Has the DUP put itself through the drama of its past 12 months entirely by mistake?
That is the almost laughable possibility arising from the Institute of Irish Studies University of Liverpool/The Irish News opinion poll, published this week.
It has the TUV on 6 per cent, half the party’s 12 per cent support in the most recent LucidTalk poll.
Polls are a snapshot in time and the Liverpool survey was conducted two weeks after LucidTalk’s, during which a great deal occurred, including the DUP trying to suspend sea border checks and first minister Paul Givan resigning.
However, the difference in the TUV figures is both enormous and persistent.
Last February, Jim Allister’s party jumped from 6 to 10 per cent in LucidTalk’s quarterly tracker poll. It has remained at between 11 and 14 per cent ever since. LucidTalk also conducted a poll last November for Lord Ashcroft, which again had the TUV on 12 per cent.
The only previous Liverpool survey, last October, had the TUV on 6 per cent.
This is the only significant discrepancy between the pollsters - they agree with each other, within their margins of error, for every other party.
Something about the difference between the LucidTalk and Liverpool methodologies is uniquely unable to agree on TUV support. Both cannot be right and if LucidTalk is wrong the implications for the DUP are horrific. It was the apparent jump in TUV support last February that spooked the DUP into reversing its acceptance of the protocol. The continued apparent loss of its voters to Allister then drove it to depose Arlene Foster, have a farcical leadership crisis and threaten to bring down Stormont from last September, until it finally cornered itself into Givan’s resignation.
If TUV support has been closer to 6 than 12 per cent the whole time, none of this was necessary and most of it was massively counter-productive. The DUP should have been trying to calm its base instead of leading a hysterical charge after lost hardliners.
High polling numbers also have an impact on the public. Perhaps the TUV has only reached 6 per cent because people think it is at 12 per cent.
LucidTalk has an excellent track record predicting elections. The main concern voiced about its online panel method is that inviting internet users to fill in questionnaires may attract the ‘politically engaged’, with weird effects that are hard to adjust for through statistically weighting.
This could mean people on the panel are unusually interested in the protocol - another divergence from the Liverpool survey, which found only one in ten unionists say the sea border is their top priority. DUP voters more exercised by the protocol could in turn be more likely to switch to the TUV.
Of course, the DUP knows this - party figures have often commented on LucidTalk’s method. But the party clearly finds the high TUV figures credible because they chime with its own soundings among the grassroots.
Like the difference between a unionist and a loyalist, the difference between the grassroots and the garden centre is ill-defined and loaded with unspoken class distinctions. Although the media almost certainly underestimates grassroots sentiment, parties may overestimate it as it comes with its own problem of attracting the politically engaged.
Most people do not attend protests or political meetings, or buttonhole their elected representatives. A Dromore Orange Hall was packed last week for a meeting against the protocol, with over 100 people in attendance and others waiting outside. It was an impressive turnout for such an event and will have made a strong impression on the DUP, TUV and UUP politicians on the stage, yet in electoral terms the numbers were meaningless. Electing an MLA takes at least 5,000 votes and however passionate their views the grassroots only have one ballot each, like everyone else.
The real issue with polling in Northern Ireland, as any pollster will confirm, is that our media market is too tiny to sustain surveys from multiple companies, facilitating the ‘poll of polls’ averages available elsewhere that expose outliers and confirm general trends.
There is a case for public funding of more polls but it is problematic. LucidTalk does not deserve to have its fairly won business undermined by taxpayer-funded competition. The DUP does not deserve taxpayer help to avoid making terrible mistakes.
So we may bear in mind a lesson from physics, alongside statistics. Very small things cannot be measured without changing them. In little Northern Ireland, that could have major consequences.
