Would Naomi Long have won in East Belfast if there wasn't a unionist pact?
As you may know we carried out a full constituency poll in East Belfast back in January with the results published in the Belfast Telegraph in February. This of course was before the Unionist pact so we also had poll estimates for the UUP, TUV, Ukip, and the PUP among others. But even taking this into account, how did our full constituency poll compare to the final result?
Our January poll produced the following result (excluding non-voters): Alliance 28.7%, DUP 34.4%, UUP 14.6%, SDLP 0.8%, Sinn Fein 1.8%, Green 1.4%, TUV 2.8%, PUP 6.5%, UKIP 1.6%, Others 7.5%.
These are the exact figures as published in the Belfast Telegraph mid-February. So we got the result correct, with the gap between the DUP and Alliance in our poll being 5.7%, and the gap between these two parties in the final result last Thursday was: 6.5%, i.e. only 0.8% difference.
So would Naomi Long have won without the pact? Let's play around with some of the figures.
If you look at our East Belfast poll and take our forecasted vote figures for PUP, TUV, and UKIP (who along with the UUP didn't run in East Belfast), and estimate that say 70% of those poll vote % forecasts passed to the DUP in the real election.
This gives a running score for the DUP's Gavin Robinson of 34.4% + 7.6% - at this stage still excluding the UUP voters. Let's try the same sort of trick with the Alliance's Naomi Long.
Her poll score in January was 28.7% and let's add-in 70% of the 'Others' which is reasonable, so the Alliance running score would be 28.7% + 5.3% - again still excluding UUP voters.
However undoubtedly Alliance attracted a small number of former non-voters to support them because of the pact. If someone is viewed as a victim then some former non-voters come out to support them.
This isn't 'smoke and mirrors', the late David McClarty noticed this effect when he ran as an independent in the 2011 Assembly election having been de-selected as a UUP candidate, and Sylvia Hermon had the same boost when she resigned from the UUP over the UNCNF arrangement with the Conservatives and ran as an independent.
So let's allow Naomi Long a small 10% uplift on her base poll score for these former non-voters turning out to support her. So the Alliance score is now running at 28.7%+5.3%+2.9%.
Then crucially we come to the UUP voters as estimated in our January poll i.e. 14.6% (a good UUP poll score by the way). Let's add these in at a 90% rate i.e. 90% of those intending to vote UUP as expressed in our January poll - after all this is reasonable as the UUP was one of the two main players in the pact, so we can assume these former UUP voters would be more inclined to go out and vote.
The big question now is how did our former UUP voters split between Alliance and the DUP. Let's assume 60% would transfer to the DUP (i.e. 60% of the 90% of the UUP poll score of 14.6% - i.e. 7.9%). This gives you a final total DUP vote % of 34.4+7.6+7.9 = 49.9%, and the DUP's Gavin Robinson's actual result last Thursday was: 49.3%.
With the Alliance's Naomi Long her final score using our January poll figures (adding in 40% of the UUP voters), would be 28.7+5.3+2.9+5.3 = 42.2%, and Naomi Longs actual result last Thursday was 42.8%. So perhaps a 60/40 split of the UUP vote in the DUP's favour vs Alliance is a fairly good estimate .
This 60/40 UUP vote split is not unreasonable, and is what both the DUP and Alliance said was the way the UUP voters seemed to be splitting in the late stages of the campaign. The calculations above, and comparison to our January poll, proves this to be true.
Of course the above is only a bit of fun as they say, and you could write the calculations and assumptions several different ways. But there's a couple of clear facts - Naomi Long gained from the pact as well, and picked up a sizable chunk of UUP voters.
In fact, she said so herself with 'We have lots of former UUP voters coming over to us'. She can hardly turn around now and say her defeat was all due to the pact! Secondly, as can be seen from these calculations, our January LucidTalk constituency poll published in the Belfast Telegraph in February must have been close to the mark, and the DUP were going to win this seat anyway, pact or no pact.
But of course, we'll now never really know for sure how it actually would have turned out without the Unionist election pact.
Bill White is Managing Director of Belfast polling and market research company LucidTalk, polling partners to the Belfast Telegraph. You can follow LucidTalk on Twitter: @LucidTalk.