Figure 3.1 Layered analyses
Source: Adapted from von Lübke (2014a, 2014b).
A multi-layered analysis
Comparison as a research methodology has often been challenged by scholars with Area Studies or Cultural Studies backgrounds.1 They question whether social phenomena in highly culturally diverse settings are amenable to a comparative research methodology and theoretical concepts that have mainly been developed in the West. However, more recent work on the methodology of comparison has refuted such reasoning and has developed new avenues to make comparative research a fruitful endeavor in non-Western research contexts (Huotari & Rüland 2014, 2018). Two promising approaches to overcome this sterile debate between disciplines and Area Studies are the Comparative Area Studies (CAS) approach (Basedau & Köllner 2007; Ahram 2011; Berg-Schlosser 2012; Huotari & Rüland 2014, 2018; Ahram, Köllner & Sil 2018) and the concept of mixed methods (Creswell & Plano Clark 2007; von Lübke 2014a, 2014b), both of which are employed in this study. Comparative Area Studies (CAS) operates at three levels: intra-regional, interregional, and cross-regional comparison (Basedau & Köllner 2007). By focusing on two conflict regions in the Philippines and Indonesia, the current study rests on the intra-regional level of comparison with a cross-case and a within-case dimension. Cross-case comparison refers to the two countries under consideration, the Philippines and Indonesia. Within-case comparison includes a synchronic and a diachronic dimension (George & Bennett 2005; Gerring 2008). The synchronic dimension refers to the counterfactual comparison of two villages with church-based projects and two villages without church-based peace projects in both project sites: Cotabato in Mindanao, Philippines, and Ambon City in Maluku, Indonesia. The diachronic dimension entails comparisons of respondent attitudes prior to church-based peace projects and after church-based peace activities.
A mixed-methods research design cuts across narrow disciplinary silos and confines and, in doing so, holds the promise of complementing different viewpoints and granting novel insights (Creswell & Plano Clark 2007: 1, 169). It is a methodology that is embedded in philosophical assumptions guided by pragmatism with the central premise “that the use of quantitative and qualitative approaches in combination provides a better understanding of research problems than either approach alone” (ibid.: 5). Pragmatism means the freedom of a researcher “to use all methods possible to address a research problem” and “to combine inductive and deductive thinking” (ibid.: 10). By opting for a mixed-methods approach, we have been strongly inspired – as indicated in the introduction – by Katzenstein’s “analytic eclecticism” (Katzenstein 2007; Katzenstein & Sil 2008). In particular, we expect that the complementary nature of different research methodologies and the different worldviews or paradigms they represent, helps to mitigate misperceptions that are inevitable when researchers study complex social phenomena in settings which culturally differ from their own background. Moreover, combining the “unit context sensitivity” of qualitative research methods with the “population context sensitivity” of quantitative methods (Pepinsky 2014) allows for better assessment of the merits and relevance of individual cases against the background of a larger population of cases. In other words, mixed methods help to map developments and measure “trends, prevalences, and outcomes and at the same time examine meaning, context, and process” (Creswell & Plano Clark 2007: 175). Apart from cross-fertilizing qualitative and quantitative viewpoints, our methodological approach also includes different spatial levels of conflict and peacebuilding. The joint focus on national and local domains – which will be outlined in the following section – also resonates well with key themes in the conflict and peacebuilding literature. It allows for a more balanced evaluation of whether peacebuilding efforts resemble a “local turn” or approximate a “third culture,” not least because a multilevel assessment places local agency squarely into the wider context of national events and discourses.
The regional focus of our research project has been placed on two Southeast Asian democracies: the Philippines and Indonesia. Both countries lend themselves well to comparative analyses of conflict transformations and the contributions of church-based activities, not least because they feature a notable degree of cross-country comparability that stems from similar sociocultural and political underpinnings.
The Philippines and Indonesia share distinctive properties beyond the fact that they belong to the same geographical region: both are emerging democracies with a distinct history of authoritarian rule, which some scholars have characterized as “defective” Third Wave democracies (Croissant & Bünte 2011); both are archipelagic states with highly diverse ethnic, religious, cultural, and sociopolitical structures; both feature numerous instances of vertical (insurgency-related) and horizontal (violent political competition and armed struggles between local elites, clans, ethnic groups, and rival insurgent groups) intergroup conflict (International Alert 2014); and (in response) both have attracted substantial interest and commitment from religious organizations in the field of peace and reconciliation.
These similar underpinnings are also reflected in aggregate indicators, such as the Bertelsmann Transformation Index (BTI). The BTI assessments indicate that the two countries evince similar economic and political governance measures: the Philippines and Indonesia rank thirty-sixth and thirty-fifth, respectively. Comparable performance levels are also evident across specific BTI indicators such as the management index, political transformation, and economic transformation. A closer look at the annual reports by Freedom House (“Freedom in the World”) shows that the levels of political freedoms and civil liberties have been comparable as well: the Philippines and Indonesia exhibit a mixed picture in their respective political institutions and practices, and hence have been categorized as “partly free” over the past four years.
To trace the ways (and the extent to which) church-based activities have contributed to the process of conflict transformation in the region, we apply – as outlined above – a specific mixed-methods approach that integrates qualitative and quantitative perspectives, combines different modes of contextualization and comparison, and transcends different spatial levels of observation. This is achieved by pursuing a strategy of multi-layered analyses (von Lübke 2014a, 2014b) consistent with the proposed methodological framework that preceded this empirical study. We are convinced that this synthesis allows us to tap into the strengths of different methods: it combines qualitative understanding of context and process with quantitative rigor and replication.
To achieve a meaningful exchange of different lines of reasoning, we distinguish the analyses along two key dimensions: contextual and comparative analyses. Figure 3.1 summarizes the main features of this guiding framework.
The first dimension relates to varying levels of analytic abstraction, which can be either low (contextual analyses with a focus on in-depth, location-specific conditions and processes) or high (comparative analyses and estimations focusing on cross-case patterns and trends). The second dimension pertains to levels of analytic scope. Here a distinction can be made between analyses that feature a higher scope (review of national/regional developments and multivariate regressions) and those that feature a lower scope (direct field observations and local surveys in selected project sites) in their observations. Accordingly, the layered framework that guides our study comprises a balanced emphasis on “contextualization” and “comparison.” The following sections will outline in further detail the four methodological quadrants of this framework (see Figure 3.1), namely: (1) in-depth field interviews and observations, (2) process tracing and media/discourse analyses, (3) systematic comparative analyses, and (4) multivariate regression analyses.2
Figure 3.1 Layered analyses
Source: Adapted from von Lübke (2014a, 2014b).
One key ingredient for gaining contextual information was conducting in-depth interviews with selected groups of local and national actors. In the Philippines and Indonesia, we conducted 45 in-depth interviews; the duration of these face-to-face consultations ranged between one and two hours. Interview respondents included Catholic bishops, Protestant and evangelical church leaders, local priests, church personnel, community elders, politicians, academics, NGO members, and media representatives. Interviews were semi-structured and covered issues including histories of conflict and atrocity, opposing forces and interests, conflict transformation processes, church-based activities (strategies, efforts, implementation, results, implications), and underpinning social, political, and economic contexts. Consistent with the overarching research questions, interviews were conducted by using an “interview guideline” that addressed general conflict characteristics, potential conflict drivers, community contexts, and project characteristics. The guideline covered, inter alia, the following questions:
This interview guideline provided a useful general framework for face-to-face inquiries. In many situations, however, interviews diverted slightly from this thematic outline in order to secure a fluid and trustful dialogue with local actors (whose narratives and voices often followed their own thematic and intertemporal sequencing). Moreover, some interviews took the form of ad hoc and informal talks, in which the above-outlined guideline would have been neither feasible nor effective. Key conversations also included informal talks with senior government staff, academics, NGOs, and media representatives, which focused on gaining selective background information rather than a complete assessment of all outlined issues. In addition, we conducted a number of group discussions, for example, with group members of “Bantay Ceasefire” (Ceasefire Watch) in Cotabato, and a select group of conflict transformation experts in Ambon. These were generally characterized by interwoven voices that transcended thematic containers.
Contextual field observations often occurred in connection with semi-structured and informal interview activities. Examples include observations at local bus shelters near the villages of Passo and Waai in Ambon; or observations during two workshops of the DEPAdev (“Democratic Party Development”) program in Davao City and in General Santos in February/March 2016.3
A second important step for contextualizing social/political arenas (and the ideas, interests, and powers that shape them) is the detailed analysis of the root causes of the conflicts under investigation and the discourses in prominent media shaping and entrenching prejudices, symbols, and myths that aggravate mutual aversions of the conflict parties and eventually promote and legitimize the use of violence.
Process tracing is a method that helps to identify causal mechanisms and thus solve the equifinality problem of variable-based research. Equifinality describes the problem of spurious correlations: A appears to cause B, although in reality this is not the case (Schimmelfennig 2006: 265). Such misinterpretations of causal effects can be avoided by tracing the causal mechanisms and pathways between independent and dependent variables: “A causes B, B then causes C, C then causes D, and so on” (Checkel 2006: 363; Bennett & Checkel 2014). Process tracing thus constitutes a series of theoretically informed intermediate steps – the intervening causal process (George & Bennett 2005: 206) – and “fine-grained explanations to a theoretical assumption.” It “places theory and data in close proximity” (Checkel 2005: 22). In our case, process tracing helps in attaining a better understanding of critical junctures and key events that inform our theoretical assumptions about the evolution of the conflicts in Mindanao and Maluku. In addition, they indicate the extents to which primordial (deep-seated religious sentiments), materialist (grievance, greed, or opportunity), and cognitive factors (ideas and discourses) have been affecting observable levels of local violence.
There is a vast strand of literature regarding the methodological application of discourse analysis (Heindl 2015; Gee & Handford 2012). In this research project the term “discourse” will build on Foucault’s understanding of the term (Foucault 1968 [2014], 1969, 1970). Foucault defined discourse as a system of representation, a system of criteria or categories which, taken together, reflect the institutionally stabilized construction of common practices, rules, and resources for generating meaning within a society (Foucault 1969, 1970). This understanding of discourse has been adopted by a wide range of discourse scholars (Keller 2011; Mayring 2002).4
In general, the analysis of political discourses aims at identifying political meanings in accessible texts. To be sure, political discourses legitimize, shape, and (re-)direct political debate, influence political decision making, justify political actions, and (re-)interpret historical memories, myths, and symbols. The focus on the analysis centers on the rules and practices that produce meaningful statements and regulate discourse. Thus, the analysis of political discourse includes the following steps: analyzing the context of the text, coding the text, and analyzing discursive statements.
The assessments of texts in prominent media sources in the Philippines and Indonesia followed a straightforward roadmap. During the preparation and execution of fieldwork activities, our research team scanned, collected, and analyzed various media materials in major newspaper outlets (including The Jakarta Post, Jakarta Globe, Republika, Philippine Daily Inquirer, Manila Bulletin, The Philippine Star, and Malaya and Minda News). Fortunately, most of these media sources feature online archives that are readily accessible. In addition, we included gray literature; for example, working papers, reports, the homepages of the Catholic Bishops Conference of the Philippines (CBCP) and the MILF, insider accounts such as that of MILF chief negotiator Mohagher Iqbal and government negotiator Miriam Coronel Ferrer, blogs, and other online sources in our source materials. The time frame for the Philippine discourse analysis was 2012 to 2016, thus covering the time between the conclusion of the Framework Agreement on the Bangsamoro (FAB) and the time of the main fieldwork in mid-2016, and for Maluku from 1999 to 2017. In the case of Maluku, we also benefited from the newspaper collection of the Ambon Crisis Center and an (unpublished) chronology compiled by Father Kees Böhm, who collected news reports and documents from 1999 until 2006 in the three Moluccan dailies: Siwalima, Suara Maluku (both Christian), and the Ambon Ekspres (Muslim).
Do communities with church-based projects evince more successful conflict transformation processes than counterfactual cases? And to what extent do underlying social, political, and economic contexts affect the success of such conflict transformations? Initial answers to these central questions of our research project arise from two modes of inquiry: the first pertains to “controlled case comparisons” in the tradition of Mill’s analytic logic; the second is the execution of a “qualitative comparative analysis” (QCA), a technique that uses Boolean algebra to investigate key patterns and interactions in a set of comparative observations. We describe both approaches below.
In our study, controlled case comparisons are based on the idea of identifying and examining counterfactual project sites. In order to establish counterfactual assessments in systematic ways, we conducted a set of paired project-site comparisons in both countries. Most similar case comparisons, for instance, which build on Mill’s “method of difference” logic, provide a useful analytical starting point. The key is to find pairs of cases that match in as many respects as possible (analytical twins) but differ distinctly in the explanatory variable. This controlled setting allows for the testing of whether “singular variations” in explanatory factors are accompanied by corresponding variations in observable outcomes and, therefore, allows for the derivation of more robust inferences (Geddes 1990; Gerring 2008). The following illustration summarizes the key ideas (Table 3.1).
As outlined in Table 3.1, the first step involves distinguishing between different Philippine/Indonesian project sites in terms of those with high levels of church-based activity (“treatment sites”) and those with no or sporadic activity (“control sites”). Based on this distinction, the second step focuses on the actual matching process. The challenge is to identify project-site pairs (location A/B and location C/D) that differ markedly in the degree of church-based activities and, at the same time, remain similar across a number of contextual factors. Ideally, paired sites – treatment sites and counterfactual sites – should roughly match in terms of exposure to prior conflicts, historical backgrounds, and cultural underpinnings.
Within this controlled setup, it is then possible to systematically link underlying variations in the presence of church-based activities to observable variations in local reconciliation outcomes. Possible effects (e.g., that efforts of religious actors coincide with improvements in local reconciliation) can be verified or refuted on the basis of field assessments.
Table 3.1 Controlled case selection
| Paired Comparison 1 Location A vs Location B | Paired Comparison 2 Location C vs Location D | |||
|---|---|---|---|---|
| Level of Activity | High | Low | High | Low |
| Prior Conflict Exposure | High | High | High | High |
| Socioeconomic Development | Similar | Similar | Similar | Similar |
| Cultural Context | Similar | Similar | Similar | Similar |
| Improvement in terms of Peace Outcomes | High | Low | High | Low |
These field assessments predominantly draw on the results of original household surveys across local communities. Primary data was generated through the use of a 20-page questionnaire that was translated into Indonesian and Filipino languages by local university counterparts, and was completed on site jointly with local village respondents.5 Survey questions covered (but were not limited to) the following aspects:
Surveys were conducted in cooperation with the Institute for Autonomy and Governance at the Notre Dame University in Cotabato City, the Philippines, and the Department of Sociology at the Universitas Pattimura, Ambon City, Indonesia. Local counterparts helped us to identify research sites for the survey that matched the selection criteria (outlined above) and allowed us to conduct systematic counterfactual analyses. Prior to survey implementation, pretests were administered in both localities, upon which the questionnaire was substantially revised. In both locations, we worked with five enumerators. Enumerators in Cotaboto had substantial professional experience in implementing surveys, while enumerators in Ambon were postgraduate academics with moderate survey experience (mostly sociology students selected by the project counterpart). Both enumerator teams were adequately trained prior to fieldwork activities, and during the data gathering process regular briefing sessions were held with enumerators and local project partners which enabled us to discuss and sort out problems that occurred during the administration of questionnaires. The survey in the Philippines was conducted in the treatment villages Nalapaan and Kalacacan (municipality of Pikit, Cotabato) and in the control villages Cupayon and Bangilan (municipality of Kabacan, Cotabato) between 14 February and 25 March 2016. In Indonesia, we conducted the survey in the treatment villages Poka and Rumah Tiga (Ambon) and in the control villages Lembra Argo and Ahuru (Ambon) between 4 May and 3 June 2016. Table 3.2 details the number of interviews conducted in each location.
In Mindanao, access to Nalapaan and Kalacacan was provided by Father Layson and his staff. They were the first entry points and arranged access for our enumerators to the secondary entry points, which were the so-called local “barangay kapitans.” The first entry point in Kabacan was the local municipality administration. The administrative staff of the office of the Mayor of Kabacan, Herlo P. Guzman Jr., arranged our survey and contacted the barangay captains in Cuyapon and Bangilan in advance. The “kapitan” is the elected local barangay chairperson. We were able to gain the consent and support of all four kapitans and their staff for our survey.
In Ambon, our field activities in the four selected communities – Poka, Lemba Argo, Rumah Tiga, and Ahuru – were conducted in close cooperation
Table 3.2 Survey location and number of interviews
| Country | Treatment Sites | Control Sites/Counterfactuals |
|---|---|---|
| Philippines | Nalapaan/Pikit, Cotabato (N = 65) | Cuyapon/ Kabacan, Cotabato(N = 30) |
| Survey period: 14–17 February 2016 | Fieldwork: 25–26 February 2016 | |
| Kalacacan/Pikit, Cotabato (N = 30) | Bangilan/Kabacan, Cotabato(N = 50) | |
| Fieldwork: 22–23 March 2016 | Fieldwork: 3–4 March 2016 | |
| Indonesia | Poka, Ambon(N = 50) | Lembra Argo, Ambon(N = 50) |
| Fieldwork: 4–6 May 2016 | Fieldwork: 16–19 May 2016 | |
| Rumah Tiga, Ambon(N = 50) | Ahuru, Ambon(N = 50) | |
| Fieldwork: 10–13 May 2016 | Fieldwork: 30 May–3 June 2016 |
with the Pattimura University’s sociology department. Professor Pariela and his team of experienced researchers provided access to local community leaders, leading religious figures, government officials, and other key contacts. The cooperation with Pattimura offered valuable opportunities to discuss and refine the project’s overarching questions, survey questionnaires, and interview guidelines. It also provided a platform for exchanging thoughts and building networks. By hosting a number of ad hoc workshops and small conferences, the sociology department successfully stimulated a fruitful exchange between Indonesian and German collaborators and important discussions with local counterparts. These interactions also paved the ground for interview opportunities with leading religious actors – including Rev Jacky Manuputty, Father Kees Böhm, Sister Brigitta Renyaan, and Bishop Peter Canisius Mandagi – who shared their insights and helped to identify four Ambonese village sites that resembled the outlined case selection criteria.
To ensure a greater level of “triangulation” and systematic replication, we also include additional measures of quantitative testing. To add further support for causal propositions and to draw more representative inferences, it is helpful to complement contextual analyses, case comparisons, and QCA techniques with multivariate statistical estimates. While our case comparisons focus on the local conditions in four carefully selected communities in Mindanao and Ambon, the statistical analysis addresses the meso level by relying on data from an entire province. The access to a complete set of subnational indicators in Maluku means that it lends itself particularly well to this effort of quantitative triangulation.
A multivariate regression exercise aimed at estimating the association of church-based activities with observable improvements in local peace (i.e., a reduction of conflict-related violence) proceeded along two basic steps. The first step entailed the compilation of a subnational dataset. Due to profound data constraints in the Philippines (basic income and household surveys are only collected on aggregate levels and are not available on provincial, district, or sub-district levels), our regression analysis focuses on the Indonesian case in Maluku.
The Indonesian dataset on church activities and local violence was compiled by combining standard surveys and census instruments which were obtained from the Indonesian National Bureau of Statistics, BPS, as well as from the World Bank office and line ministries in Jakarta. In addition, the dataset draws on perception data and existing documentation of local church-based activities in Maluku. Undoubtedly, the data on church-based project work proved to be more problematic than expected. Many data sources and much church documentation were lost during the Maluku conflict and, hence, had to be reconstructed in a cumbersome process with the patient support of leading church representatives. After combining the described sources and perception indicators, the resulting dataset comprises observations across roughly 60 communities on the sub-district level.
The second step was concerned with the actual estimation process. Here, a series of OLS analyses were carried out in which proxies of peace outcomes (lower incidence of local conflict-induced violence) were regressed against the level of church-based activities in respective sub-districts. To test for robustness, regression estimates are reported for different sets of socioeconomic controls. These control variables include measures on subdistrict population, household prosperity (economic status), and the distance to local security forces (district police posts).
The QCA technique is based on Boolean algebra and set-theoretic analyses. It provides a unique opportunity not only to examine the influence of specific conditions (i.e., church-based activities) on observable outcomes (indicators of conflict transformation), but also allows an assessment of other relevant contextual factors. This mode of inquiry is therefore helpful in identifying the effects of additional social, political, and economic factors that shape conflict transformation outcomes in local communities. The QCA research design also makes it possible to examine different combinations of these contextual or supplementary factors. This assessment of different permutations of relevant conditions provides a better understanding of reinforcing and mitigating context factors and, therefore, a more grounded perspective of the effects of church-based activities (Ragin 1987, 2006; Schneider & Wagemann 2012).
The primary survey data of the research project allowed for the testing of a number of interesting sets of conditions consistent with the second research question (i.e., “to what extent have church-based activities been affected by specific economic, political, or social contexts?”). Economic and social conditions, which are tested in the QCA analysis, include local income and education levels. Here, the guiding assumption was that communities with lower socioeconomic pressures (e.g., higher average incomes) might evince lower levels of conflict. Moreover, educational attainment is commonly seen as a conflict-attenuating factor, not least because it enhances job and life opportunities and hence may reduce the likelihood of being recruited to local militias and armed activities.
QCA analyses were also used to gauge additional sociopolitical aspects such as the prevalence of social capital, fundamentalist religious orientations, and ethnic identifications. In general, we assumed that – following Putnam (1993) – high levels of social capital imply a greater presence of cross-cutting solidarity and therefore a means of moderating local tensions. This appears to be sensible, particularly if local social capital entails “bridging” relationships (across different ethnic and religious groups) rather than merely “bonding” relationships (within a specific ethnic or religious group). Conversely, higher levels of fundamentalist religious attitudes or ethnic identification, which tend to sharpen religious and ethnic divides, were assumed to be constraining conditions.
The study rests on extensive fieldwork. An exploratory trip to the Philippines and Indonesia, including Mindanao and Ambon, was conducted in September and October 2015. During the trip, numerous interviews were conducted with politicians, scholars, peace activists, representatives of religious organizations (both Christian and Muslim), and civil society and development donor organizations. Apart from meeting resource persons, the trip also served the function of identifying partner and host institutions for the organization of surveys and data collection in both conflict regions. The local institutions eventually selected for the main fieldwork were the Institute for Autonomy and Governance at the Notre Dame University in Cotabato City and the Department of Sociology at the Pattimura University in Ambon City.
The main fieldwork took place between February and June 2016; first in Mindanao (February to April), then in Ambon (April to June). However, unfortunately, project realities on the ground in Mindanao were characterized by a fragile peace and, hence, posed numerous challenges and constraints. Here, the deteriorating security conditions made it necessary to scale down some of the projected field activities. Bomb explosions during the administration and assessment of field surveys (in Cotabato City and Zamboanga City), alleged attempts by the Islamic State to expand clandestine operations in the southern Philippines and to build up a Southeast Asian front (Jones 2017), as well as violent incidents during the 2016 general election campaigns led to a premature end to our fieldwork in Cotabato. This was consistent with international and national security warnings and a widespread withdrawal of foreign expert staff from the region.
A final visit to the project site in Ambon took place in October 2016, with the objective of completing information and data that could not be obtained in previous stages of fieldwork.
1 For a critical assessment of this literature, see Huotari & Rüland (2014, 2018).
2 This preparatory methodological framework was submitted under the title Analysis of Ecclesiastical Peace and Reconciliation Efforts: Methodological Frame and Case Selection Strategies, Jürgen Rüland, Christian von Lübke, Marcel Baumann, and Anne-Kathrin Weber, University of Freiburg, 2014.
3 DEPAdev is jointly organized by the Konrad Adenauer Stiftung (KAS) Philippines Office with the Institute for Autonomy and Governance (IAG) at the Notre Dame University in Cotabato City. The DEPAdev project rationale is that through the formation of political parties, Muslim, Christian, indigenous, and other underrepresented groups “shall be empowered for active political participation under a parliamentary system,” available at: http://depadev.com/about-us/ (accessed 1 May 2016).
4 For a discussion on critical discourse analysis, see also Jäger (1993), Fairclough (1995), Wodak (2014), and Wodak and Meyer (2015).
5 Scholars interested in working with the survey questionnaires may obtain them from Jürgen Rüland or Christian von Lübke.
Ahram, Ariel I. (2011): “The Theory and Method of Comparative Area Studies,” Qualitative Research 11(1): 69–90.
Ahram, Ariel I., Köllner, Patrick & Sil, Rudra (eds.) (2018): Comparative Area Studies: Methodological Rationales and Cross-Regional Applications, Oxford: Oxford University Press.
Basedau, Matthias & Köllner, Patrick (2007): “Area Studies, Comparative Area Studies, and the Study of Politics: Context, Substance, and Methodological Challenges,” Zeitschrift für Vergleichende Politikwissenschaft 1(1): 105–124.
Bennett, Andrew & Checkel, Jeffrey T. (eds.) (2014): Process Tracing: From Metaphor to Analytic Tool, Cambridge: Cambridge University Press.
Berg-Schlosser, Dirk (2012): “Comparative Area Studies – Goldener Mittelweg zwischen Regionalstudien und universalistischen Ansätzen?” Zeitschrift für Ver-gleichende Politikwissenschaft 6(1): 1–16.
Checkel, Jeffrey T. (2005): It’s Process Stupid! Process Tracing in the Study of European and International Politics, Oslo: University of Oslo, Centre for European Studies, Working Paper No. 26.
Checkel, Jeffrey T. (2006): “Tracing Causal Mechanisms,” International Studies Review 8(2): 362–370.
Creswell, John W. & Plano Clark, Vicki L. (2007): Designing and Conducting Mixed Methods Research, Thousand Oaks, London & New Delhi: Sage Publications.
Croissant, Aurel & Bünte, Marco (eds.) (2011): The Crisis of Democratic Governance in Southeast Asia, Basingstoke: Palgrave Macmillan.
Fairclough, Norman (1995): Critical Discourse Analysis: The Critical Study of Language. London & New York: Longman.
Foucault, Michel ([1968] 2014): “An Archeology of Discourse,” In: Johannes Angermüller et al. (eds.), The Discourse Studies Reader: Main Currents in Theory and Analysis, Amsterdam: John Benjamins Publishing Company, pp. 98–110.
Foucault, Michel (1969): L’Archéologie du Savoir, Paris: Éditions Gallimard.
Foucault, Michel (1970): The Order of Things, New York: Pantheon.
Geddes, Barbara (1990): “How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics,” Political Analysis 2: 131–150.
Gee, James Paul & Handford, Michael (eds.) (2012): The Routledge Handbook of Discourse Analysis, London: Routledge.
George, Alexander L. & Bennett, Andrew (2005): Case Studies and Theory Development, Cambridge, MA: MIT Press.
Gerring, John (2008): Case Study Research. Principles and Practices, Cambridge: Cambridge University Press.
Heindl, Andreas (2015): “Diskursanalyse,” In: Achim Hildebrandt et al. (eds.), Methodologie, Methoden, Forschungsdesign: Ein Lehrbuch für fortgeschrittene Studierende der Politikwissenschaft, Wiesbaden: VS Springer, pp. 257–294.
Huotari, Mikko & Rüland, Jürgen (2014): “Context, Concepts and Comparison in Southeast Asian Studies – Introduction to the Special Issue,” Pacific Affairs 87(3): 415–440.
Huotari, Mikko & Rüland, Jürgen (2018): “Context, Concepts and Comparison in Southeast Asian Studies,” In: Ariel I. Ahram, Patrick Köllner & Rudra Sil (eds.), Comparative Area Studies: Methodological Rationales and Cross-Regional Applications, Oxford: Oxford University Press, pp. 85–102.
International Alert (2014): Rebellion, Political Violence and Shadow Crimes in the Bangsamoro: The Bangsamoro Conflict Monitoring System (BCMS), 2011–2013, London: International Alert.
Jäger, Siegfried (1993): Kritische Diskursanalyse: Eine Einführung, Duisburg: DISS.
Jones, Sidney (2017): “How ISIS Got a Foothold in the Philippines,” available at: https://mobile.nytimes.com/2017/06/04/opinion/isis-pilippines-rodrigo-duterte.html?mwrsm=facebookreferer=http%3a%2f%2fm (accessed 23 December 2017).
Katzenstein, Peter J. (2007): “Regionalism Reconsidered,” Journal of East Asian Studies 7: 395–412.
Katzenstein, Peter J. & Rudra Sil (2008): “Eclectic Theorizing in the Study and Practice of International Relations,” In: Christian Reus-Smit & Duncan Snidal (eds.), The Oxford Handbook of International Relations, Oxford: Oxford University Press, pp. 109–130.
Keller, Reiner (2011): Wissenssoziologische Diskursanalyse: Grundlegung eines Forschungsprogramms, Wiesbaden: VS Verlag für Sozialwissenschaften.
Mayring, Philipp (2002): Einführung in die Qualitative Sozialforschung, Weinheim: Beltz.
Pepinsky, Thomas (2014): “Context and Method in Southeast Asian Politics,” Pacific Affairs 87(3): 441–461.
Putnam, Robert D. (1993): Making Democracy Work: Civic Traditions in Modern Italy, Princeton: Princeton University Press.
Ragin, Charles C. (1987): The Comparative Method: Moving Beyond Qualitative and Quantitative Strategies, Berkeley: University of California Press.
Ragin, Charles C. (2006): “Set Relations in Social Research: Evaluating Their Consistency and Coverage,” Political Analysis 14(3): 291–310.
Schimmelfennig, Frank (2006): “Prozessanalyse,” In: Joachim Behnke, Thomas Gschwend, Delia Schindler & Kai-Uwe Schnapp (eds.), Methoden der Politikwissenschaft: Neuere qualitative und quantitative Analyseverfahren, Baden-Baden: Nomos, pp. 263–271.
Schneider, Carsten Q. & Wagemann, Claudius (2012): Set-theoretic Methods for the Social Sciences: A Guide to Qualitative Comparative Analysis, New York: Cambridge University Press.
Von Lübke, Christian (2014a): “Modular Comparisons – Grounding and Gauging Southeast Asian Governance,” Pacific Affairs 87(3): 509–538.
Von Lübke, Christian (2014b): “Grounding Governance Research in Southeast Asia: A Framework for Controlled Multimethod Policy Analysis,” In: Mikko Huotari, Jürgen Rüland & Judith Schlehe (eds.), Methodology and Research Practice in Southeast Asian Studies, Basingstoke: Palgrave Macmillan, pp. 213–232.
Wodak, Ruth (2014): “Critical Discourse Analysis,” In: Constant Leung & Brian V. Street (eds.), The Routledge Companion to English Studies, London: Routledge, pp. 302–317.
Wodak, Ruth & Meyer, Michael. (2015): “Critical Discourse Studies: History, Agenda, Theory and Methodology,” In: Ruth Wodak & Michael Meyer (eds.), Methods of Critical Discourse Studies: Introducing Qualitative Methods, London: Sage Publications, pp. 1–22.