Previous PageTable Of ContentsNext Page


A Note on Methodology
The Methodological Approach to
Documentary, Testimonial and Forensic Evidence Used in This Report

Testimonial Evidence

To a large measure this report is based on testimonies obtained in Iraqi Kurdistan from eyewitnesses to (and often victims of) Anfal-related abuses. Two Middle East Watch researchers and an assistant spent a total of six months in the Kurdish areas on three separate missions between April 1992 and April 1993, conducting approximately 350 in-depth interviews. The methodology they used in obtaining this testimonial evidence is described below.

Prior to its first mission in April 1992, the research team designed a questionnaire on the basis of our understanding of Anfal, still limited at that time, and following discussions with regional experts and statisticians. This questionnaire was constructed with a view to facilitating the tabulation and quantification of data concerning the forced displacement and/or disappearance of Kurds during Anfal.

The team tested the questionnaire through a small number of interviews shortly after its arrival in the area, and immediately determined that the questions did not take account of a number of factors, including specific historical events that are instrumental to understanding the circumstances that surrounded Anfal, as well as the methodical nature of the operation. The team then revised the questionnaire drastically and began its research, only making minor adjustments to the basic design in the weeks that followed.

The purpose of the research was to find out as much as possible about Anfal and about the people who were said to have disappeared during and after the operation. The research population was divided into three groups: (1) direct eyewitnesses to Anfal-related abuses; (2) persons active in (para-)military units, either Kurdish guerrillas (peshmerga), former military officers, or leaders (mustashars) of the pro-governmentKurdish militias; and (3) staff of local and international non-governmental organizations and officials of the local Kurdish administration, all of whom were intimately familiar with the situation on the ground before, during and after Anfal.

Because of the particular nature of Iraqi policy vis-a-vis rural Kurds in the 1980s, most eyewitnesses to Anfal were to be found in the large housing complexes (mujamma'at) in the valleys of northern Iraq. After the Iraqi government withdrew from a large part of the Kurdish region at the end of October 1991, villagers in some areas began returning to their destroyed villages to farm the lands and, sometimes, to rebuild their homes. The research team visited as many of the complexes as possible, as well as some of the (partially rebuilt) villages. In all cases, the operative question was: "Where can we find the Anfals?" (Anfalakan, in Kurdish). Local residents would then guide the team to a house where Anfalakan were said to be living, and some initial questioning was done to ascertain that the people were indeed "Anfals" and not persons who had been relocated there from their villages during earlier stages of village destruction and had not been affected directly by Anfal. This method had a snowball effect: one family of Anfalakan would lead the team to another, until the team felt it had exhaustively covered a particular geographic area of Anfal.

Essentially, the team obtained eyewitness testimonies in three different ways: (1) by visiting places of residence randomly and asking for Anfalakan (the most frequent method); (2) by pursuing specific leads; and (3) occasionally, by responding to unsolicited requests to be interviewed. In the beginning, the sole criterion employed in deciding whether or not a person should be interviewed was whether the person was present in a military-demarcated area during Anfal and had lost relatives as a result of the campaign. In later stages of the research, when clear patterns had started to emerge, the search was more specifically for persons from certain Anfal areas, i.e., those about which the team had insufficient data, or those where particularly egregious abuses, like chemical weapons attacks, had taken place. In addition, a number of interviews were conducted at that time with people who had been in Anfal areas during Anfal but whose families had managed to escape unhurt, as well as with people who had experienced various forms of human rights abuses in the periods immediately preceding and following Anfal (1987 and 1989).

The team specifically sought out one sub-population of eyewitnesses: those who had been arrested in Anfal and taken to massexecution sites (then and now in areas that are controlled by the Iraqi government) from where they had managed to escape and return to safety. The testimonies of these execution survivors have proven crucial in the effort of Middle East Watch to provide evidence that the vast majority of those detained during Anfal, whose fate is officially said to be unknown, were actually killed. The team was able to locate seven such Anfal survivors, as well as one person who had survived an execution three months after Anfal. Some of these survivors did not want their identities to be known because, they said, they feared future government reprisal. One of the eight, Taymour, had already been interviewed by local Kurdish television, as well as by foreign journalists on numerous occasions. A second one, Hussein, had given testimony to the UN Special Rapporteur on Iraq, Mr. Max van der Stoel, during the latter's visit to the area earlier in 1992. Four were located through local peshmerga commanders who had heard of their stories. The remaining two were found through the testimony of one of the survivors who had been in the same group as them at the time of the execution.

Invariably, respondents were eager to tell the team what had happened to them. In almost all cases, these people had not been interviewed about their experiences before. All freely gave their names, and only some requested that their identities be concealed for publication. Apart from the small number of persons who requested that their names be changed, all names referred to in this report are genuine. The team taped most interviews on audio-cassette and took photographs (slides) of the respondents afterwards. In the case of important interviews, the team asked the respondent's permission to videotape either highlights or a full second interview (as in the case of some of the execution survivors). The team traveled with one or more Kurdish interpreters at all times. These persons were asked to provide literal translations from English to Kurdish (Surani or Kurmanji dialects) and back. In addition, some interviews were conducted by team members directly in English or in Arabic.

In virtually all cases, the team interviewed a single person at a time, although close relatives were often present during the interview. The questions covered the following topics:

1. Personal history before Anfal (personal status, family members, property, occupation, religious and tribal affiliations, etc.)

2. Information concerning the village in which the person was living before Anfal (location, size of population, main tribe, main economic activity, availability of goverment services and facilities, etc.)

3. Military activity in and around the village before Anfal, and government policies affecting the inhabitants (presence of peshmerga, government attacks, administrative and economic blockade, casualties, 1987 population census, etc.)

4. Events during Anfal (nature of government attack, circumstances of arrest, route of transport, selection process, conditions of detention, casualties, circumstances of release, etc.)

5. Living conditions after Anfal and attempts, if any, to locate missing relatives.

Usually, topics 1 and 2 followed a fairly strict question-and-answer format, while topics 3, 4 and 5 allowed for greater flexibility: the person was asked to recount events as he or she remembered them, and the team would (a) only ask questions for clarification about specific dates, locations or identities; or (b) pursue at some length a narrative of particular interest to the project; or (c) probe any contradictions that might appear in the testimony, or between the testimony and a previous one.

Due to the high incidence of illiteracy in rural Kurdistan, as well as the local population's particular way of marking time, the team encountered considerable difficulty in its attempts to establish exact dates for specific events, or particular chronologies, on the basis of interviews with individual villagers. Dates would often be related to religious feasts, for example. On the whole, though, after numerous interviews, the team was satisfied that it had obtained an accurate picture of the separate stages of Anfal and the events that transpired within these stages. Some of these dates have subsequently been substantiated in documents captured by the Kurds from the Iraqi intelligence agencies during the March 1991 uprising.

Generally, the team determined the accuracy of individual accounts by virtue of their internal consistency, their general consistency with the overall patterns that emerged during the project, including with other types of evidence, and their specific consistency with a follow-upinterview conducted, in a few cases, with the same respondent. In the case of all interviews, the team tried to obtain supporting evidence. This included personal documents that were in the possession of the respondents (for example, "movement permits" and administrative orders), or an inspection of the site described (e.g. a prison, or a village that had been subjected to a chemical attack). As a result of this procedure, a small number of interviews, or segments thereof, were discarded or left unused, either because the testimony was not deemed to be reliable or sufficient supporting evidence was not immediately available.

* * *

Documentary Evidence

In the March 1991 popular uprising in northern Iraq, Kurdish civilians and members of the Kurdish political parties stormed and took control of offices of the Iraqi government and its agencies, including the various intelligence agencies. Several of these buildings were heavily damaged, or even burned to the ground, but others survived unscathed. The Kurds thus came into possession of the inventories of many of these facilities. Matters taken include large quantities of documents, logs and registers, as well as audiotapes, videotapes, films and photographs.

In the days before the uprising was crushed by advancing Iraqi troops, the Kurdish parties succeeded in removing the majority of the documents they had captured from the towns to strongholds in the mountains. In the spring of 1992, one of the two largest parties, the Patriotic Union of Kurdistan (PUK), agreed to a tripartite arrangement in which Middle East Watch and the U.S. Senate Foreign Relations Committee were the other two partners. Under this arrangement, the PUK agreed to send the documents in its possession to the United States for research and analysis; the Senate Foreign Relations Committee agreed to turn the documents into official records of the U.S. Congress and store them in the facilities of the U.S. National Archives; and Middle East Watch agreed to conduct research on the documents for human rights purposes, including the pursuit of a genocide case before the International Court of Justice at The Hague.

The PUK cache consists of fourteen tons of documents contained in 847 boxes. The total number of pages has been estimated at over four million. In May 1992, the PUK placed these documents in the temporary custody of Middle East Watch and they were then flown, in the presence of the director of Middle East Watch, to the United States. In Washington, D.C., the documents were then handed over to the U.S. National Archives and placed in its storage facilities, while remaining under the joint custody of the PUK and Middle East Watch.

At the end of October 1992, a Middle East Watch-led team of researchers began the task of screening, cataloguing, and analyzing these documents. Important documents were pulled out, photocopied and translated, and a number of these have been included in this report. The United Nations Special Rapporteur on Iraq, Mr. Max van der Stoel, has also made use of some of the documents found by Middle East Watch, in his report to the U.N. Human Rights Commission in March 1993.

The vast majority of the documents are from two main locations: Suleimaniyeh governorate and its districts; and Erbil governorate and its districts, especially the qadha of Shaqlawa. Almost all hail from the offices of Iraq's General Security Directorate (Mudiriyat al-Amn al-Ameh) in these locations, with a smattering of documents belonging to the General Military Intelligence Directorate (Mudiriyat al-Istikhbarat al-Askariyeh al-Ameh) and the Ba'ath Party. Generally, the documents are either file folders or pages tied together with shoelaces between two hard covers. There are also a number of bound ledgers. Due to the conditions that prevailed at the time of the uprising, especially in Suleimaniyeh, a number of the documents have completely fallen apart, and individual pages have been burned, trampled upon, muddied, or, in many cases, torn. The majority of the documents are in good order, however.

All documents, handwritten or typewritten, are in good and legible Arabic. They cover a wide variety of subjects that can most easily be divided into three main categories:

(1) administrative matters concerning agency staff: salaries, vacations, promotions, gun permits, disciplinary actions, etc.

(2) Personal information: These are files containing information on agency staff, ordinary citizens, or suspected members of Kurdish resistance parties. They include secret background checks, as well as records of investigations and interrogations. A number of files containvirtual life histories, some concluding with execution orders and death certificates.

(3) Reports on events in the area, and policy statements: These two types of documents are often mixed in together, and include reports on Amn and military actions undertaken against the peshmerga; reports on peshmerga activity in a particular area; and official orders and instructions that are being passed on down through the ranks.

Although the three categories generally appear separately in the documents, it does happen that a copy of an important order concerning the government's policy vis-a-vis the Kurds will appear in a person's secret file. Sometimes, evidence of abuse is either fragmentary and embedded in a larger text that outwardly seems innocuous, or phrased in such euphemistic terms that an untrained eye would have difficulty in recognizing it. The task of the Middle East Watch-led team, then, has essentially been to sift through these tons of documents in search of hidden nuggets.

* * *

Forensic Evidence

A team of forensic investigators was sent by Middle East Watch and Physicians for Human Rights to Iraqi Kurdistan in May-June 1992. The team consisted of forensic investigators trained in forensic anthropology, archeology, and law, who had carried out exhumations of graves in several countries, including Argentina, Chile, El Salvador and Guatemala. It carried out exhumations of graves at three sites in Iraqi Kurdistan: at the village of Koreme, the village of Birjinni, and the cemetery of a complex of Anfal survivors outside the city of Erbil.

In its investigations, the team followed internationally accepted standards set forth in the United Nations "Model Protocol for a Legal Investigation of Extra-Legal, Arbitrary and Summary Executions" (the "Minnesota Protocol").1 The full results of the team's investigations arefound in Middle East Watch and Physicians for Human Rights, The Anfal Campaign in Iraqi Kurdistan: The Destruction of Koreme, January 1993, and its methodology at each site is described below.

Koreme Site:

The team undertook the exhumation of a mass grave at the destroyed village of Koreme containing the skeletal remains of twenty-six men and boys, all of whom had died by gunfire at close range in a line indicating execution by firing squad. The team archaeologist directed the survey of the destroyed village, mapping the village as it stood before destruction, using standard archeological survey techniques; in addition, the team archeologist directed the collection and mapping of cartridge casings to determine the pattern of weapons firing at the execution site. The team's lead anthropologist directed the exhumation of the gravesite at Koreme, using standard exhumation techniques to preserve the skeletons and other artifacts. Investigations were carried out at the morgue at Dohuk General Hospital to determine the number of different individuals in the grave; sex, age, and other identifying marks; and manner of death. The team's lawyers directed interviews with survivors and other villagers to give a narrative of events corroborated by scientific evidence.

Birjinni Site:

The team carried out archeological surveys and exhumations of graves at the destroyed village of Birjinni which, according to surviving villagers, had been bombed in August 1988 with chemical weapons. The team archeologist carried out standard surveys of the ruined village. The forensic anthropologists exhumed the graves of persons reported to have died from inhalation of chemical agents. The team's lawyers conducted interviews with surviving villagers to obtain their account of events. In addition, the team took soil and other samples from the craters where chemical weapons were reported to have impacted. In 1993, the British Ministry of Defense chemical weapons laboratory at Porton Down reported discovering degradation products of mustard gas and nerve agents in samples taken from these sites. This is the first instance of a chemical weapons attack being proved on the basis of chemical residues left behind at the impact site.

Erbil Site:

The team undertook exhumations at the graveyard of a complex where survivors of the Anfal were taken. The gravesite was surveyed by the archeological team in order to make determinations of the ratio of adult and child graves in the cemetery. The forensic scientists exhumed three children's graves, one of them reported to have been made by a survivor from the village of Koreme and containing his infant sister. The exhumation of that grave corroborated his account, and contained the skeletal remains of a girl about one-year old, with evidence of malnutrition.


1 Manual on the Effective Prevention and Investigation of Extra-Legal, Arbitrary and Summary Executions, 1991.

Previous PageTable Of ContentsNext Page