Belfast Telegraph

Scottish referendum: How well did the polls predict the No vote in Scotland?

By Bill White

Like them or hate them there is no doubt that the polls drove the flow of the Scottish referendum campaign. It was a great referendum for the polling industry, with regular polls by all the big-name companies contributing to the campaign agenda.

Regarding the Scottish referendum and the polls, it’s important to note that a referendum is not the same thing as an election. In a referendum people are asked, broadly speaking, to choose between the status quo and change, whereas in elections they choose between parties, candidates, and a range of policies. Referendums are more difficult for pollsters than normal elections because there are additional unknown factors.

First of all, there are a large number of people who vote in referendums who don’t vote in elections, and therefore there is no previous historical data on which to base projections. Unlike elections, there are no core voters; people who vote a particular party, come what may. There is another factor that some people, including some in the media, don’t seem to appreciate – people change their minds during a campaign. Frankly it would be a bit frightening if there weren’t changes of minds and opinion during a campaign - images of North Korea come to mind. This opinion changing also happens during normal elections, but there is much more ‘changing of minds’ during a referendum campaign, and therefore the opinion tracking has to be more fluid and rapid.

The key point for the polls during the campaign was the famous ‘YouGov Sunday’ (September 7) with that YouGov poll showing ‘Yes’ ahead, for the first time in the campaign, at  51% to 49%. This sent shock waves through the Westminster establishment and Cameron was on the first plane to Scotland the following morning. The regular Wednesday Prime Minister’s Questions show was handed over to others, so that these two main party leaders could throw their kitchen sink into the campaign. In the last campaign day Gordon Brown made an impassioned appeal for the No campaign, and the post referendum polling shows this had more impact than Dave and Ed. Brown persuaded some ‘Yes’s’ back to No, and crucially a lot of the 'don’t knows' to No. The polls tracked all this accurately and the last poll of the campaign, again from YouGov, was 54/46 for ‘No’. In fact, last Thursday night, Peter Kellner (YouGov President) accurately predicted the result on Sky News, well before any results were announced.   

Yes, the polls all probably underestimated ‘No’ as the campaign progressed, and there was undoubtedly a ‘Shy’ No factor i.e. those people who were voting ‘No’ but didn’t want to say so. But one thing that is never wrong in polling, are trends. Undoubtedly there was a growth for ‘Yes’ during the 3-4 weeks leading up to referendum day, which then faded and reversed during the last 4-5 days of the campaign. The polls accurately tracked and reported all these trends. Personally I doubt if ‘Yes’ were ever ahead, but probably peaked at about 48/52 around ‘YouGov Sunday’.

Polls are also useful in terms of tracking the key demographics behind the result. It’s interesting to note that in the Scottish referendum voters aged between 16 and 54 voted Yes by 54% to 46%. There was a clear No majority among those aged 55 to 64, and a three-to-one margin for No among those aged 65 and above. So without the 55+ vote the Union would have been over.

So how can the final result be interpreted? It’s interesting watching and listening to the results coverage and all the post-referendum analysis. I like the way journalists hype up their reporting with such terms as ‘surging’ and ‘decisive’ e.g. One reporter said the famous game-changing ‘YouGov Sunday’ poll was the ‘Yes’ campaign ‘surging’ into the lead – How can 51/49 be termed as one side surging into the lead? Then we have all the media reports stating that the 55/45 result is/was ‘decisive’ – Really?, a huge constitutional issue question produces a 55/45 result and that is decisive? I’m entirely neutral on the overall question i.e. Yes or No, but let’s be clear, a 55-45 result for the status quo i.e. the No’ campaign is anything but a win. In a happy society, at ease with itself (to use the expression that’s often quoted in Northern Ireland), the status quo, i.e. stay within the UK, should be hitting 80-90%+.

I recall a critic of David Trimble, referring to the split in the UUP over the Good Friday agreement saying 'You can’t keep running a party, or any organisation, on 55-45’ – referring to Trimble’s 55% support base in the UUP. He was right of course, and eventually Trimble was destroyed. As they manage the Scottish situation, the UK government should remember this.

One final point – don’t think the referendum issue has gone away. Many observers think we’re heading towards the ‘never-endum referendum’. If, as seems likely, the SNP win a majority again at the 2016 Scottish Parliamentary elections, and if the UK goes for an EU In/Out referendum in 2017, there will be a strong case for running another Scots independence referendum on the same day as the UK’s In/Out EU vote. Remember Scotland is overwhelmingly Yes for Europe.

Bill White is Managing Director of LucidTalk, polling partners to the Belfast Telegraph

Belfast Telegraph Digital


From Belfast Telegraph