What should we make of the latest Irish Unity poll?

First up, a confession. Strictly speaking this was not an Irish Unity poll, it was conducted on behalf of the European Movement Ireland as one of an annual series on Irish attitudes towards the EU and the policy choices it faces. But it contained a question on voting intentions in a referendum on Irish Unity. You can download the full report from https://www.europeanmovement.ie/em-ireland-eu-poll/

Second confession. Brian asked me if I would be interested in writing about it, and at first I declined. There are others, such as Philip McGuinness. who are better qualified on this subject. However I became intrigued by the results.

I will confine myself to the simple question of how much attention we should pay to these results.

The same question was asked to a sample of 1,200 people in NI and the same number in ROI.

The results were unexpected. The Northern Ireland sample responded with a whopping majority for a United Ireland.

If the poll were reported in the normal way, with the Don’t Knows excluded, a United Ireland would win by 68% to 32%.

While in the south they answered:

So again a United Ireland would win (73% to 27% excluding Don’t Knows).

The response in NI is very different from other polls on the issue, which tend to show a plurality for remaining in the UK. The response in the south for a United Ireland also looks less enthusiastic than I would have expected. Theoretically this could be because the question is very different from those normally posed. Other polls ask something like “If there were a border poll tomorrow how would you vote?” This one asks how you would vote on a “united Ireland in the EU”.

You might rightly say that a UI would of course be in the EU and that should therefore make no difference to people’s answers. But in fact wording questions differently can make a significant difference to the result obtained. None the less I would find it very surprising if it made that much difference in this case. (Although just because I am surprised does not mean that it is not so.)

The report states that the margin of error is 2.2%, meaning that in real life each figure given could be higher or lower by that amount.

Anthony Wells, one of the Uk’s foremost guru’s on opinion polling laid it down as a law of polling to remember that the margin of error statement is always followed by the words “at 95% confidence level”. This means that no matter how well conducted, on average one poll in 20 will be an outlier, or more prosaically a rogue. He repeatedly said that if you got a poll with an exciting result then you should wait to see if the result is repeated in subsequent polls before drawing any conclusions.

The report’s methodology statement does not give any information about balancing (that is not unusual). It does break down the results by age and sex, but we do not know if the sample was balanced for religious/community background or previous voting patterns.

If we are to believe the poll at least a third, and maybe up to a half, of people who have consistently voted unionist would be prepared to consider voting for a United Ireland tomorrow if membership of the EU was also on offer.

Also, that the support for a United Ireland under this condition amongst the over 55’s in NI (right hand column) is not far short of those under 34. This does not correspond with what we would expect from previous polling or from the demographic make up of these age cohorts.

Also the numbers of Don’t Knows and No’s to a UI in the Republic is so high that a really well run No campaign would have some hope of defeating the proposition.

In short, my advice is to note the poll as interesting – but neither smile nor frown (depending on your constitutional preference) unless you see the same result repeated in subsequent polling.