RESEARCHING VIOLENTLY DIVIDED SOCIETIES

From: Marie Smyth and Gillian Robinson (eds) 2001 'Researching Violently Divided Societies: Ethical and Methodological Issues'. London, Pluto Press

CHAPTER 2
Does research make any difference? The case of Northern Ireland

Marie Smyth and John Darby

Introduction

In 1999, whilst conducting a review of the Good Friday Agreement in Northern Ireland, former United States Senator George Mitchell proposed a media black-out - not for the first time - to the political parties participating. Information on the progress of the talks was withheld from the media for almost a week during the most delicate phase of the negotiations. Mitchell was attempting to preclude the negative impact of these activities in his suggestion of a media silence. And if anyone in the local or international media understood the implied judgement of their role in the earlier peace process, they gave no indication of such understanding, nor did they indicate that they were reflecting on their role or responsibility.

The media play an enormously powerful role in the shaping of political life, and one that is especially sensitive at times of political crises. Over the thirty years since 1969, Northern Ireland has had recurring experiences of political crises. In this chapter, the less prominent but nonetheless potentially significant role of researchers in the understanding and resolution of the conflict provides the focus of attention. Would George Mitchell have admitted researchers to information about the talks? Are researchers to be trusted or are they likely to behave in any more or less trustworthy ways than other observers? Does the activity of researchers make a difference? And if it does, what kind of impact can it be expected to make? What lessons can be learned for those researchers who would maximize the societal impact of their work?

Researchers tend to over-estimate the impact of their research on policy. Policy-makers tend to under-estimate it. It is hardly novel for academics to regard their fields of study as central, and their own work within the field as important. In reality, policy decisions emerge from a much broader range of influences - the particular interest of ministers, their advisers and civil servants, parliament, political exigency, lobbying groups. It is not easy to separate and grade the relative influence of each factor, any more than the direct impact of policy research, although it is sometimes possible to trace a line between research commissioned by government and legislation resulting from the research. This is uncommon but not unique. The tangible effect of intellectual debate on public policy may not always be obvious. Such debate may have more obvious influence on public discourse and political debate. One example is the ease with which all actors in the Northern Ireland peace process unconsciously borrowed and adopted terms and approaches, which had long been used in the academic debate on conflict and conflict resolution, such as a conflict resolution situation, human rights, pluralism, respect for cultural diversity, celebrating our cultural tradition.

There has never been a shortage of myths about the Irish conflict, and since the increased demand for information about the conflict since 1969, they have proliferated. Some of these myths contain an element of truth. During the early 1970s, for example, the evening cluster of visiting reporters in the bar of the Europa hotel in Belfast did lend some support to the popular view that its bar provided a more frequent source of news stories than the dangerous streets outside.

The research scene has changed considerably since those early days. It is difficult to imagine an ethnic conflict anywhere in the world that been more thoroughly researched. The most recent register of conflict-related research on Northern Ireland tracked down 605 projects under study in 1993 (Ó Maoláin, 1993). Basic data have become more readily available to researchers. A body of theory, as distinct from polemic, has emerged. More subjectively, the depth of scholarship has also improved. The province's academic institutions have become more concerned with the problems, and better equipped to tackle them. Ó Maoláin (1993) and Flackes and Elliott (1988 and 1995) - has made the information more accessible.

This chapter explores the contribution of research to the understanding and resolution of the conflict in Northern Ireland. In order to do this, a number of research case studies will be considered throughout the chapter. These are presented here in précis form and aspects of them will be discussed at greater length later in the chapter. The selection of the case studies themselves highlights one of the pervasive obstacles to discussing the subject of policy research objectively. By careful selection almost any general observation could be justified. The deficiency of many comparative studies, as Horowitz put it, may not be selection bias as much as it is the failure to be explicit (Horowitz, forthcoming, Chapter 12). Collier and Mahoney (1996) suggest that it is important to be open about the research's distinctive frame of reference so that readers can judge for themselves the boundaries within which the research is located. So, to be explicit, the case studies that follow do not attempt to cover all aspects of social policy research in Northern Ireland. They were selected to illuminate research which was undertaken to affect social policy in Northern Ireland; all relate to Northern Ireland's central sectarian division; they are intended to cover both successes and failures; they reflect the research experiences of the authors, both of whom have spent most of their academic careers working - separately - on such issues. These are the paper's frame of reference. It is, of course, possible and even likely that a different selection would have produced different conclusions.

Case studies in research

a. Local government discrimination, Research conducted by the Fermanagh Civil Rights Association

A strong body of evidence now available demonstrates that successive Unionist administrations from 1921 to the mid-1960s openly discriminated politically, economically and socially against the Nationalist minority in Northern Ireland (see Whyte 1991). Despite this, apart from periodic noises from the British Labour party (for a recent review see Rose (2000), especially the first two chapters), there was little organised opposition and even fewer research investigations into the abuses until the formation of the Campaign for Social Justice in 1964 and the Northern Ireland Civil Rights Association (NICRA) three years later (An admirable exception was a book compiled by Denis Barritt and Charles Carter in 1962. This attempted to lay out a template for Northern Ireland's political and social institutions, by means of informal surveys. It was an essential reference until the publications boom began in the early 1970s). The campaigns carried out by these organisations during the late 1960s were modelled on the American civil rights campaigns - sit-ins, marches (a familiar form of protest in Northern Ireland), rallies and publications. Alongside these emerged the first explicit attempts to adopt research techniques as a means of both informing public opinion of abuses and lobbying for their reform. In 1969 the Fermanagh Civil Rights Association, a branch of NICRA, produced a report, Fermanagh Facts investigating discrimination against Catholics in the allocation of public housing and jobs by the Unionist local council. This publication stood out from the more polemical ephemera by its thoroughness. The claims of gerrymandering and bias in employment by Unionist-controlled Fermanagh County Council were supported by detailed data. For example, they tracked down the religious affiliation of all the full-time council employees in April 1969, excluding those working in schools. The results needed no additional rhetoric. Out of a total of 370, and in a county that was approximately equally divided between the two religious communities, only 32 employees were Catholic.

The significance of the research was its comprehensive nature. It was not uncommon for critical social research in the early years of Northern Ireland's violence to be systematically attacked as incompetent or politically biased. Fermanagh Facts was able to withstand the closest scrutiny because the sample studied was the total population. It and other evidence was accepted completely and incorporated into the Cameron Report (1969), the first official investigation of Northern Ireland's peculiar practices. Indeed it had a major effect on Cameron's main finding, that 'much of the evidence of grievance and complaint which we heard, when analysed, was found, as might be expected, to be concentrated on two major issues - housing and employment. Jobs and houses are things that matter and touch the life of the ordinary man more than issues of 'one man, one vote' and the gerrymandering of ward boundaries.' (Cameron Report (1969), Paragraph 129).

b. Research on the population shifts and increasing segregation in Derry Londonderry Templegrove Action Research

In 1994, a group of people from both sides of the sectarian divide in Derry Londonderry (The name of the city is a matter of contention, Nationalists referring to it as Derry, whilst Unionists refer to it as Londonderry. Templegrove Action Research conducted research on the topic and concluded that the city should be called Derry Londonderry, hence the nomenclature in the text) became concerned about the deepening sectarian divide in the city and the movement of Protestants out of the city. A group of people from both sides of the sectarian divide (later known as the Guildhall Group) began to meet secretly to discuss sectarian divisions in the city. Some of these people were prominent in their own community and had views which would lead some on their own side to suspect them of treachery were it known that they were meeting with "the enemy." One of the authors was invited to participate. A set of common concerns emerged from those meetings, related to sectarian division. A central concern related to the exodus of Protestants from the city, and what this exodus meant for the city's future. No accurate statistics on the scale of the population shift were available, opening the way to rumour, speculation and scare-mongering. Furthermore, there seemed to be a dearth of public policy on the issue. In consultation with the Guildhall Group, which was composed of those from both sides of the sectarian divide, a two-year research project to document the scale of the movement of population, and investigate aspects of sectarian division in the city was undertaken. One of the authors acted as the researcher.

The gathering of robust data on the scale of population movement in the city of Derry area over the period since the Troubles began was one of the goals of the research. The establishment of reliable figures was aimed at undermining the rumours and the talk of 'ethnic cleansing' that in turn exacerbated sectarian tensions. A review of estimates that were in use by, for example, local political parties, revealed widely divergent methods of estimation, with most methods presenting significant problems for the reliability and validity of the estimates produced. The production of robust figures however was not straightforward for two reasons. The most obvious source of population data was the Census of Population. Censuses had been conducted in 1971, 1981 and 1991, allowing a longitudinal analysis that revealed shifts in population. But there were problems. First, the ward boundaries had changed between the three censuses, so no longitudinal analysis could be conducted using wards as the basis for measurement; and second, the 1981 census was regarded as unreliable because of a census boycott in certain areas. However, census data for the three census periods was used at grid square level to produce figures on population movement. Analysis of these data showed the following trends: an overall decrease in the Protestant population of the cityside (-7052); an increase in the Protestant population in the Waterside (+1903); a small increase in the Catholic population in the Waterside (+324); a decrease in the number of Protestants in the city as a whole (-5149). Analysis of population figures for a Catholic enclave in the city showed an increase in the percentage of Catholics and a marked decrease in the percentage of Protestants, whereas analysis of the population of a Protestant enclave showed marked decreases in all denominations, associated with dramatic decline in the total Protestant population. Overall, there was a trend towards deepening segregation.

The results were published (Smyth, M. (1995) Sectarian Division and Area Planning: a commentary on the Derry Area Plan 2011, Derry Londonderry: Templegrove Action Research, April, 1995) in monograph form, and presented at local public meetings, including those attended by politicians. The scale of the exodus of Protestants from the city was established and the data was generally accepted as authoritative, thereby ending most of the speculation and rumour. The research subsequently examined other aspects of sectarian division in the city, and public policy on segregation and sectarian division, making recommendations on policy and generating public discussion on the position of minorities within the city.

c. Research into educational provision, Standing Advisory Commission on Human Rights' (SACHR) (a government-appointed body) and the Nuffield Foundation (A United Kingdom-based charitable trust)

Barritt and Carter's first survey of Northern Ireland's community relations problem (1962) placed a heavy responsibility for Northern Ireland's divisions and intransigence on its highly segregated school system (see Gallagher 1989 for an overview). They have not been alone. By the late 1970s segregated schooling had become one of the most popular subjects for academics working on the Northern Ireland conflict and it still absorbs a large amount of research energy. In 1993, the year of the most recent comprehensive register of research on Northern Ireland subjects, 88 projects were devoted to education (O Maoláin 1993). As the research interest developed, it developed a number of different foci. One stream explored the relationship between educational segregation and socio-religious divisions. Another highlighted funding issues, both for schools attended by Catholic and Protestant pupils, and on the financial problems facing the new integrated schools. Two interventions, one a research study relating to differential funding, the other an innovative intervention by a charitable trust to support integrated schools, together throw light on how educational policy changes were made in Northern Ireland.

In 1990 the Standing Advisory Commission on Human Rights (SACHR), an independent body advising government on human rights issues, commissioned a research study of how schools were funded in Northern Ireland. SACHR's motivation for initiating the research was its growing conviction that lower academic performance from Catholic schools were caused in part by their inferior funding, and that lower performance helped to explain Catholic disadvantage in employment. Since the formation of the Northern Ireland state, Catholic schools, in return for retaining greater independence, had been provided with less public funding support than state schools, leaving the resulting state schools as de facto Protestant establishments. The SACHR research examined the per capita spending in schools, and confirmed that spending per pupil was substantially lower for Catholic children. The publication of the research was followed in 1992 by an increase in state financial support to Catholic (maintained) schools to the level enjoyed by state schools. At first sight this episode appears to demonstrate an outstanding example of how an alliance between strategic research and a public organisation can effectively lobby for change. This view, while true, needs to be tempered by regarding the matter from a broader perspective. In fact the effective lobbying had been carried out before the research had been commissioned. To decrease the differential employment of Catholics, and to respond to changing political attitudes both within government and the Catholic Church, there was already a strong pre-disposition towards altering the funding arrangements for schools, and no effective lobby against it. It was not the first time, nor the last, that research had been commissioned to justify a decision already reached in political terms.

If Catholic schools suffered vis-à-vis funding, the new category of integrated schools was in an even more difficult situation during the 1980s. These schools, which welcomed both Catholic and Protestant pupils, were permitted to open from the early 1980s, but were deprived of public funding support until each school had demonstrated its ability to attract and sustain a satisfactory school population . To many advocates it seemed a classic Catch-22 problem. The condition required for success seemed unattainable. At this point the Nuffield Foundation entered the lists. In an outstanding example of strategic intervention, Nuffield established a fund from which prospective integrated schools could borrow interest-free to purchase or build accommodation and cover essential costs until the schools qualified for government funding. The money was then repaid and recycled to support further integrated initiatives. The experiment helps to locate the relative importance of research as factor in policy-making. There was a growing body of research into the new integrated schools during the Nuffield initiative, and it certainly helped to raise the profile of the integrated issue and to inform policy developments, but the real breakthrough came from an imaginative financial intervention rather than from research.

d. Fair employment, Various researchers, and the Standing Advisory Commission on Human Rights

It is certainly true that some researchers in Northern Ireland have been associated either overtly and sometimes unwillingly with particular political parties or lobbies. Indeed it is perhaps surprising that research investigations are not more partisan. Perhaps the most controversial and bitter dispute on public policy is research into the distribution of employment and unemployment between Catholics and Protestants. Curiously enough there is little disagreement on the basic facts. Catholic unemployment rates have always been higher than those for Protestants, operating at a differential of around 2.5:1 for males, around 2:1 for females, although both differentials have been diminishing. The dispute is about the reasons for the differential. The dispute has taken on a strong political flavour, with politicians from the Democratic Unionist Party and Sinn Fein, entering the contest. The spectrum of explanations ranges between those who argue that sectarian discrimination sets the pattern and still maintains it, those who believe it is caused by demographic and geographical patterns which disadvantage Catholics and those who suggested that the principal cause was related to higher Catholic birth rates.

The third charge is probably the most controversial, as it overlaps the emotional issue of political demographics. A number of research studies and surveys have suggested that a continuation of past demographic trends could lead to a Catholic majority in Northern Ireland within a quarter century (Gallagher, AM (1991) provides an excellent review of the arguments and positions). This trend has been given added political significance by the increased residential segregation resulting from violence. In 1994 37% of electoral wards in Northern Ireland were highly - more than 90% - religiously exclusive (Clark, Liam, 'Apartheid takes root after 25 years of segregation', Sunday Times, 14 August 1994). Religious demographics is a hot political issue.

The controversy about fair employment research has cooled but not disappeared during the 1990s. The production of three reports on fair employment by the Standing Advisory Commission on Human Rights (1996), dealing respectively with the law, policy aspects and public views and experiences, has added a much-needed dimension of impartiality. But the bitterness of the dispute, the exchange of implications that the research has been influenced by sectarian bias, and the willingness of some political parties to exploit the debate for sectarian purposes, are all reminders that research is never neutral. Researchers working on controversial subjects within divided societies often find it difficult to avoid charges that they are partisan.

e. Religious imbalances in Northern Ireland's Civil Service, David Donnison, Centre for Environmental Studies

In 1973, Professor David Donnison, at the time Director of the Centre for Environmental Studies, published an article on employment in Northern Ireland in New Society. He cited a study of 477 senior Northern Ireland civil servants, from the grade of deputy to assistant principal upwards, which revealed that 95% of these officials were Protestants. These figures had not been previously published, and the article was widely copied by mainstream newspapers. Shortly afterwards, a press notice from the Northern Ireland Civil Service alleged that Donnison's figures were incorrect and claiming that 15% of the senior grades mentioned were Catholics, adding moreover, not all the remaining 85% were Protestants. (NICS 1973) The lower official figure provided for the first time solid confirmation of a serious imbalance and remained the basis for debate on public services employment until the Fair Employment Agency began producing religious breakdowns of employment practices some years later. Donnison's original figures were guesstimates, and had the desired effect of drawing more reliable data from official sources by way of challenge to his figures.

f. Research on the impact of the Troubles on the population of Northern Ireland, The Cost of the Troubles Study (COTTS)

In the period following the paramilitary cease-fires of 1994 in Northern Ireland, a new atmosphere of hope prevailed. In the atmosphere of new optimism, the position of those who had been bereaved and injured was in danger of being ignored. In the early peace process after 1994, whilst it was clear that prisoners played an essential part in the establishment and maintenance of cease-fires affording them a powerful and central role, those who had been bereaved or injured in the conflict had no such role or access to power. The constituency of those bereaved and injured was a fragmented one, divided by politics, geography, and isolated within a culture of silence and a lack of support services. The initiative to bring a group of bereaved and injured people from across the political spectrum together in order to collaborate on research and documentation of the effects of the Troubles came from one of the authors. The Cost of the Troubles Study (COTTS), the organisation that was formed, had two concerns: first there was a lack of robust evidence of need on which any new services could be based; and second that the interests of those bereaved and injured in the Troubles would be overlooked in the broader maelstrom of political change. As such, COTTS was part of the process of social construction of victim identity, and played a part in the subsequent development of victim politics, documented elsewhere (Smyth, M. (2000) 'Burying the past? Victims and community relations in Northern Ireland since the cease-fires' In Biggar, N. (ed) Burying the Past: Making Peace and Doing Justice after Civil Conflict. Washington DC: Georgetown University Press. and Smyth, M. (2000) 'The role of victims in the Northern Ireland Peace Process' in Guelke A. and Cox, M. A Farewell to Arms: From War to Peace in Northern Ireland. Manchester University Press).

The significant feature of the project was that those bereaved and injured in the Troubles held executive positions within the project, and participated not merely as informants but also as decision makers in the project as a whole.

The main goals of the project were:

In order to establish a sampling frame for a survey of Northern Ireland that was designed to examine experience and effects of the Troubles, the study compiled a database of all deaths due to the conflict since 1969. This database was then analysed to show which geographical areas had the highest death rates, and the survey was conducted accordingly. The database was then further analysed to show comparative death rates for gender, age, religion, affiliation, year and location. The results were published (Fay, M.T., Morrissey, M. and Smyth, M. (1999) Northern Ireland's Troubles: The Human Costs. London: Pluto) commercially in 1999. The study also conducted a survey of Northern Ireland investigating exposure to Troubles related events and the effects of the Troubles, and conducted in-depth interviews that were used to produce a touring exhibition, a book of personal accounts of the Troubles and a video to be used for training human service professionals.

Conclusions: Research and Policy

Patterns of research, as well as patterns of conflict itself, change constantly. Since the early 1970s there has been an extensive, if fluctuating, interest in the Northern Ireland conflict from outside Ireland - educationalists studying its segregated school system, churchmen examining the apparently denominational basis of the conflict, students of violence and its effects, and medical researchers examining the emergency procedures and surgical techniques in its hospitals. More recently there has been a shift towards comparative analysis of ethnic conflicts and approaches to conflict resolution. INCORE, for example, is currently researching peace processes in societies recently engaged in ethnic violence (Darby and MacGinty, 2000). The move towards comparative analysis is likely to continue, as those involved in peace processes across the world are looking abroad for practical guidance. This development, already well established, will certainly create new opportunities and problems for policy researchers, but it is likely that it will elaborate existing problems. Six of these problems are presented as propositions below.

1. Policy research without a dissemination strategy is a contradiction in terms

If policy research is to be effective it must start with a dissemination strategy, which takes account of the kind of research undertaken and how it can impact on the social and political situation. Reports, books or articles may be effective methods of dissemination, but only if the publication ends up on the desk of those the researcher wishes to influence. Ensuring this kind of policy impact requires the abandoning of purist notions of 'appropriate' roles for researchers. Personal contact through formal presentation of results is often essential, and both INCORE and COTTS have organised seminars for Northern Ireland politicians for this purpose. Research seminars for civil servants may achieve the same goal. Diversity of outputs, such as exhibitions that tour town halls, videos that can be used for training, and that are produced in language that is easily accessible to a wider audience can also improve dissemination. This diversification takes time and energy and, in the case of COTTS, the expenditure of money to employ professional help with publicity.

In order to render research meaningful to an audience wider than the research team, it is necessary to devise a strategy for ensuring that the findings impact on policy targets. This strategy starts at the inception of the research and continues until the publication of results and beyond. The fact that research is being carried out in the first place may be publicised in order to draw attention to a set of issues. When conducting large field surveys on issues of political sensitivity it may be wise to provide information to participating communities and constituencies about the researcher's intentions and credentials. This publicity can allow a process of drawing out any questions or doubts before embarking on the collection of data, and a means of publicising the later research findings.

Often research that is expensive to undertake and time consuming to document attracts little attention because it is written up in technical language or published in outlets with restricted circulation. University libraries in Northern Ireland, as elsewhere, are replete with research reports that have been read by few and had little influence. A dissemination strategy is required if this is to be avoided. In the case of the Fermanagh Civil Rights Association study, the research was undertaken not by a research organisation, but by a body, which wished to use research to document their concerns about discrimination, and to contribute to a wider civil rights campaign. A pre-existing network of civil rights campaigners was able to use the research to further its campaign. In the case of COTTS, a network of victims' organisations, many of whom participated in the study, were provided with copies of the work, and a touring public exhibition brought the results of the work within the reach of the general public. A public relations officer was also employed by COTTS for a three-month period in order to maximise the media coverage and dissemination of the findings.

Both of these examples are uncommissioned research, where the onus to disseminate - and the ownership of data - lies clearly with the researcher or researching organisation. Where research has been commissioned, the commissioning agent may own the data, and the prerogative to publish and disseminate results. In such cases, the researcher is in a much less powerful position, and will, more often than not, have to negotiate dissemination with the commissioner of the research, who may well decide that it is not in their interests to disseminate the work, or be willing to provide the extra funding necessary to ensure this happens.

Here, it is important to mention the role of the media in publicising research. Whilst censorship in the media at certain points in Northern Ireland's history would have precluded certain issues being effectively highlighted, in both the examples cited, the media were important (if not entirely accurate) agents of research dissemination. News media are often more interested in 'controversial' stories, in certain conditions where censorship is not severe. Critical research is likely to be more controversial and therefore more likely to be considered newsworthy - a mixed blessing, of course.

Finally, dissemination of research results does not, of itself, ensure that the research will make an impact. The ability to anticipate and plan for resistance to the influence of the research information is an important part of planning a research strategy. In violently divided societies, perhaps a disproportionate share of the population tends towards conspiracy views of social and political processes. Research critical of the status quo is often subject to criticism and attack from predictable quarters, often taking the form of attacks on the methodological soundness of the study, or on the credibility of the researcher. A 1972 report into housing by four academics at the University of Ulster produced a counter-blast by the Northern Ireland Housing Trust (then responsible, with the local councils, for housing) substantially longer that the report it attacked, detailing spelling and grammatical infelicities and typing errors as well as substantive criticisms (Birrell et al, 1972). The COTTS researchers also became aware that a government department assigned a statistician to 'vet' the study after the designated person concerned wrote to the researchers with a list of minor concerns. It is therefore important to iron out any methodological worries early in the project, and to give consideration to the way the research is presented - and the credentials of those who do the presentation. There are few enough examples of research that policy makers would have liked to ignore but couldn't, such as Fermanagh Civil Rights Association's unanswerable documentation of housing inequality. There are many more examples of research that officialdom wished to ignore, and could and did mange to do so.

2. The researcher's credibility is essential in determining the research's impact;

The orientation of the researcher is a factor in the way the research is disseminated and presented. The researcher's faith in the ability of the system to change will influence the success of the dissemination strategy. In many cases where research is addressing concerns of injustice, inequality and violence, those undertaking the work are caught in a dilemma between aiming to achieve change through their work and wishing to discredit the system responsible for the abuses they have uncovered. In both the Fermanagh Civil Rights Association and the COTTS example, this dilemma was marked.

Is it an advantage to come from outside the society one is researching? Outsider status in a researcher will usually raise concerns about trustworthiness in certain local communities within violently divided societies. Nationality and previous work are factors in this, and it is not unknown in violently divided societies for researchers to be suspected and accused of espionage. Questions about a researcher's credentials in such circumstances can be critical and potentially fatal for the researcher and colleagues, and will usually neutralise any potential impact that the research might have.

On the other hand, David Donnison's exploration of the composition of the Northern Ireland Civil Service was enhanced by his academic reputation and lent authority to the figures for the religious composition of the NICS that he published in New Society. Furthermore, Donnison was an outsider, assumed to be 'neutral' with no axe to grind in the Northern Ireland conflict. His Englishness symbolised and compounded his trustworthiness. Had he been from Northern Ireland, his data, and the motivation for publishing it, would have been suspect.

Other studies, such as Templegrove Action Research or COTTS were conducted by local researchers, and the issue of their identity was explicitly addressed in the presentation of the findings. Measures, such as co-working with colleagues from the other side of the divide, or the use of advisory group members with such perspectives, were included as part of the research's strategic attempt to produce inclusive analysis and reduce researcher bias (See Smyth, M. 'Researching Sectarianism' in Smyth, M. and Moore, R. (1996) Three Conference Papers on Aspects of Sectarian Division: Researching Sectarianism: Borders within borders: material and ideological aspects of sectarian division; and Limitations on the capacity for citizenship in Post-Cease-fires Northern Ireland. May 1996. Derry Londonderry, Templegrove Action Research). These strategies were described in the write up of the research, so that the reader could assess the methods used to achieve this end as well as the success or failure of the results of its use.

In COTTS, where a comprehensive survey of a divided population was conducted, it was necessary for the researcher to establish credentials in a wide range of fields within a divided society. With former members of the security forces, for example, who have no history of working with civilians, this presented a challenge, in more ways than one. Gaining credentials within one constituency in a divided society may entail losing credentials in another.

3. Timing is everything;

On reviewing the impact of the selected pieces of research in Northern Ireland, the issue of the timeliness of the project is central. Arguably, had Fermanagh Civil Rights Association produced their research earlier, in the absence of a growing demand for civil rights, and without the increased articulation of concern about discrimination against Catholics, the work may not have had the impact it did. The timing of a research initiative may be a function of the foresight or good judgement of the researcher or the commissioner of the research, or it may be a fortuitous accident. Timing may be a function of the need on the part of a public body to draw attention to a particular issue at a particular point in time, as when SACHR commissioned its research on educational funding. Other public bodies may also use research as a form of procrastination, to avoid being launched into what they consider precipitate action. Action can then be postponed until the research is completed, and then the results may be seen as inconclusive and more research called for, in the time-honoured fashion.

The timing of COTTS research into the human costs of the Troubles was fortuitous, in that the study had collected data not available elsewhere at a time when government was beginning to address the issues with which the study was concerned. As a result, the researchers were in a key position to influence the government's own commission into the impact of the Troubles, and the data collected was used as baseline data in preference to the official statistics available from government. Had the study been conducted earlier, the data might have been considered to be dated by government, and had it been conducted later it would have been too late to affect policy. The approach adopted by the researchers, in terms of inclusive definitions of 'victims' - which went against the earlier government positions adopted by, for example, the Compensation Agency - was readily adopted by the Bloomfield Commission, in part due to the influence of the work of COTTS. Later, the inclusive definition of victims was to come under attack from certain quarters of unionist and loyalist opinion, but by then, inclusion had been adopted as policy.

4. The indirect influence of research is greater than direct influence: there is a non-linear relationship between results and impact;

The indirect, or unintended, impact of research may, on occasion, be greater than the intended impact. An example of this is research on the labour market in Northern Ireland that showed the extent of employment and unemployment inequality between Catholics and Protestants (See Smith and Chambers, 1982. For a summary of the research see Gallagher, AM (1991), The Majority-minority Review, no. 2, Employment, Unemployment and Religion In Northern Ireland, Coleraine, Centre for the Study of Conflict). This fuelled a furious and ongoing debate (the so-called Catholic disadvantage debate) about the causes of employment inequality in Northern Ireland. This, in turn, contributed to a campaign in North America for disinvestment in firms in Northern Ireland whose employment practices were not manifestly egalitarian. However, the findings of the original research by Cormack and Osborne also led to the reworking of employment statistics by academics such as Graham Gudgin and Breen (See Gudgin and Breen, 1996) and politicians like Gregory Campbell of the Democratic Unionist Party who contest either the extent and causes for Catholic disadvantage. During a debate on the progress of measures guaranteeing equality in government in June 2000, Mr Gregory Campbell said it was a nonsense and a fallacy to keep saying people are discriminated against in Northern Ireland because they are Roman Catholic…. The reality is it's our [Protestant] community that's being discriminated against, day in, day out ('Anti-Catholic bias a fallacy: Campbell', Irish Times, June 7 2000). Thus, although a piece of research may have intended consequences, such as the raising of a debate in the first place, it may have unintended consequences such as stimulating others to rework data, in this case statistics, in increasingly creative and unorthodox ways.

5. He who pays the piper may have another tune in mind than that assumed by the researcher

The research commissioned by SACHR on how schools in Northern Ireland were financed revealed the inequity in the per capita funding for each school, which in turn was used to recommend and implement 100% funding of Catholic schools. This would suggest that the government, through SACHR, sought and found evidence for reforms or changes that they wished to implement in any case, with the research providing the justification for doing so. A similar approach may apply to research commissioned in 1998 by government into educational selection in Northern Ireland.

In such cases the commissioning body was reasonably confident that the research findings would support the proposed policy initiative. When research findings produce unexpected findings, especially if they are controversial, the outcome may be very different. Research into intimidation, carried out in 1972 by staff members of the Northern Ireland Community Relations Commission (Darby, J. and Morris, G. (1972), Intimidation in Housing, Belfast, Northern Ireland Community Relations Commission), found evidence of collusion between some police members and loyalist paramilitaries in north Belfast. The Commission established a sub-committee of its members to consider the matter of publication, which attempted to persuade the researchers to remove the controversial section and effectively delayed publications for more than a year. The report was leaked to the media by another Commission member and, after an unsuccessful enquiry into the source of the leak, the report was published in its original form (Darby and Morris 1974). The incident underscores the importance of researchers clarifying the terms of publication with the commissioning agency before research is started. When the research is directly commissioned by government, or is conducted in-house as the intimidation research was, the researchers' hands are often tied. When there is an intermediate or private funding source, flexibility is usually possible. It is probable that the intimidation report, which had a considerable effect on government policy and practices, would not have been published had it not been leaked to the media.

6. Expect your research to be spinned.

Once a piece of research is placed in the public domain, through publication of some kind, then the researcher or author loses control of how it is used. Findings and even statistical results will be presented in ways that do not accurately reflect the original findings, results will be selectively read and represented in ways that suit a particular agenda. Spinning, - the process of manipulating the research results to fit in with pre-existing agendas - often ignores the overall meaning of a set of findings, ignores complexity, and goes for a simple and inevitably partial version of the findings. Donnison's publication of unconfirmed figures about the proportion of Catholics in the senior ranks of the Civil Service demonstrates the point. After the publication, and the rebuttal, the issue became one of considering whether Donnison's estimate of 5% or the Northern Ireland Office's claim of 15% was correct; but one could be reasonably certain that the true figure fell somewhere between the two estimates. The debate can then start. The key lesson is that researchers need to be aware of the process and to present their findings in a form that diminishes the potential for misrepresentation and accentuates the most significant findings.

In the case of COTTS, the media used the data on responsibility for deaths in the Troubles to highlight the predominant role of the IRA in Troubles-related killing. However, the media misquoted the research and attributed to the IRA all deaths caused by republican paramilitaries in general. Furthermore, they chose to ignore other equally newsworthy aspects of the research, such as the consistently greater death rate amongst Catholics than Protestants. The spinning of research by the media in this case led to researchers being confronted (or congratulated) by politicians. Journalists who report the research may have a limited grasp of, for example, statistics and may misinterpret or over-generalise the results in the interests of a good story. The careful preparation of press releases, the provision of press briefings, and the cultivation of key relationships with journalists can help reduce these tendencies, although the press is unlikely to become a reliable outlet for accurate and detailed research dissemination. The Templegrove project is an example of how research can be designed to provide an antidote to a press spin - in this case about so-called ethnic cleansing in Derry Londonderry. In that project, the response of the Protestant community to the research was crucial, as was the involvement of politicians in research, so that dialogue and agreement about methods and results would generate an authoritative set of figures about population movement. In a sense, the Templegrove research was designed to be the antidote to a media spin on so-called ethnic cleansing.

Does research make any difference?

The famous Hawthorne experiments (Homans, GC (1965) Group Factors in Worker Productivity. In H. Proshansky and L Seidenberg (Eds) Basic Studies in Social Psychology. New York: Holt) showed that conducting research in itself alters that which is studied. The act of researching conveys to those who are the subject of inquiry a number of messages about the nature of their situation and the level of outside interest in it. In violently divided societies, where world attention may well be a factor in the conflict itself, it is clear that doing research is unlikely to be without effects of some kind. It seems that research may well make a difference, although the nature of that difference may not always be the difference that was intended. Violently divided societies are complicated and often unpredictable research environments and when the nature of division is the subject of research, the lack of predictability and complication are further enhanced.

References

Barritt, D. and Carter, C. (1962) The Northern Ireland Problem, (Oxford: Oxford University Press).

Birrell, D et al, 'Housing Policy in Northern Ireland', Belfast, Community Forum, 2.2.1972.

The Cameron Report (1969): Disturbances in Northern Ireland, (Belfast: HMSO).

Clark, L. 'Apartheid takes root after 25 years of segregation', Sunday Times, 14 August 1994.

Collier, D and Mahoney, J, 'Insights and Pitfalls: Selection bias in Qualitative research', World Politics, 49, 1, 1996.

Darby, J.and McGinty, R. (2000) The Management of Peace (London: Macmillan).

Darby, J. and Morris, G (1974). Intimidation in Housing, (Belfast: Northern Ireland Community Relations Commission).

Darby, J. (1997) Scorpions in a Bottle, (London: Majority Rights Group).

Donnison, D., 'The Northern Ireland Civil Service', New Society, July 5 1973.

Downing, JDH, (1974) Nothing to Hide- the Boehringer case and academic freedom in Northern Ireland (Belfast, Council for Academic Freedom and Democracy: National Council for Civil Liberties).

Fermanagh Civil Rights Association (FCRA) (1969) Fermanagh Facts, (Fermanagh: FCRA).

Flackes, WD, Elliot, F. (1994) Northern Ireland: A Political Directory 1968-1988. (Belfast: Blackstaff)

Gallagher, AM (1989) The Majority-minority Review, no. 1, Education and Religion in Northern Ireland, (Coleraine: Centre for the Study of Conflict).

Gallagher, AM (1991), The Majority-minority Review, no. 2, Employment, Unemployment and Religion In Northern Ireland, (Coleraine: Centre for the Study of Conflict).

Gudgin, G. and Breen, R. (1996) Evaluation of the Ratio of Unemployment Rates as an Indicator of Fair Employment (Belfast: Central Community Relations Unit).

Homans, GC Group Factors in Worker Productivity. (1965) In H. Proshansky and L. Seidenberg (Eds) Basic Studies in Social Psychology. (New York: Holt).

Horowitz, Donald (forthcoming) The Deadly Ethnic Riot (California, University of California Press).

O'Maolain, Ciaran (1993) Register of research on Northern Ireland, (Coleraine: Centre for the Study of Conflict).

Osborne, RD, Cormack, RJ and Miller, RL, (1987) Education and Policy in Northern Ireland (Belfast: Policy Research Institute).

Rose, P. (2000) How the Troubles came to Northern Ireland (London: Macmillan).

Smith, DJ and Chambers, G (1991) Inequality in Northern Ireland (Oxford: Clarendon).

Smyth, M. (1996) Life in Two Enclave Areas in Northern Ireland. A Field Survey in Derry Londonderry after the Cease-fires. June, 1996 (Derry Londonderry: Templegrove Action Research).

Smyth, M. and Moore, R. (1996) Three Conference Papers on Aspects of Sectarian Division: Researching Sectarianism: Borders within border: material and ideological aspects of sectarian division; and Limitations on the capacity for citizenship in Post-Cease-fires Northern Ireland. May 1996 (Derry Londonderry: Templegrove Action Research).

Smyth, M. A Report of a Public Hearing on the Experiences of Minorities in Derry Londonderry, April 1996. (Derry Londonderry: Templegrove Action Research).

Smyth, M. Two Policy Papers: Policing and Sectarian Division (with Ruth Moore) and Urban Regeneration and Sectarian Division. Templegrove Action Research, April 1996.

Smyth, M. A Report of a series of Public Discussion on Aspects of Sectarian Division in Derry Londonderry, held in the period December 1994 - June 1995. Templegrove Action Research. March 1996.

Standing Advisory Commission on Human Rights (1996) Employment Equality in Northern Ireland (Authors: Magill and Rose; McLaughlin and Quirk; McVey and Hutson), (Belfast: SACHR).

Whyte, John (1991) Interpreting Northern Ireland, (Oxford: Blackwell).