The blog below is by my long time collaborator Faha
As always, it is forensic and based upon the known facts. To be honest the speed at which things are changing here is almost overwhelming. I wish the peerless Horseman was still with us to witness how this endgame is playing out.
The title of the blog is mine inspired by Christy Moore and Faha’s final line. I’ll not spoil it for you.
Over to Faha. BD
An opinion poll was conducted from August 24th through August 28th by Deltapoll. Deltapoll was commissioned by 2 groups, Best for Britain and Our Future Our Choice, and the poll was an online poll of voters in Scotland and Northern Ireland.
LucidTalk in Northern Ireland did the actual polling for Deltapoll in Northern Ireland. The final unweighted sample was 1,199 voters which was weighted to 993 voters to adjust for over representation and under representation of certain demographics and the turnout changes in recent elections.
Voters were asked this question:
Do you consider yourselves to come from a Nationalist or Unionist Heritage?
The results were:
Unionist 47.3%
Nationalist 42.0%
Neither 9.3%
Other 1.4%
This sample appears to approximate the electorate in the 2017 Assembly and Westminster elections. The vote for unionist candidates in the Assembly election was 3.7% higher than that for nationalist candidates and in the Westminster election the vote for unionist candidates was 7.2% higher than that for nationalist candidates. The average for the 2 elections is 5.0 % which is very close to the 5.3% difference in this sample.
Voters were asked how they voted in the 2017 Westminster election and 2016 EU Referendum. They were also asked how they would vote if a new EU Referendum was held. 1% of those who voted Remain would switch to Leave and 7% of those who voted Leave would switch to Remain. This is only a small change and would increase the Remain vote from 56% in 2016 to 58.5%. However, those who did not vote in the 2016 election would vote 83% Remain and only 9% Leave if a new vote was held. This is significant and I will return to this point later.
Voters were also asked how they would vote on a referendum on a United Ireland under 3 different scenarios.
The first scenario is that the UK somehow remains in the EU and the results were:
Northern Ireland Remain in the UK
52%
United Ireland
35%
Undecided
11%
With the UK remaining in the EU support for a United Ireland is only 35%.
The 2nd scenario asked voters how they would vote if the UK leaves the EU and the results were:
Northern Ireland Remain in the UK
39.9%
United Ireland
52.9%
Undecided
7.1%
The 3rd scenario asked voters how they would vote if the UK leaves the EU and there is a hard border between the Republic of Ireland and Northern Ireland and the results were:
Northern Ireland Remain in the UK
39.8%
United Ireland
56.4%
Undecided
3.6%
With a hard border there is no change in the percentage that wish to remain in the UK but half the undecideds would switch to voting in favour of a United Ireland.
The breakdown by Constitutional Position (heritage background) is:
Nationalist
Northern Ireland Remain in the UK
1.9%
United Ireland
96.4%
Undecided
1.7%
Unionist
Northern Ireland Remain in the UK
77.1%
United Ireland
16.9%
Undecided
6.0%
Neither/Other
Northern Ireland Remain in the UK
22.2%
United Ireland
75.9%
Undecided
1.9%
In the Lord Ashcroft and Lucid Talk polls from May 2018 support for a United Ireland among nationalists was in the 85% to 90% range with many undecided and support for remaining in the UK in the 4% to 7% range. The undecided have uniformly shifted in favour of a United Ireland. What little support there was among nationalist voters to remain in the UK has almost entirely disappeared.
There has also been a marked shift among those who are neither unionist nor nationalist.
Support for a United Ireland was in the 36% to 50% range in the May polls and is now 75%. Even excluding those who are Other (foreign nationals) from this group support for a United Ireland is over 70%.
There has also been a significant increase in support for a United Ireland among those from a unionist background which is at 17% in this poll.
This is double that in the May polls. However, I wish to point out that there are 2 very different subgroups within the unionist population. This poll shows that unionists (large U) who vote for unionist parties are only 2% in favour of a United Ireland. This group is ¾ of all unionists. Small u unionists (those who vote Alliance, Green, PBP, SDLP, etc.) or who were nonvoters have a majority that are in favour of a United Ireland. This is an important distinction that has developed.
Those who are nationalists will vote overwhelmingly for a United Ireland. They will vote this way whether they are nationalists who vote for nationalist parties or vote for nonsectarian parties. Those who are neither nationalist nor unionist are also now strongly in favour of a United Ireland. The unionist population electorate differs from the nationalist electorate in one significant way. Nationalists who vote for nationalist parties are only 2% in favour of remaining in the UK. Unionists who vote for unionist parties are only 2% in favour of a United Ireland.
However, unlike their nationalist counterparts, small u unionists who do not vote for unionist parties are who are nonvoters now have a majority that are in favour of a United Ireland. Of the overall 3.6% of all voters who are still undecided only 0.1% of that are voters who voted for a unionist party. The other 3.5% are voters who voted for nonsectarian or nationalist parties or were nonvoters.
Now some may claim that small u unionists who have not voted in recent elections do not matter if they do not vote. Turnout for the EU referendum, Assembly and Westminster elections was 800,000. The turnout for a Border Poll will be much higher. The GFA referendum has a turnout of 80% of the voting age population. The Scotland independence referendum had a turnout of 85% of the voting age population. Similar turnouts for a Border Poll would have between 1,200,000 and 1,300,000 voters. That is an extra 400,000 to 500,000 voters.
The extra nationalist community voters will vote overwhelmingly for a United Ireland. Foreign national voters will vote overwhelmingly for a United Ireland. The extra voters from neither community will vote 75% in favour of a United Ireland. The extra unionist voters will have significant minority in favour of a United Ireland.
This poll also noted a strong correlation between those who wish to remain in the EU and voting for a United Ireland (88% yes vs 7% no). There was also a strong correlation between those who wish to leave the EU and voting to remain in the UK (87% vs 12%). In the 2016 EU referendum constituencies such as North Down, Lagan Valley, East Antrim, Strangford, South Antrim and East Belfast had votes in favour of remaining in the EU that were much higher than the non unionist vote in Westminster and Assembly elections.
This is an indication that there were many small u unionists (as well as some nationalist and neither voters) who voted in the EU referendum that usually do not vote. Indeed North Down had the highest turnout of all constituencies in the EU referendum while it usually has the lowest turnout in other elections.
This is only one poll. There will probably be more in November and December when the final nature of Brexit will be known.
If subsequent polls confirm that there is a majority of Northern Ireland voters in favour of a United Ireland then a Border Poll should be called for next April or May.
"We in coming days may be" 
Comparing the crosstabs of Q2 on heritage to Q6 on border poll post-Brexit illuminate the depths of the hole the DUP has dug for Unionis. Antrim is 73% Unionist, 17% Nationalist and 10% Neither and for Q6 is Union 51%, UI 37%, DK 12%. Antrim. The age demos of Q6 show under45s are overwhelmingly UI.
The writing is clearly on the wall, Brexit is both spark and catalyst. Even if Article 50 is withdrawn, or a deal is done to keep the border as is, NI has a decade at most. As these polls are making clear, a no deal Brexit is doom for NI. So good to see Unionists continue to dismiss them as outliers and point to the “authoritative NILT”.
LikeLike
Business interests in North Down did a lot to cause the creation of NI. Business interests in North Down may be the factor that causes its demise.
LikeLiked by 1 person
thanks for this excellent analysis Faha – and its very encouraging conclusions. So many are hoping you’ve called it right!
LikeLike
It’s worth looking at the demographics as well. The strongly pro-union vote is concentrated in the older half of the population.
Many nationalist bloggers don’t seem to realise that the rate of death can be estimated with considerable precision.
To do this you need a set of life tables. Here’s an abbreviated set of life tables for Scotland for illustration.
The table lets you track what percentage of a given age cohort will have died off by the time it’s reached the next age cohort.
If you know the composition of the population who vote DUP, you can forecast what percentage of them will have died within 5 years, 10 years, and 15 years.
For example, from looking at the life tables for NI, we see that the annual mortality rate for 45 year old females is around 0.15%. So if there were 1000 45 year old females who voted DUP in the last election, within 5 years since around 0.7% will have died off.
A similar calculation gives us that amongst 50 year old males who voted DUP in the last Assembly Election, around 1% will have died off within 5 years.
Since the typical DUP voter tends to be over 40, demographic attrition does enough damage to the DUP’s voter base to be noticeable every 5 years or so.
In other words, the DUP will inevitably lose around 0.5% to 1% of their voters every five years, not counting younger DUP voters growing up.
The life tables tell us that if you’re a 45 year old male, you’ve around a 50% chance of dying within 45 years.
If you’re a 55 year old male, you’ve around a 50% chance of dying within 36 years.
By the time of the next census, the DUP’s electorate will be in demographic freefall. In fact, the process started around 2010.
Of course it would be preferable to win a border poll earlier, but it’s still worth noting that Unionism will be weakened severely by 2030. This is on top of it already losing it’s majority in AE17.
(This isn’t perfect since we don’t know what percentage of the younger generation will vote DUP when they are old enough to vote, but it has merit IMHO.)
LikeLike
There is also a marked gender difference in a Border Poll with a hard Brexit. For woman 67% for United Ireland and 30% stay in the UK. For men it is 45% for a United Ireland and 47% stay in the UK. This is partly explained by the difference in identity in the sample. For women 42% were unionist, 44% nationalist and 10% Neither. For men it was 52% unionist, 39% nationalist and 8% Neither. After adjusting for the underlying preference by identity it appears that of the 17% of unionists that would vote for a United Ireland over 30% of unionist women would vote for a United Ireland but only 5% of unionist men.
LikeLike
In most opinion polls you have two levels of sampling variation:
First, the distribution of various groups within the sample (Male vs Female; Rural vs Urban; Newry vs Larne).
Second the sampling error associated with measuring the support level for each individual group.
If this is weird, it might be easier to think of it this way:
If you do an opinion poll by polling people at random, the overall support level reported is a weighted sum of the estimated support levels in each subgroup, weighted by the number of people from each group who appeared in the sample.
So the overall Margin of Error has to account for the fact the weights shift around randomly from poll, and not just the reported support levels.
LikeLike
The stats on the female vote are notable. Especially given the track record of inherent caution/conservatism by female voters on most major political questions in Europe and North America over the last several decades. Though, am I right that – anecdotally at least – anti-treaty SF in 1922-23 was believed to have a stronger female vote than pro-treaty SF, Labour and others?
LikeLike
The EU has strongly favoured womens rights/female empowerment.
The PUL ladies (or at least 30 % of them) are smart enough to know where their interests are best served.
Those organising the UI referendum campaign, would be well advised to focus on that section of unionist voters, as much if not more than the farmers and business owners.
LikeLike
Great article, excellent on the detail. I would like to see a “negative poll” analysed to the same extent and then compared and contrasted.
No one has provided a satisfactory answer yet as to why the results vary so wildly; I have heard the shy voter and don’t knows answer but would love to read more…
LikeLike
Bill White of LucidTalk who conducted the poll here, has said that the relative anonymity of their polling system has improved accuracy versus the Life and Times system of using face to face polling, for example. The accuracy of the results tends to bear that out
LikeLiked by 1 person
PBP is not a small u unionist party and Im not a supporter…
LikeLike
I would say “caveat emptor” to a certain degree, as the weighted response had almost twice as many Alliance respondents as DUP voters, and the Remain:Leave breakdown was 3:1. I don’t doubt LucidTalk’s professionalism in relation to weighting, but Newton Emerson may well have a point in stating the results are open to question.
LikeLike
That should, of course, read “unweighted response”.
LikeLike
Polyglot Paul,
You and Newton Emerson are misinterpreting how an unweighted sample and a weighted sample work in a poll. The unweighted sample is adjusted to a weighted sample to reflect the actual recent election results in Northern Ireland and the demographics based on the census. The actual responses to the questions are not altered in any manner. The DUP numbers in the unweighted sample were too low so each DUP voter was multiplied by 1.5 to give an accurate percentage. The Alliance numbers in the unweighted sample were far to high so the Alliance percentage was multiplied by 0.25 to lower that percentage. The voters were asked who they voted for in the 2017 Westminster election and of those who actually voted the weighted sample was
DUP 33%
UUP 12%
Other Unionist 4%
Total Unionist 49%
The unionist parties actually did receive 49% of the vote in 2017 so the weighted sample exactly matches the total unionist vote though the DUP is a little low and the UUP a little high.
The weighted sample for nationalist parties in the 2017 Westminster election was
SF 28%
SDLP 12%
Total nationalist 40%
The total nationalist vote was 40% in the election though the sample has SF a little low and the SDLP a little high.
There is nothing unusual in the responses given by voters who voted for unionist or nationalist parties. When voters were asked how they would vote in a Border Poll with a hard Brexit 99% of DUP voters said they would vote to stay in the UK. 100% of SF voters said they would vote for a United Ireland. Neither of those preferences by party are unexpected in any way. 83% of SDLP voters would vote for a United Ireland and 88% of UUP voters would vote to stay in the UK. Nothing unusual there. The reason a Border Poll would have 56% in favour of a United Ireland is that Alliance, Green and other non sectarian voters would vote 75% for a United Ireland and that non voters who did not vote in the Westminster election but would in a Border Poll are also 75% in favour of a United Ireland. These nonsectarian voters and nonvoters are also opposed to Brexit by an even greater margin-85%.
LikeLike
https://m.belfasttelegraph.co.uk/news/northern-ireland/united-ireland-only-winning-scenario-after-brexit-report-claims-but-unionist-sceptical-saying-it-would-be-economically-detrimental-37502745.html
LikeLike