In July 2009, The Human Rights Data Analysis Group (HRDAG) concluded a three-year project with the Liberian Truth and Reconciliation Commission to help clarify Liberia’s violent history and hold perpetrators of human rights abuses accountable for their actions. (This work was conducted by HRDAG while with Benetech.)
In the course of this work, HRDAG analyzed more than 17,000 victim and witness statements collected by the Liberian Truth and Reconciliation Commission and compiled the data into a report entitled “Descriptive Statistics From Statements to the Liberian Truth and Reconciliation Commission.” The report is included as an annex to the final report of the Liberian Truth and Reconciliation Commission (TRC) released on July 1, 2009, in Monrovia, Liberia. The Liberia Truth and Reconciliation Commission data and the accompanying data dictionary are available on our website.
The HRDAG report found that former Liberian president Charles Taylor—who was tried and found guilty in The Hague for war crimes and crimes against humanity in Sierra Leone’s civil war—also led the Liberian rebel group responsible for the largest number of violations during Liberia’s 24 years of civil unrest.
Africa’s oldest republic, Liberia was founded on the west coast of Africa in 1847 by freed slaves from the United States. It is home to approximately 4.1 million people including people from 16 main tribes, numerous smaller and sub-tribal groups, and descendants of freed slaves from the United States and the Caribbean. (Population is a 2011 estimate.)
A military coup in 1979 sparked decades of violence in Liberia that caused great suffering and mass dislocation of citizens. Since the end of the conflict, the United Nations Mission in Liberia has maintained 15,000 UN peacekeepers throughout the country.
Established by the 2003 Comprehensive Peace Agreement that ended the conflict in Liberia, the Truth and Reconciliation Commission (TRC) is currently the only mechanism for accountability in Liberia. The TRC held hearings throughout Liberia to gather the 17,000 witness statements and investigate widespread human rights violations and infringements of humanitarian law during the conflict. It is notable that the Liberian TRC collected almost 48 percent of their statements from women. The Liberian TRC was more successful in encouraging women to submit statements than truth commissions in other countries where women are often not reflected in a proportional number.
Warring factions in Liberia’s civil war subjected the civilian population to severe human rights abuses including summary execution, sexual violence, forced displacement and property destruction. As part of its broader mandate to “promote national peace, security, unity and reconciliation,” the TRC sought to determine whether these violations form part of a systematic pattern or policy of abuse. Since the international community has not pushed for the creation of an ad hoc tribunal or other legal mechanism for bringing justice to the citizens of Liberia, the TRC’s role is key in recommending next steps in Liberia to ensure accountability and end years of impunity.
A significant challenge for all truth commissions is establishing the magnitude, pattern and relative levels of responsibility for “what happened” during a period of mass human rights violations. The TRC requested HRDAG’s assistance from Benetech to develop a data collection and analysis process to address key questions about the conflict and violations that occurred.
HRDAG provided training and support to help the TRC accurately and defensibly quantify information about human rights. The HRDAG team analyzed the data to examine overall patterns and trends of violations experienced or perpetrated by the statement givers. Together, the aggregated statements magnify the voices of the victims and provide a body of empirical data that supports the process of acknowledgement and accountability in Liberia.
The HRDAG Analytic Study
The HRDAG analytical study is based on statistical analysis of victim and witness statements collected by the TRC that correspond to human rights violations from January 1979 to October 2003. The report also considered other information about human rights abuses during this period collected from local and international NGOs.
The purpose of the HRDAG report is to outline and interpret the nature and extent of the violations, the behavior of the perpetrators and characteristics of the victims reported to the TRC in statements. The report presents information about the statement-givers and the testimonies they provided the TRC. Subsequent sections analyze the recorded acts of violence in-depth—over time, by county, by victim characteristics, perpetrating groups and violation types. The report also analyzes statement-giver responses to supplemental questions and statements provided by Liberians living outside the country. The HRDAG Liberia report concludes with a look at the implications of the findings and suggestions for further analysis.
The HRDAG report was drawn from 17,160 of the 17,416 statements entered into the TRC database. The analysis excludes 256 statements because the statement givers reported no violations with the TRC’s mandate period, or because the county or country where the statement was taken was not recorded.
The 17,160 statements contain information about 86,647 victims and 163,615 total violations including 124,225 violations suffered by individual victims; 39,376 suffered by groups; and 14 by institutions. The HRDAG report also analyzed 1,165 statements from diaspora Liberians; the reports contained information about 6,398 victims and 10,154 violations. The majority of statements collected from diaspora Liberians were provided by statement-givers in Ghana.
The analysis in the HRDAG report does not necessarily represent all the violence that took place in Liberia during the period of study. The information is incomplete because some victims did not choose to give statements due to illness, fear, intimidation, or because they live in remote areas and were not contacted by statement-takers.
Despite these challenges, the TRC was able to document tens of thousands of violations, more than any other previous truth commission. It is worth noting that the Truth and Reconciliation Commission in South Africa collected approximately 21,000 statements in a country nearly 14 times the size of Liberia.
Insights Provided by the HRDAG Report
The HRDAG quantitative analysis provides insight into many features of the Liberian conflict, including the duration of the violence. The report shows that violations reported to the TRC spiked in 1990 with the second most violations reported in 2003 and the third greatest in 1994.
Forced displacement was the most commonly reported violation comprising approximately one-third of reported violations. The TRC also documented more than 28,000 killings, the second most commonly reported violation after forced displacement. The HRDAG report includes analysis of reported violations by type for different perpetrator groups. It also looks at patterns or reported violations by select type over time.
The report includes a quantitative analysis of 1,165 statements from members of the Liberian diaspora collected in collaboration with U.S.-based Advocates for Human Rights. Patterns in violations in diaspora statements reveal notable differences to patterns in violations reported in statements collected in Liberia, including a near lone spike in reported violations in 1990 and a small second rise in 1996. Violations reported in diaspora statements overall took place overwhelmingly in Montserrado County, where the capital city of Monrovia is located. When reported, the victim tribe in diaspora statements was predominantly Krahn.
The HRDAG analysis also reveals broad distribution of violations geographically throughout the country. But the report notes that these results could be an artifact of how the TRC deployed statement-takers across Liberia’s 15 counties. The high numbers of violations reported in rural Bong and Lofa counties are notable, however, given the higher proportion of statements collected in Montserrado County. This suggests that many statement-givers in Montserrado reported events that occurred in other counties. It could also reflect the high level of forced displacement during the conflict and a pattern of migration to the capital city by people forced to leave their homes.
The report also reveals surprising information about the age of victims of the violence during the period studied. The HRDAG analysis indicates that among killings reported to the TRC, men of an increasingly older age were at greater risk for being killed or subject to looting violations than younger men. In contrast, the data suggests that young men, between the ages of 15 and 19 in particular, were at greater risk for forced recruitment as a combatant. A possible interpretation of the killing and forced recruitment information is that perpetrators avoided young people for killing, targeting them instead for forced recruitment into their ranks.
The process of analyzing the TRC statements illustrates some of the challenges posed by missing data, particularly data about age. Due to incomplete information about the age of victims, HRDAG analysts cannot be certain that age-based patterns and estimates of relative risk represent the true patterns in the statements given to the TRC. Further, this analysis shows only the direct effects of violence, ignoring the higher rates of death among the very young and old that often accompanies forced migration.
The information from the statements includes very few reports of rape for which the victim’s age is known. But the data does indicate that the majority of reported rapes for which the victim’s age is known were committed against adolescent women, rather than against socially taboo groups such as older women or very young children. Not surprisingly, the majority of rapes reported to the TRC, in which the victim’s gender was known, were committed against women. But the data also shows that a small number of men were also raped during the violence.
The proportion of rapes with female victims aged 15-19 represents more than five times the proportion of women aged 15-19 in the general population. In contrast, the data indicates that relatively more men than women were victims of sexual abuse. The definition of sexual abuse included stripping the victim naked. During the Liberian conflict, this tactic was employed by many perpetrator groups as a means of humiliating victims.
The analysis shows that men in general, and men in a number of age categories, are overrepresented for killing, assault, torture, forced labor and forced recruitment violations. Women are underrepresented, except in those age 70-74 who had a higher representation among people killed.
The HRDAG report analyzes the degree to which specific tribal groups were subject to violations. It also provides information about violations attributed to a specific group of perpetrators. The report presents the total number of violations attributed to each group of perpetrators—and the percent of total violations reported to the TRC for each group.
The data indicates that the National Patriotic Front of Liberia (NPFL) headed by former rebel leader Charles Taylor, is responsible for more than three times the number of reported violations as the next closest perpetrator group, the Liberians United for Reconciliation and Democracy (LURD). In 2003, the LURD besieged the city of Monrovia in an attempt to dislodge then-president Taylor. The NPFL was identified as the perpetrator of approximately 40 percent of the violations reported to the TRC. Taylor was found guilty in April 2012 of all eleven charges levied by the Special Court, including terror, murder and rape. In May 2012 he was sentenced to 50 years in prison. Reading the sentencing statement, Presiding Judge said, “The accused has been found responsible for aiding and abetting as well as planning some of the most heinous and brutal crimes recorded in human history.”
Beginning at its inauguration in June 2006, the TRC partnered with Benetech’s Human Rights Data Analysis Group to develop a data collection and analysis process to address key questions about human rights violations and the nature of conflict in Liberia.
At the time, that TRC was the ninth truth commission that HRDAG had assisted to incorporate information technology and scientific methods. HRDAG has helped these commissions establish analytical objectives, collect data, design and implement information management systems, conduct statistical analysis, integrate quantitative findings and provide follow-up support.
While at Benetech, HRDAG’s Kristen Cibelli worked for three years with TRC staff in Liberia to develop a coding process by which “countable units”—violations, victims and perpetrators—are identified in statements and transcribed to coding forms. This process converts the qualitative information contained in the statements into data that can be used for quantitative analysis.
The coding process allowed the HRDAG team to count violations by county, by year, etc., to precisely analyze the patterns of human rights violations reported to the TRC. For example, what distinguishes “rape” from “sexual abuse”? The two categories must be defined clearly so that people doing the coding apply the definitions in a standard way. The definition must be so clear that if the same narrative statement is assigned to the entire coding staff, they will classify it in precisely the same way. HRDAG refers to these definitions as the “controlled vocabulary.”
To create this “controlled vocabulary,” the TRC defined 23 violation types based on the nature of the violence in Liberia and the TRC’s analytical objectives. These violations include forced displacement, killing, assault, rape, forced labor, forced recruitment and robbery. The HRDAG analytic report presents the total number of reported violations for each violation type and the percent of all reported violations for each type.
When more than one person is working on coding statements, it is important to monitor inter-rater reliability (IRR). IRR measures whether different coders, given the same source material, produce the same quantitative output (e.g., the same number of victims and the same number and type of violations). High levels of IRR, or agreement between the coders, ensure that the information entered into the database is more than the individual interpretations of each of the coders.
This process is crucial to the quality of data analysis as the coding team expands. In September 2007, the coding team increased from three coders to eleven, and then in May 2008 to sixteen. In Liberia, the coding team has achieved an overall average of 89 percent agreement on coding exercises throughout their work on TRC statements. By social science standards, this is a high rate of IRR.
Representing the Complexity of Human Rights Violations
Coded information about the violations in Liberia were entered into the TRC database and then extracted for analysis by HRDAG statisticians. This information was then used to produce the statistical report. The process is designed to conform to scientific norms as well as internationally accepted statistical and technological best practices. It allows the raw information, carefully collected by the TRC in statements, to be accurately represented in a defensible analysis.
But there is a considerable amount of complexity that must be managed when counting human rights victims and violations. Consider these facts:
- Victims can suffer many violations
- The violations can happen at many different times and places
- Each violation may be committed by one or many perpetrators
- Each perpetrator may commit one or many violations
A person can be a witness, victim and/or perpetrator within a sequence of events. A data model must be able to accurately reconstruct which victims suffered which violations committed by which perpetrators. Simplifying these points leads to distorted statistical results that can be attacked by opponents of the process. HRDAG has developed the “Who Did What to Whom?” data model to capture and maintain the complex relationships between different elements, roles and events.
The most effective way of managing the relationships between different interdependent pieces of information is with a relational database. HRDAG developed Analyzer, a database tool based on the “Who Did What to Whom?” model which is specifically designed to organize human rights data for statistical purposes. Analyzer manages the challenges involved in structuring and quantifying human rights data.
Different projects need to analyze different variables according to their specific human rights situation. HRDAG worked closely with the TRC to identify and add custom data fields needed for the TRCs work. The TRC hired a Database Manager and an initial team of three Data Entry Clerks when the customized Analyzer database was installed in October 2007. Two additional Data Entry Clerks were hired in December 2007 and six in March 2008 in order to increase the speed of data entry.
The TRC database server and computers were set up on a network separate from that connecting other workstations at the TRC and they were not connected to the Internet. Maintaining the database network independently of the rest of the TRCs network and off the Internet strengthened its security and prevented infection from viruses.
The TRC Database Manager conducted backups of the database to ensure that the database could be recovered in case of theft or failure of the TRCs database server. Copies of the database backups were stored on-site as well as encrypted and sent securely via the Internet for remote storage. The data from coded statements captured in Analyzer was securely backed up and transmitted to Benetech’s HRDAG for final processing and analysis.
Analysis: Patterns of Reported Victims and Violations
When documenting human rights situations, different statements may describe the same event. For instance, the same killing may have been reported by multiple statement-givers. Victims documented in the testimonies are counted once in each of the counties in which they suffered a violation. Victims with several reported violations in more than one county could be counted more than once.
As a result, an unknown amount of duplication of reported violations exists in the database. Duplicates were not identified or systematically removed from the TRC’s data. However, HRDAG statisticians applied an approximate matching exercise to assess the effect of duplication on the patterns of reported killing violations. They concluded that the duplication present in the TRC statements does not alter the qualitative conclusions about the patterns of killings across time, age, sex, or perpetrator.
Applying Responsible Data Processing Principles
The data extracted from the TRC database was reformatted from the database to be read into R, a statistical tool used to generate the analysis, graphs and tables presented in the HRDAG report. HRDAG uses R in conjunction with LATEX, SWeave (LATEX plus R), make, and Subversion (version control software) in an infrastructure based on the HRDAGs data processing principles of transparency, auditability, replicability and scalability.
Transparency means that other HRDAG team members or reviewers from outside of HRDAG can follow each step of the HRDAG’s work. Auditability means that it is possible to track each step of the analytic process and its subsequent output, facilitating testing. Replicability means that the analysis can be re-run by another HRDAG-team member, reviewer or independent third-party, at any time. Scalability means that, because of the transparency of the project structure and analytic process, HRDAG can bring other team members into the project with minimum overhead and maximum efficiency at any time, as well as accommodate growing amounts of data.
The principles that underlie the HRDAG analytic process enabled the team to rapidly reproduce its analysis in response to feedback and requests from the TRC and the addition of more statements to the database. It also ensures that results are transparent for review by TRC colleagues and peer reviewers and can withstand close scrutiny by commentators.
In addition to narratives about violations, the TRC asked statement-givers a series of questions about the impact of the conflict on the statement-giver and their views on what Liberia needs to support the process of reconciliation and recover from conflict. The HRDAG report presents responses to these supplemental questions broken down by county.
From the data collected, the HRDAG report infers that respondents name “poverty” and “destroyed source of livelihood” as the main economic impacts of the conflict followed by “academic backwardness.” Across all counties, between 50 and 70 percent of respondents were willing to meet with the perpetrator who caused their suffering. This openness to reconciliatory measures suggests success for possible future reconciliation initiatives. Statement-givers across the country unanimously agreed and recommend that a practice of “Forgive and Forget” would foster the process of reconciliation in the country followed by “Peace Programs.”
Finally, statement-givers were asked to provide personal recommendations to the Government of Liberia and to the TRC respectively. Results indicate that the main expectation towards the Liberian government in all counties is “Good Governance,” followed by “Reconstruction,” “Job Opportunities,” and “Forgive and Forget.” As for recommendations to the TRC, a broad majority of respondents agreed that the TRC should “Carry out its mandate,” with “Reparations” ranking as the second highest priority.
By supporting the effective capture, preservation and analysis of statements relating to human rights violations, the TRC has been able to tell a broader truth about Liberia’s conflict. HRDAG provided the expertise to transform information into scientifically-defensible knowledge to create a clear historical record and help end impunity for the perpetrators of human rights abuses.
An anonymized version of the TRC’s data from statements collected in Liberia, and among diaspora Liberians, has been published and is available on the HRDAG website. HRDAG encourages scholars and other analysts to extend the analysis and compare statistical results from other sources of data with the information reported by the statement-givers.
The results of this analysis could impact potential recommendations of the TRC for prosecution, amnesty, reparation and reconciliation or forgiveness. The findings will also offer a new perspective on the history of Liberia. Providing a contextual description for how and why such a set of violations occurred provides a deeper understanding of the possible causes lying behind the patterns of violations.
The findings in the HRDAG report will also be of value to scholars, lawyers, historians, journalists, human rights and civil society groups and the families of the victims for generations to come.
> Kristen Cibelli, Amelia Hoover, and Jule Krüger (2009). “Descriptive Statistics From Statements to the Liberian Truth and Reconciliation Commission,” a report by HRDAG at Benetech and Annex to the Final Report of the Truth and Reconciliation Commission of Liberia.