Methods & Frequently Asked Questions

1 PROJECT BASICS

1.1 What is included in the TJET website, and what is not?

Our website contains data on the history of democratic transitions, violent intrastate conflict spells, and transitional justice mechanisms in all countries of the world, from 1970 to 2020. We also include population-wide survey data on transitional justice where we possess it.

Regarding transitional justice, TJET includes information on the following mechanisms:

• Domestic criminal prosecutions for human rights and conflict-related crimes
• Foreign criminal prosecutions initiated by one state against nationals of another state
• International criminal prosecutions by ad hoc tribunals, hybrid tribunals, and the ICC
• State-sanctioned truth commissions
• State-sanctioned reparations policies
• State-sanctioned amnesty offers and laws
• State-sanctioned vetting policies, particularly those labeled lustration
• United Nations investigations

TJET does NOT include information on the following mechanisms:

• Judgments by the International Court of Justice (ICJ) or by regional human rights courts
• Commissions of inquiry that do not meet the definition of a truth commission (see Section 2.1.3)
• Non-state truth-telling procedures or memorialization projects
• Reparations that are required by court rulings
• Comprehensive peace agreements or peace accords
• Instances of security sector reform
• Traditional or customary forms of justice

1.2 What’s new about the TJET database?

TJET modifies and extends previous efforts at data collection, most notably that of the Transitional Justice Research Collaborative (TJRC). There are six new features of TJET that distinguish it from our prior work.

1.2.1 Comprehensiveness

The TJET database is global, and it does not exclude countries because they fail to meet certain ‘rules’ of inclusion. Accountability mechanisms are initiated in a variety of contexts–during and after democratic transition, during and after intrastate conflicts, in stable democracies, and in stable autocracies–so we present data on these mechanisms in all of these contexts, wherever we observe them. In short, the TJET project leans toward inclusion and comprehensiveness.

1.2.2 Flexibility

Transitional justice researchers are interested in different mechanisms and policies, and they often are concerned about different contexts. For instance, they might inquire about TJ mechanisms in civil war, TJ mechanisms amid democratic transitions, or TJ mechanisms that occur in stable democracies. This is a sample selection issue because it involves a choice of which set of cases to compare. Previous projects have confronted this sample selection issue either by narrowing their data collection efforts to specific mechanisms or to specific contexts. For example, they might only collect data on truth commissions across all contexts, or collect data on all TJ mechanisms but only in specific contexts (like post-conflict or democratic transition settings). Our strategy is instead to collect as much data about mechanisms across contexts, and allow the user to customize sample selection. Our database is relational: the data are stored in many different tables linked to one another. Using the TJET database, a researcher can select the sample of cases they are interested in and query our database to produce relevant qualitative and quantitative data. In this way, the TJET data are flexible and responsive to the researcher’s needs.

1.2.3 Macro-level Comparison

Drawing on our extensive observational data, TJET is using Bayesian techniques to produce a global human rights accountability index. This index consists of three different components that address demand for, and supply of, accountability: legacy of violence, regular legal accountability, and transitional justice. Transitional justice is further broken into three attributes: truth, criminal prosecution, and reparation. Using these various global indices, one can make determinations of which countries have the highest need for human rights accountability, and which countries have produced the most comprehensive transitional justice policies to date. These human rights accountability measures will be added to this website in the near future.

1.2.4 Micro-level Granularity

TJET pushes beyond the collection of macro-level institutional data. One goal of our project is to combine the study of state-sanctioned transitional justice institutions with micro-level perceptions of justice and accountability. To that end, we are providing extensive results from transitional justice perceptions surveys in five countries.

1.2.5 Gender-attentiveness

Whereas our previous data collection efforts were unable to distinguish which TJ mechanisms included provisions meant to address gender-based inequities, TJET pays particular attention to whether mechanisms are attentive to gender – meaning the socially constructed roles, status, and identities of girls, women, boys, men, and gender-diverse people in society. Our data now include a number of gender-based measures, though what constitutes gender-attentiveness varies by type of mechanism. (See Section 2.4.)

1.2.6 Transparency

TJET is committed to data transparency. On our website, we make available our coding manuals, our raw data tables, our code for assembling and cleaning data (see Section 3.1), our country-year dataset for analyses, and our Reference Library. We are glad to answer any questions that users have about the website, and we plan in the future to move toward expert- and crowd-sourced error correction.

Please provide us website feedback by email: tjetdata@gmail.com

1.3 What type of data are available on the TJET website?

The data collected by TJET on mechanisms are events-based. They consist of observations made about institutions, practices, and perceptions. A variety of qualitative and quantitative data are available on the TJET website.

First, on the Map page, you can find counts of transitional justice mechanisms by country for cases you filter using the left-hand drawer panel called “Select Data.” For example, if you are specifically interested in truth commissions that occurred amid democratic transitions in Africa, you may filter the data to produce information on that specific sample, by selecting “truth commissions” under Mechanisms, “democratic transition” under Contexts, and “Africa” under Regions.

Second, on the country pages accessible from the map or from the table below the map, you can obtain information on regime transitions, conflict spells, and mechanism-level transitional justice data for the specific country of interest. For several focus countries, we also make available survey data collected by our team members over the last two decades. (See Section 1.2.) The transitional justice data shown on each country page can be downloaded at the bottom of the page (in Excel format).

Third and finally, on the Downloads page, you can access (in .csv format) TJET’s extensive country-year dataset with measures constructed from our database, as well as our raw mechanisms data, which include hundreds of fields of information.

1.4 How can I suggest corrections or additions to the TJET database?

TJET is devoted to fully open and free data. We are also committed to error correction. The purpose of this site is not to present an authoritative or frozen database, but to offer a starting point on which to build a more complete and updated repository of information on transitional justice around the world. As such, the TJET database is intended to evolve based on user feedback. If you find a factual or compositional error on this website, including any database coding issues, please report this to TJET using the corrections form on the Contact Us page.

1.5 What information is presented on the country pages?

On each country page, you will find data on regime transitions, intrastate conflict history, and five types of transitional justice mechanisms: amnesties, criminal prosecutions, reparations policies, truth commissions, and vetting policies (see Section 2.1). The country pages are not meant to be fully developed case studies. The purpose of these pages is to contextualize and describe the TJET data relevant to the country.

For data on democratic transition, we draw on other well recognized data projects, including Varieties of Democracy, Polity5, and the Box, Miller and Rosato dataset. (See Section 2.9.) For data on intrastate conflict, we rely on the Uppsala Conflict Data Program (UCDP), which is both the most reliable and most extensive database on low-level to full-scale armed conflict. (See Section 2.10.) For data on transitional justice, we rely on our own coding, though we draw on a number of other secondary sources. (See Section 3.2.) Citations are provided on the Downloads page.

1.6 What are TJET Focus Countries?

TJET Focus countries are ten countries that TJET highlights for more extensive data presentation and analysis. These include Cambodia, Central African Republic, Colombia, Cote d’Ivoire, Ethiopia, Guatemala, Iraq, Myanmar, Sri Lanka, and Ukraine. For most of these countries, we include transitional justice perceptions survey results, as well as additional analysis and commentary.

1.7 What is the date range of the TJET database?

TJET presents data from 1970 to 2020. The year 1970 was chosen because it follows a wave of decolonization in the 1960s, and it precedes by a few years the first efforts to initiate human rights criminal prosecutions and truth-seeking processes. While we do possess data on mechanisms that have been established since 2020, we do not present data that post-dates 2020 because we have not systematically updated our database for all countries beyond that year.

2 CONCEPTS AND CATEGORIES

2.1 What transitional justice mechanisms are included in the database?

The TJET database includes five different types of mechanisms.

2.1.1 Amnesties

We broadly define amnesties as any legislative, constitutional, or executive provision granting immunity for criminal activity. This includes legal offers and legal acts preventing prosecutions for activities ranging from crimes against the state to terrorism, treason or human rights violations. Though pardons are distinct from amnesties we also include information on the mass release of political prisoners and pardons for previous human rights violations.

2.1.2 Criminal Prosecutions

Human rights prosecutions, or trials, encompass a variety of actions made by a number of actors. We define a human rights criminal prosecution as the use of domestic, foreign, or international courts of law to bring criminal procedures – including preliminary trial processes, trial hearings, or verdict and sentencing – against state-agent and rebel perpetrators of human rights violations.

TJET distinguishes domestic, foreign, and international human rights prosecution types. Domestic prosecutions are criminal indictments and proceedings in a state’s civilian or military justice system for accused who are nationals of that state. These would include trials like that of Guatemala’s former president Efrain Rios Montt for genocide. Foreign prosecutions are criminal proceedings of foreign nationals that take place in a state’s justice system, either under universal, territorial, or passive personality jurisdiction, as Spain’s case against Augusto Pinochet or the many prosecutions under way in Ukraine for Russian nationals charged with war crimes. Finally, international prosecutions are proceedings that occur in hybrid tribunals, international tribunals, or the International Criminal Court (ICC). These would include the trial of Charles Taylor conducted by the Special Court for Sierra Leone, the Slobodan Milosevic trial at the International Criminal Tribunal for the former Yugoslavia, and the Dominic Ongwen trial at the ICC. Altogether, TJET includes four types of domestic human rights prosecutions, three types of foreign prosecutions, and three types of international prosecutions. (See Section 2.5.)

Intrastate Conflict Prosecutions include trials of rebels or members of the opposition for human rights violations, but also for crimes other than human rights violations or core crimes. The reason is that rebels are often tried for treason, terrorism, or other crimes against the state, and this is often germane to during- and post-conflict justice considerations. For example, in 2002, the United Kingdom tried and convicted Real IRA member Colm Murphy for plotting to commit a bombing at Omagh in 1998. Technically, these offenses were classified as a form of terrorism, not a human rights violation. But due to the high number of casualties, and the nexus of the bombing to peace negotiations between the Unionists and Republicans in Northern Ireland, this trial was certainly relevant to post-conflict justice. TJET ultimately identifies two types of conflict-related criminal prosecutions, those that occur in the context of violent intrastate competitions between state and rebel forces, and those that follow unsuccessful coups, bloodless coups, or other low-level acts of violence.

2.1.3 Truth commissions

We operationalize a truth commission (TC) as a formal, state-sanctioned, temporary body that investigates a pattern of past human rights abuses and aims to issue a final report of its findings. Some TCs examine one particular event of human rights abuse, as opposed to a pattern of events, e.g. the UK’s Bloody Sunday Inquiry and the US’s Greensboro Truth Commission. In these instances, we pay particular attention to those commissions that, despite the focus on one event, have mandates that additionally explore wider patterns of violence in the region under investigation. Furthermore, TJET also presents data on truth commissions that do not issue final reports, or issue final reports that remain unavailable to the public. Lack of a final report does not disqualify a TC from our database; lack of a final report is instead reflected in our data about TC outcomes. Finally, we do not collect information on all commissions of inquiry, especially those that are not relative to the perpetration of past harms.

2.1.4 Reparations

We follow the 2005 United Nations Basic Principles on Reparations (“Basic Principles”), which includes five forms of reparation: restitution, compensation, rehabilitation, satisfaction, and guarantees of non-recurrence (GNR). TJET compiles data on policies that aim to provide the first three forms of reparation, not the final two. The reason is that the satisfaction and GNR could feasibly include a wide range of activities, like public apologies and development projects, that are beyond the scope of our analysis, or reforms that are coded elsewhere in our data base, such as vetting policies.

TJET collects data on reparations policies created by a state through executive action, domestic law, a domestic truth commission, or a domestic peace agreement to repair a population for human rights violations committed by state or armed opposition actors. We do not include reparations policies created by anyone other than a state; policies created by a state government at the municipal level; policies that address matters unrelated to human rights abuses; policies that were never intended to be implemented (e.g. a draft bill that was never passed into law); reparations that are awarded as a result of a civil or criminal trial; or reparations that are paid from one state to another state.

2.1.5 Vetting

Vetting is an official policy that deprives categories of perpetrators access to public sector jobs on either a temporary or permanent basis as the result of past human rights abuse. For our project, vetting sanctions are either dismissal from existing employment or ban from future employment or the declassification of secret files implicating the individual or group. Thus, vetting policies are operationalized as dismissal from existing employment; or ban from future employment; or declassification of secret files; or some combination of dismissal, banning, or declassification.

We use the term “vetting” because the word is often employed to describe similar processes in different regions and because “vetting” refers to a narrow set of policies with specific sanctions. In contrast, the term “lustration” usually refers to vetting and declassification processes in Eastern Europe, and the terms “purge” and “disqualification” usually refer to policies with a broader scope of targets and/or sanctions than we intend to code.

Our vetting data is currently complete only for Eastern European and Post-Soviet countries, where the vast majority of these policies were established. For the rest of the world, TJET is currently performing an update beyond the year 2010.

2.2 What are the different types of amnesties, reparations, truth commissions, and vetting policies?

TJET classifies these mechanisms into four different types: transitional, conflict, transitional & conflict, and not transitional or conflict. These types are based on the nexus between the transitional justice mechanism and democratic transition or intrastate conflict. A transitional TJ mechanism is established during or after a democratic transition, and is meant to address crimes by state agents of the former regime (for example, Argentina’s 1983 CONADEP truth commission and 1987 Due Obedience Amnesty Law). A conflict TJ mechanism is established during or after intrastate conflict, and is meant to address crimes committed by government or rebel forces (e.g. Nepal’s 2008 Interim Relief Program and 2014 Truth and Reconciliation Commission). Sometimes, mechanisms meet both of these criteria at the same time, in which case they are classified as transitional & conflict (e.g. Liberia’s 2003 vetting policy). And if TJ mechanisms meet none of these criteria, they are classified as not transitional or conflict (e.g. Bahrain’s 2019 Independent Commission of Inquiry).

2.3 What is the difference between the truth commission process and outcome indices?

Several of TJET’s truth commission measures are split into separate process and outcome indices, including for instance, the victim-centered truth commission measure. Process indices are specific to the operation of a commission and are coded from the year of its creation. Outcome indices measure attributes that are only relevant once the commission has completed its work, so these are coded from the year the truth commission issued its final report. For analyses, the process and outcome indices can be added together within a given year.

2.4 How does TJET incorporate gender into its data collection efforts?

TJET measures how “gender-attentive” transitional justice mechanisms are. By “gender-attentive,” we mean mechanisms that explicitly address gender injustices and gender-based violations. A gender-attentive mechanism can include but is not at all limited to sexual and gender-based violence (SGBV). We include a battery of gender-related questions for coders to consider when they are researching each prosecution, reparations policy, or truth commission. What ultimately makes a particular mechanism gender-attentive differs depending on the mechanism. (See our paper “Gender and Transitional Justice: Explaining Global Trends” for more detail.)

2.5 What are different types of criminal prosecutions in the TJET dataset?

TJET collects information on international, foreign, and domestic types of criminal prosecutions. Each of these we classify into subtypes.

%%{init: {"flowchart": {"htmlLabels": false}} }%%
flowchart TB
  A(PROSECUTIONS) --> B("` **International Prosecutions** `")
  A(PROSECUTIONS) --> C("` **Foreign Prosecutions** `")
  A(PROSECUTIONS) --> D("` **Domestic Prosecutions** `")
  B --> E(ICC)
  B --> F(Ad Hoc)
  B --> G(Hybrid)
  C --> H(Universal Jurisdiction)
  C --> I(Territorial Jurisdiction)
  C --> J(Passive Personality Jurisdiction)
  C --> K(Active Personality Jurisdiction)
  D --> L(Regular human rights)
  D --> M(Transitional human rights)
  D --> N(Intrastate conflict)
  D --> O(Opposition)

International prosecutions have three subtypes, distinguished by the kind of court in which they are initiated. The first are prosecutions that are undertaken by the International Criminal Court, which was established by the Rome Statute and began operating in 2002. This subtype of trial begins when the ICC Office of the Prosecutor issues a summons or arrest warrant to a person accused of grave international crimes. The second subtype of international prosecution are those that occurred in ad hoc tribunals, those established in the mid-1990s for the former Yugoslavia (ICTY) and Rwanda (ICTR). Unlike ICC prosecutions, these tribunals had direct, rather than complementary, jurisdiction for a set period of time. The third subtype are hybrid tribunals, which undertake prosecutions in domestic jurisdictions, but usually with the support of international personnel, resources, and bodies of law. These include, among others, the Special Court for Sierra Leone (SCSL) and the Extraordinary Chambers in the Courts of Cambodia (ECCC), among others.

Foreign prosecutions are those wherein nationals of one country are tried by domestic courts in another country. Most of the time these involve the application of universal jurisdiction, which is our first subtype of foreign prosecution. Universal jurisdiction is “a legal principle allowing or requiring a state to bring criminal proceedings in respect of certain crimes irrespective of the location of the crime and the nationality of the perpetrator or the victim.” Several states exercise universal jurisdiction in their criminal courts, as allowed by legislation, or by the text of certain treaties. Spain’s case against Pinochet involves Universal Jurisdiction, both because Spain had passed legislation permitting such trials and because the Torture Convention also permits foreign prosecutions for torture.

In a second subtype of foreign prosecution, State A exercises territorial jurisdiction to prosecute nationals of State B for crimes committed in State A. For example, in 2016 an Argentine court convicted a Uruguayan colonel Manuel Cordero for the crime of disappearing Uruguayan citizens in Argentina. Of course, this occurs relatively often, but when such prosecutions involve atrocity crimes, we consider them foreign trials. This has happened in contexts like the former Yugoslavia, where for example, Serbs might be tried in Croatian or Bosnian courts for crimes committed during the wars in the 1990s.

A third subtype of foreign prosecutions involves passive personality jurisdiction, wherein State A prosecutes nationals of State B for victimizing nationals of State A. For example, in 1990 a French court sentenced in absentia Argentine Captain Alfredo Astiz for kidnapping and torturing to death two French nuns in Argentina. This is increasingly common as governments limit universal jurisdiction, arguing that trials of foreign nationals should have some connection to the prosecuting state. In the fourth sub-type of foreign prosecutions, states try permanent residents or new citizens for atrocity crimes they committed abroad prior to immigration, which is an exercise of active personality jurisdiction. While these resemble domestic prosecutions, we still view them as having a foreign relations aspect.

Domestic prosecutions are by far the most numerous type of trial in the TJET database. This category includes all instances where a state’s domestic courts hold its own nationals criminally accountable for human rights or conflict-related crimes committed within the territory of that state In reviewing thousands of event records from around the world, TJET has identified four common subtypes of domestic prosecutions, which are classified on the basis of their nexus to regime change or intrastate conflict, as well as the charges faced by the accused.

Nexus to democratic transition or UCDP intrastate conflict

	Nexus	No Nexus
Human rights crimes	(2) transitional human rights prosecutions	(1) regular human rights prosecutions
Conflict crimes	(3) intrastate conflict prosecutions	(4) opposition prosecutions

2.5.1 Regular Human Rights Prosecutions

Regular human rights prosecutions are trials of state agents–most often police or military forces–for human rights violations they committed in the routine execution of their security or custodial functions. These often involve common physical integrity violations committed by state agents, including torture, killing, sexual violence, or excessive force.

These acts need not have any nexus to a democratic transition or intrastate conflict. For example, when Colombian police officers Carlos Augusto Diaz and Carlos Danilo Posada were put on trial in 2012 for setting fire to a 15-year-old boy sleeping under a bridge in Bogota (TJET Trial ID 6730), we categorize this as a regular human rights prosecution because it had no nexus to a democratic transition or to the ongoing civil war with FARC. However, when members of the Colombian military were tried for killing indigent workers and pretending they were FARC guerilla fighters (the “false positives” scandal, e.g. TJET Trial 5446), these prosecutions are classified as Intrastate Conflict Prosecutions (see below).

2.5.2 Transitional Human Rights Prosecutions

When state agents are put on trial during or after a democratic transition for harms they committed prior to that transition, we classify these events as transitional human rights prosecutions. Included in this category are the paradigmatic cases of post-autocratic criminal justice, like the Trial of the Juntas in Argentina (TJET ID 427). This prosecution took place in 1985, following a democratic transition in 1983. The army generals who were put in the dock, all influential members of the autocratic military regime that ruled from 1976-1983, were charged with numerous crimes, including torture andkidnapping and murder, which took place in the years prior to the transition to multi-party democracy. This is a post-autocratic human rights prosecution. So too are the hundreds of prosecutions that have occurred since 2006, following the nullification of the 1987 Due Obedience amnesty law in 2003. Though they re-started over 20 years after the democratic transition, these trials are also for crimes that were committed during the autocratic period.

However, not all human rights prosecutions in Argentina are classified as post-autocratic. For example, in 2015, ten police officers were tried for the “Massacre at Quilmes,” wherein four teenagers were killed during a disturbance inside a police station (TJET Trial 6986). While the alleged perpetrators were state agents, they were charged with crimes that occurred well after democratic transition, so this trial is categorized as a regular human rights prosecution.

2.5.3 Intrastate Conflict Prosecutions

TJET relies on the Uppsala Conflict Data Project to match prosecutions to the appropriate conflict contexts. Sometimes conflict actors, including members of state armed forces or their rebel counterparts, are tried in domestic courts for their actions in the context of intrastate conflict. These trials may occur during conflict or post-conflict. When targeting state agents, conflict prosecutions could involve planned and unplanned abuses, excessive uses of force, or crimes of opportunity by government forces. In addition to the Colombian “false positives” case mentioned above, one could find numerous additional examples of intrastate conflict prosecutions. For example, in 1996, eight soldiers and one police officer were prosecuted in Sri Lanka for raping and murdering an 18-year-old Tamil student named Krishanthi Kumaraswamy, as well as murdering her mother and brother (TJET Trial 2544). This violence was a byproduct of the enhanced military presence in the context of civil war between the state and the separatist Tamil Tigers (LTTE). This is an intrastate conflict prosecution of state agents.

Intrastate conflict prosecutions can also target rebel forces, for planned and unplanned violence against civilians, crimes against the state, terrorism, and other crimes. For example, in 2002, 11 Bosnian Serbs were put on trial in Bosnia & Herzegovina for the 1995 murder of Croat Priest Tomislav Matanovic and his parents, which occurred in Prijedor in the context of war between the Bosnian government and the breakaway Republic Srpska (TJET Trial 779). The rebel Bosnian Serb forces had control of Prejidor during the abduction and murder of Matanovic, and they were held criminally accountable following the end of the war, For this reason, TJET further categorizes this prosecution as post-conflict justice. However, if the trial had occurred amid ongoing conflict, it would be labeled during-conflict justice.

2.5.4 Opposition Prosecutions

Some armed conflicts do not reach the UCDP’s 25 battle-death threshold, and some of these conflicts involve prosecutions. We refer to these as opposition prosecutions, and they often involve state agents and/or opposition members who are put on trial for conducting a bloodless coup, for attempting an unsuccessful coup, or for trying to launch an insurrection that failed in its initial stages. For instance, starting in 2012, 103 suspects were charged in Turkey with overthrowing the democratically elected Islamist Prime Minister Necmettin Erbakan in 1997. The coup involved the military simply pushing out the Erbakan coalition government without force. This “postmodern coup” thus did not amount to a violent intrastate conflict, but it was nevertheless prosecuted later as a coup. This is an opposition conflict prosecution.

2.6 What are international and hybrid tribunals?

International criminal tribunals are ad hoc or permanent courts established to prosecute cases arising under international criminal law. These tribunals have direct or complementary jurisdiction over territories in which international crimes are committed.

Hybrid tribunals are courts of mixed composition and jurisdiction, encompassing both national and international aspects, usually operating within the jurisdiction where the crimes occurred.

The following international and hybrid tribunals are included in the TJET database:

International Tribunals
International Criminal Court (ICC)
International Criminal Tribunal for Rwanda (ICTR)
International Criminal Tribunal for the former Yugoslavia (ICTY)

Hybrid Tribunals
European Union Rule of Law Mission in Kosovo (EULEX) & United Nations Interim Administration Mission in Kosovo (UNMIK)
Extraordinary Chambers in the Courts of Cambodia (ECCC)
Iraqi High Tribunal
Special Court for Sierra Leone (SCSL)
Special Criminal Court (SCC), Central African Republic
Special Panels for Serious Crimes (SPSC), Timor Leste
Special Tribunal for Lebanon (STL)

2.7 How are prosecutions matched to countries?

We match prosecutions to countries by nationality of the accused. If Croatian courts try Croatian nationals, these would be counted as Croatian domestic trials. If Croatian courts try Serbian nationals, even for crimes committed in Croatia, we would count these as foreign trials that act on the country of Serbia. And if the International Criminal Tribunal for the former Yugoslavia tries Croatians, we count these as international trials that act on the country of Croatia. In short, prosecutions are assigned to countries based on whether their nationals are accused of crimes.

There are two important caveats. First, the distinction between domestic, foreign and international is a political, not a legal, distinction. It is not based on strict understandings of legal jurisdiction. It is perfectly within Croatia’s rights to hold a Serbian foreign national criminally accountable for crimes the Serbian national committed on Croatian soil; that is territorial jurisdiction. However, we believe it is important to distinguish this process from the process wherein Croatian courts hold their own citizens accountable for human rights crimes. The first involves foreign relations, while the latter does not.

Second, we make an exception to our matching rule for trials conducted by the ICC. When the ICC becomes involved in a situation, we match the trial proceedings to the country in which that situation occurs, instead of matching to the nationality of the accused individual. The reason is that ICC involvement in a situation country is meant to have an effect in that country. It would be deceptive to claim that the ICC had been involved in Rwanda because two defendants, Callixte Mbarushimana and Sylvestre Mudacumura, were issued arrest warrants for their actions during hostilities in neighboring Democratic Republic of Congo (DRC). Our methodological argument is that we should study the impact of these ICC arrest warrants on the DRC, not on Rwanda.

2.8 How does TJET determine when a prosecution has concluded?

It is often difficult to establish from the source materials when a prosecution has concluded. Therefore, we need to make assumptions about the end years. We consider a prosecution finalized when there is evidence of a verdict and no evidence of ongoing appeals. In many cases we find information regarding the start of proceedings without updates about the conclusion of proceedings. In such cases, we apply a five-years rule: if there has been no new information on a trial for five years since its last activity, we code the end year of the trial as the last year of activity. For instance, if there was evidence of an appeal in 2005, but we could not find any updates on that appeal between 2005 and 2010, then we code the end of the trial as 2005.

2.9 How does TJET construct its list of democratic transitions?

We base our coding of democratic transitions on three prominent scholarly democracy datasets: Polity5, the Boix, Miller and Rosato dataset (BMR), and Varieties of Democracy’s Episodes of Regime Transformation (VDem-ERT). (See citations on the Downloads page). We specifically register transition start years and consider transitions to be open-ended until a new democratic transition begins. A transition begins when at least two out of the three democracy datasets indicate a significant change from autocracy to democracy.

For Polity5, we code a transition year when the Polity5 project codes a transition or when the polity score increases by three or more from one year to the next, and the subsequent polity score is positive. For BMR, we code a transition year when the project codes a transition to electoral democracy. And for VDem-ERT, we code a transition year for the first year of a democratization episode which results in a leveling up in the VDem Regimes of the World ordinal regime type measure. In the majority of cases, these three coding rules result in the same transitions but they often do not agree on the transition start year. In the latter cases, we code the transition start year once two out of of the three indicate a democratic transition. We adjust these coding rules for newly independent states (where we take into account the regime type of predecessor states), for micro states (for which only BMR provides data), and after 2018, when the Polity5 data end. For 2019 and 2020, we rely on BMR and VDem-ERT, and after 2020, we use only the VDem-ERT data to determine new democratic transitions.

2.10 How is intrastate conflict defined?

TJET uses UCDP data for coding conflict-related transitional justice. UCDP defines armed conflict as “a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths in one calendar year.” For more information, see the UCDP Frequently Asked Questions. TJET includes only intrastate conflict, i.e. between a government and a non-governmental party, including internationalized ones.

3 DATA COLLECTION METHODS

3.1 Where can I find the details on how the TJET data are assembled for the website and downloads?

All R code for processing our raw data and preparing the data shown on the website and included in the downloads is provided in our public our public GitHub repository; this makes our pipeline from our development database to our website production database completely transparent. Of course, the credentials for accessing our databases are NOT shared in the repository. Users who want to replicate the transitional justice measures in the TJET country-year dataset should consult the GitHub repository.

3.2 What sources does TJET use for coding transitional justice mechanisms?

Key sources, such as the US State Department’s Country Human Rights Reports, and other databases and datasets that we consulted in assembling the data presented on this website are cited on the Downloads page.

We also provide an extensive, searchable Reference Library on transitional justice, which includes the additional sources used in coding specific mechanisms. For amnesties, reparations policies, and truth commissions, this reference library is searchable by the respective record IDs in the TJET database. For instance, a user interested in a specific truth commission can locate the citations of sources used in coding the TJET data on the commission in question. We have also begun to tag additional sources for coding trials by trialID and accusedID, but this work is incomplete and ongoing.

3.3 How does TJET locate new data on transitional justice mechanisms?

TJET compiles data from a variety of sources with the goal of making information on transitional justice as comprehensive and easy-to-access as possible. In general, our method for collecting data involves finding ‘leads’–which are reports of events that have occurred–and then following those leads with additional research. Leads come from the following three types of sources.

Other databases. Not all of the data presented on TJET are original to the project, though we perform independent research on every single record entered into our own database. Particularly for data prior to 2010, we rely on a series of previous databases and collections, including the Transitional Justice Research Collaborative (TJRC) which was assembled by some current TJET members along with other colleagues in the past. We also draw on the Transitional Justice Database (TJDB), Trial International’s Universal Jurisdiction database, UCDP’s Post-Conflict Justice Database, Louise Mallinder’s Amnesty Law Database, Luke Moffet’s Reparations Database, and the Encyclopedia of Transitional Justice.

News updates. Particularly for data after 2010, we do country-specific searches for news stories on amnesties, reparations policies, truth commissions, and vetting policies around the world. We conduct those searches in Factiva and Nexis Uni, and they provide several leads to follow with additional research.

Human rights reports. Data on amnesties, truth commissions, reparations policies, and vetting procedures are easier to discover through news updates than are data on prosecutions. To find leads on criminal prosecutions, we rely heavily on State Department Human Rights Reports, from which we produce a sourcebook file that contains clipped text containing any reference to criminal prosecutions. Over the course of the project, our research assistants have read and clipped over 5,000 State Department country reports. Where available, we also rely on reports produced by local and regional experts. For example, to retrieve data on all Chilean human rights prosecutions, we reviewed the extensive newsletter documentation created by Observatorio. And to update data on criminal prosecutions in Sri Lanka, we used Bhavani Fonseka’s Elusive Justice & Emblematic Cases in Sri Lanka (2023). Our Argentine data draws on the documents of the Argentine Procuraduría de Crímenes contra la Humanidad. These are just three of many examples.

Some data do not require us to track down leads and follow up with additional research, because we know that we possess the full universe of cases. For example, international criminal prosecutions conducted by the ICC, the ICTY/ICTR, and other hybrid tribunals nearly always have websites that contain information on all criminal cases initiated. In these instances, the sourcing and data collection stages are compressed into a single step because we do not need to look beyond the case files to gather relevant information.

3.4 How does TJET collect and code information on transitional justice mechanisms?

Once TJET obtains new leads on transitional justice mechanisms, we move to data collection and ‘coding’ – a term that we use simply to mean the conversion of information into our project’s data format. This proceeds in three steps.

Coding manual review. For each transitional justice mechanism, TJET seeks dozens of pieces of information that correspond to concepts of theoretical interest. These pieces are defined, explained, and separated into fields using coding manuals. (See Downloads page.) Coders are trained to deeply familiarize themselves with the coding manuals, which will guide their search for additional information.

Additional source gathering. When looking for information prompted in the coding manuals, coders first turn to whatever primary documents are available. For instance, when researching truth commissions, coders would find, download and read mandating laws, as well as final truth commission reports that are available online. Should no primary documents exist, coders are trained to search for additional secondary source material found in Factiva, Nexis Uni, and Google. Any journalistic sources used to compile information are linked in Microsoft Word files, saved to a Dropbox folder, and entered into a Zotero reference library.

Data entry preparation. Once the coder has exhausted their search for additional information on any transitional justice event, and saved their notes in a Microsoft Word file, they are now ready to transfer the information into our structured database via TJET Airtable data entry forms. This preparation stage often involves one more step, which is scanning our database for information TJET has already gathered. For example, some of the fields in our coding manual pertain to the nexus between the mechanism and democratic transition or intrastate conflict. To know whether any mechanism is related to these phenomena, coders review our lists of transitions and conflict spells to match the information they have collected to these contents.

3.5 How does TJET enter information into a qualitative dataset?

TJET uses an online database platform called Airtable to input information on transitional justice mechanisms. Airtable is equipped to handle relational database projects that link together different grids, or spreadsheets, of information organized into columns and rows. The TJET project on Airtable has over 30 grids of data. Some of these grids correspond to specific mechanisms, like truth commissions, while others contain supportive or background information on countries, democratic transitions, intrastate conflict dyads, or cases before the International Criminal Court.

When a coder is ready to add a new transitional justice mechanism into one of these grids, in effect creating another row in one of the grids, they access a form that contains all of the information explained in the TJET coding manuals. They open the form and select the options that correspond to the information that they assembled during the collection and coding phase. Once they have finished filling out the form, they submit it. The data thus enter the TJET database.

Two additional items are worth noting. First, coders will almost never possess all pieces of information prompted in the coding manuals. Where information is absent, coders simply answer that they “do not know.” Missing data is thus quite common in TJET’s mechanism datasets. Second, when a row of data is entered, it always receives a unique identifier, which makes it possible to locate and match to any given event or item in the database.

3.6 How does TJET detect and correct errors?

Any database will have errors, and the TJET database is no exception. We distinguish between incidental and systematic errors, and we devote significant effort to eliminating both types.

Incidental errors. These include coding accidents, such as selecting the wrong field in a dropdown menu, cutting and pasting incorrectly, or simply making typos when inputting items like dates. Some of these errors we discover in the analysis phase. For example, if we see that a single event in the database is coded as having occurred in the year 2982, we log the error and fix it right away. Other incidental errors, like the accidental classification of a trial as a transitional human rights prosecution (THRPs), we fix by doing batch edits. For instance, we might produce a spreadsheet of data containing only the subset of prosecutions classified as THRPs, and have a number of coders check and double-check each row to make sure that it fits our criteria for that category.

Systematic errors. These include errors that are produced by the nature of our data collection methods. One kind of systematic error to which TJET is susceptible is the creation of duplicate records. TJET is built on previous efforts by the Transitional Justice Research Collaborative (TJRC), and overall, our data collection has proceeded in four main waves over time. Coders in the latest wave are bound to encounter news coverage of mechanisms that are already included in the database. When this happens, they often inadvertently add a new mechanism event, instead of modifying an already-existing mechanism. Knowing this, we scour our data for duplicates, marking those cases that might be in need of de-duplication. In the event that we deem a case to be a duplicate, we do our best to merge together the separate entries. However, we never delete any single row from our database. When we are finished with a duplicate, or we find other mechanism entries that do not ultimately meet our criteria, we simply ‘invalidate’ them, meaning that we categorize them as invalid. When we move to dataset preparation and measurement, we exclude all invalid cases.

A second type of systematic error involves the evolution of project concepts, or the revision of previous categories. An example of the former is what constitutes a “gender-attentive” transitional justice mechanism, which we define as a mechanism that confronts instances of sex- and gender-based violence (SGBV). What is included in SGBV is now well-defined by our coding materials, but it was not in previous iterations, which were primarily focused on instances of rape. When TJET altered the definition of SGBV, it created systematic errors in our database. This was corrected by assigning a team to revisit all SGBV-related columns in each grid to make sure that they were coded accurately.

For an example of the latter type of issue, where previous categories are altered, we can point to trial outcomes. Initially, our trial outcome variable consisted of five options: “pending”, “acquitted”, “guilty”, “dismissal”, and “don’t know”. Now, it has 26 possible options, including “acquittal overturned”, “amnesty applied”, and “guilty overturned”, among others. One of the most recent changes was to add “death of accused” as a trial outcome option. Previously, it had been lumped together with dismissal. The creation of a “death of accused” option in trial outcome meant that our prosecution data now contained error; to fix this, the team had to review all of those categorized as “dismissal” and change them to “death of accused” where appropriate.

3.7 How does TJET produce its quantitative datasets?

Though the TJET project is often considered quantitative in nature, the process of quantification begins only at this step, following data location, coding, data entry, and error correction. The raw data produced and saved in the Airtable relational database are mostly qualitative. Recorded there are properties that belong to single transitional justice mechanisms. In our raw database, rows correspond to single amnesties, prosecutions, truth commissions, etc. This mirrors the formatthe simplified tables that are presented on the country pages. Only through a complex and meticulous process are these data transformed into numerical values.

Because TJET is a project driven by the need to create baselines for comparison across countries, our main quantitative dataset is formatted into the country-year unit of analysis, meaning that every row in our the dataset corresponds to a country and year pairing. This allows us to analyze trends and conduct statistical analysis for comparative research. Transforming information on single mechanisms into values that vary by country over time is a challenging undertaking, and all of the decisions that we make to create our measures are explained in our codebook available on the Downloads page.

An example can help illustrate TJET’s process of operationalization. In our raw data on truth commissions, we collect information on the reports issued by each TC, including whether the report recommended prosecutions, reparations, vetting of state agencies, or other judicial and legal reforms. We then create a 0-4 scale based on those four components. In our quantitative dataset, that scale carries forward from the year of the report through the rest of the country panel to which the TC belongs. The value of that scale thus continues until the year 2020, unless another TC occurs in the country, at which point the values are reset and recoded to reflect the outputs of the successor TC. This measure allows us to analyze the effects of TC recommendations on other outcomes of interest, such as democratic institutions reforms in the country.

3.8 How does TJET calculate its legacy of violence index?

In order to comparatively assess the potential demand for transitional justice within countries over time, TJET created a legacy of violence index for 174 countries. For any year, values on the legacy of violence index capture how widespread human rights violations have been in a given country since 1949, compared to the level of human rights violations of all other countries since 1949.

This index uses and builds on the well-established latent physical integrity measure first introduced by Schnakenberg & Fariss (2014), and the most recent version of which was published by Fariss, Kenwick & Reuning (2020). (See the Downloads page for citations.) This measure was designed to assess the protection of core, non-derogable human rights – the rights to life and bodily integrity – in practice over time.

The legacy of violence index is a cumulative and discounted average of the latent physical integrity measure since 1949. It is cumulative because for each country in the dataset it includes every available prior year in the annual calculation. It is discounted because it weighs the recent past more heavily than the distant past.

Two methodological details are worth noting. First, to provide as much history as possible, we back-filled time-series for countries which gained independence later where other data are available. For instance, for post-Soviet states which (re)gained independence in the early 1990s, we used the latent physical integrity measure for the Soviet Union, and similarly incorporated the measure for predecessor countries in other cases where possible. For countries which merged – West and East Germany and North and South Yemen – we averaged the values for the two predecessor countries. Moreover, where the latent physical integrity measure was missing for some country-years, we used a fixed effects linear regression model to impute observations based on physical integrity data available from the Varieties of Democracy (VDem) project, which sometimes provides data for countries before they became independent. Second, as both the latent physical integrity and the VDem data are derived from measurement models, we incorporate their measures of statistical uncertainty into the index calculation.