Fear of Reprisal – Researching Intra-Communal Violence in Northern Ireland and South Africa1
Colin Knox and Rachel Monaghan
Published in R.Lee and E. Stanko (eds) Researching Violence: Essays on Methodology and Measurement. London: Routledge, 2003: 157-176
Colin Knox and Rachel Monaghan School of Policy Studies
University of Ulster Newtownabbey Northern Ireland BT37 0QB
1 The authors wish to acknowledge the support of ESRC funding (grant L133251003) in carrying out the work for this chapter and research assistance of Dermot Feenan in the early stages of the project.
Telephone: 028 90366378/90366667 Email: firstname.lastname@example.org
Fax: 028 90366847 18th February 2002
Measuring the impact of violence is fraught with problems, not least because of the definitional problems around what constitutes violence and how this might be quantified. If one considers violence perpetrated by paramilitaries in Northern Ireland or vigilante groups in South Africa then the task becomes even more difficult. This chapter will examine methodological problems associated with gaining access to, and undertaking research in, the dangerous arena of paramilitary violence in Northern Ireland and vigilante violence in South Africa. It will consider the difficulties in obtaining reliable information on the levels of violence particularly from official police sources, and the way in which this type of crime is classified. Other statistical sources such as pressure group data will be examined and problems highlighted with the use of documentation gleaned from tendentious organisations. Data about the impact of violence from primary research are also fraught with problems. Victims, for example, can be reluctant interviewees for fear of paramilitary/vigilante reprisal. A
‘victim’ of paramilitary/vigilante ‘punishment’ may be an erstwhile perpetrator of violence. Should one treat internecine turf wars between paramilitaries in Northern Ireland in a similar way to violence meted out in pursuit of their political goals? These and other issues make the measurement of the impact of violence difficult methodological questions. This chapter will therefore highlight these problems and examine how we managed them within this study of intra-communal violence.
The research context South Africa
South Africa and, more tentatively, Northern Ireland are emerging from bitter ethno- national conflicts in which violence and crime characterised the transition to peaceful political settlements. The collapse of apartheid in 1989, lifting the 30-year ban on the African National Congress (ANC) and the subsequent release of Nelson Mandela, created a climate for political negotiation and change in South Africa. This paved the way for an interim constitution, the first multi-racial democratic elections in 1994 and led to the Government of National Unity. The ANC’s success in the June 1999 elections gave the party an overwhelming mandate to accelerate Thabo Mbeki’s programme of ‘transformation’ aimed at tackling the significant socio-economic problems facing South Africa: unemployment, AIDS, crime and education. The legacy of political resistance, often violent, deployed to make the townships ungovernable during apartheid has created a culture tolerant of citizens taking the law into their own hands. Although the number of political killings dropped sharply from about 2,500 in 1994 to fewer than 240 in 1999 (South African Institute for Race Relations 2000), Mbeki in his inauguration speech regretted that some South Africans were ‘forced to beg, rob and murder to ensure that they and their own do not perish from hunger’. This is reflected in a rising tide of other kinds of violent crimes. Rape, car-jacking, serious assault, housebreaking and common robbery, have been increasing since 1996, and the trend has been sharply upwards since 1998. About a third of all reported violent crimes in 1999 were violent, and the number increased by over 9 per cent on 1998. The savagery of the crime wave is captured in reports that one in every two South African women will be raped during their lifetime, the average South African is eight times more likely to be murdered than the average
American, and one policeman is killed each day - 1,400 have died since the ANC came to power (Weekly Mail and Guardian 2000). Accordingly, the public response is that ‘brutality should be met with brutality. The rich surround themselves with razor wire and private security guards, and the poor resort to vigilantism’ (The Economist 1999: 23).
Vigilantism is undertaken by organised groups such as Mapogo a Mathamaga2 in the Northern Province, People Against Gangsterism and Drugs3 (PAGAD) and the Peninsula Anti-Crime Agency4 (PEACA) in the Western Cape. In addition, taxi associations in some townships have become involved in ‘crime solving’ for a fee. All of these groups stand accused of using corporal punishment and violence in responding to crime. Indeed Mapogo’s leader, John Magolego asserts that public flogging, ‘is the African way of stopping crime. The criminal must lie on the ground, and we must work on his buttocks and put him right’ (Quoted in Soggot and Ngobeni 1999). Alleged suspects are usually beaten until they confess or provide information as to the whereabouts of stolen goods or moneys. Mapogo has also been accused of throwing suspects into crocodile-infested waters, while taxi-drivers in Guguletu (a
2 Mapogo a Mathamaga was established in 1996 and has some 50,000 members who pay a monthly subscription to the organisation in return for protection against crime.
3 PAGAD was formed towards the end of 1995 and targets drug dealers and gangsters. Between March and July 1998 they targeted 86 alleged drug dealers and succeeded in killing 24.
4 PEACA is based in Khayelitsha, a township near Cape Town. It was established in August 1998 by ex-combatants of the liberation struggle who came together to fight crime and its members number 1,500.
township in Cape Town) are implicated in dragging alleged criminals behind vehicles.
Further to its crime solving activities, Mapogo has moved into the area of crime prevention by offering services usually provided by private security firms such as the protection of property and patrolling; takers have included schools and churches.
Furthermore, in some cases spontaneous mobs form to mete out justice to alleged criminals. In some instances, those present convene kangaroo courts5 but this is not always the case. The justice meted out is often of an extremely brutal nature and deaths are common. Examples in the townships include the stoning to death of three youths found stealing chairs from a church in Pimville, Soweto (Khangale 2001); the severe beating of an alleged rapist by women in Chatsworth (Durban) (Williams 1999) and the near necklacing6 of a man accused of armed robbery in Orange Farm (Johannesburg) (Ndaba 2001). In some cases, members of a suspected criminal’s family or the family home are targeted. Their houses are either destroyed or burnt and often the families subsequently leave the area as a result of intimidation (Ntabazalila 1997 and interview with the mother of an alleged rapist and murderer, November 1999). In February 1997, the mother of an alleged criminal was stoned to death for the deeds of her son by a crowd of 4,000 in the township of Mamelodi, near Pretoria (Amupadhi 1997).
5 Kangaroo courts are characterised by the assumed guilt of the accused, denial of due process and instant punishments, usually violent in nature.
6Necklacing involves placing a petrol-filled tyre around the victim’s neck which is then set alight.
Northern Ireland’s transition to ‘peace’ has been more recent and capricious. The signing of the Belfast Agreement in April 1998 and its subsequent endorsement in referenda by its electorate (71.2 per cent) and voters in the Irish Republic (94 per cent) heralded a political solution to the seemingly intractable problems that bedevilled the province for 30 years. The British and Irish Governments formally resolved their historical differences through the general and mutual acceptance of the principle of consent – Northern Ireland is part of the United Kingdom, and will remain so, as long as the majority wishes. Signatories to the Agreement affirmed their
‘total and absolute commitment to exclusively democratic and peaceful means of resolving differences’. This has created the impression that violence has been eschewed in Northern Ireland. Over four years later the evidence suggests this is far from true. It is the case that the worst manifestations of the conflict, sectarian killings and bombings, are declining. In 1999/2000, for example, seven civilians were murdered, the lowest figures since the ‘Troubles’ began, and the first year that there were no security force fatalities (Royal Ulster Constabulary7 2000). This, however, ignores an insidious and ongoing level of paramilitary violence inflicted on working- class communities referred to as ‘punishment’ beatings and shootings or the informal criminal justice system. Paramilitary groups see themselves as community protectors;
their actions aimed ostensibly at maintaining ‘law and order’ through tackling petty crime such as car theft, joyriding8, burglary and drug dealing. Up to the end of 2001
7 On 4th November 2001, as part of the policing reform process, the Royal Ulster Constabulary changed it name to the Police Service of Northern Ireland.
8 Joyriding is a term used, somewhat perversely, to describe the theft of cars subsequently used for reckless, high-speed chases usually by teenage boys in urban areas. It has resulted in a number of fatalities and serious injuries.
Police Service of Northern Ireland (PSNI) statistics show that there have been 2,564 shootings (an average of 88 per year) and 1,802 beatings (an average of 90 per year) since 1973 and 1982 respectively, when figures were first recorded9.
Paramilitaries exact community ‘justice’ using baseball bats, hammers, hockey and hurley sticks, iron bars, pickaxe handles and steel rods. Other forms of ‘punishment’
include dropping heavy concrete blocks on limbs and using power tools on bones.
Surgeons in the fracture clinic at the Royal Victoria Hospital in Belfast, for example, report that ‘following the cessation of violence there has been an increase in the level of injuries occurring in those undergoing paramilitary punishment’ (Nolan et al. 1999:
8). Their study of treating victims showed that those who had been shot with pistols, resulting in open injuries, suffered much less damage to soft tissue and bones than those who had been beaten. The brutal reality is that it is ‘better’ to be shot than beaten.
Informal criminal ‘justice’ and vigilantism
Three principal reasons are advanced for the existence of the informal criminal justice system in Northern Ireland and vigilantism in South Africa (Knox, 2002). First, in both countries there is an absence of an adequate policing service. In Northern Ireland, particularly in republican areas, the Police Service for Northern Ireland has no legitimacy, and their communities would not normally involve the police in dealing with crime in their areas. Republicans claim that the PSNI are prepared to tolerate at best, or encourage at worst, crime in their communities as a way of
9 The figures show that loyalists were responsible for 44.5 per cent of the shootings and 47 per cent of the beatings, republicans carried out the remainder.
undermining the ‘republican struggle’. Police are therefore willing to ‘trade’ dropping charges of joyriding, drug dealing or burglary in return for low-level intelligence gathering on so-called known republicans. In loyalist areas, objections to involving the police are more to do with keeping the PSNI out of communities where drug dealing, racketeering and illegal drinking dens and clubs are commonplace. In South Africa, the police are seen as ineffective, lacking not only basic literacy skills but with an estimated 25 per cent of its members unable to read and write (Randall 2000).
Their crime solving capabilities are also unimpressive not least because experienced personnel were trained in the policing methods of the old apartheid regime. For example, of the 2.2m crimes reported to the police in 1998, almost half were considered ‘undetected’ because evidence was insufficient or the suspect had disappeared. Half a million more were withdrawn, leaving 524,000 cases that reached court (The Economist 2001).
Second, in Northern Ireland there is a rising level of crime including ‘anti-social behaviour’, petty and violent crime. This is evidenced in crime and victimisation statistics, which show that those from an unskilled social class background are most vulnerable and feel their quality of life is particularly affected by fear of crime (Northern Ireland Office 2000; Louw and Shaw 1997). In 1998/99 the number of violent crimes (defined as offences against the person, sexual offences and robberies) increased by 21.2 per cent (to 21,452) and by a further 12.6 per cent in 1999/00 (PSNI 2000: 75). A Home Office report on international crime statistics showed Northern Ireland’s percentage increase in recorded crime (28 per cent) as second only to South Africa where it rose by 37 per cent in 1998 (Barclay and Tavares 2000: 3). In the absence, therefore, of an adequate/legitimate police service and/or because people are
discouraged from seeking police involvement, some people in communities turn to paramilitaries or vigilantes to secure a prompt, visible and effective response to crime in their areas. Hence, a proportion of local people living in fear of crime endorse paramilitary ‘punishment’ beatings and shootings and vigilante stoning or sjambokking (whipping) of alleged criminals or ‘anti-social’ elements.
Third, the formal criminal justice system within both countries is perceived as slow, ineffectual, and soft on crime. The view of a focus group participant in a loyalist community illustrates this:
I know young lads who were put on probation for stealing cars. The first week they went to the Probation Board they talked about the consequences of their actions for victims. The next week they were taken go-karting, then deep-sea fishing. The average mother cannot afford to send her kids to these activities. I then heard one young fella who hadn’t been involved in crime ask ‘how do you join the probation club for week-ends away?’ People see these young lads who have committed quite serious crime being taken by the hand without punishment for their actions against the community.
(Interview with focus group, Shankill – Belfast, November 1999)
The formal justice system is not only perceived as failing the victims of crime but also those involved in criminal activities. The case of Brian Connolly, a deceased joyrider provides an example of this failure of the formal system of justice. Brian was a passenger in a stolen car (allegedly travelling at 100mph) that crashed into the family car of Charmaine and Justin Watson killing the young couple in July 1999 and
orphaning their two children. Brian’s mother believes the formal system let her son down badly in that if he ‘got into trouble, he never seemed to be really punished. He’d often get bail or maybe just a three-month sentence. He also became very good at working the system. He knew what he’d get away with’ (Carmel Donnelly quoted in Walker 1999: 5) and had 109 prior convictions, many of them for motoring offences (Harper 2001). South Africa’s criminal justice system can also be said to be under- performing in that, on average, fewer than 9 per cent of recorded crime result in the conviction of the perpetrators. For more serious offences the conviction rate is even lower. For example, in 1999 the rate for car hijacking was 2 per cent, aggravated robbery 3 per cent and rapes 8 per cent (Schönteich and Louw 2001). Thus in societies where violent conflict has been the norm, it is not surprising that the time taken to process offenders, the necessary safeguards in the legal system, and the standard of proof required for conviction is seen as no match for summary justice meted out by paramilitaries and vigilantes.
Researching paramilitary violence and vigilantism
Commentaries on political violence in Northern Ireland and South Africa concentrate on two broad levels – firstly, trying to establish the facts or data about the levels, distribution and sources of violence (Murray 1982; Poole 1993; Fay et al. 1999;
Coleman 1998; Minnaar et al. 1998; Truth and Reconciliation Commission 1998; and Jeffrey 1997) and secondly, examining the causes of, or motivation for, violence (White 1993; O’Duffy 1995; Patel 1997; and Bornmann 1998). What is largely absent, however, is research on the nature of the relationship between paramilitary or vigilante groups and the communities over which they exert social control. Kennedy, for example, points out that although the ‘brute facts of communal violence are well
known...what is less well known is the degree of ‘internal’ paramilitary repression (in the form of beatings, shootings and mutilations) which developed in the shadows of the larger conflict’ (Kennedy 1995: 67). The same can also be said of our understanding of vigilantism in South Africa.
There are, however, exceptions including Cavanaugh’s (1997) ethnographic study undertaken in loyalist and republican communities in Belfast and Brewer et al. (1998) study of the role played by local communities in civil unrest and crime management.
In terms of post-apartheid South Africa, much of the literature deals with popular justice or community courts (utilising non-violent methods of punishment such as community service) as opposed to vigilantism or kangaroo courts. Examples would include the work of Nina (1995) and his examination of popular justice and civil society, especially within the context of street committees in Guguletu (Cape Town), and the work of Schärf (1992), which explores how best community courts can be adapted to the formal justice system.
These issues combined to frame the research upon which this chapter is based. Our discussions here centre on methodological conundrums associated with conducing qualitative research involving victims of ‘punishment’ beatings (n = forty), focus groups within communities (n = four) ‘controlled’ by the paramilitaries in Northern Ireland, victims of vigilante attacks (n = six) and focus groups within communities (n
= six) where vigilantes operate in South Africa. Of the four focus groups conducted in Northern Ireland, two took place in loyalist areas and two in republican areas.
Furthermore, an area containing a restorative justice project was chosen and an area without such a project (one of each in a loyalist and republican area respectively).
This allowed for comparisons to be made at inter and intra-community levels. We offer these personal reflections on researching paramilitary and vigilante violence in the hope that they have lessons for those undertaking ‘sensitive fieldwork’ elsewhere.
We are mindful of Brewer’s conclusions on police research in Northern Ireland. He argued that sensitivity is highly situational and researchers need to consider ‘what they believe to be controversial and sensitive but also what their respondents, potential gatekeepers, and the community at large might consider to be sensitive about the research’ (Brewer 1993: 143). Similarly, Alty and Rodham (1998) suggest that research within sensitive areas requires flexibility and demands practical solutions that are not always linked to traditionally recognised ethical dilemmas. We consider these issues in no particular priority under four key headings: reliable and valid information, accessing victims, personal security, and dissemination and engagement with stakeholders.
The search for reliable and valid information on violence
In Northern Ireland the PSNI collates statistics on the number of paramilitary-style shootings and assaults that are reported to them. These statistics are thought to underestimate the magnitude of the problem by as much as 30-50 per cent, according to one former pressure group (Families Against Intimidation and Terror - FAIT), not least because those who have been subjected to beatings are reluctant to go to the police through fear of reprisal. There is also no information available on charges brought against perpetrators. Detection rates are subsumed within ‘violence against
the person’ statistics but are described by the Police Authority for Northern Ireland10 (PANI) as ‘relatively low’.
Other statistical sources collated by pressure groups such as the former FAIT and the Northern Ireland Human Rights Bureau (NIHRB) are also problematic. FAIT was formed in 1990 by individuals affected or concerned by the continuation of
‘punishment’ attacks on those allegedly involved in ‘anti-social behaviour’ or petty crime. The organisation self-described as anti-sectarian and non-political, sought to
‘heighten political and public awareness of intimidation and mutilation within Northern Ireland’s controlled communities, to provide support for primary and secondary victims, and to speak-out on behalf of those unable or afraid to voice their own objections’ (FAIT 1999: 1). As part of its activities it established a database on paramilitary violence including ‘punishment’ beatings, shootings, intimidation, exiling and petrol bombings. Its figures for the number of shootings and beatings however were obtained from the PSNI, merely replicating the official statistics. On examination of the database we were able to obtain it is not clear from the data if any distinction was made between ‘punishment’ and other sectarian or racial attacks. The numbers of individuals exiled or intimidated out of their homes, as counted by FAIT, show a total number of people forced to be re-housed. The organisation counted each person in a household as an individual exiled or intimidated. A family of four
10 The Police Authority for Northern Ireland had key responsibility to secure the maintenance of an efficient and effective police service. It held the Chief Constable to account for his actions in certain areas (e.g. objectives, performance targets, and budgets). A new Policing Board replaced it in November 2001.
therefore counts as four intimidations or exiles although only, in effect, one family member may have been intimidated or told to leave the area by a paramilitary group.
FAIT also experienced internal divisions that resulted in questions being asked about the group’s credibility (Oldham 1998; Clarke 1998). For example, Vincent McKenna, a self-proclaimed former member of the Provisional Irish Republican Army (PIRA) who became the group’s development officer, named publicly two individuals allegedly involved in the Omagh bombing11 at the UK Unionist Party’s annual conference (Murphy 1999; Fergus 1999). A short while later McKenna left FAIT and formed a new human rights group, the Northern Ireland Human Rights Bureau in May 1999. The reliability and validity of the figures gathered by the Bureau have been questioned by the Northern Ireland Association for the Care and Resettlement of Offenders12 (Campbell 1999) as has the credibility of the organisation following reports that McKenna had been investigated in connection with child sex abuse allegations. McKenna claimed the PIRA was waging a campaign of intimidation against him, including two bomb attacks at his home and the vilification of his character by accusations of child sex abuse. However, McKenna was later arrested and found guilty of 32 charges of sexually abusing his daughter in the Republic of Ireland and received a three year custodial sentence (Harper 2000). The overall credibility of an organisation affects the credibility of their ‘numbers’. As a
11The Omagh bombing of 1998 was the worst terrorist atrocity in Northern Ireland’s bloody history where 29 innocent civilians were killed in a busy rural town centre.
12The Northern Ireland Association for the Care and Resettlement of Offenders is a voluntary organisation operating in the field of criminal justice which is part-funded by government sources.
consequence of the controversy surrounding FAIT, the research team used official PSNI statistics alongside developing its own comprehensive database compiled from daily newspaper reports of paramilitary attacks. The database was subsequently used to analyse the frequency, nature and geographical concentration of attacks over a two- year period using GIS mapping techniques. It also acted as a source of qualitative data, in that each incident recorded brief details of the attack/assault.
During the research project, the best known pressure/support groups in this area of violence, FAIT and the Northern Ireland Human Rights Bureau, lost credibility. Yet the information they generated played a key role in the public debate on the issue. The Conservative Party, for example, drew heavily in a House of Commons debate on information provided by FAIT. This was at a juncture in the Belfast Agreement when the early release of paramilitary prisoners was the focus of heated discussion. The Conservative Party spokesperson on Northern Ireland (Andrew MacKay) set the scene for the debate on
mutilations, torture and beatings by giving some statistics that have been prepared by Families Against Intimidation and Terror, a non-sectarian organisation that is doing marvellous work to help the victims of such mutilations and beatings, giving them hope, succour and, at times, protection and, sadly, at other times spiriting them out of the country to safety…FAIT has more accurate and up-to-date figures, than, sadly, the Royal Ulster Constabulary can possibly have.
This illustrates how the multiple layers of ‘counting’ and recording violence are ignored in a partisan political agenda. The credibility of statistics gathered by self- appointed spokespeople may be enhanced through publicity in official (Hansard) and media sources.
Like Northern Ireland, South Africa collates crime statistics. Unfortunately, the South African Police Service (SAPS) does not maintain separate records for vigilante attacks, therefore no official statistics are available. Even if these statistics were available it is unlikely that they would be valid and reliable. As the Institute of Security Studies notes ‘crime statistics are usually regarded with caution: in South Africa they are treated with outright scepticism’ (Louw 1998: 11). A Committee of Inquiry into the Collection, Processing and Publication of Crime Statistics was appointed in May 1997 by the Minister of Safety and Security and found that there was an absence of an ‘information culture’ within the SAPS, which highlights the problems affecting crime statistics. They found, for example, that many SAPS officials did not make optimal use of crime statistics in their daily policing activities, which in turn affected data input and the quality of the output statistics accordingly.
Furthermore, there was insufficient training, a high turnover of skilled personnel and inadequate physical and human resources (Louw 1998). In light of these findings the Minister placed a year-long moratorium (June 2000 - June 2001) on the release of crime statistics on the grounds they were unreliable and inaccurate. Critics suggest that the ban had more to do with masking rising levels of violent crime and the government’s poor performance in fighting crime.
Clearly with the absence of national crime statistics it is difficult to assess the scale of vigilante attacks in South Africa. Having said this, a small amount of information has been compiled by the Human Rights Committee (HRC). The HRC is an independent human rights non-governmental organisation that has been monitoring and reporting on human rights abuses and political violence in South Africa since 1988. In its monthly publication Human Rights Report instances of vigilante attacks were detailed in the Political Rights section. These data, although providing some information, are not a comprehensive survey of vigilante activities. In addition the Report is no longer published.
Given the nature of the crimes under review here, it is impossible to obtain accurate data on the scale of the problem. Official sources under-estimate the level of paramilitary/vigilante crime because of victims’ fears of reporting incidents to the police (in the case of Northern Ireland) or systemic data collection problems which provide a useful excuse for politicians embarrassed by lawlessness (in the case of South Africa). When this is overlaid with an overt political agenda to discredit, for example, the outworking of the Belfast Agreement or to criticise the ANC Government, there is no incentive for state bodies (police or criminal justice) to improve the quality of their data and provide their opponents with information to use against them.
Speaking with victims of violence
One major difficulty in the present study was securing access to those who have been subjected to paramilitary ‘punishment’ or vigilante attacks. Making contact with community organisations with which the researchers had previous experience seemed
a useful starting point. Whilst community workers knew victims of such attacks, their role in brokering contact met with limited success. What became clear during the course of this work in Northern Ireland were the contradictions within communities.
The very same community organisations may have intentionally or unintentionally referred victims to paramilitaries. It should not have been too surprising therefore that
‘punishment’ victims felt unwilling to co-operate with community organisations they suspected of ‘running to paramilitaries’ about them. This was not wholly unproductive, we did interview, for example, a father and son who had been beaten in a mistaken identity case and other community volunteers/workers who had themselves been beaten or shot.
A typical account given by a victim illustrates the sensitivity of the topic:
I was walking down the road and I heard a screech of brakes and I looked behind me, and I knew right away because of the big figures in the car. So I tried to run down an entry (alley) way and they ran on down the front of the street and got to the bottom of the entry before I had. I ran into them and they hit me a punch and I went on my back and they just pointed the gun at me and told me to lie there. Well they didn’t shoot me there, they took me away in a car…to the area I lived in. They took me into a house there, I was tied up, put a hood over my head and I was made to stand facing the wall. I asked could I sit down and there was no answer so I was going to try and sit down myself. I tried to get down like that and they just laid in and booted me all over the fucking show and pulled me back up onto my feet. They took me out, made me wait for another wee short while in the entry way, brought me out onto a wee grass
verge, told me to lie down and kneecapped me. After a couple of weeks I got out of hospital and went home. Two days after I got out of hospital they came to the door and said “you’ve got ‘til 12 o’clock the next day to get out of the country”.
(Interview with victim of ‘punishment’ shooting, December 1999).
Other problems arose even if victims agreed to talk to us. Some interviewees expressed concern over who might see interview transcripts. For example, one victim who had received three ‘punishment’ beatings and been told to leave his local area said ‘I’m scared. I don’t know what to say, or who’s going to hear it. I’m scared in case anybody hears it and knows [my story]’. We assured our interviewees that the audio tapes would be wiped and the interview transcripts anonymised and kept in a secure location. This was not always convincing to interviewees nervous about their security, despite their willingness to participate. In some cases we knew victims’
obvious reluctance to reveal stories in too much detail.
When access was proving particularly difficult and slow, however, one suggestion was to visit victims admitted to hospital for injuries resulting from paramilitary
‘punishment’ attacks, the researcher’s equivalent of ambulance chasing. Media coverage of these incidents often carried footage of victims in hospital beds. Whilst this had the clear potential to raise awareness of the horrific nature of these incidents on a scale beyond the capacity of our research, there was also the danger of sensationalism. Our own predisposition was therefore to reject this possibility for access, on the grounds that it could exacerbate the victim’s distress recounting the
incident so soon after the event, and we suspected payments were made for the interviews which may have influenced the motivations of the respondents13.
The most productive source of access, however, proved to be via the Probation Board for Northern Ireland, an executive non-departmental public body whose aim is to help prevent re-offending. Their day-to-day business brought them into contact with young people, a number of whom had been ‘disciplined’ by the paramilitaries for ‘anti-social behaviour’. Support for the research was secured at the senior level of the organisation and a letter of endorsement sent to area probation managers. Thereafter, the researchers made contact with local probation officers for referrals. Their professional interest in the topic and access to a ‘captive’ client base proved fruitful.
A number of probation officers had to deal with the consequences of paramilitary violence for young people and were keen to assist with research that could help to address its causes. Their brokerage role also carried certain credibility and cultivated trust with those willing to be interviewed which would have been difficult for us as researchers to secure.
Gaining access to victims in South Africa also proved difficult. To this end an informal working relationship was entered into with the Centre for the Study of Violence and Reconciliation (CSVR), a Johannesburg-based non-governmental organisation. CSVR were engaged in a programme of research examining the nature and extent of violence during South Africa’s transition from apartheid rule to
13 We deliberated over payments-for-access in our project and decided to reimburse interviewees at a standard rate for expenses incurred in attending, such as travel costs, child-minding fees and lost earnings.
democracy. One strand of this research focused on revenge violence and vigilantism.
Even with this relationship only six interviews were secured with individuals who had been victims of vigilantism. This may in part be explained by reluctance on behalf of victims to come forward through fear of further attacks. Indeed there have been cases of community members paying the bail of alleged criminals so that they are released from police custody into the community and subsequently killed. One interviewee who was abducted from his home by a crowd of 50 community members and taken to a local school hall to appear before a kangaroo court explains:
The fact of the matter is you either plead guilty or not guilty and then if you plead not guilty you are not given the opportunity to state why you are pleading not guilty. And they say “but you did this” and whilst you are busy trying to explain something, somebody will come from nowhere and start slapping you and saying “but tell the truth”. So there is a lot of intimidation, harassment and complete abuse of your personal dignity and esteem so that you end up cowering in a corner. The guys have guns, AK-47s, sjamboks, knobkieries, pangas14. You are defenceless in that type of situation. You are completely at their mercy.
(Interview with victim of vigilantism, October 1999).
The brokerage roles of organisations with which victims have had prior contact and where a relationship of trust had been established, help in gaining access. The credibility of intermediary bodies, not seen as organs of the state, was of prime
14 Sjamboks are whip like implements, knobkieries are wooden sticks and pangas are large knives.
importance. When the state itself is integral to the nature of the conflict and its various manifestations such as communal violence, these pre-requisites are all the more significant. Given the problems of accessing victims, the forty interviewees in the research emerged through a process of snowball sampling, intermediary agencies and personal contacts with community organisations. This is clearly a non-probability sample and it is legitimate to question the quality of data arising from this methodological approach. The research did, however, triangulate victims’ stories through focus group work within communities, interviews with statutory and voluntary bodies dealing with paramilitary/vigilante violence, feedback from political parties and secondary data sources (newspapers, periodicals and community-based newsheets). The nature of the attacks, the interrogation processes and the violence used were independently repeated in the accounts of a number of the interviewees. An
‘ideal-type’ sampling framework simply isn’t available to researchers in areas of high sensitivity – being mindful of the consequences of using alternative strategies on data quality and finding ways to validate findings are essential in these circumstances.
Personal security of researchers studying political violence
The personal security risks associated with this type of research are high. Lee (1993), referring to Yancey and Rainwater (1970), described two kinds of danger that may arise during the research process: the ‘presentational’ and the ‘anonymous’. The former occurs when the researcher’s presence or actions evoke aggression, hostility or violence within the setting. The latter, when the researcher is exposed to otherwise avoidable danger, simply because of the dangerous research environment. Both kinds of dangers apply to studying paramilitary violence and vigilantism. Paramilitaries in Northern Ireland are well practised in dealing with ‘touts’ or those passing
information, alleged or otherwise, to the PSNI. This extends to those ‘speaking out’
against them. Suspected informers are shot, there are no exemption clauses for academic researchers. For example, a former member of the PIRA, Eamon Collins, was murdered in January 1999 following his evidence against former PIRA colleagues in court and a book, revealing their operations and depicting its members as ‘a sadistic conspiracy of ageing, pot-bellied drink-induced egos’. Likewise in South Africa, University of Cape Town academic Dr Ebrahim Moosa had his home bombed after criticising the vigilante activities of PAGAD. Of more direct relevance to this research was the attempted assassination of Queen’s University Belfast Professor of Comparative Politics, Adrian Guelke, in September 1991. The Ulster Freedom Fighters (UFF)15 broke into his home and shot him in the side but failed to kill him because their guns jammed. The terrorists claimed that he was an intelligence officer for the IRA and involved in importing arms from the Middle East. It subsequently transpired from a journalist’s investigation that Professor Guelke’s South African background and research on violence therein, had led to loyalist paramilitaries being approached. He was, he suspects, set up by an outsider. Guelke explained:
How or why I fell foul of some person in South Africa to the extent that I became a target for assassination, I do not know. Perhaps my writing about South Africa’s supply of arms to Loyalist paramilitary organisations gave offence, or a brief investigation I carried out into extreme right-wing violence in South Africa may have been the cause. There were a number of possibilities. From my experience of Northern Ireland I know how utterly
15The UFF is a loyalist paramilitary organisation and a cover name used by the Ulster Defence Association (UDA) to claim sectarian killings.
trivial the reasons someone becomes a target can be. In general, campaigns of violence are rarely conducted with precision, whatever their ultimate purpose.
(Guelke 1998: 196)
Guelke’s first hand experience provides an ominous warning against complacency in undertaking fieldwork into paramilitary violence. A surveillance network operates on behalf of the paramilitaries tasked with ‘keeping their eyes open’ for unusual activity.
As a precautionary measure we ‘informed’ political representatives of the paramilitaries (Sinn Féin, the Progressive Unionist Party and the Ulster Democratic Party) about our research so that they were aware of its authenticity. In essence this amounted to securing their ‘approval’ in the event of questions being asked about its precise purpose. A key aspect of convincing the political representatives of our bona fides was the source of funding for the research. Had the research received government funding, particularly from the Northern Ireland Office16, political/paramilitary ‘endorsement’ for the fieldwork would have been impossible.
Researchers would be seen by republicans, for example, as part of the ‘British establishment’ and therefore, by definition, engaged in intelligence gathering under the guise of academic research. This would undoubtedly have put us at risk in undertaking fieldwork on such a sensitive topic. The fact that the research had been funded through the Economic and Social Research Council (ESRC) attributed significant independence to our motives and we offered this information to
16The Northern Ireland Office is the department of the Secretary of State for Northern Ireland which remains responsible for matters not devolved to the Northern Ireland Assembly. These include matters such as policing, security policy, prisons, criminal justice, international relations and taxation.
participants as part of negotiating access. Suspicion of the researchers’ motives may not arise solely from paramilitaries. The security forces may have concerns about researchers (without a police record) being spies on behalf of paramilitary organisations.
In South Africa, we sought the support of respected community members in the townships we wished to study. These local residents provided invaluable assistance with our fieldwork, not only acting as guides and ‘gatekeepers’ but also as ‘sponsors’
thereby ensuring safe passage in the townships. In most cases, ‘gatekeepers’ arranged to meet us at the edge of the township and travelled with us in the car. Meetings and interviews were held during the day, thus travelling to and from the townships in darkness was avoided.
Suspicion of ‘outsiders’ is intense in this type of research. In Northern Ireland, the perceived religion of the researcher is likely to be a key factor in the minds of interviewees. Respondents look for ‘clues’ to religious affiliation that has become intrinsic to social interaction in Northern Ireland. The most obvious, although not fail- safe, is the researcher’s name. Beyond that, area of residence, birthplace, accent/language, appearance, school attended, interest in particular sports and, somewhat bizarrely, whether one is ‘Catholic or Protestant-looking’. Coming from the
‘other’ community may condition the response of interviewees or put the researcher at some risk given the sensitive topic under review and the nature of the questions posed. In South Africa, this was less of a consideration given that we were non-South African researchers interested in a South African phenomenon. To understand paramilitary ‘policing’ in Northern Ireland or vigilantism in South Africa requires probing questions about motives, methods, support for paramilitaries’/vigilante
actions within their communities, and the exploration of alternative ways of dealing with ‘anti-social behaviour’. To the suspicious interviewee this may smack of information gathering reinforced by our preference to tape-record the interview for the purposes of data analysis. For the interviewer in these circumstances, Lee contends that ‘complete neutrality is probably impossible’ and when researchers proclaim their neutrality ‘they are, in fact, concealing their own sympathies’ (Lee 1995: 23). This is a difficult and sometimes dangerous balancing act.
It is only sensible in undertaking fieldwork to observe cautious security protocol.
Being aware of the constituency in which one is interviewing is crucial - staunchly loyalist and republican enclaves are dotted throughout most towns/cities in Northern Ireland while South African townships tend to be on the fringes of urban areas.
Sensible security planning can involve working out entry and exit routes, opting, where possible, for safe(ish) locations to conduct interviews (for example, administrative headquarters of political parties linked to paramilitaries and the offices of non-governmental organisations), taking taxis to venues as opposed to using personal transport with car registration details, doing fieldwork in pairs, informing other members of the research team of your schedule. The venue for conducting interviews is particularly important. Participants and researchers must feel safe. There are areas in Northern Ireland, for example, where one community or the other feels threatened, given the territorial nature of segregated space. Finding ‘neutral’ venues isn’t always easy and hence the use of Probation Board offices, where a number of the interviewees had official appointments, provided secure settings. Whilst highlighting the importance of sensible personal security measures, these must be kept in perspective. Here we concur with Punch, that researchers of controversial topics must
not ‘become over-sensitive so as to avoid dubbing the setting or topic virtually unresearchable’ (Punch 1989: 181).
Feeding back findings: Dissemination and engagement with stakeholders
Our final heading concerns the dissemination of the research findings and the extent to which one engages with stakeholders. One month prior to the release of the findings, separate seminars were held in two of the four focus group areas (one in a republican area and the other in a loyalist area). The research team considered it important to give something back to participants in the project and to the communities studied. In both seminars the findings of the research were well received. A findings leaflet was produced and sent to everyone who had participated in the research (copies of the leaflet were sent to Probation Board officers to pass on to those individuals who had been referred to us) and to other interested parties. The project website also contained a summary of the research and its findings.
We did not anticipate how the media would portray the findings. Their reaction was generally supportive of the research but honed in on criticism levelled at the Northern Ireland Office (NIO) with headline items such as ‘NIO slammed over punishment attacks’ and ‘NIO lacking courage on thug menace’. The NIO reacted angrily and denied it was ‘indifferent’ to paramilitary attacks and refused to ‘comment in detail on a report officials were not given the opportunity to see before publication’ (Graham 2000: 7). This raises the interesting question as to whether organisations, particularly those which have been criticised in research, should be afforded a prior ‘right to reply’. If so, how should this feature in the findings and dissemination process, or how might this be anticipated in advance to give the organisation time to prepare their
‘press release’? Furthermore the NIO stated ‘What we have seen amounts to a series of unsubstantiated, generalised and politically tendentious assertions’ (Irish News 2000: 5). What they had ‘seen’ at the time of issuing this statement were newspaper accounts of the research and not the final report. In fact, the NIO were interviewed as part of the research and the project team also briefed the NIO on ‘punishment’ attacks prior to a debate on the topic in the House of Commons. In addition, a NIO official attended a research seminar where preliminary findings were presented.
The Northern Ireland Office’s response may have been conditioned by their anticipation of how the research could/would be used in the wider political context.
Politicians drew selectively on the research to condemn the government and its response (or lack thereof) to the informal criminal justice system. Such was the extent of this public discourse that the report was debated on the floor of the Northern Ireland Assembly. The motion put to the Assembly was:
The Assembly notes with grave concern the contents of the ‘Informal Criminal Justice Systems in Northern Ireland’ report on punishment beatings by paramilitary organisations; deplores and condemns the Government’s inadequate response to the report; and calls on the Government to bring forward measures to ensure that those responsible are made amenable to the law.
(Northern Ireland Assembly 2001a)
The debate itself highlighted the politicisation of the research findings in which some members of the legislative assembly used the criticisms levelled at the NIO over the issue as a blanket condemnation of the outworking of the Belfast Agreement. One
contributor argued that the research report ‘has presented cogent evidence that the Good Friday Agreement is failing’ (Northern Ireland Assembly 2001b: 47). This enormous extrapolation from a research report which dealt with a single (but significant) issue of criminal justice to a declaration that the Belfast Agreement had failed demonstrates the way in which the findings were manipulated for political purposes. The motion was passed on a strictly partisan vote – the Ulster Unionists, Democratic Unionists and Alliance parties voted in favour; Sinn Féin, the Progressive Unionist Party (PUP) and the Northern Ireland Women’s Coalition voted against; and the Social Democratic and Labour Party (SDLP) abstained. The findings of the research assumed a degree of political import never intended or anticipated by the researchers. The timing of the report, when the Northern Ireland Assembly still was still faltering, may have been unfortunate in that pro- and anti- Belfast Agreement parties attempted to seize the initiative over any issue in which there was political capital. This volatile political milieu would have been difficult to predict and take into account in the dissemination of the research.
Two broad methodological themes are considered in the conclusion to this chapter, first data inputs and outputs in the form of valid and reliable information sources, their translation into research findings and dissemination, and second the specific concerns of access and security in conducting fieldwork. In a comparative research project of this type the political context is an important and influencing factor on the data sources available. In South Africa and Northern Ireland, the state and its organs are protagonists in the conflict and hence ‘official’ sources of data may not represent an accurate picture of the problem. This may be for reasons that the state is not
perceived as a neutral arbiter and hence there is under-reporting by victims of communal violence or that the state attempts to suppress information which would reveal its poor stewardship. Either way, official statistics are not, in themselves, reliable and valid sources of information. When looking for alternatives, however, researchers need to be aware that pressure groups, by definition, have their own agenda and hence information emanating from these sources may be tendentious. The
‘politics’ of information is not only apparent at the research input stage but also relevant for outputs or at the point of presenting findings and their dissemination. This research took on a life of its own beyond the formal presentations of its findings. One reflective question is the extent to which this process can or should be managed so that political spinning is minimised. Project data had to rely largely on quantitative reporting from official sources, and the development of our own databases expanded on the detail of each paramilitary assault. The problem of under-reporting remains however, and is a more fundamental issue associated with fear of paramilitary/vigilante reprisal and lack of confidence in the security forces.
In terms of the fieldwork, this chapter draws attention to two specific issues, access and security. In the former, the time taken to establish contacts, convince them of the bona fides of the researchers, and gain access to victims took much longer than anticipated. When brokering organisations or community groups are used to secure access, they too have baggage as gatekeepers which researchers must be aware of in terms of the interviewees they assist with. Once access was secured the key problems encountered were suspicion of the researchers’ motives, how the information would be utilised and, importantly, who would have access and how it would be stored. In a highly sensitive research topic these are not unusual methodological concerns.
Perhaps more unexpectedly for the research team was the potential to re-traumatise interviewees in the course of retelling their stories and possible repercussions of their own exposure to highly distressing accounts of brutal and violent incidents. In one incident the victim of a particularly brutal paramilitary beating became highly emotional in recounting the incident. He had spoken during the course of the interview of his anger at the attack on him, concern at the disfigurement to his limbs and extreme nervousness about the potential for future paramilitary attacks. In his heightened state of anxiety he dropped his trousers to authenticate the injuries sustained to his legs and broke down in a state of emotional distress. As researchers, perhaps naively, we were unprepared and untrained in dealing with the minor number of cases presenting with symptoms of psychological anguish. Although conducting research in two volatile contexts such as Northern Ireland and South Africa presents obvious dangers, our experience was that with sensible precautions and an acute awareness of the political sensitivities, prior planning and attention to potential problems minimised the risk.
A key consideration for this research was the policy implications of its findings. Since those ultimately charged with implementing policy recommendations (the Northern Ireland Office) were the focus of criticism within the research, this presented difficulties. Ultimately this is a balancing act in convincing key stakeholders that the criticism levelled is constructive and trying to gain their ownership of, and endorsement for, changes recommended. Sensationalist press reporting did not help in the pursuit of this goal.
Alty, A. and Rodham, K. (1998) ‘Pearls, Pith and Provocation. The Ouch! Factor:
Problems in conducting sensitive research’, Qualitative Health Research 8, 2: 275- 282.
Amupadhi, T. (1997) ‘Police worried about the rise in mob action’, Weekly Mail and Guardian (Johannesburg), 14 February.
Barclay, G. C. and Travares, C. (2000) International Comparisons of Criminal Justice Statistics 1998, London: Home Office.
Bornmann, E. (1998) ‘Group membership as a determinant of violence and conflict:
The Case of South Africa’, in E. Bornmann, R. van Eeden and M. Wentzel (eds.) Violence in South Africa: A variety of perspectives, Pretoria: HSRC Publishers.
Brewer, J.D. (1993) ‘Sensitivity as a problem in field research: a study of routine policing in Northern Ireland’, in C. Renzetti and R. Lee (eds.) Researching Sensitive Topics, London: Sage.
Brewer, J.D., Lockhart, B. and Rodgers, P. (1998) ‘Informal social control and crime management in Belfast’, British Journal of Sociology 49, 4: 570-585.
Campbell, B. (1999) ‘The dark secrets of a peace hero’, The Independent of Sunday (London), 26 September.
Cavanaugh, K.A. (1997) ‘Interpretations of political violence in ethnically divided societies’, Terrorism and Political Violence 9, 3: 33-54.
Clarke, L. (1998) ‘Terror victim group’s head quits in row’, The Sunday Times (London), 27 September.
Coleman, M. (1998) A Crime Against Humanity: Analysing the Repression of the Apartheid State, Cape Town: David Philip Publishers.
The Economist (1999) ‘Mandela’s Heir’, 29 May.
- - (2001) ‘Africa’s Great Black Hope - A Survey of South Africa’, 24 February.
Families Against Intimidation and Terror (1999) Media Information Package, Belfast:
Fay, M.T., Morrissey, M. and Smyth, M. (1999) Northern Ireland’s Troubles: The Human Costs, London: Pluto Press.
Fergus, L. (1999) ‘Fait ‘naming’ furore’, News Letter (Belfast), 9 February.
Graham, I. (2000) ‘NIO slammed over punishment attacks’, New Letter (Belfast), 28 November.
Guelke, A. (1998) The Age of Terrorism and the International Political System, London: I.B. Tauris.
Harper, S. (2000) ‘Self-styled ‘champion of abused’ is a liar and a paedophile’, Belfast Telegraph, 22 November.
-- (2001) ‘Catalogue of tragedy from our spiralling car crime’, Belfast Telegraph, 24 March.
Irish News (2000) ‘NIO ‘indifferent’ to paramilitary attacks’, Irish News (Belfast), 28 November.
Jeffery, A. (1997) The Natal Story: 16 Years of Conflict, Johannesburg: SAIRR.
Kennedy, L. (1995) ‘Nightmares within nightmares: paramilitary repression within working-class communities’, in L. Kennedy (ed.) Crime and Punishment in West Belfast, Belfast: The Summer School, West Belfast.
Khangale, N. (2001) ‘15 held for mob killings’, Sowetan (Johannesburg), 5 March.
Knox, C. (2002) ‘‘See No Evil, Hear No Evil’: Insidious Paramilitary Violence in Northern Ireland’, British Journal of Criminology 42, 1: 164-185.
Lee. R. (1993) Doing Research on Sensitive Topics, London: Sage.
-- (1995) Dangerous Fieldwork, Thousand Oaks, CA: Sage.
Louw, A. (1998) ‘The Problem with Police Statistics’, Nedcor ISS Crime Index 2, 3:
Louw, A. and Shaw, M. (1997) Stolen Opportunities: the impact of crime on South Africa’s poor, ISS Monograph Series 14: 12-14.
MacKay, A. (1999) ‘Terrorist Mutilations (Northern Ireland)’, House of Commons Hansard Debates, 27 January.
Minnaar, A., Pretorius, S. and Wentzel, M. (1998) ‘Political conflict and other manifestations of violence in South Africa’, in E. Bornmann, R. van Eeden and M.
Wentzel (eds.) Violence in South Africa: A variety of perspectives, Pretoria: HSRC Publishers.
Murphy, C. (1999) ‘FAIT in internal row over claims on bombing’, Irish Times (Dublin), 9 February.
Murray, R. (1982) ‘Political violence in Northern Ireland 1969-1977’, in F.W. Boal and J.N.H. Douglas (eds.) Integration and Division: Geographical Perspectives on the Northern Ireland Problem, London: Academic Press.
Ndaba, B. (2001) ‘Photographer saves suspect from necklacing’, The Star (Johannesburg), 27 February.
Nina, D. (1995) ‘Reflections on the Role of State Justice and Popular Justice in Post- Apartheid South Africa’, Imbizo, Issue 3/4: 7-14.
Nolan, P. C., McPherson, J., McKeown, R., Diaz, H. and Wilson, D. (1999) ‘The Price of Peace: the personal and financial cost of paramilitary punishments in Northern Ireland’, unpublished paper. Belfast: Royal Victoria Hospital.
Northern Ireland Assembly (2001a) Order Paper OP25/00. Online. Available HTTP:
<www.ni-assembly.gov.uk/order25-00.htm> (accessed 24 January 2001).
-- (2001b) Official Report (Hansard), 23 January. Online. Available HTTP:
<www.ni-assembly.gov.ukrecord/010123.htm> (accessed 24 January 2001).
Northern Ireland Office (2000) Fear of Crime and Victimisation in Northern Ireland, Belfast: Northern Ireland Office.
Ntabazalila, E. (1997) ‘Crowd demolish alleged rapists’ shack’, Cape Times (Cape Town), 30 January.
O’Duffy, B. (1995) ‘Violence in Northern Ireland 1969-1994: Sectarian or Ethno- National?’, Ethnic and Racial Studies 18, 4: 740-772.
Oldham, J. (1998) ‘Anti-terror spokesman accused of following his own agenda’, News Letter (Belfast), 5 August.
Patel, D. (1997) ‘Taxi wars in South Africa: Can there be peace?’ in A. Minnaar and M. Hough (eds.) Conflict, violence and conflict resolution: Where is South Africa heading? Pretoria: HSRC Publishers.
Poole, M. (1993) ‘The spatial distribution of political violence in Northern Ireland: an update to 1993’, in A. O’Day (ed.) Terrorism’s Laboratory: The Case of Northern Ireland, Aldershot: Dartmouth.
Punch, M. (1989) ‘Researching police deviance’, British Journal of Sociology 40, 2:
Randall, E. (2000) ‘Lack of schooled cops plagues SAPS’, The Sunday Independent (Johannesburg), 19 February. Online. Available HTTP: <.http://www.iol.co.za/news>
(accessed 19 February 2000).
Royal Ulster Constabulary (2000) Report of the Chief Constable 1999/00, Belfast:
Schärf, W. (1992) Community Courts, Cape Town: SJRP/LEAP.
Schönteich, M. and Louw, A. (2001) Crime in South Africa: A country and cities profile. Briefing presented at the Institute for Security Studies, Pretoria, 15 March.
Soggot, M. and Ngobeni, E. (1999) ‘We must work on their buttocks’, Weekly Mail and Guardian (Johannesburg), 14 May. Online. Available HTTP:
<http://www.sn.apc.org/wmail/issues/990514/NEWS33.html> (accessed 14 May 1999).
South African Institute of Race Relations (2000) South Africa Survey 1999/2000, Johannesburg: SAIRR.
Truth and Reconciliation Commission (1998) Truth and Reconciliation Commission of South Africa Report, Cape Town: CTP Book Printers.
Walker, G. (1999) ‘A mother torn by grief and anger’, Belfast Telegraph, 4 August.
Weekly Mail and Guardian (Johannesburg) (2000) ‘A Society of Rapists’, Online.
Available HTTP: <www.mg.co.za/mg/news/2000apr1/7apr-rape.html> (accessed 7 April 2000).
White, R.W. (1993) ‘On measuring political violence: Northern Ireland, 1969 to 1980’, American Sociological Review 58: 575-585.
Williams, M. (1999) ‘Police release kangaroo court victim’, The Mercury (Durban), 23 February.
Yancey, W. and Rainwater, L. (1970) ‘Problems in the ethnography of urban underclasses’, in R.W. Habenstein (ed.) Pathways to Data, Chicago: Aldine.