Abstract
This article presents recent research on the collection and organization of unarmed civilian peacekeeping (ucp) data and proposes a pathway to the creation of a rich, regularly updated database on ucp. The article sets out by giving an overview of the wealth of data on conventional (military) peacekeeping and related research, and raises the question about the need for a similar data wealth for ucp. Secondly, it describes the inventory of current ucp data, to highlight advances in knowledge as well as data collection challenges and critical information gaps. In the main part, the article then sets out what a more comprehensive data collection on ucp would need to take into account and what such a dataset could look like. To do so, it also draws on databases regarding nonviolent campaigns and other fields (health, aviation) to demonstrate the potential for future data collection methodologies and research, and to consider safeguarding issues. Finally, this article suggests four strategies to better structure existing ucp data in order to collect missing information on various research themes. The benefits of the resulting rich database, it argues, would be greater visibility of ucp, the production of data useful for various research programs, and insights to improve field practice.
Introduction
Modern unarmed civilian peacekeeping (ucp), understood as “the practice of deploying unarmed civilians before, during, and after violent conflict, to prevent or reduce violence, to provide direct physical protection to other civilians, and to strengthen or build resilient local peace infrastructures” (Oldenhuis et al., 2017, p. 30), has been practiced by different ucp organizations for almost 40 years, and the community of international, national and local organizations using ucp strategies is growing. A qualitative study by Julian and Gasser (2019) comparing ucp and armed peacekeeping suggests that, except for the rebuilding of security institutions, ucp organizations have successfully carried out all core tasks of armed peacekeepers in some place and at some point, even if at a smaller scale than the large and well-funded UN missions. This should make ucp an interesting object of systematic study, to understand its functioning, effectiveness, and wider effects on security and peace, and to also more systematically compare it to armed protection practices to which it claims to be a practical critique and a viable alternative. Yet, research on ucp among scholars of security, peace and conflict, and international intervention (including the sub-field of the protection of civilians) has remained an exception. A main reason for this lack of scholarly engagement with ucp, we believe, is the absence of a rich, regularly updated database on ucp that would give the practice of ucp greater visibility, make it more readily accessible for various research programs, and not least provide insights to improve field practice.
To work towards filling this gap, this article presents recent research on the collection and organization of ucp data and proposes steps towards the creation of rich, regularly updated data on ucp. Particularly, we ask: If we wish to more rigorously study unarmed civilian peacekeeping, what additional data do we need to collect? How should this data be organized? And how can the data be used efficiently but without causing harm for individuals or organizations? We generated these questions after having reviewed the breadth of research on activities that involve violence (including conventional military peacekeeping), which are much more plentiful and better resourced in comparison to ucp (Eck, 2005). This data gap between conventional (that is, violent) and nonviolent practices to protect civilians from armed violence is mirrored in research budgets as well, exemplified by comparing the quantity and breadth of evaluation data between United Nations peacekeeping operations (unpko) and ucp. Of the many questions these points may raise, the main ones we discuss here are: what types of data exist on armed peacekeeping operations (pko), what questions do this data help answering, and would these questions also be relevant for ucp? What existing data can a future ucp database build on, and how can or should this (scarce) existent base be expanded to provide rich, rigorous data for the type of in-depth quantitative studies, visibility and insight on unarmed civilian peacekeeping that violence-based forms of peacekeeping already enjoy?
We believe that the disparity between research on armed and unarmed forms of peacekeeping can serve as a catalyst to expand data specifically for ucp. In fact, as we will show in detail in this article, the comparison of data collection methods can go beyond unpko, to include research into nonviolent actions in a broad sense (which shares some nonviolent methods with ucp) and other fields such as health and aviation (to learn from good practice around data safety), to gain more insight on how to improve ucp data collection. We also make an argument for the development of new categorization schemes and terminologies, and a critical inquiry into the best use of both quantitative and qualitative data collection, which could facilitate using a wider set of analytical methods, from descriptive statistics and data mining to more advanced modeling and hypotheses testing (Kauffmann, 2009, 2020c).
With this in mind, this article provides an overview of the data on ucp that is currently available and introduces some research proposals.1 It starts with an overview of methodologies and data on unpko conducted in order to generate ideas on how to expand the data collected for ucp. This is followed by a review of the most advanced database on ucp to date, built at Selkirk College in Canada. Specifically, we discuss the lessons learned in its creation, such as methodological and data collection challenges, and limitations of the data. Building on these lessons learnt and on research into other datasets on nonviolent campaigns, we then suggest steps that future database building on ucp should take and also reflect on safeguarding issues that may arise and how they could be handled (taking lessons from aviation and health databases). We conclude with a summary of the four central components of the important endeavor of considerably improving data on ucp, which include: the improvement of the quantity and quality of ucp data; the dissemination of ucp data; synergy building between ucp data efforts; and the prioritization of research efforts while maintaining the diversity of methods. We close with a final reflection on the reasons why all of this matters: namely, how more extensive and rigorous data can improve research in ucp (and nonviolence in general) in the short term, and with its insights potentially support a reduction in violence in the long term.
Data on UN Peacekeeping
The study of traditional peacekeeping operations (pko) is well established and builds on a breadth of existing data on unpko, which has provided the foundations for a plethora of quantitative research studies and insights. This is not least due to the fact that the UN and several institutes published many datasets on unpko, generally under liberal licenses such as Creative Commons (United Nations, 2022a). Additionally, there are more than a hundred datasets on violent actions and armed conflicts (Eck, 2005, 2008). In the meantime, there is just a handful of (quite incomplete) datasets on nonviolent activities such as ucp and non-violent action.
Two of the unpko datasets are the Peacekeeping Fatalities Dataset and the Troop and Police Contributions Dataset. The former provides comprehensive data on unpko fatalities that occurred since 1948 (Henke, 2019). This Peacekeeping Fatalities Dataset is updated monthly; it includes the following data for each incident: a unique casualty identifier, the incident date, the mission acronym, the type of casualty and the cause of incident (accidents, illness, malicious acts…), the country of origin of the affected personnel, the geographical area code of the incident (M49 code), the type of personnel involved (9 categories), and the type of appointment. The Troop and Police Contributions Dataset has monthly data by country, with the following variables: a unique id, country of origin, area of deployment, rank, gender, mission acronym, category of personnel (United Nations, 2022b).
Thanks to such widely available and comprehensive datasets, unpko has been widely studied using quantitative methods, with a flourishing number of publications using these data. A first series of quantitative studies on unpko used mainly data aggregated as country-years or conflict spells (Diehl, 1988; Doyle and Sambanis, 2000; Fortna, 2004). Later analyses disaggregated data, using monthly or daily datasets, and subnational or even point-based2 datasets (Hultman, Kathman and Shannon, 2013, 2014; Costalli, 2014; Di Salvatore and Ruggeri, 2017, 2017).
Table 1 shows estimates3 of the number of salient publications in this field, showing how the field of quantitative studies on unpko has grown since the 1990s.
The wealth of quantitative studies on unpko
| Number of quantitative studies on unpko | |
|---|---|
| 1990-1993 | 1 |
| 1994-1997 | 2 |
| 1998-2001 | 4 |
| 2002-2005 | 10 |
| 2006-2009 | 13 |
| 2010-2013 | 26 |
| 2014-2017 | 33 |
In total, at least 90 quantitative studies on unpko were published in 18 years; we are aware of only two quantitative studies on ucp in the same period (Janzen, 2014, 2015). The availability and richness of unpko data facilitated the study of many dimensions of traditional, armed peacekeeping. Central questions scrutinized by quantitative studies of unpko include the following, which we list with sample publications:
What are the motivations of states contributing to unpko? Do developing states contribute with the aim of benefiting financially – through reimbursement? (Coleman and Nyblade, 2018)
Given the variety of pko missions, how can we measure effectiveness? (Di Salvatore and Ruggeri, 2017)
Do pko protect civilians? (Hultman, Kathman and Shannon, 2013) Why does their protection efficiency vary? (Haass and Ansorg, 2018)
Do pko reduce the numbers of soldiers killed in the battlefield? (Hultman, Kathman and Shannon, 2014)
Do pko achieve to end civil war? (Doyle and Sambanis, 2000) Do they achieve to keep peace after civil war? (Fortna, 2004)
What is the incidence of peacekeeping personnel fatalities, injuries, and kidnappings? (Bromley, 2018)
What is the impact of pko on criminal violence? (Di Salvatore, 2019)
To illustrate the value of this research, we highlight the findings of this last study, which established that:
[classic armed] peacekeeping missions, even when effectively reducing conflict, can inadvertently increase criminal violence. First, less conflict opens up economic opportunities […], thus increasing violent competition among criminal groups. Second, demobilized combatants are vulnerable to turn to crime […]. Cross-national and subnational empirical analyses show that large UN military deployments result in higher homicide rates, whereas UN police, overall, moderate this collateral effect.
Thus, the question that could be raised with regard to unarmed civilian peacekeeping is: does ucp impact criminal violence in the same way? We find a similarly critical question by Lakey regarding the outcomes of nonviolent campaigns: “When a [nonviolent] campaign is experiencing especially heavy repression, what are the benefits/liabilities of inviting third party nonviolent intervention, as compared with violent intervention as in the case of Libya in 2011?” (Lakey, 2011). Currently, data to test hypotheses are missing for ucp, while entire datasets have been gathered to answer those questions with respect to unpko.
So, how can we generate, collect, and systematize the necessary ucp data to investigate all of the questions on this list? In order to address this question, we first review the ucp database at Selkirk College and the lessons that were learned in the process of setting it up, before we turn to the steps that future databases will have to take to overcome the limits of this valuable but ultimate limited attempt at quantifying ucp and its effects and effectiveness.
The ucp Database at Selkirk College
In the early 2010s, research in ucp was gaining ground; practitioners, educators and researchers were beginning to gather data in more coherent and systematic ways, but a repository of basic data on ucp activities and organizations was missing. In 2014, the Mir Centre for Peace at Selkirk College (Canada) embarked on the task of gathering as much descriptive data on global ucp activities as possible and to organize the data in an accessible and usable format.
Thus, the ucp Global Database was begun and has since attempted to track global ucp activities. In the nine years since its inception, the database has continued to track global ucp activity as the broader field of ucp scholarship and collaboration has grown considerably. In this section, we first focus on the original intent of this database, and then describe some of its findings.
In addition to gathering basic descriptive data, such as names of ucp organizations, places (countries, territories, or regions) and years of activity, the database had greater aspirations. In the initial methodology, surveys were sent out to all known ucp organizations, which were identified through an extensive mixed-methods search. The intention was to gather descriptive data on many other variables, such as type and length of training, number of ucp personnel deployed per year, ucp practices utilized, core values, budget, and expenditures, as well as country of origin of ucp personnel, and number of injuries and deaths among them. Additionally, short narratives were gathered for each region to describe the context in which each organization was working, and to highlight a success story of how ucp made a difference in reducing violence. Despite concerted efforts, it proved to be difficult to glean accurate and comprehensive data, as between 10-20 percent of ucp organizations did not respond to the questionnaire (some were no longer operating at the time), and those that did often lacked the historical data. Therefore, the database remained focused on three basic descriptive variables: names of organizations, years of operation, and locations of operations.
Despite the realization that detailed and comprehensive ucp data were not forthcoming, there were, nonetheless, some valuable things learned. Firstly, the database seems to indicate a steady increase in both the number of ucp organizations in operation, and in the regions in which they operate. For example, for the year 1990, seven organizations operating in six countries or regions were identified. By 2022, the number of identified organizations has increased to 61, operating in 29 countries or regions.4 Secondly, the database seems to indicate that ucp organizations are concentrated in certain places. Four countries or territories (Guatemala, Colombia, Palestine, and the United States) have hosted more than ten organizations each. Moreover, while the number of organizations has remained virtually unchanged in Guatemala, Colombia, and Palestine, the United States witnessed an increase from four organizations in 2010 to 13 in 2022.
While the published database essentially contains information on the three variables for which information could be collected for all ucp organizations, there is some significant information that was gathered beyond this basic descriptive data. For example, efforts to establish the database resulted in information, for the first time, to estimate the total number of ucp personnel deployed (although data were incomplete – not all organizations were able to provide this data). Since the total number of ucp personnel killed was already known, we were now able to provide an estimated mortality rate for ucp work – moreover, we could compare them to mortality rates for conventional (armed) peacekeepers. The result was that conventional armed peacekeepers were twelve times more likely than ucp personnel to be killed while on duty (Janzen, 2014).5 This statistic sheds light on the question of knowing whether sending unarmed personnel into volatile and violent contexts is dangerous.6 More research is needed to further explore the relative danger (or safety) of the work of ucp personnel compared to armed actors (ranging from military peacekeepers to law enforcement agents) to better understand the perceived benefits of confronting violence with nonviolence. While evidence is gaining momentum with regard to the effectiveness of nonviolence as a political tool (Chenoweth and Stephan, 2011), more research is needed to specifically explore the relative effectiveness of ucp. The ucp database has created a starting point to gather data to critically explore this question.
In the process of collecting data for the ucp database, other trends and patterns became evident, lending themselves to classifying organizations into a typology based on pertinent variables: whether the organization deploys local or international staff (or both); whether the ucp activities replace conventional military peacekeepers, law enforcement agents, or paramilitary groups; and what type of violence reduction is the main focus, for instance: reducing crime, protecting human rights defenders, or preventing ceasefire breaches (for a detailed proposal for the framework for ucp, see Janzen (2023)). Placing ucp organizations into a typology based on different variables seems to reveal one important trend: earlier organizations tended to be international in scope, sending volunteers from the global north to volatile contexts in the global south. In recent years, the majority of ucp organizations are community-based, training local personnel to address community-identified priorities.
These interesting insights notwithstanding, it is clear from the experience at Selkirk described above that the gap between the original intent of the ucp database and its current state is wide. This is not least due to the limited capacity or interest of some of the ucp organizations at the time to systematically store historical data, and also some reluctance with regard to the sharing of sensitive data. The potential lost opportunity is even more striking when one compares the limited datasets of the ucp Global Database with data available on unpko and the resulting possibilities for the systematic and comparative study of traditional pko s. So how could or should a future ucp database be built in order to enable researchers to study ucp in ways that are similar to the intensive study of classic armed pko s? This is the question we will turn to in detail in the remainder of this article.
Expanding ucp data and analysis
Expanding ucp data, be that in the form of an expansion of the existing database or through the creation of new datasets and databases, would require us to catch up with other peace research fields. Instead of dividing ucp researchers according to their methods, we believe this can be better achieved with mixed methods research, “formally defined here as the class of research where the researcher mixes or combines quantitative and qualitative research techniques, methods, approaches, concepts or language into a single study. […] Its logic of inquiry includes the use of induction […], deduction […] and abduction” (Johnson & Onwuegbuzie, 2004). Such an ambitious objective would require a widening of the inclusion criteria, and a consideration to the methodology of typologies, terminologies, and aggregation, in order to collect not only more data, but also better data – in the sense of well-organized data of various types.
Widening inclusion criteria
ucp has changed since early experiences like Gandhi’s Shanti Sena project (Weber, 1996) and Muller’s work on ucp concepts (Muller, 1997); in addition to the increasing number of ngo s implementing ucp (Janzen, 2014), states and a wide variety of intergovernmental organizations (igo s) implement various subsets of ucp methods, often using different designations for similar methods.
Being a pioneering work, the Selkirk College ucp database used strict inclusion criteria. Organizations had to demonstrate activities ‘on the ground’, such as protective accompaniment, protective presence, interpositioning, interactive monitoring (e.g. ceasefires and truces with volunteers on the ground), and creating buffer zones or safe spaces. The organization also had to have an explicit commitment to nonviolence, and ucp had to be a core activity. Lastly, the organization had to be active after 1990. Thus, organizations such as Amnesty International which is primarily concerned with information gathering and advocacy, the United Nations and the osce in whose work ucp is a component but which do not have a core value of nonviolence, and Watch Guard Chicago, a community-based organization aimed at reducing violence but which does not have nonviolence as a core value, were excluded from the database.
To have a broader understanding of ucp, data collection would benefit from wider criteria than the ones originally used for the Selkirk College ucp database. This would include actions by states, intergovernmental organizations (igo s), and nongovernmental organizations (ngo s) including community organizations. According to Oldenhuis et al (2017, pp. 47–48),
many actors are unarmed civilians who provide protection with methods, principles, sources of guidance, and skills that are similar to the key components of the ucp spectrum, [including igo s] like the African Union (au), the Organization for Security and Co-operation in Europe (osce), the European Union (EU), the UN High Commissioner for Refugees (unhcr), and the UN Office of the High Commissioner for Human Rights (ohchr) […]. UN peace operations (managed by the UN’s Department of Peacekeeping Operations, dpko) include police and other civilians whose work concerns protection. Whether these organizations would be placed at the heart, the edges, or outside of the ucp spectrum is to some extent arbitrary.
Similarly, Weiss and Minear (1999) cite some ohchr missions, EU Monitoring Missions and some duties carried out by UN Volunteers.
Organizing ucp methods
A second task to address in order to establish a richer ucp database concerns the categorization and organization of ucp methods. Table 2 presents four sets of categorizations of ucp methods reviewed by Beer (2021, p. 69; Beer cites: Burrowes, 2000; Hunter and Lakey, 2003; Beer, 2005; Dudouet, 2015), along with four other sources (Muller, 1997, p. 70; Mahony, 2006, p. 47; Oldenhuis et al., 2017, p. 117, 2021, p. 145), to illustrate the problem that existing ucp method categorizations may pose in the task of building a more comprehensive but also usable database. Most importantly, the comparison reveals a lack of convergence among various categorizations of ucp methods. There are differences in naming, in grouping, and in level of abstraction, that is, what is considered a top-level category and what is a sub-group; what is a mission or a strategy and what is an actual method.
Variety of categorizations of ucp methods
| Author | Year | ucp concept | Number of categories | Categories of methods |
|---|---|---|---|---|
| Beer | 1995 | Third-party nonviolent interventions (tpni) | 11 | Accompaniment Delegations Economic sanctions such as boycotts Election monitoring Human rights observation and investigations Information gathering and polling Media Mediation Peace walks/boating Physical interposition Rescue teams/humanitarian assistance |
| Muller | 1997 | Civilian intervention | 5 | Observation Information Interposition Mediation Cooperation |
| Burrowes | 2000 | Cross-border action on behalf of indigenous nonviolent movements | 9 | Local nonviolent campaigns Mobilization actions Nonviolent humanitarian assistance Nonviolent witness and accompaniment Nonviolent intercession Nonviolent solidarity Nonviolent reconciliation Development |
| Hunter & Lakey | 2003 | tpni | 4 | Interposition Observing/monitoring Protective accompaniment Presence |
| Mahoni | 2006 | Protective presence | 5 | Sustained multi-level diplomacy Conscious visibility Active encouragement Convening and bridging Public advocacy |
| Dudouet | 2015 | Nonviolent intervention | 6 | Promoting Capacity-building Connecting Protecting Monitoring Pressuring |
| Oldenhuis et al. | 2017 | Unarmed civilian protection | 4 (with 10 sub-categories) | Proactive engagement Monitoring Relationship building Capacity enhancement |
| Oldenhuis et al. | 2021 | Unarmed civilian protection | 5 (with 12 sub-categories) | Same as Oldenhuis et al. (2017), plus: Advocacy |
There are groups of synonyms in Table 2. One such group, for instance, is monitoring/observation/witness. Four categorizations (Hunter & Lakey, 2003; Dudouet, 2015; Muller, 1997; Oldenhuis et al.: 2017 & 2021) consider “monitoring” (or a similar concept) to be a single method, while Burrowes (2000) attaches “monitoring” to “accompaniment”, and Beer (1995) splits this concept into at least three top-level methods (“election monitoring”, “human rights observation and investigations”, “information gathering and polling”). Mahony (2006) puts some of these methods into a classification of five types of missions (not methods): Human rights monitoring, Complex peacekeeping mission, Ceasefire monitoring, Electoral monitoring, and Humanitarian agency presence. Oldenhuis et al. (2021) consider that the five Nonviolent Peaceforce (np) top-level methods can be split into 12 sub-categories; they add one category (advocacy) only in the newer version of the np handbook, which illustrates the fact that no categorization scheme is perfect; np’s evolution is a strong point in favor of a list of ucp methods both highly disaggregated and inclusive.
The problem with the categorization schemes of Table 2 can be illustrated with Figure 1. Let us imagine there are 16 basic ucp methods, named after letters A to P.
Four possible categorizations of 16 ucp methods
Citation: Journal of Pacifism and Nonviolence 1, 2 (2023) ; 10.1163/27727882-bja00018
According to categorization (a), two items H and L form a single method; this is similar to “Observing/monitoring” for Hunter and Lakey (2003). In categorization (c), L+O+P form a single method; this is similar to Oldenhuis et al. (2021) where Proactive engagement is Protective presence plus Protective accompaniment plus Interpositioning. The problem of using 4 or 5 general concepts rather than 16 basic concepts is that 50 research teams working on ucp data may end up with 50 different categorization schemes – note there are 82 billion ways7 to categorize 16 basic methods into groups of methods!
A similar problem occurred in the field of armed conflict data, with more than a hundred definitions of types of armed conflicts, and a huge number of different datasets (Eck, 2005). Reconciling these definitional differences would be fighting a losing battle, because diverse research teams have diverse data needs (Eck, 2008). Instead, a plea was made in favor of disaggregating conflict data into individual conflict events (ideally with daily data), with a detailed description (and several attributes) for each event (Kauffmann, 2008). This led to increased popularity of daily datasets such as acled (Raleigh and Carlsen, 2008), from which anyone can use the attributes to filter and group individual events into more general categories such as “civil wars” or “armed conflicts”. Learning from this, we discuss how to do this in the ucp field.
Disaggregating ucp data
Based on what has been discussed, ucp data can be disaggregated in three ways: semantic disaggregation, temporal disaggregation, and spatial disaggregation. Semantic disaggregation involves using many sub-categories. It renders more transparent and reproducible the disciplinary cognitive processes that are involved in creating categories of highly constructed objects whose definitions are sometimes abstract. Temporal disaggregation breaks down ucp campaigns into actions over one to a few days, which is required to account for rapidly changing situations in unstable countries (Kauffmann, 2007). Spatial disaggregation provides a high resolution of geolocation. With disaggregated data, verification of facts is facilitated, and several types of subsequent aggregation are possible, depending on the research paradigms (Kucera et al., 2011).
Gene Sharp’s typology of 196 non-violent methods highlighted the advantages of semantic disaggregation. The typology is used by the Global Nonviolent Action Database (gnad), which contains more than 1,300 nonviolent campaigns. The data have been further supplemented, with international efforts to further temporally and spatially disaggregate the data (Kauffmann, 2020b). Each campaign is a case study and a set of data allowing quantitative analyses (Kauffmann, 2017). About 21 campaigns are Third Party Nonviolent Interventions (tpni); we list below three of them as an illustration of the type of campaigns that fall into this category:
1969: Shanti Sena (Indian Peace Brigade) intervenes in deadly Hindu/Muslim riots in Ahmedabad.
1984-1989: Peace Brigades International (pbi) protects and aids Guatemalan Mutual Support Group (gam).
2009: Afghan policewomen form human chain to protect women’s rights activists in Kabul.
According to the gnad, 20 different methods were used during these 21 tpni campaigns (see Table 3).8
Some methods used by Third Party Nonviolent Interventions
| Category | Subcategory | id (Sharp) | id (Beer) | Method name (Sharp) | Freq. |
|---|---|---|---|---|---|
| Protest and Persuasion | Public Assemblies | 50 | 92 | Teach-ins | . |
| Economic Noncooperation | Ordinary Industrial Strikes | 106 | 163 | Industry strike | + |
| Multi-Industry Strikes | 117 | 174 | General strike | . | |
| Political Noncooperation | Citizens’ Noncooperation with Government | 129 | 186 | Refusal of assistance to enforcement agents | . |
| Citizens’ Alternatives to Obedience | 134 | 191 | Nonobedience in absence of direct supervision | . | |
| 136 | 193 | Disguised disobedience | . | ||
| 138 | 195 | Sit-down | + | ||
| 140 | 197 | Hiding, escape, and false identities | ++ | ||
| Action by Government Personnel | 147 | 203 | Deliberate inefficiency and selective noncooperation by enforcement agents | . | |
| International Governmental Action | 152 | 209 | Delay and cancellation of diplomatic events | + | |
| Nonviolent Intervention | Physical Intervention | 162 | 226 | Sit-in | + |
| 171 | 238 | Nonviolent interjection | +++ | ||
| 172 | 239 | Nonviolent obstruction | + | ||
| 173 | 289 | Nonviolent occupation | + | ||
| Social Intervention | 174 | 287 | Establishing new social patterns | +++ | |
| 180 | 258 | Alternative communication system | + | ||
| Economic Intervention | 182 | 264 | Stay-in strike | + | |
| 184 | 297 | Defiance of blockades | . | ||
| 192 | 283 | Alternative economic institutions | . | ||
| Other | Accompaniment, observation/monitoring and presence (gnad name) | (not in Sharp list) | 334 | Solidarity accompaniment (Beer typology) | +++ |
Frequency codes: “.” very rare; “+” rare; “++” frequent; “+++” very frequent.
We can find other tpni cases in the gnad by searching for campaigns that combine various methods of Sharp’s typology, for example combining three out of: (120) Withholding or withdrawal of allegiance; (140) Hiding, escape, and false identities; (171) Nonviolent interjection; (173) Nonviolent occupation.9 Similarly, while the gnad definition of tpni includes the first four elements of the Selkirk College ucp database (protective accompaniment, protective presence, interpositioning, monitoring), it excludes the fifth (creating buffer zones/safe spaces). However, the gnad has Sharp’s method “68. Sanctuary” with examples of “creating safe spaces”.10
We have discussed these examples in detail here because they illustrate that, with a detailed typology of methods (198 methods, in this case), differences in definitions of general concepts are not an obstacle for research, since researchers can easily re-organize detailed data into more general concepts. Hence, creating a detailed list of ucp methods would facilitate data collection and use. For instance, the abstract “rumor control” concept could be related to the following basic methods:
Monitoring traditional media
Monitoring numeric media
Monitoring social networks
Monitoring medias with automated methods (Natural Language Processing)
Monitoring word of mouth
In-place factchecking
Verification by triangulation of written information
Documenting disinformation (in an information system)
Tracing disinformation flows
Exposing occurrences of specific rumors
Awareness campaigns about rumors
Training on rumor control techniques
Some organizations and researchers may eventually group the last three methods into “capacity building”; others may put all these methods into “rumor control”, a category in np’s circle. But if all reporting on rumor control activities used the same basic terms, data collection and sharing would be improved.
A multilingual ucp terminology database
To implement the above, we need an international data collection effort, and consequently detailed, multilingual definitions of methods11 and sub-methods. More generally, a multilingual ucp glossary is required. This work could be based, among other things, on Sharp’s list of 198 methods, already translated into 30 languages (Sharp, 2002a), and other works (Muller, 2014; Beer, 2021). This would help the exchange of content between research institutes and researchers of several countries; on the operational level, such work would also benefit international ucp missions. This could mimic unterm, a terminology database with translations in eight languages of more than 80,000 terms relevant to the UN system (United Nations, 2022c).12
Table 4 shows a sample of what such a ucp glossary could look like, displaying two of the 198 methods of Sharp (Sharp et al., 2005, pp. 51–64) with their id s, as well as the corresponding id s and definitions in the icnc typology (Beer, 2021), along with translations in Spanish (Sharp, 2012), French (Sharp, 2009), Arabic (Sharp, 2002b) and two Russian versions: 1) one in which the translator coordinated their work with Sharp and used a periphrasis (Sharp, 2005, p. 69); and 2) a more concise translation (Guseinov, 1991). The table shows that translating concepts between languages is sometimes difficult, and that a ucp glossary certainly requires definitions (not just translations) to be useful. Historical examples of uses of a method (such as in Beer, 2021) would help distinguish related concepts, for example nonviolent interjection (the “Tank Man” near Tiananmen Square in Beijing on June 5, 1989) and nonviolent obstruction (bulldozer blockades in Serbia in 2000).
Possible structure of a glossary of methods
| Method name (English) | Nonviolent interjection | Nonviolent obstruction |
|---|---|---|
| id (Sharp) | 171 | 172 |
| id (Beer) | 238 | 239 |
| Spanish | Interposición no violenta | Obstrucción no violenta |
| French | Interposition nonviolente | Obstruction nonviolente |
| Russian (1) | ||
| Russian (2) | ||
| Arabic | ||
| Definition (Beer) | Physical intercession of someone’s path, blocking their ability to execute their task in an attempt to make them reconsider (i.e., bulldozers or soldiers) | Physical blocking of a space, creating a big enough obstacle to actually prevent passage |
Collecting more data
Expanding the types of ucp data to collect would help answer more questions. Below is a non-exhaustive list of topics that data should be collected for:
Number of headquarter and field personnel for each year, each country (or region, district…), and if possible, with details such as nationality, function, part-time/full time, employed/volunteer, length of deployment, age range, level of experience, gender, etcetera.
Types of actions carried out, with places, dates, frequency, staff, and other resources used.
Incidents (injuries, deaths, kidnappings, illnesses, others…) for staff and protected people.
Costs (breakdown by year, country, with budget lines).
Trainings delivered or followed (by ucp staff): place, duration, number of attendees, topics covered, official recognition.
Job descriptions.
Text documents.13
Multimedia sources (movies, videos, radio broadcasts, images) showcasing ucp organizations and their missions, using a multimedia catalog.14
Other data that ucp organizations believe might be relevant.
Ensuring the security of sensitive ucp data
The expansion of the ucp database ultimately aims to better understand ucp in order to save lives. Consequently, this should be done taking into account the “Do No Harm” principle: the lives of the “beneficiaries” and ucp professionals should not be endangered by the research process. Hence, privacy and security considerations should be given the highest priority. We look at the management of sensitive personnel data in another field, then present some techniques to protect sensitive ucp data.
Sensitive personnel data
One may ask whether collecting detailed data on ucp staff is too intrusive. Looking at other professions may shed light on this. The Federal Aviation Administration (faa) database gathers detailed data on pilots, including name, birthdate, certificate number, information on skills, and performance, while defining strict rules of access for various entities (US Federal Aviation Administration, 2021). In addition, intimate psychological data are collected to study the mental health of pilots and questions related to “their lifestyle, psychological requirements for medical certification, psychological problems amongst air crew, personality factors, disruption to personal relationships, reactions to incidents and accidents, alcohol and drug misuse, suicide by aircraft and environmental challenges” (Bor, Field and Scragg, 2002). Similar research is carried out on UN Blue Helmets (International Peace Institute, 2022), albeit in a less advanced fashion. Thus, one may ask about ucp practitioners: does their mental health have an impact on their behavior with respect to risk exposure? Should such data be included, then strong data protection measures should be taken.
Pseudonymization
Some useful concepts for managing sensitive data are defined in the EU General Data Protection Regulation (gdpr). Article 4 defines pseudonymization as
the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information, provided that such additional information is kept separately and is subject to technical and organizational measures to ensure that the personal data are not attributed to an identified or identifiable natural person.
Pseudonymization allows linking data at the individual level, without disseminating the identity of individuals. In law enforcement, the effective practice of combining restricted-access databases produced by various institutions for research purposes is not recent; examples in the British judiciary systems exist that use data collected since 1963 (Francis, Crosland and Harman, 2002), with the aim of combining information from the police, the courts, the prison service and the probation service, with procedures using mainly automatic matching algorithm, complemented by manual matching for a few cases.
There are also numerous examples of sensitive data managed through pseudonymization in the health sector. For instance, in the context of the covid-19 pandemics, the French Government15 developed secure protocols to merge three medical databases, on hospitalization (si-vic), testing (si-dep), and vaccination (vac-si). In each of the three databases, information on first and last name and date of birth of people hospitalized, screened, or vaccinated is used to create a pseudonym, a non-meaningful character string uniquely identifying each person. Nominative data are then deleted from the databases before being transmitted to drees. Only the pseudonym is transmitted (it is impossible to find out the identity of people). The pseudonym is built with the same algorithm in the three databases, allowing matching data about tests, hospitalizations, and vaccinations (dress, 2021). The algorithm yields an extremely low probability of wrong matching due to homonyms (this risk could be reduced by adding the place of birth in the algorithm). These datasets are then made available to researchers with some conditions, and to the public in an aggregated format (see below). The latter aggregated datasets were downloaded 400,000 times in a year (dress, 2022). Similar methods could be used to manage data on ucp staff who worked in various ucp organizations.
Data aggregation and data degradation
Data aggregation is a process in which detailed data pertaining to groups of individuals, events, or other categories are summarized by statistics such as average, median, or quantiles. For instance, instead of publishing the fact that there are, say, five ucp personnel aged 25, 30, 35, 40 and 45 in a given city, data aggregation implies publishing only the mean age (here, 35) for this city. Such aggregation allows analysts and the public to draw some conclusions without disseminating individual data. Under strict conditions, some researchers may receive some of the disaggregated data.
Data degradation is another technique to store sensitive data. In the drees covid-19 database, birthdates are replaced by age categories, preventing tracing back the individuals. Similarly, aggregated ucp data may store age categories of ucp personnel. Similar techniques exist for geographical data (spatial cloaking). The map below, a rural area in Somalia, shows spatial data randomized with the geographic information system application qgis. The original data set (two points) is shown with question marks. The randomized data (X crosses) lie near the original points, within a two-kilometer radius.16
Let us imagine the point in Kobon is a local ucp office, and the point near Bar is the house of a community leader in need of protection. The randomization does not preserve the exact location of the two places, but allows answering questions such as: In which regions and districts are ucp offices located? What is the average distance and travel time from local ucp offices to the houses of people in need of protection? What is the security impact of a deteriorated security situation in a given place? For instance, if ucp staff usually drive through Mraara and Sunguuni to go to Bar, they may need to change their itinerary because of security issues such as riots or fighting in Mraara and go through Sinjibaar instead. Hence, randomizing place locations may protect staff while allowing to answer research questions with little loss of precision.
Hiding the true location of ucp actions
Citation: Journal of Pacifism and Nonviolence 1, 2 (2023) ; 10.1163/27727882-bja00018
Anonymous reports
Anonymous reporting is a way to get feedback from people who might be reluctant to share valuable information if their identity is revealed. This principle guided the nasa Aviation Safety Reporting System (asrs), established in 1976, to collect, analyze, and distribute de-identified safety information provided through confidentially submitted reports from aviation personnel. Tens of thousands of reports are submitted every year (Connell, 1994) without breaching the identity of reporting people, and thus the asrs has released thousands of safety alerts (Connell, 2017). It might be worth thinking about ways in which implementing a similar data collection scheme could benefit the safety of ucp operations.
Conclusion: Future prospects
While the collection, organization, and dissemination of ucp data has begun, this article has outlined the fundamental gaps and, using data and research from conventional peacekeeping, health care and other sources, has proposed a number of directions for the future of ucp data. In closing, we summarize the four central components of this important endeavor: the improvement of the quantity and quality of ucp data; the dissemination of ucp data; synergy building between ucp data; and the prioritization of research efforts while maintaining the diversity of methods.
Improving the quantity and quality of ucp data. In addition to providing researchers and the public with basic data, an expanded ucp database would provide information to assist researchers in answering research questions, such as: Why is there a trend toward more community-run ucp organizations? Is ucp safer than conventional military peacekeeping? What is the best method to measure the economic efficiency of ucp? (Boulogne, 2013) How could ucp be improved? Since the Selkirk College database was first published, the landscape of ucp practice and research has changed significantly. Efforts to create ucp networks have led to a higher degree of mutual awareness and information sharing. It is therefore likely that the present is a promising time to revisit the amount and type of data that can be gathered. Initial goals of determining annual counts of deployed ucp personnel, organizational methodologies, and underlying principles, budgets, etc., may now be more readily achievable.
Improving the dissemination of ucp data. In terms of accessibility, an improved future ucp database should take into account recent developments in the field of databases on international relations and armed conflicts (Kauffmann, 2020a), in particular with respect to interoperability and interactivity. Interoperability and use of open standards enable greater dissemination and integration of data from multiple sources, and greater use in related research programs (Kauffmann, 2008). This also facilitates the maintenance and evolution of the information system. Interactivity meets the diversity of user needs. Several means can be mobilized, such as simple and advanced search engines (including full-text search on qualitative data), interactive representations (statistical charts, maps, and timelines) in easy-to use web dashboards, and export to statistical software. This focus on the diversity of user needs reflects the mixed research method: teams of researchers may have members with various methodological orientations, which should complement the variety of entry points to the various data types.
Building synergies in ucp data. One way to foster ucp data collection and research is to develop international partnerships among institutions. The model could be inspired by independent research centers that regularly publish databases (such as sipri, idmc, and acled). The development of the ucp database should be accompanied by meetings (seminars, conferences) and trainings, whose themes could include, among others, the computerization and structuring of data on nonviolence and ucp; the quantitative and qualitative use of these data; data protection and cybersecurity;17 the use of bibliographic and document management systems (dms); and new technologies as a nonviolent means of action. Bridging the gap between research and practice is also essential. At some point, sociological or anthropological methodologies, with ucp researchers who observe and participate in various programs and training during several weeks (Almeida, 2008, p. 57) could prove helpful.
Creating a network, pooling resources, and ensuring the continuous updating of the main database would help. Examples in other fields showed that this might save money. For example, the cost of building the faa pilot database (software engineering and data entry – 8600 hours yearly) (US Federal Aviation Administration, 2021, p. 31056) demonstrated that “the annual recurring industry cost savings will more than offset the recurring annual costs”, counting only administrative costs and benefits (i.e. excluding the potential safety benefits). ucp data will be mainly supplied by institutions involved in ucp, but semi-automatic data collection (from news and reports) is possible through Natural Language Processing tools18 with human validation involving universities. This network could take the form of a “ucp observatory”, similar to some observatories at the UN (United Nations, 2022a) or at other institutions (International Peace Institute, 2022; ngo Monitor, 2022).
Prioritizing the research efforts while maintaining the diversity of methods. If the set of data described in this paper is too wide, initially it might be desirable to get such data for at least a shorter period (say, 2016-2017, or some earlier years – for safety reasons), for a few countries, and for a limited number of organizations; however, we believe the initial effort should keep the objective of getting highly detailed data in various forms (tabular data, texts, images…), within this limited scope. One objective is to build a dataset showing how detailed data could be, and to face a larger variety of data collection and management problems, and thus to learn from them. ucp, contrary to armed pko, is less often carried out by states or igo s; as such, ngo s often lack the statistical culture that states and igo s have. The proposed gradual approach would help bring the culture of data into ngo s. This process would also allow ngo s to experience the advantages of mixed methods research. Hopefully, this would help these organization understand ucp better, and thus facilitate improvements within organizations.
The utilization of ucp, as a legitimate practice to reduce violence and save lives, has significantly increased in the past several decades. This article has argued that the data needed to document ucp trends and activities require significant expansion and improvement, in order to facilitate research and dissemination, and ultimately to facilitate best practice. The suggestions proposed in this article, based on current data challenges and experience gleaned from related disciplines, highlight specific methodologies through which this may be accomplished. The practice of ucp, in all its forms, holds great promise for our collective humanitarian goal of breaking the cycles of violence and promoting social justice. It is imperative that our knowledge of ucp is rigorous and effective.
Notes
Some ideas were taken from the irnc research program (irnc steering committee, 2019). We are grateful to the participants of a seminar on ucp data held online on December 4th, 2019 (hosted at Université Grenoble Alpes), to the editors of this issue and to two anonymous reviewers.
In point-based datasets, the geographical dimension of events is referenced not by the polygons in which the events occurred (administrative areas such as countries, provinces, municipalities…) but by the coordinates (latitude, longitude) of the estimated location.
Table 1 shows a crude estimate, based on the number of publications after 1988, year of publication of Diehl’s seminal study; publication list from Di Salvatore & Ruggeri (2017) which, like any review, is not comprehensive.
Note that the data collection process was biased towards the outcome of a perceived increase in the two indicators: 1) Organizations which were active in 1990 but inactive after 2014 would not respond; 2) Because of limited historical knowledge in organizations (i.e. due to staff turnover), lack of resources or other reasons, responding organizations may be biased towards reporting more accurately recent activities (and countries) than old activities. In addition, for both indicators, the publication of missions on web pages is likely to be more frequent for 2022 than for 1990. Geographical and linguistic proximity between the MIR Centre for Peace and the United States (in which many organizations were created) may reinforce this bias. (Hence, technically, it might be that the real number of organizations and regions actually decreased between 1990 and 2022). Based on these potential biases and the limitations of the data, the results are, at best, a good approximation.
Still, some may argue that ucp staff were possibly sent to less risky places than armed personnel (thus explaining the difference in death rate). Better data (timing and location of deployments) are required for a more robust assessment (in which local risk may be assessed with the daily acled dataset).
According to a 2015 Canadian survey, the general public was supportive of the concept of ucp, but the main concern was the safety of sending unarmed civilians to do the work of military peacekeepers. See Janzen (2015).
Or 8.28×109, computed using Bell’s number (Bell, 1938), allowing exclusion of 1 to 14 methods.
The gnad uses Sharp’s typology. Table 3 includes codes from the ICNC typology (Beer, 2021) which adds 148 methods to Sharp’s 198-method list.
For instance:
International Committee for the Nanking Safety Zone (made of US, British, Danish and German nationals) (Han, 2012a).
Actions in South Korea supported by Amnesty International, Human Rights Watch Korea and other local organizations (Hopkins-Hayakawa, 2011).
Three organizations worked jointly to protect and save 5000 Jews during World War II in the French commune of Le Chambon-sur-Lignon: the Presbybterian Church, the Secours Suisse aux Enfants (Swiss Aid to Children), the Scout Movement (Han, 2012b).
Broad categories might rely on the aims of various ucp interventions, and sub-categories on more specific actions.
unterm covers many subjects relevant to peace keeping/building but lacks many ucp terms.
A ucp database should be linked with an online document management system (dms), compatible with reference management software (Zotero, JabRef…) supporting annotation sharing, with full text documents where possible (or excerpts and summaries). ucp data entries should be linked to documents in the dms, to bridge the gap between qualitative and statistical heuristics. Similar endeavors are made in mixed quantitative/qualitative databases that summarize or reproduce text sources, such as Bell & Badanjak (2019), and Sundberg & Melander (2013).
Such as the open source Mediathread platform or similar (Estrada et al., 2017), possibly with automated pre-tagging of shots (Bartolini, Patella and Romani, 2010).
Statistics department of the French governmental social and health services, drees (Direction de la Recherche, des Études, de l’Évaluation et des Statistiques).
The two-kilometer value is arbitrary; it can be increased (if more privacy is required) or decreased (if more precision is required).
In particular for researchers working in dictatorships whose work could endanger their safety, but also for other researchers.
Such as ConflitBERT, the Europe Media Monitor, gdelt, icews, Phoenix, etc. Validation (human control of the automatically collected data; tagging) could involve Master’s students.
References
Almeida, D.A. (2008). Shanti Sena: Unarmed Peacekeeping. University of Rhode Island (Open Access Master’s Theses). Available at : https://digitalcommons.uri.edu/theses/1103.
Bartolini, I., Patella, M., and Romani, C. (2010). ‘shiatsu: semantic-based hierarchical automatic tagging of videos by segmentation using cuts’, AIEMPro ’10: Proceedings of the 3rd international workshop on automated information extraction in media production, p. 30.
Beer, M. (2005). Mainstreaming Peace Teams. Nonviolence International.
Beer, M. (2021). Civil Resistance Tactics in the 21st Century. Washington, D.C.: International Center on Nonviolent Conflict. Available at https://www.nonviolenceinternational.net/nv_tactics_book.
Bell, C., and Badanjak, S. (2019). Introducing pa-x: A new peace agreement database and dataset. Journal of Peace Research 56 (3), pp. 452–66.
Bell, E.T. (1938). The Iterated Exponential Integers. Annals of Mathematics 39 (3), pp. 539–57.
Bor, R., Field, G., and Scragg, P. (2002). The mental health of pilots: An overview. Counselling Psychology Quarterly, 15 (3), pp. 239–56.
Boulogne, É. (2013). Rapport de recherche sur une évaluation adaptée de l’Intervention Civile de Paix. IRNC. Available at http://irnc.org/IRNC/Textes/614.
Bromley, S.L. (2018). Introducing the ucdp Peacemakers at Risk dataset, sub-Saharan Africa, 1989-2009. Journal of Peace Research 55 (1), pp. 122–31.
Burrowes, R.J. (2000). Cross-border Nonviolent Intervention: A Typology. In: Y. Moser-Puangsuwan and T. Weber, eds., Nonviolent Intervention Across Borders: A Recurrent Vision, Honolulu, Hawai’i: University of Hawai’i Press, pp. 45–72.
Chenoweth, E., and Stephan, M.J. (2011). Why Civil Resistance Works: The Strategic Logic of Nonviolent Conflict. Columbia University Press.
Coleman, K.P., and Nyblade, B. (2018). Peacekeeping for profit? The scope and limits of “mercenary” UN peacekeeping. Journal of Peace Research 55 (6), pp. 726–41.
Connell, L.J. (1994). Pilot and Controller Communications: Incidents Reported to the nasa Aviation Safety Reporting System. sae Transactions 103, pp. 2104–10.
Connell, L.J. (2017). ‘nasa Aviation Safety Reporting System (asrs)’, 17 October. Available at https://ntrs.nasa.gov/citations/20170010289.
Costalli, S. (2014). Does Peacekeeping Work? A Disaggregated Analysis of Deployment and Violence Reduction in the Bosnian War. British Journal of Political Science 44 (2), pp. 357–80.
Di Salvatore, J. (2019). Peacekeepers against Criminal Violence—Unintended Effects of Peacekeeping Operations? American Journal of Political Science 63 (4), pp. 840–58.
Di Salvatore, J., and Ruggeri, A. (2017). Effectiveness of Peacekeeping Operations. Oxford Research Encyclopedia of Politics.
Diehl, P.F. (1988). Peacekeeping Operations and the Quest for Peace. Political Science Quarterly 103 (3), pp. 485–507.
Doyle, M.W., and Sambanis, N. (2000). International Peacebuilding: A Theoretical and Quantitative Analysis. The American Political Science Review 94 (4), pp. 779–801.
dress (2021). Exploitation des appariements entre les bases si-vic, si-dep et vac-si : des données pour la période du 20 septembre au 17 octobre 2021 à partir d’une nouvelle méthodologie. Available at https://drees.solidarites-sante.gouv.fr/communique-de-presse/exploitation-des-appariements-entre-les-bases-si-vic-si-dep-et-vac-si-des.
dress (2022). Données statistiques publiques en santé et social. Available at https://data.drees.solidarites-sante.gouv.fr/pages/accueil/.
Dudouet, V. (2015). Sources, Functions, and Dilemmas of External Assistance to Civil Resistance Movements. In: K. Schock, ed., Civil Resistance, University of Minnesota Press (Comparative Perspectives on Nonviolent Struggle), pp. 168–200.
Eck, K. (2005). A Beginner’s Guide to Conflict Data: Finding and Using the Right Dataset. ucdp Paper. University of Uppsala. Available at http://www.pcr.uu.se/publications/UCDP_pub/UCDP_paper1.pdf.
Eck, K. (2008). An Overview and Typology of Conflict Data: the Advantages of Data Diversity. In: M. Kauffmann, ed., Building and Using Datasets on Armed Conflicts, Amsterdam: IOS Press, pp. 29–40.
Estrada, L.M., Hielscher, E., Koolen, M., Olesen, C.G., Noordegraaf, J., and Blom, J. (2017). Film Analysis as Annotation: Exploring Current Tools. The Moving Image: The Journal of the Association of Moving Image Archivists 17 (2), pp. 40–70.
Fortna, V.P. (2004). Does Peacekeeping Keep Peace? International Intervention and the Duration of Peace after Civil War. International Studies Quarterly 48 (2), pp. 269–92.
Francis, B., Crosland, P., and Harman, J. (2002). The Police National Computer and the Offender Index: Can They be Combined for Research Purposes? Findings (170), pp. 1–4.
Guseinov, A.A. (1991). 198 методов ненасильственных действий (198 Methods of Nonviolent Actions).
Философские науки (Philosophical Sciences) (12), pp. 33–40.Haass, F., and Ansorg, N. (2018). Better peacekeepers, better protection? Troop quality of United Nations peace operations and violence against civilians. Journal of Peace Research 55 (6), pp. 742–58.
Han, S. (2012a). Foreigners establish Safety Zone and intervene to save civilians during Nanking Massacre, 1937-1938. Global Nonviolent Action Database. Available at https://nvdatabase.swarthmore.edu/content/foreigners-establish-safety-zone-and-intervene-save-civilians-during-nanking-massacre-1937-1.
Han, S. (2012b). French Christians save Jews from Nazi holocaust, Le Chambon, 1940-1944. Global Nonviolent Action Database. Available at https://nvdatabase.swarthmore.edu/content/french-christians-save-jews-nazi-holocaust-le-chambon-1940-1944.
Henke, M.E. (2019). UN fatalities 1948–2015: A new dataset. Conflict Management and Peace Science 36 (4), pp. 425–42.
Hopkins-Hayakawa, S. (2011). South Koreans protest land seizure for United States military base expansion, 2005-2007. Global Nonviolent Action Database. Available at https://nvdatabase.swarthmore.edu/content/south-koreans-protest-land-seizure-united-states-military-base-expansion-2005-2007.
Hultman, L., Kathman, J., and Shannon, M. (2013). United Nations Peacekeeping and Civilian Protection in Civil War. American Journal of Political Science 57 (4), pp. 875–91.
Hultman, L., Kathman, J., and Shannon, M. (2014). Beyond Keeping Peace: United Nations Effectiveness in the Midst of Fighting. The American Political Science Review 108 (4), pp. 737–53.
Hunter, D., and Lakey, G. (2003). Opening Space for Democracy: Training Manual for Third-party Nonviolent Intervention. Philadelphia, PA: Training for Change.
International Peace Institute (2022). Peacekeeping Observatory, International Peace Institute. Available at https://www.ipinst.org/program/peace-operations-and-sustaining-peace/peacekeeping-observatory.
irnc steering committee (2019). irnc 2020-2024 Research Program on data related to Unarmed Civilian Peacekeeping and nonviolence.
Janzen, R. (2014). Shifting Practices of Peace: What is the Current State of Unarmed Civilian Peacekeeping? Peace Studies Journal 7 (3), pp. 46–60.
Janzen, R. (2015). Incorporating unarmed civilian peacekeeping into Canadian foreign policy: what do Canadians think? Canadian Foreign Policy Journal 21 (1), pp. 15–27.
Janzen, R. (2023). A Typology for the Various ucp Practices. In: E. Furnari, R. Janzen and R. Kabaki, Unarmed Civilian Protection – A New Paradigm for Protection and Human Security, (Spaces of Peace, Security and Development), pp. 27–40.
Johnson, R.B., and Onwuegbuzie, A.J. (2004). Mixed Methods Research: A Research Paradigm Whose Time Has Come. Educational Researcher 33 (7), pp. 14–26.
Julian, R., and Gasser, R. (2019). Soldiers, civilians and peacekeeping – Evidence and false assumptions. International Peacekeeping 26 (1), pp. 22–54. https://doi.org/10.1080/13533312.2018.1503933.
Kauffmann, M. (2007). Short term and event interdependence matter: A political economy continuous model of civil war. Peace Economics, Peace Science and Public Policy 13 (1), pp. 1–19.
Kauffmann, M. (2008). Enhancing Openness and Reliability in Conflict Dataset Creation. In: M. Kauffmann, ed., Building and Using Datasets on Armed Conflicts. Amsterdam: IOS Press (nato Science for Peace and Security Series – E: Human and Societal Dynamics), pp. 107–32.
Kauffmann, M. (2009). Méthodes statistiques appliquées aux questions internationales. Paris: L’Harmattan.
Kauffmann, M. (2017). Resilience to oppression and to violent conflict escalation through nonviolent action. In: European Commission, The 2nd International Workshop on Modelling of Physical, Economic and Social Systems for Resilience Assessment, Ispra, pp. 82–91. Available at https://www.researchgate.net/publication/334400567_Resilience_to_oppression_and_to_violent_conflict_escalation_through_nonviolent_action.
Kauffmann, M. (2020a). Databases in Defense Studies. In: D. Deschaux-Dutard, ed., Methods in Defence Studies, Routledge, pp. 129–50.
Kauffmann, M. (2020b). Le numérique au service d’un idéal de non-violence. In: Stéphane Hessel: Les perspectives d’un engagement. lgdj (Transition & Justice, 27), pp. 97–109.
Kauffmann, M. (2020c). Quantitative Analyses in Defense Studies. In: D. Deschaux-Dutard, ed., Methods in Defence Studies, Routledge, pp. 109–28.
Kucera, J., Kauffmann, M., Duta, A.-M., Soler, I.T., Tenerelli, P., Trianni, G., Hale, C., Rizzo, L., and Ferri, S. (2011). Armed conflicts and natural resources. Brussels: European Commission. Available at https://publications.jrc.ec.europa.eu/repository/handle/JRC64271.
Lakey, G. (2011). Global Nonviolent Action Database: Uses of the Database. Available at https://nvdatabase.swarthmore.edu/content/uses-database-0.
Mahony, L. (2006). Proactive presence: Field strategies for civilian protection. Geneva: Henri Dunant Centre for Humanitarian Dialogue.
Muller, J.-M. (1997). Principes et méthodes de l’intervention civile. Paris: Desclée de Brouwer.
Muller, J.-M. (2014). Dictionnaire de la non-violence. Paris: Editions du Relié.
ngo Monitor (2022). About. Available at https://www.ngo-monitor.org/about.
Oldenhuis, H., Carriere, R., Duncan, M., Furnari, E., Frisch, A., and Wagstrom, T. (2017). Unarmed civilian protection: Strengthening civilian capacities to protect civilians against violence. 1st ed. Unitar, Nonviolent Peaceforce.
Oldenhuis, H., Furnari, E., Carriere, R., Wagstrom, T., Frisch, A., and Duncan, M. (2021). Unarmed civilian protection: Strengthening civilian capacities to protect civilians against violence. 2nd ed. Geneva: Nonviolent Peaceforce. Available at https://nonviolentpeaceforce.org/ucp-manual/.
Raleigh, C., and Carlsen, J. (2008). The acled Data Project: Purpose and Methods. In: M. Kauffmann, ed., Building and Using Datasets on Armed Conflicts, Amsterdam: IOS Press, (nato Science for Peace and Security Series – E: Human and Societal Dynamics), pp. 161–68.
Sharp, G. (2002a). From Dictatorship to Democracy: A Conceptual Framework for Liberation. 4th ed. (2010). Portypool, Wales, London: Green Print. Available at https://www.aeinstein.org/from-dictatorship-to-democracy/.
Sharp, G. (2002b). men aledketatewreyh ela aledyemqerateyh etar teswery lelther [From Dictatorship to Democracy, Arabic ed.]. 2nd ed. Available at https://www.aeinstein.org/free-resources/free-publications/arabic/.
Sharp, G. (2005). Ot diktatury k demokratii: Konceptual’nye osnovy osvobozhdenija [
От диктатуры к демократии: Концептуальные основы освобождения . From Dictatorship to Democracy]. Available at https://www.aeinstein.org/free-resources/free-publications/russian/.Sharp, G. (2009). De la dictature à la démocratie [From Dictatorship to Democracy]. L’Harmattan. Postface by M. Kauffmann.
Sharp, G. (2012). De la dictadura a la democracia: un sistema conceptual para la liberacion [From Dictatorship to Democracy]. Samos Lugo: SENECA. Available at https://www.aeinstein.org/free-resources/free-publications/espanol/.
Sharp, G., Paulson, J., Miller, C.A., and Merriman, H. (2005). Waging Nonviolent Struggle: 20th Century Practice And 21st Century Potential. Boston: Extending Horizons Books. Available at https://www.aeinstein.org/waging-nonviolent-struggle/.
Sundberg, R., and Melander, E. (2013). Introducing the ucdp Georeferenced Event Dataset. Journal of Peace Research 50 (4), pp. 523–32.
United Nations (2022a). Peacekeeping Open Data Portal, United Nations Peacekeeping. Available at https://peacekeeping.un.org/en/open-data-portal.
United Nations (2022b). Troop and Police Contributions, United Nations Peacekeeping. Available at https://peacekeeping.un.org/en/data-troop-and-police-contributions.
United Nations (2022c). unterm. The United Nations Terminology Database. Available at https://unterm.un.org.
US Federal Aviation Administration (2021). Pilot Records Database. Part ii: Rules and Regulations. Federal Register 86, (110), pp. 30753–1085.
Weber, T. (1996). Gandhi’s Peace Army: The Shanti Sena and Unarmed Peacekeeping. Syracuse, N.Y: Syracuse University Press.
Weiss, T., and Minear, L. (1999). Volunteers Against Conflict: United Nations Volunteers. Tokyo/New York: United Nations University.
