Be highly cautious with Northern Ireland “polls”!

I am a big fan of the polling company Lucid Talk which deserves credit for trying to break into a market which simply does not exist in Northern Ireland – namely polling.

The company itself has been consistently clear about its methodology, and has openly warned that much of what it does is not “polling”, or at least is not to be compared to what is referred to as “polling” in Great Britain.

It is not for me to promote a particular company, but the simple fact neither Lucid Talk nor anyone else will be able to deliver polls of real value until they are paid properly to do so.

Polling is in fact an incredibly complex thing. Most obviously, it has to be properly weighted – to take account of voting intention in the past, voting likelihood, social background, gender, age and so on. This in itself requires an enormous amount of research – I have already cautioned that even in Scotland it may be way off with regard to the forthcoming referendum.

With more limited resources, the trick really is not to overstate your case. My own company did an exit poll (purely for research purposes) at the last Assembly Election, after which we were able confidently to predict that Anna Lo would top the poll in Belfast South; the Ulster Unionist vote was down in some places but not in others; the Alliance vote was up in some places but not in others; the DUP and Sinn Fein vote was roughly stable; and (the late) David McClarty would be elected. Here’s the thing, though: with the exception of the latter, I would have confidently predicted all of those things anyway! In other words, I was reading the poll through my own instinct before declaring what I felt it meant – a vast amount of interpretation utterly unacceptable (rightly) to a professional pollster.

As it happened, I was fairly confident (but not confident enough to state overtly) that Steven Agnew would be elected. I reckoned (again not with huge certainty) but had got wrong that David Ford would be on 18% (he was on about 14%) – however, this turned out to be not far off as we did that particular section in Ballyclare, where the Alliance vote in the Local Election did increase markedly (by an amount which, if repeated across the constituency, would have seen it reach 18%). But then, I also reckoned that Mr McClarty would take the seat from the DUP not from the Ulster Unionists – our poll in Coleraine was actually quite good for the Ulster Unionists, when in fact their two candidates came last and second last. A mixed bag at best, in other words…

One thing that I would like to research again that stuck with me was that 29% of those polled had changed their vote in the Assembly Election from the General Election the previous year. In fact, it was striking how many could not remember how they had voted in the past (indicating no affiliation to a particular party). That one is maybe to be pursued next year – but it’ll be a voluntary activity, I suspect.

If we want real polls of real value, frankly, someone is going to have to pay proper money for them.

Advertisements

6 thoughts on “Be highly cautious with Northern Ireland “polls”!

  1. Martin Lemon says:

    I remember that poll- did it not have Alliance on about 35% and SF on 5%? I found both of those figures astonishing, I mean it is Ballyclare after all!

    • That would be about right in some parts of Ballyclare!

      I can’t remember the actual figures, because it was all about *trend*. Certainly in Ballyclare there was a significant number who voted Empey in 2010 and Ford first pref in 2011.

      • Martin Lemon says:

        Well, 3 days from polling, can you see past Anderson, Dodds and Nicholson? Being completely honest! 😉

      • Martin,

        I never could see past them, wearing my “objective” hat!

        The one thing I would say is that I haven’t polled it at all and I’ve knocked doors in only two constituencies – nothing like enough to make a definitive call! In particular, my sense is that something odd is happening again at the hardline Unionist end (like, say, another big vote for Allister) but I have no actual evidence (or means of determining it)!

        Just a final point, reporters are inclined to underestimate the Ulster Unionist vote, because that vote is now predominantly rural, particularly in border constituencies (and reporters are based in Belfast). In places like Upper Bann or Fermanagh, Nicholson will likely be well ahead among the Unionist candidates.

        But we’ll see soon enough!

  2. Monster Sock says:

    Surely there will be a big vote for Lo in Nationalist areas, now that’s she’s self defined as one

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: