The one thing upon which pollsters usually agree is that their work is correct at the time of asking the question. Beyond that, everything else is subject to dispute.
Last Tuesday, the Belfast Telegraph splashed with a poll showing just 29% of people in Northern Ireland would vote for Irish unity. (It was part of a much bigger study of political attitudes, led by the University of Liverpool).
Naturally enough, given unification is now one of the central issues in Irish politics - and most other polls on the subject show support is significantly higher - it was a bold claim to make.
Yesterday, LucidTalk asked a similar question and found a completely different answer.
In a poll for The Detail, voters in Northern Ireland were asked: ‘If there was a border poll how would voters in Northern Ireland vote?’ 46.8% came back for the constitutional status quo and 45.4% for a united Ireland.
How do we explain such a vast difference?
Well, for starters, the framing of a polling question is all important. In the BelTel/ Liverpool University poll last week respondents were asked: ‘If there was a border poll tomorrow, how would you vote?’
Unless I’ve missed something, there isn’t going to be a border poll tomorrow. Neither is anyone of whom that I am aware calling for one in the immediate short term either. So, if you are offering people a hypothetical scenario that isn’t attached to a specific timeframe or context, they are won’t to provide you with, well, a hypothetical answer.
That said, a poll last September conducted by Lord Ashcroft asked the identical question: ‘If there were a “border poll” tomorrow, how would you vote?’
The contrast could not be starker. 46% wanted to see Irish unity then, with 45% backing the status quo. Given it’s the same question, has there been a massive reversal of fortune for United Irelanders in the past four months?
Not if the Irish general election result is any guide.
Something else must explain the variance. The sample size of the two polls was similar (2,000 for the poll in the BelTel, 1500 for Ashcroft). Yet, while last week’s poll was conducted by face-to-face interview, Ashcroft’s was based on an online sample. (As indeed was yesterday’s LucidTalk Poll - sample: 1,896 respondents).
Of course, there is a healthy debate among the stat-wonks of the polling industry about the accuracy of the various methods of collecting data. Do face-to-face surveys elicit more truthful responses than telephone or internet samples? Or do they disproportionately survey people who are available in the daytime and not in work, such as the elderly?
Given we already know that older people in Northern Ireland are significantly more likely to back the constitutional status quo - and given there is often a reluctance from respondents to voice their real opinions in face-to-face polls - might the result of the BelTel/ Liverpool University poll be skewed in that way?
In evidence to a House of Lords inquiry into political polling back in 2018, YouGov, perhaps Britain’s preeminent polling company, argued that online was the most accurate method:
‘Using quota sampling from a panel of volunteers who we already hold extensive demographic data…[allowing]for more detailed quotas to be set on who was interviewed, ensuring greater representativeness on more variables. Online interviewing also reduces or removes the interviewer effect (that is, people being embarrassed to give answers seen as socially undesirable to a live interviewer)…’
For argument’s sake, though, let’s take the BelTel’s 29% figure at face value. What’s so unusual is that it’s not much higher than the share of the vote Sinn Fein routinely receives in actual elections. The party has averaged 25% of the vote in the last three major elections in Northern Ireland, (two Westminster and one assembly).
Are we to believe that hardly anyone else in Northern Ireland – either in the SDLP, Greens, People Before Profit or Alliance - would support Irish unity, given the chance to make up the difference? Basic intuition suggests that isn’t the case.
Indeed, as the Belfast Telegraph reported last week: ‘…92% of Sinn Fein and 81% of SDLP voters say[ing]they would support Irish unity in a border poll.’
The other point worth making is that 29% is similar to the level of support Scottish independence had a year out from the 2014 referendum. Of course, that figure quickly grew, ending up at 45% - just short of a majority. What happened is that once the concept of independence became a realistic possibility, support among Scots increased dramatically. Rather than facing decades in the foothills, the growing likelihood of a border poll in the next few years may see support grow as quickly in Northern Ireland as it did in Scotland.
Three other variables that will impact on future levels of support for a border poll are worth bearing in mind.
Firstly, the demographic profile. The issue of Irish unity isn’t going to decline in relevance. The Ashcroft pollshows that a majority in all age groups under the age of 65 now supports Irish unity. (Precisely the sort of people who are not likely to be answering face-to-face surveys on a wet Wednesday in Belfast City Centre). In that respect, the future belongs to United Irelanders.
Secondly, any border poll also includes the supplementary issue about whether people in Northern Ireland want to become part of the European Union again. This is now a powerful secondary motive for many people who may not be primarily motivated by Irish unity, per se.
This explains why polls on this subject that include this context tend to show much stronger support. Indeed, the LucidTalk poll from yesterday shows 47.9% of respondents would support Irish unity ‘as a pathway back to membership of the EU for Northern Ireland.’
Thirdly, we haven’t yet internalised the effects of Brexit. With no prospect of an easy ride in subsequent negotiations about a future trade deal, (certainly if Michel Barnier and the French Foreign Minister Jean-Yves Le Drian are to be believed), support may grow among the ‘persuadables’ – pragmatic unionists in the main who might opt to secure the family farm or engineering business by throwing their lot in with the rest of Ireland and thereby retaining farm subsidy payments and unimpeded single market access.
All considered, the weight of existing polling evidence suggests last week’s BelTel/Liverpool University poll is an outlier. The fieldwork was carried out by a Belfast company called Social Market Research, however they don’t appear to have published their data - which is the usual form for a polling company. Given the contentiousness of its headline finding, this is somewhat remiss.
Criticisms aside, we should welcome as much serious inquiry as possible, given it’s a sign that the concept of Irish reunification has moved mainstream. There will be disagreements about individual polling figures and methodologies – there always are – and about timelines and demands for a border poll, but the debate is truly alive and kicking.
Kevin Meagher is the author of ‘A United Ireland: Why unification is inevitable and how it will come about’
1 Comment
While polls should be treated with scepticism I find the arguments here a little weak.
1. The wording of the Ashcroft /Tel. poll is directly related to voting intentions at this moment is time. To dismiss it as “hypothetical” is meaningless.
2. To conflate southern support for SF with northern attitudes to united ireland is at best disingenuous. In fact the retrenchment of support for United ireland could be explained as a consequence of southern vote: moderate unionists and nationalists would not vote yes in a situation where SF were in government in the south.
While I agree that the poll is suspect, so too is quite a lot of this analysis.