Stan the Statistician <<Last | Next>> | Current Stan | Archive Stan
Work in the media? Struggle with statistics? Stan's irreverent (and often irrelevant) review of the latest media reports, news and gossip may not help at all... Stan the Statistician

Stan # 104: BREXIT decided by a simple majority June 2016


With so much data flying around, Stan has not slept for more than a few hours since Thursday. Taking things in chronological order as you would expect me to, the polls were not wrong at all. It is very irritating to listen to journalists saying over and over that the polls got it wrong because they failed to predict the result just as they failed to predict the result of the General Election. The polls did not fail to predict the results simply because polls do not predict. It is journalists who predict and then try to blame mistakes on the polls. A prediction makes for a better story and then another story when the prediction is wrong.

Polls take a snapshot of opinions at a point in time. If a poll was conducted say three months before the Referendum which said that the Remain camp was going to win, would this ‘predict’ the result? If it did, then it would suggest that interventions by such as Barack Obama and Tony Blair as well as the vigorous campaigning for three months, would be a waste of time because no one would change their minds and the undecided would remain undecided.

The difficulty pollsters have these days is obtaining a representative sample. Not everyone has a landline; not everyone is in the house to take a call; not everyone will answer questions to strangers on the phone. Not everyone will respond to Internet polls; not everyone has access to the Internet. Not everyone looks at a site which features an on-line poll.

The final four polls which came across my desk showed two had the Brexiteers in front and two the Remainers. With the size of the samples, the margins of error meant that the result was too close to call. It is unfortunate that we statisticians use the term, ‘margin of error’ as those poor souls who have not studied statistics think that, ‘margin of error’ means mistakes have been made. What it really means, of course, is that given the sample size (and other factors) the result of a survey could be a bit higher or lower. For example, a result of 51% for remain with a ‘margin of error’ of +/- 2% could mean that the result is 53% or 49% - that it will be within that range (well 19 times out of 20 usually). With a simple majority deciding the matter, this makes a significant difference. The margin of error can be reduced by increasing the sample size but to get a really accurate result, the sample size would have to increase significantly and this costs money. Those who pay usually prefer to spend their budget on more polls rather than one deadly accurate poll which could change within days (but they would never know).

Turning now to the results themselves, it was clear that there were some geographic and demographic differences (which the polls did reveal very well). London, Scotland, Northern Ireland, Gibraltar, Manchester, Liverpool and Cardiff were all for remain whereas almost every other area was for leave. The polls showed that the younger you are, the more likely you were to vote to remain.

As I predicted, Scotland would vote to remain but that the majority of the electorate would vote to leave and that this would trigger a demand for another ‘once in a lifetime’ referendum on Scottish independence. Statistics show that the average life expectancy in Scotland is much lower than in the rest of the UK but having two referenda in such a short space of time does rather suggest that the life expectancy in Scotland is even lower than the current statistics show.

The SNP are flexing their muscles and demanding a second referendum on Scottish independence but what would have happened if the vote had been a narrow victory for remain but England and Wales and voted to leave – the overall result being changed by Scotland? Would the English and Welsh have demanded a referendum on Scottish independence because their democratic rights had been ignored? Indeed, as I have said before, if the consensus is that the Crimea’s vote to join Russia was unlawful because the whole of the Ukraine didn’t vote, shouldn’t the same principle apply to the UK? Shouldn’t we all have a vote on Scottish independence? It’s something Nicola Sturgeon might like to consider as, from what I hear, most of the English would like to be rid of the troublesome Scots. It would certainly save the English subsidising the Scots through the Barnett formula.

If Scotland does vote for independence in a second referendum and wish to be part of the EU, they would have to apply as a new country and could not expect to join under the same terms as the UK. Of course, if that should happen, Scotland would be one of 28 countries (with five pending) in the EU so would be subject to the rule of Brussels and would no longer be independent! So, the Scots would vote for independence only to lose it not to one country where they have MPs and have a voice but to the EU where their voice would be drowned by those of all the other countries. It appears to be a kind of death wish.

By then, a number of countries could be exiting the EU with Greece, Cyprus, Austria and France among the favourites. Would Scotland then want to push for membership of a club in which everyone one was leaving? Would Scotland then want to re-join the UK? Would England and Wales want them back? If so, on what terms? Rats normally leave a sinking ship not fight to board it.

If Scotland or Northern Ireland voted to leave the UK, England and Wales would need to establish border controls. Maybe we could seek advice from Donald Trump on refurbishing Hadrian’s Wall. At least the extra money we may have to spend on foreign holidays to Europe could be recouped by the return of duty free.


Check in again at my desk soon!

<<Last | Next>> | Current Stan | Archive Stan Who we are
What we do
Our work
Contact us