A gay marriage referendum in Northern Ireland? Beware the shy 'No' voters
There has been plenty of recent debate about Northern Ireland's position on same-sex marriage, it now being the only place in these islands were same-sex marriage is still not recognised.
Many debates in the media have referred to a recent NI Ipsos Mori poll showing 68% of the NI public being in favour of legalising same-sex marriage which approximately matches LucidTalk's own polling on the issue over recent years.
But that brings us back to the ancient question with polls (including our own polls!) - can they tell us what would happen if a real live NI-Wide same-sex marriage referendum took place?
Perhaps to help us answer this question we should look at the most recent experience with the same-sex marriage issue and polls - the Republic of Ireland same-sex marriage referendum in May.
In that, nearly all the pollsters overestimated the Yes (to same-sex marriage) vote by several points.
Not to the extent that the polls could be said to be wrong, as they were all within the margin of error of the final result, and predicted the trends quite accurately.
But the key point was that they all showed the same overestimation of the Yes vote suggesting that there were 'shy' No's at play - people who were reluctant to say to a pollster that they were voting No, probably because it wasn't, and still isn't, the media trendy view to take.
But in the privacy of the polling booth on referendum day, these shy No's made their voice heard!
However, as I said in a previous article, it's interesting to note that one S. Ireland polling company - RedC, got the result exactly spot-on with their 'Wisdom of Crowds' polling methodology.
This approach, which was used in an effort to uncover 'shy No' voters, meant they were able to accurately predict a Yes - 62% to No - 38% result the day before the referendum (matching the final result). This suggested that there were indeed 'shy No' voters, that weren't picked up in the standard poll analysis used by the other polling companies (including Mori).
LucidTalk use a similar 'Wisdom of Crowds' methodology with some of our poll-projects here in Northern Ireland (NI), including last year's European election (May 2014), and this year's Westminster election - both of which produced poll predictions pretty well in line with the final results. For these projects we asked three questions - the main question (a) What way is the respondent planning to vote themselves, but then also (b) what way they think their relatives, friends, and co-workers will vote, and (c) What do they think will be the final result of the referendum (or election, seat etc.)?
This 'Wisdom of Crowds' method is extremely useful in a country or region like NI, which has suffered from a conflict situation, which has a conservative culture, and were polling is relatively new. All these factors lead to many people feeling uncomfortable about expressing their views freely - not only to pollsters, but even to family and friends. Questions like (b) & (c) above put the poll respondents more at ease, allowing them to express their views more comfortably. As such, it's these (b) & (c) answers that can show indirectly what way the respondent is planning to vote, and it is these clues that we build-in to our forecast modelling.
A good example of this 'shy' factor is with parties like the TUV, and best shown by Jim Allisters performance at the 2014 European election. Our 'Wisdom of Crowds' polling models allowed us to predict the Jim Allister surge in that Euro election - which all the other pundits missed. However, even with our modelling we still ended up underestimating Jim Allister's vote indicating there were more 'shy TUV' voters out there, even more than our models could predict. In fact, in our polling projects over the past several years we've found that NI has a very high proportion of shy/reluctant poll respondents - much more than in S. Ireland or Great Britain.
So as a fun exercise we took all of the same-sex marriage poll results from the past 2-3 years, including the recent Mori poll, and projected them into our 'Wisdom of Crowds' poll models. These models use data and patterns from the recent 2015 general election and 2014 European election, and includes shy/reluctant voter analyses, age-group likelihood to vote (older people vote more than younger people), community background, etc. What we found was that if there was a NI same-sex marriage referendum, the final result could be tight, very tight indeed - indeed our forecast came out 54% Yes, 46% No. There is of course an error factor with this, because as well as our own polls, we are taking other polling companies research results (as well as our own) and projecting them into our own models. However, we still believe it's a useful exercise, and it does tell a story.
Of course, close result referendums are a pollsters dream - so we say bring it on! Looking at our NI polling experience, I would be very surprised if a NI Same-Sex marriage referendum produced a result any more than 55%-45% either way - with 'Yes' (to same-sex marriage) having the edge, but it's tight!
There's the old line 'Be careful what you ask for, you may get it'! So I would say to all those pro same-sex marriage supporters out there clamouring for a NI-wide referendum on the issue - be careful what you ask for, as Yes you may get it, but not the result that you want or expect!
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.