Last month polls indicating that 63% of people in Northern Ireland would vote for “a United Ireland in the EU” made the headlines.

This suggests a much higher level of support for a United Ireland (UI) than many other surveys of local people which, broadly speaking, tend to see a plurality opposing a UI. Such outlier findings became a major topic of conversation in local media as well as a committee at the Dáil.

European Movement Ireland (EMI) commissioned the poll and published its findings. EMI is an EU-focused organisation and told FactCheckNI that its primary focus is on the EU and not the constitutional question (on which they have no particular view) and that these results “should [be] interpreted as indicative of general attitudes rather than as a precise measure of voting behaviour under defined conditions.”

This article examines the results from the specific question in the poll on a UI and EU membership, provides some context about how these findings should be understood, and also looks at polling more generally. Our analysis includes how:

  • The framing of a question can affect people’s responses.
  • When polling questions are not extremely clearly worded – or are unusually worded – this increases the risk of misinterpretations, a risk acknowledged by EMI.
  • There is no such thing as perfect polling, and that there may even be a risk that the EU focus of this question meant a loss of interest of some people that may have affected results.

For more information, read on.

  • The poll

Last month European Movement Ireland (EMI) published the findings of its Island of Ireland EU Poll 2026, a survey asking various questions relating to the EU as well as some wider political context.

Responses were taken on both sides of the border and many of the findings were broken down on a Northern Ireland/Republic of Ireland basis.

However, in NI at least, one question in particular made headlines. The Northern Ireland results are on the right hand side:

Figure 1 – source: EMI polling

  • In the news

The Irish News article’s headline said: “Poll finds more than 60% would vote for united Ireland”. Posting on X, the newspaper added in more context: “A majority of people on the island of Ireland would vote for Irish unity as part of the EU, according to a poll”

The Belfast Telegraph ran with the headline: “Majority on the island in favour of Irish unity as part of EU – poll”

The Journal reported the findings as “All Ireland poll shows cross-border support for Irish unity and EU membership”

  • Framing

With 63% of respondents in Northern Ireland saying they would vote for “a United Ireland in the EU”, this poll caught the eye.

Broadly speaking, surveys about voting intention with regards to a United Ireland show the public here to be in favour of the status quo. Ulster University’s CAIN service has summary results of various such polls throughout the 2020s, while a FactCheckNI analysis from 2023 goes a little further back. Most of those results will not have included a reference to the EU in the poll question.

EMI acknowledged this apparent variance from the norm. CEO David Geary was quoted in the three news articles mentioned above as saying the “higher support than other unity polling [is] likely due to the EU framing.”

This same word – “framing” – was used when representatives from EMI spoke to the Dáil’s Joint Committee on European Union Affairs on 13 May. 

In response to a question from Independent TD Paul Gogarty asking whether EMI has “any views on the discrepancy between the height of that poll versus other standard polls on a united Ireland”, Deputy CEO for Policy and Research Lorna Hayes said:

“I think it is very much linked to the framing of the question … We have to ask these questions in the context of the EU so we are not just asking about a united Ireland, we are asking if people would be in favour of a united Ireland in the EU. That framing really matters. With the other questions in relation to Northern Ireland, we can see that people may assess their response differently when they are thinking about a united Ireland in the EU. Part of the higher response may be that they could see it as a vehicle to return to the EU, for some people anyway.”

  • Closer look at the question being asked

Framing, in this case, just means the wording of the question.

Most polls might ask something like: If a referendum were held tomorrow on a United Ireland, how would you vote?

The EMI poll asked: If a referendum were held tomorrow on a United Ireland in the EU, how would you vote?

The simplest way to read this question is that it asks respondents whether they would vote for a United Ireland if that UI was going to end up as part of the EU.

However, there are some risks of confusion here. Firstly, based on pre-existing/current arrangements, a United Ireland would be a member of the EU automatically. This extra detail in the question – “in the EU” – is a redundancy.

The less clear a question is, the more open it is to misinterpretation. FactCheckNI saw more than one person on social media suggesting that this question is actually asking whether people believe that if there were to be a United Ireland it should be in the EU.

Of course, we should all be cautious about taking too much meaning from individual comments online. But, at the same time, this was an online poll. It quite literally involves random people on the internet reading a question and then answering it on the basis of what they think it means.

We asked EMI about possible misinterpretations of this question, and put to them that plausible readings – or, if you prefer, plausible misreadings – including the ones noted above as well as the idea that this could be a question about a joint UI/EU referendum.

  • EMI response

EMI responded by stating that they are a non-partisan and independent organisation that takes no view on the constitutional future of the island of Ireland. Their stated purpose is to encourage debate and track public sentiment on EU issues. 

They also noted that this polling was conducted to high standards (we’ll have a quick look at polling methodologies later) and that EMI’s name did not appear on the poll questionnaire. They said further (any bolding is ours, not theirs):

In relation to the wording of the question on ‘a referendum on a united Ireland in the EU’, we recognise that, as with many questions on complex constitutional issues, respondents may interpret phrasing in different ways, particularly across different jurisdictions. 

The inclusion of ‘in the EU’ reflects both our organisational remit and the current policy context. The European Council indicated in 2017 that, in the event of Irish unification, a united Ireland would be part of the European Union. While most polls on unity do not specifically mention the EU, we include this dimension to better understand how EU sentiment interacts with views on constitutional change. This is in line with our focus on EU affairs and our work examining how EU-related factors may shape perspectives on constitutional issues in Northern Ireland. Other polling, such as the ARINS project, has also tested sentiment on how membership of NATO or the Commonwealth influences positions on unity.    

The bottom line is that our poll does not track attitudes toward constitutional change but rather attitudes toward EU issues. Given that a united Ireland would automatically be part of the EU, we believe it is important for us as an EU organisation, to monitor sentiment in this area.  Unlike other polling, our question is not rooted in legacy and reflects the reality that a united Ireland will be in the EU. The results should therefore be interpreted as indicative of general attitudes rather than as a precise measure of voting behaviour under defined conditions. 

We recognise the complexity and nuanced views in Northern Ireland on the EU & unity, particularly post Brexit.  As such, by including the EU dimension, we expect our question to attract responses from a range of perspectives, which respondents are free to hold and express.  We also acknowledge that the level of support recorded in Northern Ireland is higher than in other polls. One possible contributing factor is the consistently strong pro-EU sentiment identified in our research.  It may also reflect that some respondents who would not traditionally favour unification view it differently when considered in an EU context. However, our poll also shows that a larger proportion of respondents would vote to rejoin the EU through the UK.  

  • Polling and methodology

The EMI polling was conducted online by Amarach Research between 26 and 31 March this year, using “nationally representative” samples of 1,200 respondents on both sides of the border.

All participants were aged 18+ and results were weighted in both jurisdictions to reflect the demographics of each respectively. The findings’ margin of error was +/- 2.2% at a 95% confidence interval. This is all standard stuff and describes good practice in opinion polling (see here, here and here).

Different polling organisations use differing methodologies. NILTS conducts telephone interviews and surveys, whereas the organisation LucidTalk uses an online opinion panel. The wording of questions varies, and the timing of polls can be impacted by external social, economic and political events. Individual polls are a snapshot of a moment in time and may be out of date within a few weeks. 

Even the order of questions within a poll, although that might seem innocuous, can affect outcomes. Exactly the same question asked within two differing bundles of questions (or the same questions in a different order) in otherwise similar circumstances can see different results in both (see here and here).

There is no such thing as a perfect poll. If nothing else, all polls are weighted towards people who are willing to participate. In this case, there is a further possible wrinkle. In her submission to the Dáil, EMI Deputy CEO Lorna Hayes also said:

“We have worked really carefully with Amárach and its partners in the North to look at the samples. We are really happy. We felt the religious community breakdown is consistent with other samples. We do blind polling in Northern Ireland so perhaps there is an element of people seeing that it is about the EU and choosing not to participate because they are not as interested. That could potentially influence the findings a little.”

Could this be important? Maybe, maybe not – and it’s impossible to know.

  • Analysis

One of the main points of discussion around this poll is that the results are quite different from the broader pattern of surveys concerning how NI would vote in a unity referendum.

Alongside the debate around the result, it has to be said plainly that public sentiment can change. Just because previous polls indicated one thing does not mean that this cannot be different now or in future.

With regards to the EMI question on Irish unity and the EU, the main potential issue explored in this article is about possible misinterpretations.

Any question can be misinterpreted. However, questions with more straightforward wording and with fewer variables are less likely to be misunderstood/read in different ways by different respondents. The first question in this EMI survey was:

Should Ireland remain a member of the EU?

That is extremely clear cut.

A more straightforward question about a United Ireland might read: If a referendum were held tomorrow on a United Ireland, how would you vote?

But given EMI’s “framing” around the EU, and its organisational purpose to highlight and explore EU-related issues, it is understandable that they are keen to continually reinforce this in their questions. Perhaps a better way to phrase their unity question would have been:

Based on pre-existing agreements, if a United Ireland was established tomorrow it would be part of the EU. With that in mind, if a referendum were held tomorrow on a United Ireland, how would you vote?

This clearly delineates the premise from the question and, as a result, should be less open to misreading. But it is incredibly wordy!

  • Conclusions

Concerns about the misreading of questions become concerns about interpreting the findings of a poll. When misinterpretations are less likely, findings should be more solid.

That does not mean that this poll question (which is just one of almost 20 questions in the entire survey) can or should be discounted. However, all this context needs to be taken into account – including the fact that EMI’s central focus is on EU issues not constitutional change and, according to the organisation, these results “should [be] interpreted as indicative of general attitudes rather than as a precise measure of voting behaviour under defined conditions.”