and cultural life (e.g., education and learning, health, spiritual). The third
aspect was the benefits that people gained from such networks. The fourth was the
network’s accessibility of the network, judging whether it was available only to a
privileged group in the community, or to all.
Channel: Cooperation and Collective Action
Cooperation and collective action is a channel through which social capital was
expressed. In this study, cooperation and collective action referred to the voluntary
gathering of villagers to address common issues, solve problems, or improve the quality
of life. The issues addressed varied, covering several aspects of the people’s lives, such
as helping families in distress, organizing religious ceremonies or rituals, repairing
village roads, building village temples, and constructing or repairing village tap water
systems. The team considered many aspects when assessing strength of village
cooperation and collective action. These included the number of people involved, the
degree of cooperation, the inclusiveness of diversified groups, the effectiveness of the
cooperation or the extent to which the action met the desired targets, and the equal
distribution of benefits gained from collective action.
Channel: Information Sharing and Communication
Information sharing and communication is composed of communication within
the community and communication with those outside the community. To understand
how these channels operate, the research team considered types of information flow
within and across several categories. These included information flow within the village,
with those outside the village, and concerning different realms, such as production,
marketing and development.
Outcome: Social Cohesion
Social cohesion refers to the degree to which a village tends to include all its
members. A society with high social cohesion usually has a tolerance for differences of
religion, ethnicity, language, place of origin, wealth, occupation and incidence of
communicable disease, such as HIV and tuberculosis. Beyond tolerance of differences,
the social inclusion of marginalized groups is important. Cohesive societies should be
able to manage conflicts, either private or public. In a society with high levels of social
cohesion, it is expected that its members would be sociable and actively participate in
social and cultural events. They should have a sense of safety and stability and are
hopeful for the community’s future.
Outcome: Empowerment
Empowerment is an outcome indicator of social capital. The research team
evaluates the level of empowerment based on various factors. The first is the village’s
8
capacity to plan, monitor and evaluate its development projects and development
organizations. In addition, when rural villages are closely linked to the state and market,
it is beneficial for the village to be able to deal with various sectors of the society
effectively. The multi-party mechanism was an important feature of SIF. Multiple parties
(such as the village, government agencies, non-government organizations, private profit
organizations, and the academia) came together in a forum to discuss and act on rural
issues, problems, and development. Other important indicators of the level of
empowerment are the villages’ ability to influence and control the government, to sustain
development, and lastly to engage in effective political participation and political action.
III. Case study of a CDD operation: Thai Social Investment Fund
The Thai Government established the Social Investment Funds as a US$120
million component of a World Bank loan designed to provide relief from the Asian
financial crisis. Channeled through the Social Fund Office of the Government Savings
Bank, the SIF provided resources for local and community grassroots organizations to
implement their development projects. In addition to addressing the social and economic
problems of the vulnerable sectors of Thai society, the SIF attempted to transform the
crisis into an opportunity.
The long-term objective of the Social Investment Fund was to enhance
community-learning capacity for sustainable development through community
empowerment. The purposes of the Social Investment Fund include the following:
1. Revive grassroots society through the use of decentralized procedures so that
communities and localities can participate in development activities;
2. Enhance community organization and local administration capabilities in
administration and management for long-term self-reliance;
3. Promote the emergence of self-sufficient economic systems and strong
community economies; and
4. Stimulate widespread participatory social development by supporting the
development of civic societies and good governance in the long run.
To achieve these objectives, SIF built on available social capital to support highly
desirable reforms towards decentralization, better governance, community empowerment,
and the forging of broad development partnerships involving civil society. SIF gave
resources directly to community organizations. To receive grants, communities followed
sub-project procedures for proposals, management, and monitoring. In addition to
tangible assets that resulted from community development projects, the process of
participating in the SIF was intended to help communities learn by doing, initiating a
process to build institutional capacity and social capital that would strengthen the
community in the long run.
From September 1998 to August 2002, SIF provided funding support into 5
categories (see Table 1) in the amount of 4,401 million baht to projects in 76 provinces.
Based on different menus, eligible project types included:
9
1. Community economy, such as demonstration centers, community grounds and
savings group networks;
2. Community welfare and safety, such as elderly healthcare centers, AIDs shelters
and day care centers;
3. Natural resource preservation and management, environmental protection, and
cultural preservation, such as reforestation, forest fire prevention, flood
prevention, waste water and waste management, and folk museums;
4. Community capacity building and networking, such as community information
center and local handicrafts development; and
5. Emergency community welfare designed to finance immediate community
welfare assistance needs arising from the crisis.
Table 1. Number of SIF Projects and Amount of SIF Funding Support by Menu
Menu
Menu 1:Community economy and community occupation
Menu 2:Community welfare and safety
Menu 3:Natural resource management and cultural preservation
Menu 4:Community capacity building and networking
Menu 5:Community welfare for the needy
Total
Number of
Project
3,184
1,207
790
2,236
457
7,874
Amount of Funding
Support (million baht)
778.08
354.22
193.68
1,059.65
2,015.94
4,401.57
Identified through an outreach campaign, villages were given information about
the SIF and the menu of options. With support and guidance from SIF staff, village
organizations then prepared proposals for funding through one of these menus. The SIF
reviewed these proposals and decided which SIF would fund. Villages then implemented
the sub-projects themselves, with different types of support from SIF depending on the
menu from which the sub-project was chosen.
IV. Evaluation methodology
Overview of Mixed Method Evaluation
Many operations that adopt a Community Driven Development approach have the
objective to build social capital. Further, CDD approaches explicitly work with existing
community social capital. They generally involve competition among village groups that
likely favors villages able to put together better proposals. Given this approach there are
selection and impact effects. Through a selection effect, CDD approaches may act as
mechanisms to identify and reward communities better endowed with social capital.
Since CDD operations require collective action for program participation, communities
with more social capital likely can participate in CDD operations more readily. Second,
through an impact effect, participation in CDD procedures can directly enhance social
capital, in that they help communities identify and develop ways to collaborate more
effectively.
10
Ideally, researchers would have baseline social capital information to know
beforehand how treatment villages that participated in a program differed from control
villages that did not. However, as with many development operations, particularly those
CDD operations aiming to change social capital, the Thai SIF did not have social capital
baseline data. To address this common challenge, the research team innovated with a
rigorous, pragmatic approach to understanding how a community driven development
program worked with and enhanced village social capital.
The research methodology combined quantitative matching techniques with
qualitative field research to identify how and why villages that participated in the Thai
SIF differed from others that did not. The methodology used existing household survey
data to match each of 72 sample villages that participated in SIF to six potential
comparison villages in the same province. Field teams of highly qualified qualitative
researchers then consulted with local authorities and collected additional information
about these potential comparison villages to pair each treatment village with its most
similar match. They then scored each village on a one to five scale for each social capital
variable. To establish statistically significant differences between treatment and
comparison villages, the team analyzed these scores, considering differences in means
and regressions on each social capital indicator. Finding several significant differences,
the team then asked field research teams to provide context for those quantitative
findings, judging whether the observed differences were due to SIF selecting villages
well endowed beforehand with certain social capital characteristics or whether SIF itself
changed those characteristics. As a result, the research allows nuanced conclusions about
how SIF selected villages based on some social capital indicators and had impact on
others. Because field researchers could discuss how SIF had these effects, these
conclusions are more likely to be useful for operations.
Propensity Score Matching
As summarized above, the first quantitative step of this mixed method evaluation
was to match treatment and proposed comparison villages by propensity scores. Through
its Socio Economic Survey (SES), Thailand collects household surveys every two years.
The team based the propensity score analysis on the 1998 and 2000 SES surveys before
the SIF was active. These large sample surveys visited 2112 villages, of which 201 later
participated in SIF4.
The team then created a propensity function that summarizes the relative
importance of chosen indicators in determining whether or not a village participated in
the SIF. It aggregated household data to create village level variables. Using probit
analysis, it identified how much influence the variables had on the probability that a
village participated in SIF. Table 2 illustrates the regression results of the pre4
Though it is not technically difficult (it entails finding overlap between a list of survey enumeration areas
and a list of villages that participated in a CDD operation), this step of identifying SIF villages in existing
household surveys is crucial. To get a large sample of CDD communities for which there is household
data, both the CDD operation and the survey sample need to have covered many villages. If either the
household survey sample or CDD village coverage is small, the number of communities for which
household data is available gets unfeasibly limited.
11
Table 2. Probit regression results of participation in SIF
Village average variable
Age of household head
Age of household head squared
Years of education
Years of education squared
Number of household earners
Head and spouse present in the household
Owns house and land
Professional, technical and managerial profession
Female headed households
Number of children
Employed in the private sector
Economically Inactive
Mean log per capita expenditure
Standard deviation log per capita expenditure
Coefficient
Z
-0.005
0.0001
0.416**
-0.024*
0.058
-0.104
-0.774***
0.123
-0.131
0.173*
-0.355
-0.380
0.573***
-0.864***
-0.110
0.390
2.050
-1.750
0.620
-0.270
-2.940
0.280
-0.340
1.960
-0.650
-0.970
3.330
-3.300
Note the regression includes province dummies. * denotes significance at 10%, ** at 5%, and *** at 1% level.
intervention variables – they are not influenced by participation in the SIF program. The
variables that have proven to be the most important determining which villages
participated in SIF are poverty indicators such as mean per capita household expenditure,
household head characteristics, and education. Province dummies are also included to
control for unobserved heterogeneity at the provincial level.
Figure 3. Pre-match Kernel Densities of
participation propensity
SIF Villages
Figure 4. Post-match Kernel Densities of participation
propensity (6 nearest neighbor within provinces)
Control Villages
SIF Villages
Control Villages
10.0645
4.74254
0
.00171
-.053342
1.04508
-.053342
Pr(sif)
1.04508
Pr(sif)
12
Given this relationship between SES variables and SIF participation, the team
imputed for each SES-sampled villages the propensity to participate in SIF. SIF
treatment villages had higher propensity scores than the average SES village, which
indicates that SIF villages are not “typical”. To identify comparison villages that are
equally atypical to treatment villages, the team matched each of the 201 SIF villages with
six non-SIF villages in the same province that had the closest propensity scores. The
quantitative analysis produced a set of treatment communities, each one matched to six
possible nearby comparison villages. Figure 3 shows kernel densities of SIF versus nonSIF villages, demonstrating that SIF villages had higher propensity scores. Figure 4
shows the distributions of propensity scores for SIF villages and their six nearest
propensity score comparators within provinces.5
Site Selection and Field Matching
Qualitative interview teams investigated social capital characteristics of matched
treatment and comparison villages. Thailand has a strong academic tradition of
conducting qualitative field research using multidisciplinary teams.6 When assembling
the field researchers, the principal investigators were themselves part of this tradition and
were able to involve well-trained and credible field researchers. In the Thai SIF case, the
Table 3. Number of Sample SIF Villages by Sub-regions
Region
Northeast
Central
North
South
TOTAL
Sub-region
North
Central
South
Sub-Total
Central
East
West
Sub-Total
Upper North
Lower North
Sub-Total
SIF Villages with SES data
39
20
29
88
28
7
4
39
17
8
25
12
164
5
Sample of SIF Villages
18
9
12
39
12
3
2
17
7
4
11
5
72
If matching were done without regard to whether comparison villages were in the same provinces, these
propensity score distributions become much more similar. However, because provincial characteristics are
likely important for social capital controls and field logistics, the team chose to restrict comparison villages
to be in the same province as their treatment village.
6
The principal investigators responsible for the field work are leading Thai researchers on village social
characteristics. Much of the field work was conducted using a team from the University of Khon Kaen,
which Professor Robert Chambers noted “did most to establish [participatory rural appraisal’s] credibility,
emphasizing the management of multidisciplinary teams and the techniques and value of semi-structured
interviewing (PRA Note 24, IDS, Sussex).
13
field research teams identified a sample of 72 SIF treatment villages.7 As represented in
Table 3, these selected treatment villages were distributed similarly to SIF villages
overall.
As noted above, the propensity score matching analysis generated six possible
comparison villages to choose from for each treatment village. Field teams used
physical, social and economic indicators to determine the appropriate pair for each
treatment village. The information used for this second stage matching included:
urbanization, distance to nearest town, access to infrastructure, type of terrain, ethnicity,
religion, longevity of settlement, population, out-migration level, and land ownership
structure. This additional information came from the National Statistical Office, the
National Committee on Rural Development and village basic minimum needs data.
Further, to verify the appropriate matched comparison village, field teams met with
officials of the community development, agricultural extension, irrigation land reform,
credit and local development offices, as well as with local NGOs.
Using this field information, field researchers identified 72 comparison villages
matched to the 72 treatment villages. Appendix D Tables 2 and 3 compare differences in
means between treatment and comparison villages for all the village physical, social and
economic variables field teams collected. For nearly all of these variables, the two
groups have similar mean values, suggesting that the combination of propensity and field
matching identified appropriate comparison villages.
Despite this overall comparability, a few variables are significantly different
between treatment and comparison villages: these include the prevalence of a humanmade irrigation system, prevalence of a pre-school nursery, the percentage of people
completing secondary education, and the number of village experts in development. The
first two of these variables, concerning irrigation systems and nurseries, suggest different
pre-existing social capital characteristics between treatment and comparison
communities. Accordingly, we will discuss this issue below when we consider
attributing social capital differences to the CDD operation’s selection effect as opposed to
the impact effect.
Qualitative Field Data Collection
Qualitative data is normally associated with case studies or rapid appraisals that
involve small sample sizes. This research faced major challenge to develop a data
collection methodology and analytical framework that would enable the team to analyze a
large volume of qualitative data to produce credible and representative findings, without
losing its richness, and contextual nature.
Based on the strengths of the field research team and the interest in taking full
advantage of the contextual information, the field research team developed data
collection instruments that captured the many social capital dimensions discussed in the
7
The field team sampled 72 treatment villages from the 201 identified because of budget and logistics
constraints. Due to security concerns, they excluded villages in three Southern provinces affected by
conflict.
14
analytic framework above. The Semi-Structured Interview Guide (Appendix A) poses
discussion questions organized according to the social capital dimensions described
above. Adapted for the Thai case, it is inspired by the Integrated Questionnaire for the
Measurement of Social Capital established by the World Bank. Following those semistructured discussions, each researcher completed an Interviewer’s Rating Form
(Appendix B1), scoring each village on a one to five scale on each social capital
indicators.
The research team paid considerable attention to reducing errors that could enter
due to researcher’s subjective scoring. These efforts included combining field
researchers with differing perspectives, extensive piloting and training, anchoring
vignettes and asking villagers to validate social capital scores.
To provide differentiated perspectives on social capital scores, while all research
team members had significant experience in qualitative research methodology, field
teams reflected a range of different government, civil society, and academic backgrounds.
The researchers individually scored each village on each social capital indicator and then
compared scores. Where there were differences the team would discuss and reach
consensus on a rating. Analysis by Pritchett and Narayan (1995) suggests this is an
effective way to reduce subjective errors. If necessary, additional interviews were
conducted to verify results. These consolidated ratings forms were recorded on a
Consolidated Ratings Form (Appendix B3), where field teams could record discrepancies
and explain through illustrations the rationale for final consolidated rankings.
Field teams were deeply involved in piloting field instruments and were
extensively trained. During field pilots, the research team recognized the need to build a
shared understanding of social capital, the indicators used, and the rationale behind the
indicators. At an initial team leader’s workshop, field team leaders developed a shared
understanding and vision for the social capital analytic framework. Finally, to increase
consistency and validation the same person ultimately trained all field research teams.
When considering approaches to discuss less tangible social capital indicators,
such as trust, field instruments include anchoring vignettes. These short examples
included in the semistructured interviews would frame discussion, facilitating
comparisons across villages and research teams.
As a final means to validate field teams’ social capital scores, six key informants
per village were asked to rate their village on the same social capital dimensions after
having been interviewed (Appendix B2). These random villager ratings were later
compared with the ratings given by the research teams. As presented below, these
random sample of village scores were used to check for the consistency of scores reached
by field teams.
Using all these validation techniques, teams visited 144 sampled villages,
spending on average three days in each village. They conducted interviews with key
informants of village leaders and regular citizens. They interviewed 3 to 5 leaders per
15
village including both formal and informal leaders. In treatment villages this included
people who had important roles in the SIF project. The teams also interviewed 9 to 12
regular citizens, seeking to choose key informants representing major groups, including
distribution by economic status, gender, age, socio-cultural group, housing location and
beneficiary status with respect to development projects.
To verify and augment other data collected on village resources through the SES
and other sources, the field teams worked with the village head to complete a “Village
Resource Profile” (Appendix C). The questionnaire collected quantitative data on
variables that might impact the level of social capital in the sample communities, both
SIF and non-SIF. These include data on human resources and cultural characteristics, the
natural resource base, and development experience. For villages which received SIF
funding support, data on the SIF menu and type of support were also collected.
Ex-Post Quantitative Analysis
Researchers spent three days in 72 treatment villages and 72 comparison villages,
speaking to between 12 and 18 key informants per village. Thus, field teams collected a
large amount of information. While the field team noted extensive illustrations and
explanations for their rankings, the rankings themselves provide a detailed summary of
social capital indicators for each village. They also lend themselves to quantitative
analysis to understand patterns in those indicators. Further, because each village was
originally chosen from the Socio Economic Survey sample, the team had extensive data
on non-social capital variables. Combining this SES data with the Social Capital data,
the research included extensive analysis of the links between socioeconomic
characteristics, participation in SIF, and social capital variables.
There are three parts to quantitative analysis of this village level data. First, the
team identified significant differences between treatment and comparison villages on
each social capital variable. Second, it analyzed the structural model of social capital
presented in the conceptual framework, discerning whether the relationship between
social capital dimensions differs systematically between SIF treatment and comparison
villages. Finally, it used regressions to find determinants of different social capital
variables, using both SES variables from before SIF started and participation in SIF as
potential explanations for social capital outcomes.
Debriefing Workshop for Field Researchers
The quantitative analysis of qualitative field data and existing SES data show
important differences between villages that participated in SIF and comparison villages
that did not. However, because of the lack of social capital baseline data, it was not
possible to identify in the quantitative data selection from impact effects. Further, it is
hard to draw operational conclusions from analysis of differences in means.
Field researchers were a source of very useful additional information. The study
team brought together these researchers for a debriefing workshop. Field researchers
considered the quantitative analysis showing which variables differed between SIF and
comparison villages. They discussed which of these differences were likely due to
16
characteristics that existed before SIF started and which resulted from direct SIF impact.
These qualitative case study examples helped to attribute differences to selection or
impact and gave a better sense for practical implications of SIF operations.
V. Differences between SIF and non-SIF comparison villages
As noted above, field research teams spent extended time in 72 treatment and 72
comparison villages, arriving at consensus rankings for each village on many social
capital indicators. To assess whether and how a CDD operation might have impacted
social capital, the team compared mean scores for treatment villages with those of
comparison villages. Table 4 presents those mean differences, identifying where SIF
village social capital rankings were distinguishable from comparison villages to a
statistically significant degree.
Villages that participated in SIF scored higher on several social capital
dimensions than matched comparison villages. Several social capital stock variables
were higher where SIF was active. Considering solidarity and trust, treatment villages
showed a greater sense of self sacrifice for common benefits and trust among close
neighbors. Their groups and organizations demonstrated greater diversity of leadership
and were better able to learn and adapt to new opportunities. The largest distinction was
in the two groups’ networks and linkages: a summary indicator of several network and
linkage variables was significantly higher in SIF villages than in comparison villages.
Likewise, aggregate indicators of both horizontal and vertical linkages were higher where
there was a CDD operation. Horizontal networks of organizations served multiple
purposes and generated multiple benefits to network members and surrounding
communities. Similarly, SIF villages had a broader, more diverse set of vertical linkages
that were more easily accessible to village members, served multiple purposes, and
generated more diverse benefits.
SIF treatment villages also scored significantly higher on channels through which
social capital was transmitted. Specifically, SIF villages showed a greater diversity of
types of cooperation than comparison villages. In treatment villages, government
officials seemed more accountable, responding to villagers’ voices to a greater degree.
SIF treatment villages differed from concerning social cohesion outcomes.
treatment villages were ranked significantly lower than matched comparison villages on
tolerance for differences. Where SIF had operated, village members showed less
tolerance for community members different than the majority in the village.
17
Table 4. Mean scores in SIF and Non-SIF Villages
Social capital indicators
Mean
3.786
3.813
Mean
3.759
3.771
Paired
differences
Mean
0.027
0.042
3.764
3.806
-0.042
-0.4
3.972
3.764*
3.750
3.759
4.167
3.847*
3.639
3.931
3.556
3.792
3.747
4.083
3.708
3.806
0.042
0.208
-0.042
0.013
0.083
0.139
-0.167
0.4
2.0
-0.4
0.2
1.1
1.7
-1.3
3.380
3.389
-0.009
-0.1
3.530
3.507
3.500
3.514
3.635
3.583
3.542**
3.625
3.792
3.306
3.292
3.333
3.292
3.558
3.653
3.472
3.667**
3.514
3.486
3.646
3.577
3.708
3.450
3.340
3.347
3.333
3.604
3.444
3.296
3.778
3.889
3.338
3.333
3.375
3.306
3.419
3.472
3.375
3.347
3.347
3.556
3.549
3.514
3.583
0.080
0.167
0.153
0.181
0.031
0.139
0.246
-0.153
-0.097
-0.032
-0.042
-0.042
-0.014
0.139
0.181
0.097
0.319
0.167
-0.069
0.097
0.064
0.125
0.9
1.6
1.4
1.6
0.3
1.3
2.1
-1.6
-0.8
-0.3
-0.3
-0.3
-0.1
1.4
1.5
0.9
2.7
1.4
-0.6
1.0
0.8
1.1
SIF Villages Non- SIF Villages
1. Solidarity and Trust
1.1 Solidarity
1.1.1 Collective Actions when Community
Members have Problems
1.1.2 Voluntary Cooperation for Common Benefits
1.1.3 Self Sacrifice for Common Benefits
1.1.4 Overall Community Solidarity
1.2 Trust
1.2.1 Trust within Kin Group
1.2.2 Trust among Close Neighbor
1.2.3 Trust in Community Leaders
1.2.4 Trust in Community Groups and
Organizations
2. Group and Organization
2.1 Strength of Membership
2.1.1 Inclusion of Diversified Groups
2.1.2 Voluntary Contribution of Members
2.2 Strength of Leadership
2.2.1 Availability
2.2.2 Diversified Capability
2.2.3 Honesty
2.2.4 Voluntarism and Sacrifice
2.3 Level of Participation
2.3.1 Decision Making Process
2.3.2 Consultation and Debate
2.3.3 Inclusiveness of Diversified Groups
2.4 Organizational Capacity
2.4.1 Effectiveness
2.4.2 Adaptability
2.4.3 Learning Ability
2.4.4 Sustainability
2.4.5 Transparency
2.5 Level of Benefits
2.5.1 Responsive to Needs
2.5.2 Benefit Sharing
18
t
0.4
0.5
3. Network and Linkages
3.1 Strength of Horizontal Linkages of Individuals and
Households
3.1.1 Breadth
3.1.2 Multi-dimensionality
3.1.3 Benefits
3.1.4 Accessibility
3.2 Strength of Horizontal Linkages of Groups and
Community
3.2.1 Breadth
3.2.2 Multi-dimensionality
3.2.3 Benefits
3.2.4 Accessibility
3.3 Strength of Vertical Linkages
3.3.1 Breadth
3.3.2 Multi-dimensionality
3.3.3 Benefits
3.3.4 Accessibility
4. Cooperation and Collective Action
4.1 Size of People Involved
4.2 Degree of Cooperation
4.2.1 Scale of Cooperation
4.2.2 Diversity of Types of Cooperation
4.2.3 Common-benefit Motivation
4.2.4 Level of Contribution
4.2.5 Outside Resource Tapped
4.3 Inclusiveness and Diversified Groups
4.4 Effectiveness
4.5 Equal Benefit Sharing
5. Information Sharing and Communication
5.1 Within Community
5.1.1 Between Leaders and Villagers
5.1.2 Among Villagers
5.1.3 Between Groups and Organizations
5.1.4 Equal Access Information
5.2 With Outside Community
5.2.1 Villagers' Access to Production and
Marketing Information
5.2.2 Villagers' Access to Development
Information
5.2.3 Leaders' Access to Production and
Marketing Information
19
3.505**
3.362
0.142
2.2
3.785
3.757
0.028
0.4
3.708
3.861
3.847
3.722
3.736
3.764
3.792
3.736
-0.028
0.097
0.056
-0.014
-0.3
1.2
0.7
-0.2
3.510**
3.340
0.170
2.2
3.542
3.500**
3.528*
3.472
3.219**
3.306**
3.194*
3.250*
3.141**
3.864
4.083
3.733
3.903
3.694**
3.917
3.736
3.417
3.817
3.931
3.750
3.546
3.632
3.778
3.903
3.306
3.542
3.460
3.403
3.250
3.375
3.333
2.990
3.028
2.972
3.014
2.931
3.793
4.014
3.619
3.792
3.444
3.873
3.690
3.319
3.681
3.875
3.778
3.500
3.638
3.750
3.958
3.222
3.620
3.362
0.139
0.250
0.153
0.139
0.229
0.278
0.222
0.236
0.210
0.071
0.069
0.115
0.111
0.250
0.043
0.046
0.097
0.136
0.056
-0.028
0.046
-0.006
0.028
-0.056
0.083
-0.078
0.099
1.6
2.7
1.7
1.4
2.2
2.3
2.0
1.8
2.0
1.1
0.8
1.5
1.1
2.8
0.4
0.4
0.9
1.2
0.7
-0.3
0.7
-0.1
0.2
-0.7
0.8
0.6
1.3
3.389
3.444
-0.056
-0.5
3.403
3.264
0.139
1.3
3.639
3.592
0.047
0.6
5.2.4 Leaders' Access to Development Information
5.2.5 Voicing of Problems and Needs to
Government Agencies
5.2.6 Villagers' Voices Responded by Government
Agencies
5.2.7 Adequacy and Timeliness of Information
Received
6. Social Cohesion
6.1 Tolerance of Differences
6.2 Social Inclusion and Marginalized Groups
6.3 Conflict Management Ability
6.3.1 Personal Conflicts
6.3.2 Communal or Public Conflicts
6.4 Sociability
6.5 Sense of Safety and stability
6.6 Hope for Better Future of the Community
7. Empowerment
7.1 Capacity Building
7.1.1 Planning
7.1.2 Monitoring and Evaluation
7.1.3 Strength of Multi-party Mechanism
7.2 Ability to Influence and Control Government
7.2.1 To be More Responsive to People's Need
7.2.2 To be More Accountable to People
7.3 Ability to Sustain Development
7.4 Political Participation and Action
7.4.1 Participation in Local and Nation Election
7.4.2 Join or Support Political Parties
7.4.3 Voice Problems to Government, Mass
Media or Public for Changes
3.792
3.639
0.153
1.5
3.347
3.208
0.139
1.2
3.111*
2.903
0.208
1.9
3.542
3.486
0.056
0.6
3.924
3.875**
3.903
3.576
3.639
3.557
4.319
3.944
3.931
3.426**
3.306
3.236
3.167
3.243*
3.264
3.222
3.563
3.569**
3.565
3.183
3.282
3.957
4.028
3.903
3.688
3.764
3.611
4.306
3.861
3.917
3.280
3.167
3.141
2.986
3.072
3.097
3.042
3.361
3.361
3.495
3.239
3.127
-0.033
-0.153
0.000
-0.111
-0.125
-0.054
0.014
0.083
0.014
0.146
0.139
0.095
0.181
0.170
0.167
0.180
0.202
0.208
0.069
-0.056
0.155
-0.7
-1.8
0.0
-1.1
-1.2
-0.4
0.2
1.0
0.2
2.1
1.4
1.3
0.8
1.8
1.7
1.5
1.6
2.1
1.2
1.0
-0.8
3.394*
3.214
0.180
1.7
Note: ** signifies that difference between SIF and Non-SIF villages is significant at the 5% level and * at 10% level
Finally, several empowerment indicators were stronger in SIF treatment villages.
For example, SIF villages were judged better able to sustain development activities,
demonstrating in the eyes of field research teams better capacity to make productive use
of development opportunities outside of SIF funding. Further, SIF villages were more
empowered to voice problems to government, mass media and the public. These villages
also showed greater strength of a multi-party mechanism, with greater tolerance for
political diversity.
Consistency of rankings across researchers
Individual researchers independently scored social capital variables. Starting with
these scores, each research team arrived at consensus rankings, which are the basis for
this analysis. While individual researchers may systematically give lower or higher
20
rankings, it would be problematic if a researcher’s ranking were, for example, half of the
time higher than her team members and half the time lower. To consider consistency of
rankings, the team performed Analysis of Variance for Repeated Measures to compare
each researcher’s scores for each indicator. There were no significant differences between
scores, implying consistency and intersubjective validity in the scores between field
teams.
Structural Social Capital Differences
Comparing SIF treatment and matched comparison villages on each social capital
indicator highlights several statistically significant differences. In addition, the
conceptual framework implies a structural model of relationships between social capital
variables based on stocks, channels and outcomes. Using Multi-Group Path analysis, the
team tested whether the relationships between those variables were systematically
different between treatment and comparison villages.
Figure 5 Social Capital Structural Relationships for SIF villages
R2 = .60
Solidarity
and Trust
.40*
R2 = .37
Social Cohesion
.25*
Cooperation and
Collective Action
.70*
.63*
.40*
Group and
Organization
.31*
.41*
.42*
R2 = .42
Information
Sharing and
Communication
.45*
.35*
.35*
Network and
Linkages
Empowerment
R2 = .47
SIF Village Model
Summarized in Figures 5 and 6, there are different structural relationships for SIF
treatment villages and non-SIF comparison villages. For instance, for both treatment and
comparison villages, groups and organizations is the only stock indicator influencing the
cooperation and collective action indicator. Villages with strong organizations are more
internally cooperative and better able to mobilize members to work together. However,
in SIF villages this connection is stronger than in comparison villages, as evidenced by
the larger path coefficient (0.63 vs. 0.42). The influences for information/communication
differ more significantly between SIF treatment and comparison villages. While
21
solidarity / trust supported information flow in both types of villages, in SIF treatment
villages networks and linkages also influenced it, while groups / organizations impacted
it in comparison villages.
Figure 6 Social Capital Structural Relationships for non SIF villages
R2 = .49
Solidarity
and Trust
.38*
R2 = .19
Social Cohesion
.20*
Cooperation and
Collective Action
.76*
.42*
.27*
Group and
Organization
.55*
.30*
.40*
R2 = .40
.14*
Information
Sharing and
Communication
.60*
Network and
Linkages
.63*
Empowerment
R2 = .50
Non-SIF Village Model
The structural analysis also highlights important differences between the
determinants of social capital outcome indicators in SIF treatment vs. comparison
villages. In each set of villages, social cohesion was influenced by both trust and
collective action. Apart from these determinants of social cohesion, networks mattered in
SIF villages, while information flow influenced it in comparison villages. The
determinants of empowerments differed completely in SIF vs. comparison villages. In
SIF villages, information flow and social cohesion influenced empowerment, whereas in
non-SIF villages, network and cooperation / collective action influenced it.
Empowerment resulted from different processes in these SIF treatment as opposed to
comparison villages. Taking all the connections together, the structural model for SIF
villages differed to a statistically significant degree.
Regressions on Social Capital Determinants
To buttress the analysis of differences, the study also included regressions. The
regressions were of the form YN = α+ β SES + γ SIF + ε where YN is each of the social
capital variables as a dependent variable, SES is the set of socio-economic variables, and
22
SIF is an indicator for treatment communities. The independent control variables8
included the expenditure level, inequality in the village, the share of households working
in agriculture and of households owning their own land for farming, and the average
years of education in the village.
Presented in Table 5, those regression results verify that SIF villages differed
from matched comparison villages on several social capital indicators, even when
controlling for these socio-economic characteristics. Nearly all the social capital
variables, except one, tolerance of differences, were affected positively by participating
in SIF. Of the 20 village social capital variables whose means differed to a statistically
significant degree in SIF versus comparison villages, 17 remained significantly different
after introducing these controls, as measured by whether a dummy variable on SIF
treatment was statistically significant. The social capital indicators where there ceased to
be a significant impact of SIF treatment included trust among close neighbors, the
tolerance of differences, and the capacity to voice problems to authorities. However, SIF
had a significant impact on several social capital indicators that did not emerge from the
differences in means analysis. These included the organizational capacity and
effectiveness of organizations, information sharing and communication outside the
village, village planning capacity, and ability to influence and control government.
Table 5 also provides information about how other village socio-economic
characteristics affect with social capital. More villagers working on agriculture was
associated with higher social capital indicators, as evidenced by the many positive
significant coefficients on the variable measuring share of agricultural workers.
Conversely, when more households own their land for farming, social capital is lower.
Together, these two findings suggest that villages where many people are tenant farmers
tend to have higher social capital. In general, village expenditure levels had few effects
on social capital: the only significant effects were that higher consumption villages in the
sample exhibited more trust of close neighbors and lower consumption villages were
more able to cooperate. However, more inequality in the village was associated with
higher social capital in many dimensions. When coupled with the agriculture findings
this evidence suggests that social capital norms operate as a means of providing
insurance, so that particularly in places with tenant farming and unequal consumption,
cooperative norms are more prevalent. Finally, the evidence suggests that more
education was associated with less social capital, so cooperative norms and networks are
maintained and valued by the less well educated.
8
More precisely, the independent control variables were the mean of log household per capita
consumption, the standard deviation of log household per capita consumption, the village average of the
household head’s years of schooling, the share of head of households working in agriculture and the share
of households that owned their own farms.
23
Table 5 Regression results
Log mean
per capita
expenditure
Log SD per
capita
expenditure
Share of
worker in
agriculture
sector
Years of
education
household
head
0.236**
-0.130
0.712**
1.360***
-0.062
-0.737***
0.151
0.298**
0.351***
0.167**
0.253*
-0.108
-0.148
-0.080
0.023
0.374
0.332
0.495**
1.143***
0.829*
0.526
0.997**
-0.048
-0.184**
-0.036
-0.068
-0.786***
-0.789***
-0.757***
-0.570***
0.200**
-0.201
0.615**
0.952**
-0.068
-0.588***
0.272**
0.186*
0.263**
0.304**
0.250**
0.276*
0.240*
-0.167
-0.078
0.045
-0.011
0.175
0.034
-0.008
0.462
0.558*
0.403
0.608
0.193
0.434
0.338
1.172**
0.931**
1.073**
1.402***
1.652***
0.591
0.557
-0.073
-0.089
-0.147*
-0.098
-0.207**
-0.125
-0.148*
-0.581**
-0.605**
-0.558**
-0.623*
-0.600**
-0.620*
-0.344
0.231**
-0.412***
0.053
2.2***
-0.000
-0.734***
0.27**
0.10
0.76**
0.52
-0.20**
-0.39
-0.139
0.181*
0.093
-0.026
0.474*
0.447*
0.493
1.122***
-0.043
-0.145**
-0.089
-0.580**
0.232*
0.043
0.612
2.011***
-0.194**
-1.143***
0.225*
0.196
-0.178
0.348
0.871*
-0.075
-0.347
-0.119
0.338
1.347***
-0.215**
-0.549*
0.233**
0.220*
-0.281*
-0.219
0.757**
0.905**
0.702
0.315
-0.083
-0.048
-0.603**
-0.353
0.225**
0.100
0.523***
0.162
-0.171**
-0.298
0.222*
-0.090
0.560*
0.923*
-0.135*
-0.656**
0.213*
0.137
0.402
1.295**
-0.208**
-0.723**
Social Capital indicators Significant
SIF
from difference in means analysis
1.1.3 Self sacrifice for common
benefits
1.2.2 Trust among Close Neighbors
2.2.2 Diversified Capability
2.4.3 Learning Ability
3. Network and Linkages
3.2 Strength of Horizontal Linkages
of Groups and Communities
3.2.2 Multi-dimensionality
3.2.3 Benefits
3.3 Strength of Vertical Linkages
3.3.1 Breadth
3.3.2 Multi-dimensionality
3.3.3 Benefits
3.3.4 Accessibility
4.2.2 Diversity of Types of
Cooperation
5.2.6 Villager’s Voices Responded by
Government Agencies
6.1 Tolerance of Differences
7. Empowerment
7.1.3 Strength of Multi-party
Mechanism
7.3 Ability to Sustain Development
7.4.3 Voice Problems to Government,
Mass Media or Public for Changes
2.4 Organizational capacity
2.4.1 Effectiveness
5.2 Information sharing outside
community
7.1 Capacity building
7.2 Ability to influence and control
government
***denotes statistical significance at 1%, ** at 5%, and * at 10% levels.
24
Own
farming
land
Separate regressions provide information on the impact of particular types of SIF
involvement. In some treatment villages, SIF provided training to build the capacity of
local organizations, directed towards making them more effective. SIF also provided
network support, helping to facilitate connections between local implementing
organizations. With data on whether or not training or network support occurred in
treatment villages, it is possible to ascertain whether there is marginal impact of these
specific modes of support on social capital outcomes. Figure 7 presents the findings of
regressions structured to isolate the marginal impact of these components on different
social capital indicators. The variables arrayed along the bottom summarize those social
capital variables for which either training or network support were statistically
significant.9
Network support had significant impact on seven social capital variables: social
inclusion of marginalized groups, organizational capacity, information sharing outside
the community, strength of horizontal linkages, strength of vertical linkages, level of
benefits of groups and organizations, and equal sharing of benefits from cooperation and
collective action. When SIF brought village groups together to share experiences, the
resulting comparisons had important impact on village social capital.
SIF villages that participated in capacity building also had significantly higher
social capital scores than those that did not. The variables affected include: effectiveness
of cooperation and collective action, information sharing outside the community, the
strength of horizontal and vertical linkages, political participation in elections, strength of
group memberships, ability to influence and control government, and the level and equal
sharing of benefits. The quantitative analysis shows that SIF villages differed from
comparison villages on several social capital dimensions. Further, through analysis of
relationships between training and networking activities, insights emerge into how SIF
may impact different village characteristics. However, without baseline information, it is
difficult to use quantitative techniques to distinguish selection from impact effects.
9
These regressions included the social capital indicators as dependent variables and indicators for SIF
treatment and either training or network support as independent variables. Accordingly, because the
training or network support variables are interaction terms, the statistically significant coefficients reported
in Figure 7 reflect the marginal effect of these types of support, above and beyond whatever differences are
attributable to SIF treatment overall.
25
Figure 7. Effects of Training and Network support in SIF villages
Effects on social capital of different types of support in
SIF villages
Network support
Coefficient value
Training support
0.6
0.5
0.4
0.3
0.2
0.1
0
Social inclusion Effectiveness of Orgnaizational
of marginalized cooperation and
capacity
group
collective action
Information
sharing with
outside
community
26
Strength of
Strength of
Political
horizontal
vertical linkages participation in
networks and
election
linkages at
community level
Strength of
membership in
group and
organization
Ability to
Level of benefits Equal sharing of
influence and of groups and
benefits from
control
organization cooperation and
government
collective action
VI. Attribution and Operational Implications
To identify whether SIF selection or impact explains observed social capital
differences, field researcher discussed the social capital variables that differed between
treatment and comparison villages in a debriefing workshop. As noted above, these
researchers were well trained in participatory rural appraisal and spent several days in
each of the 144 villages, collecting qualitative data about social capital, development
initiatives, and SIF through interviews with key informants. Thus, they could give
credible insights to understand whether observed social capital differences were there
before SIF or arose due to SIF activities. Further, they could suggest how SIF procedures
selected and impacted village social capital characteristics.
SIF treatment villages were more likely to show self-sacrifice for common
benefit. Field researchers agreed that treatment villages likely started out with more of
this type of solidarity. In Thailand, there are longstanding social norms of village culture
wherein village members must maintain village “face”. SIF participant villages had
enforced that norm more successfully. One sign of their success would be an ability to
organize village members to prepare a successful SIF proposal. Those administering SIF
would be more likely to approve projects from villages where this norm was respected
and in evidence.
Researchers felt that differences in self-sacrifice should not be fully attributed to
SIF selection effects. SIF procedures required community contributions for
implementation. Thus, SIF participation reinforces the value of self-sacrifice for
common benefit, which helps to explain why self-sacrifice was more prevalent in SIF
treatment villages. Researchers attributed these differences in solidarity both to selection
and impact effects.
Trust among close neighbors is a longstanding characteristic of villages that takes
a long time to develop. For example, there is a long and strong tradition of “kum” or
focus on village solidarity that researchers found in greater evidence in SIF villages,
particularly in the Northeast of the Thailand, While SIF villages generally exhibited
higher trust than comparison villages, researchers attributed these differences to traits SIF
could not impact in its short period of operation. SIF selected villages where neighbors
already trusted one another. Such trust would allow those villages to put together
stronger proposals.
The analysis above shows that SIF treatment villages had greater diversity of
leadership capability. Research teams attributed most of this difference to SIF impact.
To prepare and implement a SIF sub-project, villages need effective leaders who can
convince and inspire fellow community members. Moreover, these leaders must be
informal or outside the formal administrative structure. SIF supports the emergence of
diversified leaders. While there were likely potential leaders in many villages, SIF helps
them emerge, encouraging them to explore channels outside of formal administrative
procedures to get things done.
27
Organizations in SIF
villages learned new approaches
more easily than elsewhere. Again,
research teams attributed these
differences to the impact of SIF
operations. SIF presented villages
with several menus. Villages
decided which opportunities would
be most appropriate to their
interests. Further, given that
village organizations initiated SIF
projects, they learned by doing
these operations. Finally, SIF
focused on transferring knowledge
and experience among villages and
grass-roots organizations. That
sharing opened up organizations’
interest in learning about what
approaches work best. Thus,
observed differences in village
learning ability were not based on
SIF selection effects, but rather
resulted from SIF involvement.
Kokgong village in Udornthani province is a
remote village. It would have been difficult for the
village to achieve the development that has occurred
without participating in the SIF program. The village is
located in a new frontier, and is surrounded with sugar
fields. Its inhabitants have a different ethnicity from the
natives in the area. They are ethnic Phu Tai, who
migrated from Sakonnakorn province. They were able
to preserve much of their ethnic culture, including skills
in silk and cotton weaving. An officer of the Tambon
Administration Organization helped the villagers in
drafting a project proposal to get funding support from
SIF for a cloth weaving project. SIF gave not only
financial, but also other technical support.
The products of the weaving group made
possible by SIF’s support won much recognition from
local government agencies. Group leaders were \ invited
to participate in several training sessions, such as those
on group management (planning, accounting),
production (dyeing, packaging), marketing, as well as
on extended activities, such as how to start a savings
group. Through their participation in these sessions,
they came to know and got acquainted with other
weaving and production groups as well as resource
persons from GOs, NGOs and the private sector. These
new contacts enabled them to have wider market
channels. It is notable that such a small project offered a
range of learning opportunities to both group leaders
and members, resulting in more diversified capabilities.
As a central operational
tenet, SIF sought to build
information networks among
villages and organizations. Those
efforts seemed quite productive.
As discussed above, nine network
and linkage indicators were significantly different between treatment and comparison
villages. Field researchers attributed these observed differences to SIF impact, rather than
to selection. SIF encouraged learning connections between organizations and villages to
understand what worked best. For example, SIF organized and financed study tours
among villages, so they could share approaches as to what worked and what did not.
Further, when organizations identified, planned and implemented sub-projects, they
gained opportunities to work with other similar organizations and to interact with village
authorities.
SIF villages showed more evidence of cooperating on a diverse set of activities
than comparison villages. Field researchers attributed these differences to pre-existing
village characteristics that enabled villages to be more successful in organizing SIF subproject proposals. As further evidence of this selection effect, cooperative activities
prevalent in SIF villages tended to be more traditional and culturally based, so that a
pattern of village members working together on such activities likely existed before SIF.
28
In Buddhist villages belief in the Law of Karma led to social tolerance:
“In the village there was only one disabled. He got welfare money from the Tambon
Administration Office. If there is common work that requires labor, he is exempted. But he
usually joins the festival and religious activities. Villagers would teach their children not to
tease him, or insult him. They believe that the bad karma would fall on those who treat those
who are inferior to them badly.”(Nontasaeng village, Udornthani, non-SIF)
“There is a dumb man in our village. We teach our children not to tease him. He is dumb
because it is his karma. That’s how we teach our children.” (Sijaenoi village, Udornthani,
non-SIF)
In a Muslim village, discrimination is considered a sin, and caring of the marginalized
has become a regular practice:
“In our village there are helpless elders, the disabled, people with AIDS, and those who are
paralyzed. Most of the villagers follow our religious principle that all Muslims are brothers
and sisters. When somebody is in need, we should help each other. Discrimination is a sin.
Every Friday, after our religious practice, we will pay a visit to these people. Some walk,
others bicycle. Our ex-headman is paralyzed. We feel sorry for him. We also visit the man
with AIDS.(Natub village, Nakornsithamarat, non-SIF)
People with different origin or ethnicity are accepted, some were elected leaders:
“The in-laws from the northeast region, both male and female, were well accepted. They
were elected village leaders” (Wangkran village, Supanburee, non-SIF)
(Note:The village is in the central region, with different ethnicity from the northeasterner)
An old woman in a village in the Southern region said : “ I used to live by the railway,
collecting garbage. The village head and the people in this village invited me to live in the
village, at the headman’s house. People in the village hired me to do various jobs, such as
cutting grass. The headman’s neighbors gave me food everyday. Some people bought me
soap, powder and clothes”.(Klongpiplae village, Nakornsithamarat, non-SIF)
“In the village we have a rule that one should not do a job in competition with the disabled.
The blind in our village works as a massager, and another disabled person who cannot walk
makes fish nets”. (Nadom village, Udornthani, non-SIF)
“Every member of our cooperative shop is competent and can replace each other. Everyone
can take care of the shop. Even if the seller is not present in the shop, the buyers may take the
goods and just put the money on the counter. When the co-op leader is away, members, or
the housewife group, or school teachers can make decisions or carry out the work without
any probems.” (Thon village, Udornthani, SIF)
As additional verification that SIF selects villages with a proclivity for collective
action, note significant differences between treatment and comparison villages
concerning the prevalence of human-made irrigation systems and of pre-school nurseries
(See Appendix D Table 3). Organizing community members to maintain irrigation
systems is a quintessential form of collective action that requires community capacity to
work together over a long time horizon. The percentage of land under irrigation is a
strong proxy for social capital. Likewise, if the community organized itself to provide
29
pre-school nurseries, it suggests strong cooperation. Both qualitative and quantitative
evidence suggests that SIF villages began with a greater capacity to cooperate to
undertake joint activities.
SIF villages shared information with those outside of the community more
effectively. Further, in SIF villages government officials were more responsive to
villagers’ voices. Field research teams attributed these differences to SIF impact.
Implementing SIF sub-projects, organizations needed to share information and work
together. Resulting partnerships improved communication with other communities and
with local government officials. As Figure 7 illustrates, network support which SIF
provided had a significant impact
on information sharing: when
Chiengwae village at Udornthani province is a
organizations compared notes on
good example of how SIF support, civic networks and
their operations, they learned from
multi-party forums have made people become more
the examples of their peers.
confident and able to act together to voice and solve their
problems.
The village proposed a project to SIF with the aim
to solve and prevent drug usage among village youths, and
the project was approved. Under this project, village
leaders collaborated with the village school in several
activities, such as sports, drug education, rehabilitation,
and occupational training. This project allowed the village
to establish a ‘civic network against narcotic drugs in
Udornthani province’, and was able to expand the network
to other provinces.
This civic network later became involved in
organizing civic forums to solve environmental problems
arising from a large scale irrigation project. The part of
the project that affected the village was the construction
of a feeder canal from Nong Harn lake to the sub-district
land areas. The canal however obstructed the natural
waterways and caused floods in both farms and settlement
areas. The civic network therefore organized civic forums
with several local organizations. Together they requested
financial support from the Tambon Administration
Organization (TAO), and was able to get a grant to solve
the problem. They also mobilized people in the subdistrict to contribute their labor to construct new feeder
canals and drainages to change the waterways. This effort
was successful, and the network expanded. Now, people
can use the forum of this network to plan and discuss
issues that arise.
10
As noted above, SIF
treatment villages were less
tolerant of differences than
comparison villages.10 Field
researchers attributed this reduced
tolerance to SIF activities. They
argued that preparation and
implementation of SIF subprojects creates an atmosphere
focused on achieving and
adhering to project goals. In this
atmosphere, SIF villages were less
tolerant of including those of
lower capabilities, whether those
people be poorer or of a different
language or ethnic group.
These findings correspond
to other research on community
driven development approaches in
the region. Notably, research on
the Kecamatan Development
Project (KDP) in Indonesia
investigated the effect of CDD
operations on local level conflict.
It found that CDD operations
introduced sources of conflict into
The finding that treatment villages had lower social cohesion scores than treatment villages is statistically
significant in the difference in means. However, it is not significant in the regression findings, suggesting
that socio-economic characteristics explain part of this difference, rather than SIF participation.
30
the community: it can be disruptive to make available extra resources through new
channels. However, because that operation stressed clear and transparent operating
procedures, it also provided a means to manage those conflicts without them becoming
too heated. In support of the Indonesia CDD research, there is evidence of reduced
social cohesion in Thai SIF villages as opposed to comparators. However, field
researchers’ explanation for the effects on cohesion in SIF treatment villages suggests a
different source of potential tension, one which may operate through social exclusion.
SIF villages appeared more empowered than comparison villages. They showed
greater ability to sustain development, were more effective in voicing their problems to
authorities and advocating for change, and demonstrated stronger multi-party tendencies.
Field researchers attributed these differences largely to SIFimpact, rather than to
conditions that existed beforehand. As noted beforehand, SIF promoted networks within
and among villages. These allow villagers to articulate political voice, for networks form
to design and implement initiatives outside of formal government procedures. Further, as
noted above, SIF operations generate opportunities for informal leaders to emerge and
build confidence in their capacity to get things done. In that they may counterbalance
traditional formal authorities, these leaders also provide an avenue for village voice and
empowerment. The evidence collected in this study suggests that SIF operations had a
direct impact on village empowerment.
VII.
Conclusions
This research reflects applies an innovative ex-post assessment of how the Thai
Social Investment Fund worked with and changed village social capital. It contributes
social capital and community driven development operations in three ways. First, it
provides a compelling example of the value of breaking social capital into interrelated but
distinct dimensions and adapting those dimensions to the country or regional context.
Second, it offers an innovative, practical approach to ex-post evaluation of social capital
and community driven development operations, making full use of quantitative and
qualitative techniques to assess whether and how the Thai SIF selected and changed
social capital traits. Finally, it identified several social capital aspects that differed
between SIF treatment and matched comparison villages and explains the source of those
differences.
Disaggregated Social Capital Indicators
Referring to the norms and networks that enable collective action, social capital
can apply to a broad set of village traits. Thus, it is more precise and operationally useful
to separate social capital into component parts. With this disaggregation, one should
identify proxies and indicators adapted to the local context.
In this study, that approach worked well. First, partnering with accomplished
qualitative researchers well-versed in Thai village dynamics, the team established
dimensions of social capital relevant to the Thai context. Those consisted of stock
dimensions that are characteristics of village institutions and norms, channel dimensions
that are means through which social capital operates, and outcome dimensions which
31
refer to the ends to which social capital assets are applied. Within these three categories,
the research investigated the stock variables of solidarity and trust, groups and
organizations, and networks and linkages; the channel variables of cooperation and
collective action, and information sharing and communication; and the outcome variables
of social cohesion and empowerment. To capture these, the research identified indicators
valuable for understanding social capital across Thailand’s different regions.
Using the Thai SIF as an example, this study shows how CDD operations work
with existing social capital characteristics in communities. But many CDD operations
also have as development objectives to change social capital characteristics. With several
social capital indicators, we can better discern CDD operations’ selection from impact
effects. Evidence from this study shows that CDD operations likely act as selection
mechanisms among communities, allowing those well endowed with particular social
capital characteristics to receive program funding. Further, as discussed below, CDD
operations likely change certain village dynamics. Presenting several dimensions of
social capital and separating selection from impact effects, this study offers an example
for more precise thinking about how CDD operations operate and what they can achieve.
Practical Methodology for Rigorous Ex-Post Impact Assessment
Ideally, researchers would have baseline social capital information to know exante how treatment villages differed from control villages. However, as with many
development operations, particularly those CDD operations aiming to change social
capital, the Thai SIF did not collect baseline information about disaggregated social
capital indicators for treatment and control villages. The research team innovated with a
rigorous, pragmatic approach to understanding how a community driven development
program worked with and enhanced village social capital.
The research methodology combined quantitative matching techniques with
qualitative field research to identify how and why villages that participated in the Thai
SIF differed from others that did not. The methodology used existing household survey
data to match each of 72 sample villages that participated in SIF to six potential
comparison villages. Field teams of highly qualified qualitative researchers then
consulted with local authorities and collected additional information about these potential
comparison villages to pair each treatment village with its most similar match. After
three person teams spent three days in each village, they scored each village on a one to
five scale for each variable. To establish statistically significant differences between
treatment and comparison villages, the team analyzed these scores, considering
differences in means and regressions on social capital variables. Finding several
significant differences, the team then asked field research teams to put those quantitative
findings in context, judging whether the observed differences were due to SIF selecting
villages well endowed beforehand with certain social capital characteristics, or whether
SIF itself changed those characteristics.
The methodology is promising. It is rigorous, despite the lack of baseline data. It
gives insights into the selection and impact effects of CDD on social capital variables.
And it offers useful operational implications. Baseline data that includes social capital
32
indicators will eventually yield even stronger evidence on CDD operations and social
capital. However, absent such solid baseline information, this research demonstrates
opportunities to draw credible, operationally useful conclusions from current experience.
To take advantage of those opportunities, evaluators need to be pragmatic. They should
not be doctrinaire about what constitutes an acceptable standard of evidence, so that they
fail to learn from other valid evidence. It is imperative to learn as much as possible from
as many different sources of data as possible, combining quantitative matching methods
based on household data with qualitative research methods that can give insights into
causation and other subtle interactions within communities.
While the methodology is promising, certain conditions made it particularly
effective. First, the government expressed strong interest in understanding how to
enhance village social capital: this generated a serious engagement from government
officials and high-quality academic researchers to define social capital in the Thai
context. Second, Thailand collects extensive household data at regular intervals. The
overall sample for the Socio-Economic Survey is a repeated cross-section of 20,000
households visited every two years. Therefore, there was extensive overlap between
villages that had been sampled for the SES and those where the SIF later operated. This
overlap allowed for quantitative matching analysis that enhanced evaluative rigor.
Finally, Thailand has a strong academic tradition of multidisciplinary research to collect
qualitative evidence, particularly pertaining to village characteristics and their
implications for development efforts. This research benefited greatly from partnership
with very well-trained, experienced Thai field researchers. The Thai researchers had
credibility to understand social capital dimensions in Thai villages, to score villages
involved in the research according to social capital indicators, and to attribute observed
differences to either selection or impact effects.
A CDD Operation’s Selection and Impact Effects on Social Capital
The research produced interesting and relevant findings about the how the Thai
SIF built on and enhanced village social capital. First, after correcting for numerous
village characteristics through extensive matching and regression analysis, SIF villages
differed from matched comparison villages on several social capital indicators. Some
differences resulted from SIF processes selecting villages better endowed with certain
social capital characteristics. Specifically, to begin with, SIF villages had more trust
among neighbors and stronger traditions of cooperation and collective action.11 This is
not surprising: participation in SIF requires community members to work together and
commit to implementing proposals that entail a degree of self sacrifice for common
benefit.
This selection process likely happens “naturally” in that the requirements for SIF participation, i.e., the
need to have group meetings to discuss the village priorities, to commit village resources as cofinancing for
the effort, etc. However, it may be that SIF administrators were biased in their selection of which villages
ultimately received funding and intentionally made resources available to villages with a reputation for
having greater solidarity or trust. Even with this possibility, these villages demonstrate to other nearby
villages the value of those characteristics.
11
33
Because it follows a procedure that selects villages well-endowed with social
capital, SIF can be viewed as regressive. If trust and cooperation are valuable attributes,
one could argue that development programs should target villages that lack those
attributes and seek to enhance them. While this argument is logical, this research
suggests it may not be practical. Ex-ante, it is may not be easy to identify with clarity
low trust or particularly uncooperative villages. Further, if CDD operations rely on social
capital characteristics for sub-project success, those approaches should not be targeted to
villages that lack them. Finally, it is not entirely evident how to change deep seated
norms of trust and solidarity, particularly within villages with no history of them.
Ironically, CDD operations may offer indirect ways to affect trust and solidarity
norms. SIF rewards villages with higher trust, cooperative traditions, or self-sacrifice
norms. If those characteristics are socially desirable, SIF reinforces their value: SIF
villages demonstrate that working together effectively attracts resources and benefits
villagers. SIF identifies effective villages and creates appropriate incentives for villages
that are not so well endowed with cooperative norms.
The study also provides evidence of village characteristics associated with social
capital. For example, regardless of their participation in SIF, villages where more people
work in agriculture and fewer own their farms were better endowed with several
dimensions of social capital. This evidence suggests a spirit of cooperation emerges from
working together on farm necessities. And, while income levels made little difference,
villages with more unequal income generally had more social capital, suggesting an
insurance motive in that cooperative norm. Since more education was associated with
less social capital, there is evidence that these cooperative norms and networks are
maintained and valued by the less well educated.
Finally, the paper provides evidence that SIF had an impact on several social
capital variables. While SIF selected villages that already had a greater norm of selfsacrifice, it also enhanced that characteristic by demonstrating the benefits of sacrificing
to improve village welfare. SIF activities helped build local leadership, through its
support for networks and training and through a procedure that encourages leaders to get
things done outside of the formal government system. The intervention itself had impacts
on many dimensions of network and linkages within and among villages and how
information flows between villagers and government officials. These impacts likely
result from concerted efforts to create horizontal links between organizations and vertical
links to formal authorities. The research found that SIF activities may have reduced
social cohesion, particularly through its focus on reaching clear objectives, which seemed
to exclude those viewed as least effective. Finally, through its focus on networking and
leadership strengthening, SIF changed relations with government officials, empowering
villagers to identify alternative sources of influence, express their voice to those officials,
and sustain development activities through their own initiative.
Through invitation of the Thai government, this research has identified
differentiated and valuable social capital indicators for Thailand and the region, has
developed a pragmatic methodology for understanding the relationship between village
34
social capital and community driven development operations, and has identified several
ways that the Thai SIF selected villages with strong social capital and enhanced several
social capital dimensions. Through these successes, it opens up opportunities for a
broader and deeper understanding of social capital and CDD in the region.
35