Negative attitudes to minorities and immigrants in Northern Ireland: a very British problem

Northern Ireland has seen an increase in the number of violent protests over the last few years, many of which have been organised online with explicitly anti-immigrant or racist motivation. Whilst the majority of commentators have described them as racist violence or rioting, likening recent behaviour to a pogrom, some politicians have invoked widespread "legitimate concerns" to explain, if not to justify these acts.

The legitimacy of these vaguely-defined concerns, given the relatively low immigration numbers involved, has been questioned by other writers. Yet it is evident that many of those rioting two weeks ago did hold them, legitimately or otherwise.

This article puts the question of legitimacy to one side and examines how widespread negative attitudes to minorities and immigrants are here. It asks whether they are concentrated in certain demographics and examines commonly-held assumptions – that those with low incomes, less education and those who have never lived outside of Northern Ireland are more likely to hold negative views of minority groups and immigrants.

It finds, uncomfortably but undeniably, that whilst these are important factors, a stronger predictor of negative attitudes than any of them is national identity. In short, the evidence shows that hostility towards minorities and immigrants in Northern Ireland is a very British problem.

The evidence – the Northern Ireland Life & Times survey

Great Britain is extensively polled. According to this tracker on wikipedia, there were 27 polls conducted in May 2026 alone that asked people living in England, Scotland and Wales who they would vote for at the next general election. This is before you even consider polling on other topics, which receive less attention but are still regularly polled.

Northern Ireland (NI) is different. Partly because the big Westminster Parties don't have a presence here (more on the Northern Irish parties in just a moment), the big polling companies usually won't include Northern Irish people in their panels. The one native polling company – LucidTalk – conducts voting intention polls each quarter for the Belfast Telegraph (which generally bolt on one or two other topics), as well as custom polls for paying clients. But there is nothing like the amount of polling data on topics such as immigration that there is in Great Britain.

There is however, the Northern Ireland Life and Times (NILT) survey. This major annual survey, run by the ARK team of researchers located across Queen's University Belfast and Ulster University, has measured public attitudes towards immigration and ethnic diversity (amongst many other topics) in Northern Ireland every year since 2005. The ARK team publish research updates based on this survey, with two recent updates relevant here:

In March 2025, Katy Hayward and Paula Devine examined population-wide attitudes to immigration over this period, observing a mixed picture, with "a particular rise in expressions of support for immigration after 2020" being followed by a "drop in inclusivist attitudes between 2022 and 2023".

In a follow-up report in February 2026, Philip McDermott similarly reported "the persistence of prejudice amidst the potential for change". This report also cut the data by age and found some differences in attitudes along this dimension.

Aside from these updates, no analysis to date has really dug into the wealth of NILT data to understand what motivates opinion at a more granular level. The survey asks literally hundreds of questions on both background and attitudes across a wide variety of subjects, so there is a huge potential for this given the volume of data available. And crucially, unlike its British counterpart, the British Social Attitudes survey, ARK makes this data freely and readily available on the NILT website.

The analysis presented here is based on aggregated responses from the 2020–2025 editions of the NILT survey.

A first cut – political party support

Perhaps the most obvious place to start when examining socio-political attitudes is support for political parties themselves. As mentioned above, Northern Ireland has its own peculiar set of these, five of which are split out explicitly by NILT and have enough data to support robust analysis. These parties are Sinn Féin, the Social Democratic and Labour Party (SDLP), the Alliance Party, and the Ulster (UUP) and Democratic (DUP) Unionist Parties.

Two other parties have not been included here. The Green Party of NI is split out as an option in the NILT survey, but it doesn't consistently get enough responses to allow for robust statistics, whilst the Traditional Unionist Voice (TUV), currently polling level with three of the five parties listed above, is not separately broken out by NILT from the "Other parties" grouping.

By splitting the data by party, and examining the average response to various attitudinal questions concerning minority ethnic groups, immigration and societal mixing, we can profile the average attitudes of each party's voters across these questions:

The first four sets of questions here relate to attitutes to minority ethnic groups, with three specific groups named and a final question relating to all other minorities. The fifth group of questions relates to attitudes to refugees and asylum seekers, and the last group examines attitudes to mixing within the two traditional communities, most commonly labelled as Catholic and Protestant.

Perhaps the first remarkable observation from the chart above is that the vast majority of dots sit above the 50% mark – that is to say, a majority of respondents have generally supportive attitudes across most questions and parties. Indeed, the average response to many questions sits well above this mark across the political spectrum.

Considering the recent riots, which had a clear goal of intimidating minority residents in the areas affected, it is notable that even DUP voters, the most negative on all questions except one, still demonstrate 73% and 85% agreement with the statement "I would willingly accept [a person of group X] as a resident in my area" for Muslims and other minority ethnic groups respectively.

Nevertheless there are clear differences across parties. Alliance and SDLP voters are consistently the most positive on all questions relating to minorities and immigrants, with Sinn Féin next, followed by the two unionist parties, the UUP and DUP.

The only group of questions with a significantly different pattern in terms of ordering is the last one, referring to mixing of the two traditional communities. Here, Sinn Féin swap places with the UUP in having a more negative attitude to community mixing in neighbourhoods and at school. And SDLP voters are notably more sceptical when it comes to mixing in school. Only Alliance voters, perhaps unsurprisingly, have consistently positive attitudes across all three questions here.

Digging deeper into underlying drivers

Looking at difference in attitudes by political party support is interesting, but it doesn't tell us much about the deeper personal motivations behind this variation. One way to do this is to look at other aspects of background and identity that may be more fundamental and ingrained than political party support – an active choice that can (theoretically at least) be changed at a whim.

At the top of this article, I mentioned some commonly-held assumptions about what the most important drivers of negative sentiment in this area might be – education and income levels, whether people have lived outside of Northern Ireland and so on. These dimensions are all captured in the NILT data, so one approach would be to examine them one by one to see how attitudes differ as they are varied.

However a more systematic approach is to take the full set of data and apply a statistical test to ask which variables, out of all of those captured, are most predictive of differences in the outcomes we are interested in.

To do this, I picked the single outcome that felt most relevant to me given the recent racist riots in Belfast – the willingness to accept Muslims as neighbours. I then applied a chi-squared test to every background or identity variable in the NILT data, and examined a statistic called Cramérs V to see which were most predictive of variance in this outcome. The following variables come out on top:

This table shows the significance (p-value) and strength (Cramér's V) of association between each background variable and the attitude in question. The p-values are infinitesimally small for the first five variables, indicating practical certainty of a significant relationship for these factors. The Cramér's V can be interpreted as the extent to which the variables are associated, as a percentage of their overall variance. This was used to rank the variables.

As you can see, political party support is the most highly associated with acceptance of Muslims as neighbours. However, the next three highest factors were not level of education or income, but three closely related variables relating to national or community identity in the Northern Irish context.

The highest educational qualification held by respondents does appear next, but only in fifth place, with other variables such as income level and whether the respondent had ever lived outside of Northern Ireland coming further behind.

National identity is the most important driver

Of the three identity factors in the top 5, it is intriguing to see that Irish or British national identity is the most predictive, albeit marginally. This question is also unique in that respondents are asked to place their identity on a spectrum, from Irish only to British only, that includes mixed identities. For these two reasons, I chose to drill into this variable for further analysis.

As for the outcomes or attitudes being examined, here I also needed to focus by narrowing down the set of survey questions to a manageable number. There are broadly three categories of attitude that I was interested in:

1. attitudes to race or ethnicity itself, relating to people already living in Northern Ireland
2. attitudes to immigration and asylum, relating to those seeking to come to Northern Ireland
3. attitudes to mixing amongst the traditional binary of broadly Protestant and Catholic communities

For each of these attitudinal categories, I picked a representative question, based on its availability (whether it was asked every year or not across the full period of survey data), and how representative its responses were of those to similar questions in its category. The average responses to these questions, by national identity group, are shown in the chart below:

Again, a positive take on these charts is that every column is over 50%, and most are well above. The interest lies in the differences. Whilst only 5% of respondents identifying as Irish only would not willingly accept a Muslim neighbour, this rises to over four times that figure – 22% – amongst those who identify as British only. Those identifying with a mix of Irish and British identities are in the middle, with higher acceptance as the degree of Irishness increases.

Attitudes to refugees follow a similar pattern. As with the first chart, the biggest drop comes between those identifying as "more British than Irish" and "British not Irish". Here support for the idea that we have a duty to protect refugees falls steeeply from 68% to just 51%. All other identities have support in the 70s.

As observed in the political party analysis, the third question has a slightly different pattern. Perhaps unsurprisingly, people who identify as purely Irish or British have the lowest desire to live in mixed community neighbourhoods, whereas those identifying as equally Irish and British have the highest.

Examining differences within national identity groups

With a clear association between national identity and our selected attitudes established, the next thing we can do is examine differences within each identity group. We can do this by cross-tabulating national identity with a second variable, enabling us to test whether commonly-held assumptions about attitudes are supported by evidence, and if so how strong their associations are, compared to national identity.

Education level

The first of these assumptions is that people with less education have more negative attitudes to ethnic minorities, immigration and community mixing. Looking at the NILT data, roughly 50% of the respondents had a degree. I therefore created a binary variable which aggregated all other levels of qualification, to give a cut on has degree / no degree. The results are below:

What is striking here is not that education level affects attitudes (which it does), but that degree-educated people identifying as British have significantly more negative attitudes than people without a degree who identify as Irish.

Income

The next assumption we can test is that those on lower incomes have more negative attitudes. The NILT's SRINC question asks respondents to self-categorise into three income brackets: high, middle or low. Given that very few respondents choose "high", I decided to create a binary variable here too, with "low income" and "medium or high income" as the two levels. The corresponding cross-tabulated results are below:

Again, a similar pattern can be observed here. Low income is associated with more negative attitudes, but this is a less powerful driver than national identity. As with education, people self-reporting a medium or high income who identify as British are significantly more likely to hold negative attitudes than those reporting a low income who identify as Irish. These two groups are only close when it comes to their preferences for living in a mixed community neighbourhood, where the difference is too small to be significant.

Time spent living outside Northern Ireland

Another factor which can change people's attitudes is time spent living in other countries. Here, the relevant question asked by NILT is whether respondents have lived outside of NI for more than six months, with just over a third (34%) responding yes. The cross-tabulation of this factor with national identity is below:

A familiar pattern is emerging. Time spent living abroad does have a significant association with increased postitive attitudes, but again, this effect is less strong than national identity: whilst those who identify as British who have lived abroad for over six months are much more likely to have positive attitudes than those with British identities who have stayed in Northern Ireland, they are still more negative than almost every other group on the British–Irish spectrum, irrespective of whether they have lived outside of NI.

Age

The NILT survey asks respondents their exact age and provides a variable that aggregates their responses into age bands. Here I aggregate the data further, to give a binary variable of under 45 / 45 or over.

Before analysing this cut, I expected that younger respondents would have more progressive, positive attitudes on average. The results are below:

These results surprised me. Whilst under-45s do have slightly more positive attitudes towards acceptance of Muslim neighbours, they are significantly less likely to agree that we have a duty to protect refugees, or to state a preference for living in a mixed neighbourhood. Further research into the reasons behind this could be instructive.

Church attendance

The last variable I examined relates to a feature of Northern Irish society that retains more influence here than in the rest of the UK – church attendance. A large majority of the Northern Irish population identify as Christian (79.7% as of the 2021 census), and many attend church regularly. Given that several Christian teachings (as well as those of other religions) concern how to treat foreigners, such as this instruction in Leviticus...

When a foreigner resides among you in your land, do not mistreat them. The foreigner residing among you must be treated as your native-born. Love them as yourself (Leviticus 19:33-34)

...I was interested to see whether there would be significant variance in the attitudes under analysis here, based on whether and how often respondents attended church.

To analyse this, I built a three-way variable. Non-religious respondents were not asked whether they attended church, so they have their own category. For those reporting a religion, I split the data into regular attenders (once a month or more), and those who identified as religious but attended services less frequently than once a month. The respective attitudes of these groups are displayed below:

It is striking that the group with the most positive attitudes across most questions is consistently those with no religion, indepedent of national identity. Amongst those identifying as religious, regular church attendance does seem to be associated with slightly higher acceptance of Muslim neighbours and significantly higher agreement of a duty to protect refugees. Regular church-goers who identify as British are, however, the group least likely to prefer living in a mixed community neighbourhood.

Further analysis of the differences in social attitudes across these three groups could provide some fascinating insights – watch out for that analysis in an upcoming post.

Conclusions, concerns and a question

The Northern Ireland Life and Times survey is an invaluable resource for understanding social attitudes in Northern Ireland, particularly given the lack of other reliable polling data.

Here I have used its data to examine some assumptions about attitudes to minority ethnic groups, immigrants and community mixing, and found some striking evidence that one factor, above all the other usual suspects, is the strongest predictor of negative attitudes: national identity.

Put bluntly, those identifying as British have significantly more negative attitudes on these topics than their Irish counterparts.

As someone who identifies with a mixture of British and Irish identity, but is probably more British than Irish, this is an uncomfortable finding, and it may be very uncomfortable for the dwindling tribe of progressive unionists who feel proud that the UK has historically welcomed foreigners to its shores. But it is undeniable.

Nevertheless, there is a positive take here too: putting aside their relative differences, a majority across all identities, demographic categories and political parties do hold positive attitudes, despite what the loudest voices on social media may suggest.

What can be done to ensure that this remains the case is a question for us all.

Sample sizes

With contentious topics such as those covered in this article, it is critical that analysis is based on robust evidence. The table below shows the sample sizes used to calculate each dot or column in this article. The mean weight of survey responses used was N=763. The lowest was N=143, with 90% of all calculated outputs being based on at least 273 weighted survey responses.

As always, all the code and data behind this article are available on github. If you have any questions, please leave a comment or get in touch at jack@believethedata.org.

Finally, if you've enjoyed reading, please consider subscribing via the button at the bottom, or following the blog for updates on Bluesky.