The contributors
	Introduction Richard Breen, Paula Devine and Lizanne Dowds
1.	Public support for democratic values in Northern Ireland Bernadette C. Hayes and Ian McAllister
2.	Who wants a United Ireland? Constitutional preferences among Catholics and Protestants Richard Breen
3.	Informal care in Northern Ireland Eileen Evason and Gillian Robinson
4.	Women and work Janet Trewsdale and John Kremer
5.	Race, ethnicity and prejudice in Northern Ireland John D. Brewer and Lizanne Dowds
6.	Satisfaction with health services in Northern Ireland Ann Largey and Ciaran O'Neill
7.	National identity Karen Trew
8.	Gender discrimination among married full-time employees in Northern Ireland: some initial findings John Spencer, Mary Trainor and Boyd Black
9.	Community relations, equality and the future Tony Gallagher
Appendix 1: Technical details of the survey …………….Alan McClelland
Appendix 2: Using NISA data
Appendix 3: The questionnaires
Index

Section Headings:

Introduction

Since the partition of Ireland under the terms of the 1921 Anglo-Irish Treaty, the constitutional position of Northern Ireland has been a central source of political and social division within the province. Individual preferences about the future of Northern Ireland not only differ sharply between the Unionist and Nationalist poles, but these preferences themselves have now come to take on an unusually central role in determining what this position shall be. The UK government has maintained that the constitutional position of Northern Ireland is a matter to be decided by the people of Northern Ireland. As a consequence the importance of understanding exactly what constitutional preferences people hold and why they hold them has received a new impetus. In this chapter we use the Northern Ireland Social Attitudes (NISA) survey data to examine a number of questions concerning the constitutional preferences held by people in Northern Ireland.

We begin by looking at how these preferences have changed since the NISA surveys began in 1989. Then we examine the results of the 1994 survey in greater detail in order to say something about what distinguishes those who hold different constitutional preferences. Clearly the most important characteristic that distinguishes those who favour the retention of the Union with Britain from those who favour a united Ireland is religion. However, we also examine differences within each of the two major religious groups. How much variation is there in constitutional preferences among Catholics and among Protestants, and how might such variation as exists be explained? In regard to this latter question there are two hypotheses that we will seek to test. The first is that variation in constitutional preferences is due to differences in the socio-economic and demographic characteristics of individuals. So we might expect, for example, that middle class or better educated Protestants will be less unionist than working class Protestants and that middle class Catholics will be less nationalist than their working class counterparts. The second hypothesis is that, particularly among Catholics, variation in constitutional preferences has less to do with the individual's own socio-economic position and is more strongly linked to his or her perceptions of the position of the Catholic community vis-a-vis the Protestant community. In other words, strength of Nationalist feeling depends less on whether a Catholic is well-educated, middle class and so on and more on the degree to which he or she thinks that Catholics are discriminated against in Northern Ireland.

Constitutional Preferences: Change over Time

The data we use come from all of the NISA surveys from 1989 through to 1994. In each of these years - with the exception of 1992 when there was no NISA survey -respondents have been asked the following question about their constitutional preferences:

'Do you think the long-term policy for Northern Ireland should be for it to remain part of the United Kingdom or to reunify with the rest of Ireland?'

Other possible responses are not provided in the question format but are recorded. Table 1 shows the responses to this question in each NISA survey since 1989. Every year the great majority of people - around 70 per cent in the period 1989-1993 and somewhat less in 1994 - chose the response 'Remain part of the UK' while between a fifth and a quarter favoured the reunification of Ireland. The numbers expressing support for any other option were very small. The option of an independent Northern Irish state received the support of no more than one per cent of respondents in each survey. Around five per cent of respondents each year said that they did not know which long-term policy they would prefer but hardly anyone refused to answer the question.

Our most recent information comes from the survey carried out in the spring of 1994. This was after the Downing Street Declaration of 1993 but before the cease-fires of autumn 1994 and it is perfectly possible that constitutional preferences might have shifted in some way in response to these later events that will not be known until the 1995 survey results become available. What the data do show, however, is stability or a very slight increase over the 1989-1993 period in the percentage favouring Northern Ireland's remaining within the UK (though this change is not statistically significant) and thus a decrease in the popularity of the united Ireland option. But, in the 1994 data, we see something of a reversal with a considerable increase (from 20 to 27 per cent) in the number preferring a united Ireland and a drop in the percentage wishing to retain the Union.

Community Background and Constitutional Preferences

As we might expect, the major factor that explains variation in constitutional preferences is community background or religion, distinguishing Catholics from Protestants. Table 2 shows the preferred long-term policy for Northern Ireland for each of three groups. These are Catholics, Protestants and those who are of another religion (neither Catholic nor Protestant) or who have no religion or who refuse to state their religion (this group is labelled 'Others').

Table 2 shows that in excess of 90 per cent of Protestants wish to see the Union retained, and there has been little change in the position over the five years. No more than six per cent of Protestants favour a united Ireland. Similarly, the 'Others' who cannot be placed in either of the main religious groups are overwhelmingly in favour of retaining the Union, although their support has fluctuated over time. As we might expect, this group is relatively small, making up, at most, 14 per cent of the sample.

The major change over the period 1989-1994 is found among Catholics. Up to 1993 the trend here was towards a decline in nationalism and an increase from 32 to 36 in the percentage favouring the retention of the Union. However, the latter fell to 24 per cent in 1994 and the percentage favouring Irish unification increased from 49 in 1993 to 60 in 1994. This is a substantial change and is responsible for the overall increase in the percentages preferring a united Ireland shown in Table 1. In summary: over the period Protestant preferences have not changed but among Catholics the gradual drift towards the Union was sharply reversed in 1994.

National and Constitutional Identity

Apart from the direct question that asks people about their preferences for long-term policy in Northern Ireland there are three other items in the questionnaire that give a less direct insight into individual constitutional preferences. The first of these is what we term 'national identity'. This is captured by the following question:

Which of these best describes the way you usually think of yourself?

and the possible responses are 'British', 'Irish', 'Ulster', 'Northern Irish', 'Sometimes British, sometimes Irish' and 'Other'. As very few people (less than half of one per cent) choose the 'Other' response, we omit them from our analysis. In other research (Breen, 1994) it has been shown that, as far as constitutional and political preferences are concerned, there is no difference in the responses of those who claim to be 'British' or 'Ulster'. Similarly there is no difference in the responses of those who claim to be 'Northern Irish' or 'Sometimes British, sometimes Irish'.

Accordingly, in Table 3 we show, for the four surveys for which this information is available, the percentage distribution of Protestant and Catholic respondents across three categories: (a) 'British' or 'Ulster', (b) 'Northern Irish' or 'Sometimes British, sometimes Irish', and (c) 'Irish'.

In contrast to the direct measure of constitutional preferences reported in the previous table, Table 3 shows remarkable stability over time. Around four-fifths of Protestants consider themselves to be 'British' or 'Ulster', with the great majority of the remainder picking either the label 'Northern Irish' or the 'Sometimes British, sometimes Irish' option. Three per cent or less choose the description 'Irish'. Among Catholics just over 60 per cent regard themselves as 'Irish', just under 30 per cent as 'Northern Irish' or 'Sometimes British, sometimes Irish', with around 12 per cent considering themselves 'British' or 'Ulster'.

Table 4 reports another measure of identity, which we call 'constitutional identity'. This is the response to the question:

Generally speaking, do you think of yourself as a Unionist, a Nationalist or neither?

and the possible replies are 'Unionist', 'Nationalist' or 'Neither'. Table 4 shows that among Protestants around three-quarters pick 'Unionist' with the rest responding 'Neither'. Catholics differ from Protestants in this respect in two ways. First, Catholics are almost equally split between the labels 'Neither' and 'Nationalist'. Secondly, in contrast to the stability over time evident in the Protestant replies, the balance of responses between these two alternatives shifts substantially from one year to the next. So, for example, in 1993 40 per cent claimed to be Nationalists, while in 1994, 54 per cent did so. What is perhaps most striking - though unsurprising - about the responses shown in Table 4 is that no Protestants call themselves Nationalists and no Catholics call themselves Unionists.

Party Support

Patterns of political party support are primarily structured by the position of the party on the constitutional issue. This means that community background or religion is a very strong predictor of party support. However, within the main Unionist and Nationalist political blocs there is party competition within the Protestant and Catholic communities, respectively. This takes place on the basis of different policy approaches to the constitutional issue but also on the basis of conventional left/right support. Thus, Sinn Féin and the DUP receive a greater share of their support from the working class than do the SDLP, Alliance or the Ulster Unionists (Evans and Duffy, in press).

Table 5 shows the extent of the religious divide in patterns of party support by reporting the percentages of Catholics, Protestants and Others who support Unionist parties (Democratic Unionist Party (DUP), Ulster Unionist Party (UUP) and Other Unionist Parties), the Alliance party, and Nationalist parties (Social Democratic and Labour Party (SDLP) and Sm Fein). (A small percentage of respondents support other parties or support no party but they are excluded from Table 5). Here there is remarkable stability over time, with Protestants and Catholics presenting almost a mirror image of each other. Just under 90 per cent of Catholics support Nationalist parties while just under 90 per cent of Protestants support Unionist parties. Within each community support for the Alliance party runs at about 12 per cent.

Thus far then, we have looked at the position, over the 1989-1994 period, of Catholics and Protestants on four variables - their preferred long-term option for Northern Ireland, their national and constitutional identity and their party support. There are several points to note about the responses to these questions. First is the consistency over time in the pattern of responses, particularly among Protestants. However on two issues - long-term policy for Northern Ireland and constitutional identity - this stability contrasts with the relatively high degree of change among Catholics. Second, not only does the Catholic community display somewhat more year to year variation, but there is also clearly much more variation among the Catholic community within each year. So, in Table 2, we see that whereas virtually all Protestants favour the retention of the Union there is a substantial minority of Catholics - as high as 36 per cent in 1993 and as low as 24 per cent in 1994 - who also favour this option over that of a united Ireland. In addition, while very small percentages of Protestants favour any option for Northern Ireland other than retention of the Union or reunification, and very few of them say that they do not know what option they prefer, as many as 15 per cent of Catholics favour some other option or say they do not know. Similarly, while 80 per cent of Protestants choose the national identity label 'British' or 'Ulster', Catholic responses are a little more evenly divided with almost 40 per cent seeing themselves as something other than 'Irish'. In Table 4 we see that while only one quarter of Protestants claim not to be Unionists, as many as 60 per cent of Catholics (in 1989 and 1993) claim not to be Nationalists. However, this variation among Catholics is not evident in party support. Despite their apparently greater degree of heterogeneity on the other items we have examined, Catholics, to the same extent as Protestants, do not stray beyond the traditional boundaries of sectarian politics.

Constitutional Preferences, Identity and Party Support

In the remainder of this chapter we confine our attention to the 1994 NISA data with three aims in mind. First, to look at the relationships between the direct measure of constitutional preferences and the three less direct measures - namely the two identity items and party support. Second, we examine the degree to which socio-economic and demographic factors such as age, level of education and social class, might be related to constitutional preferences. Finally we look at the relationship between the constitutional preferences people hold and their perceptions of the standing of the Catholic community relative to that of the Protestant community in Northern Ireland - or, for short, perceptions of Catholic disadvantage. The purpose of these analyses is, as noted in the introduction, to assess the extent to which individual socio-economic position shapes constitutional preferences and the extent to which the latter are shaped by perceptions of the position of the Catholic community as a whole, vis-a-vis the Protestant community.

Table 6 shows the relationship between constitutional preferences, our two measures of identity and party preference. So, for example, we see that 94 per cent of those who describe their national identity as 'British' or 'Ulster' favour retention of the Union with Britain. The figures in this table are as we might have expected, though some features are worth drawing attention to. It is notable that all those responses to the two identity questions and party preference item that fall on the unionist side (that is, national identity 'British' or 'Ulster', constitutional identity unionist and preference for one of the two unionist parties) are associated with very high levels of support for the Union - indeed virtually everyone who gives one of the unionist replies is in favour of the Union. This might seem like a tautology, until we consider the nationalist replies. So, for example, people who consider themselves 'Irish', far from all being in favour of Irish unity, are divided in a ratio of three to one between preferring a united Ireland and preferring retention of the Union. Similarly, while it is true that almost all Sinn Fein supporters favour a united Ireland, the balance among those who prefer the SDLP is once again three to one in favour of a United Ireland. Indeed, far from the pattern of SDLP support being the mirror image of support for, say, the Ulster Unionist party, it actually mirrors that of the Alliance party. On the evidence presented here, support for the Alliance party might be characterised as unionist with a substantial minority of nationalists, while support for the SDLP might be characterised as nationalist but with a substantial minority of unionist supporters.

The picture that Table 6 presents in many ways replicates our earlier findings. The Protestant community is very homogenous in its constitutional preferences, the way its members view their national and constitutional identity, and in their support for Unionist parties. In other words, Protestants tend to cluster very tightly around a characteristic 'unionist' position. The Catholic community, on the other hand, is more heterogenous. While it is clearly the case that the great majority of Catholics consider themselves Irish and support Nationalist parties, there is, nevertheless between 25 and 40 per cent of their number (depending on the item in question) who depart from the characteristic 'nationalist' position.

Socio-economic and Demographic Influences on Constitutional Preferences

Table 7 shows the percentage of all respondents and of Protestants and Catholics who prefer the retention of the Union, according to a number of socio-demographic variables. These are gender, age, educational qualifications and social class. Our measure of social class is the Goldthorpe schema (Goldthorpe, 1980) in which the 'Salariat' comprises managers, professionals and those in higher white collar occupations, while the 'Routine non-manual' class includes all other white collar workers. The 'Petty bourgeoisie' is made up of all employers and the self-employed (see Erikson and Goldthorpe, 1992, Chapter 1 for a fuller description).

Base: All Protestants and Catholics who prefer retention of the Union or a united Ireland (excludes those who give any other reply to constitutional preference question and excludes those not recorded as either Protestant or Catholic).

What is most evident in Table 7 is the absence of any very clear relationships between these socio-economic or demographic factors and people's constitutional preferences. The exceptions to this are gender (Catholic men are significantly more likely to favour a united Ireland than are Catholic women) and education where, for all respondents, possession of a degree is associated with lower levels of support for maintenance of the Union. Young Catholics also emerge as less supportive of the Union. However, it is very obvious that there are no substantive or statistically significant differences among people at the other levels of education nor is variation in constitutional preferences linked in any consistent way to social class or age differences (with the single exception noted above). We also examined the relationship between constitutional preferences and household income and here we found a significant relationship among Protestants such that those with a higher income tend to be less likely to want to see the Union retained. However, among Catholics there was no relationship.

In general terms, we have to conclude that socio-economic and demographic factors play little part in explaining why people of either religious group have particular constitutional preferences. If anything, these factors play a more significant role among Protestants than Catholics, but one might well be more impressed by the weakness of their effect rather than their strength. Analyses of NISA data for 1989, 1990 and 1993 (Breen, 1994) also found that socioeconomic and demographic factors explained more of the variation in preferences among Protestants than Catholics.

Perceptions of Disadvantage and Constitutional Preferences

The NISA data contain two items that yield direct measures of respondents' views as to the extent of disadvantage suffered by Catholics relative to Protestants in Northern Ireland. The first of these is a question that asks:

Thinking of Catholics - do you think there is a lot of prejudice against them in Northern Ireland nowadays, a little, or hardly any?

In our analysis we score the responses 0 = hardly any, 1 = a little and 2 = a lot. The second is a set of questions that allows us to allocate respondents to one of five categories - Catholics have a much better chance of getting a job than Protestants; Catholics have a somewhat better chance; members of both communities have the same chance; Protestants have a somewhat better chance of getting a job than Catholics; Protestants have a much better chance of getting a job. Scores on this measure run from 0 = Catholics have a much better chance to 4 = Protestants have a much better chance.

Thus the first item measures the respondent's belief in whether prejudice against Catholics exists and the second focuses more directly on inequality in the labour market. Table 8 shows the correlation between these two measures and constitutional preferences. A correlation between two variables can take an absolute value between zero and one. A value near zero shows that there is no relationship between the two variables, whereas the closer the correlation is to one the stronger the relationship between them. In this case we see that our overall measure of prejudice and our measure of labour market disadvantage are both significantly related to constitutional preferences, with the latter being rather more strongly related. The direction of the relationship is as we would expect: people who believe that there is prejudice against Catholics and those who believe that Protestants have a better chance of getting a job are more likely to prefer a united Ireland. If we then look at these correlations within each religion, we find that the relationships are weaker. However, among Catholics the relationships with beliefs about labour market disadvantage and with beliefs about anti-Catholic prejudice are still significant. In other words, those who believe that Protestants have a better chance of getting a job and/or that prejudice against Catholics exists are more likely to prefer Irish unity than those who believe that both communities have similar chances of getting a job or who believe that there is little or no anti-Catholic prejudice.

Note: a positive correlation means that a higher score on the disadvantage measure is associated with a preference for a United Ireland.

These results suggest that, among Catholics, variation in constitutional preferences is not sensitive to variation in individuals' socioeconomic and demographic position but it is sensitive to individuals' perceptions of how Catholics fare, relative to Protestants. Breen (1995) shows also that constitutional preferences are related to people's beliefs about the impartiality or otherwise of the security forces. Those Catholics who prefer a united Ireland are more likely to believe that the security forces favour Protestants. Unfortunately we do not have measures of this in the 1994 NISA survey.

The fact that individual position is of so little importance considerably weakens the thesis that upward mobility on the part of individual Catholics will tend to make them less nationalist in their preferences. Rather, our results (and those of Breen, 1994) suggest that what is important is how Catholics perceive the position of their own community rather than themselves individually. In this sense, then, a weakening of nationalist sentiment is more likely to be found where there is a perception of a general improvement in the position of the Catholic community. This may, of course, involve a degree of upward social mobility, but this is perhaps less important than a perception of equality of treatment and of condition between the two communities.

Catholic Unionists

The Protestant community is strikingly homogenous in its constitutional preferences. This suggests that internal diversity among Protestants should not be overemphasised. Although there is some variation among Protestants in their national identity, between those who see themselves as 'British' and those who consider themselves 'Ulster', this seems to have little significance for constitutional preferences. Among those who consider themselves 'Ulster' exactly the same distribution across the constitutional preference categories exists as among those who consider themselves 'British'. Similarly, the often heard view that a substantial body of middle class Protestants are more likely to accept a united Ireland, or some compromise that weakens the Union, receives little support from our analysis. It is certainly the case that those Protestants who express a preference for a united Ireland are predominantly well educated and have higher incomes, but they are very few in number. Around 90 per cent of Protestants support the Union.

Table 9 looks at the national identity, constitutional identity and party support among three groups - Protestants, Catholics who favour a united Ireland and Catholics who favour retention of the Union. The consistency of Protestants in their responses to the four items is once again evident. There is also a high level of consistency in the responses of those Catholics who favour a united Ireland. However, among the third group - which we might call 'Catholic unionists' -national identity, constitutional identity and party support are sometimes in conflict with the preference for retention of the Union with Britain. They are much more evenly divided across the three national identity categories than are Protestants or Catholics who favour a united Ireland, and three-quarters of them profess to being neither a Unionist nor a Nationalist. Perhaps the most striking difference is to be observed in political party support. Whereas the two Unionist parties are supported exclusively by supporters of the Union, of the two Nationalist parties by far the larger (the SDLP) is supported by a high percentage of Catholics whose preferred long-term constitutional preference is for the retaining of the Union with Great Britain. As Table 6 shows, they account for roughly one quarter of SDLP supporters.

Conclusion

At present it appears that around two out of three adults in Northern Ireland favour a long-term policy of retaining the Union with Britain. Among Protestants about six per cent favour a united Ireland in the long term, while among Catholics, around one quarter say they want to see the Union preserved.

Forecasting is a notoriously unreliable business, but this has not deterred much speculation surrounding the issue of a future Catholic majority in Northern Ireland. Extrapolating from the results of the 1991 census, it has been suggested that Catholics may form the majority in Northern Ireland some time in the next century. However, even putting aside the uncertainties attached to purely demographic forecasts, the constitutional and political implications of a Catholic majority are quite unknown. Given the current pattern of Catholic constitutional preferences, a Catholic majority is far from being the same thing as a majority in favour of Irish unity. Indeed, assuming 70 per cent of Catholics supported Irish unity (and that no Protestants did) a majority in favour of unification would require that the adult Catholic population be more than two and a half times larger than the Protestant population. But, of course, it is unlikely that the current pattern of constitutional preferences among Catholics (and possibly among Protestants) will be maintained. The evidence we have suggests some shift among Catholics (and no change among Protestants) between 1989 and 1993 towards greater support for the Union, though this trend is not statistically significant. However, in 1994 this trend was dramatically reversed with a very marked decline in the percentage of Catholics in favour of the Union. More generally, changing, but unpredictable, economic or political circumstances may lead to further change in constitutional preferences, but the direction of such changes cannot be predicted.

In a recent paper, Ruane and Todd (1992) have warned against overemphasis on the potential for middle-ground politics in Northern Ireland. Our results reinforce this, at least in respect of the Protestant community where relatively little variation in constitutional preferences can be seen. Among Catholics there is greater variation and here, of course, we would point particularly to the group we labelled 'Catholic unionists'. On the one hand this might be seen as indicative of greater potential for some kind of middle-ground politics, but (and as the corollary to this) it can also be seen as an indicator of a possible volatility in constitutional preferences that makes predictions particularly unreliable.


References

BREEN, R. 1994. 'Constitutional preferences, socio-economic status and perceptions of disadvantage among Catholics and Protestants in Northern Ireland', unpublished report to the Central Community Relations Unit.

BREEN, R. 1995. 'Beliefs about the Treatment of Catholics and Protestants by the Security Forces' in Breen, R., Robinson, C. and Devine, P. (eds), Social Attitudes in Northern Ireland: The Fourth Report, Appletree Press, Belfast.

ERIKSON, R. and GOLDTHORPE, J.H. 1992. The Constant Flux, Clarendon Press, Oxford. EVANS, C. and DUFFY, M. 1996. 'Beyond the sectarian divide: the social bases and political consequences of Nationalist and Unionist Party competition in Northern Ireland', British Journal of Political Science, in press.

GOLDTHORPE, J. H. with LLEWELLYN, C. and PAYNE, C. 1980. Social Mobility in Modern Britain, Clarendon Press, Oxford.

RUANE, J. and TODD, J. 1992. 'Diversity, Division and the Middle Ground in Northern Ireland', Irish Political Studies, 7: 73-98.