“There are lies, damned lies and statistics”
(Sir Benjamin Disraeli, 1895 [i])

When the BBC published an article looking at the breakdown of local voting in the EU referendum on 6 February 2017 (Rosenbaum & Greenwood 2017), I am not sure even they were aware of the amount of interest it would pique. Although the results were interesting (depending on what side of the Brexit fence you sat and/or whether you appreciate statistics), what was more interesting was both the sheer scale of interest and comments it generated in such a short time (Over 3700 by the end of day[ii]) and the types of comments. Comments that effectively reflected the polarisation of the Brexit vote. In some ways all rather predictable considering what the results strongly suggested. That there was a strong relationship between age and whether an individual voted leave or remain (older people generally voting leave), but more contentiously there was a strong relationship between the level of education and voting leave or remain. Those having a degree or higher more likely to vote remain.

While the volume of responses was probably a lot to do with automated algorithms promoting the article to the BBC’s home page when it was first published and then keeping it there due to the volume of interest it was generating. Comments naturally abounded… ‘who are you calling stupid!’, ‘why is the BBC publishing such rubbish?’, ‘Waste of tax payer’s money’. ‘BBC is biased’. To generalise some of the themes of the content. As well as a few commenting that they knew individuals with a degree who voted leave, so it can’t be true.

However, this was not first post-referendum analysis and just adds weight to findings from an earlier survey by Lord Ashcroft (Ashcroft 2016) taken after the EU Referendum. It also implies the BBC is not being politically biased, particularly as Lord Ashcroft is Conservative, and as the Conservative party is collectively rooting for Brexit, these findings are not such good news for them.

Statistics… an interesting subject. Oh well for some anyway J. The problem is not that the actual numbers are doing the lying, it’s how you use them. Unfortunately for many of us, lots of people cite a statistic as given proof or worse still accept it themselves as given proof of something. Which is the point behind Disraeli’s comment. Take the actual EU Referendum vote for instance. 52% voted ‘Leave’. That’s a given fact, except…. what it really states is that on the 23 June 2016, 52% of those who voted, voted for the UK to leave the EU. Nothing more, nothing less. What it does not state and can NOT state is whether if this was repeated you would get the same result. If the UK held a referendum again it could just as easily have a greater percentage of voters voting to leave, as it could have a greater percentage of voters now voting remain. The only certainty is that on the 23 June 2016 the percentage of voting was 52% in favour of the UK leaving the EU.

What this statistic also does not mention is who did not vote. Those who were not able to vote because of age or other reasons (28.9% of the population) and those who could have voted but decided not to (19.9% of the population). If the 52% is looked at again in the light of the estimated total UK population at the time of 65 million (World Population Review 2016) it’s closer to 27%. With 25% voting remain. The interesting thing is that the smaller the actual sample the less accurate the prediction is for the total population, so it is not possible to state that 52% also represents 52% of the total population.

image002Chart showing percentage of voting in UK EU Referendum as a proportion of the estimated UK 2016 population.

So what’s with the furore over the BBC’s breakdown of the local results of the EU Referendum (supported by the results from Lord Ashcroft’s poll)? The results do not suggest that those who voted for Brexit are less intelligent. If you read where the data is drawn from, it’s the percentage of the population in each area who have a degree or higher (graduates). Those areas where a higher proportion of the population are graduates the more ‘likely’ it was to vote remain. Two things. ‘Likely’ does not conclusively mean that an area with lots of graduates voted remain, it also does not mean all graduates voted remain, this is fairly evident in the graph. One can see the distribution is slightly scattered, but at the same time there is a rather linear pattern, what is known as a correlation (think “co – relationship” one thing exists in relationship to another).

_93993281__eu_ref_educated_population
Graph comparing percentage of graduates against percentage leave vote of 1070 wards (Rosenbaum & Greenwood 2017)

So what does this mean and what age got to do with it? One possibility is that thanks to successive governments more and more individuals have had the chance to go to university. So a greater percentage of each successive generation has been able to take advantage of a university education. The amount of intelligence in a population has not changed in each generation, just a higher number of individuals who have a university education. The thing about university education is that it’s not just the subject that’s taught, it’s how to study, how to independently analyse, to not necessarily accept information at face value and to look at the wider implications.

Either way, the government has a problem on their hands. How to appease two rather polarised sides, both those who voted for the UK to leave the EU and those who voted to remain. Appeasing just the brexiters would be foolhardy considering how many also voted to remain and probably feel just as strongly, they also have to take into account that the younger generations were more likely to be in favour of remain. Interesting times ahead, indeed. The Conservatives can only hope.

References

Ashcroft, M.A.P., 2016. How the United Kingdom voted on Thursday… and why. Available at: http://lordashcroftpolls.com/2016/06/how-the-united-kingdom-voted-and-why/ [Accessed January 9, 2017].

Lee, P.M., 2012. Lies, Damned Lies and Statistics. History of Statistics pages, University of York. Available at: https://www.york.ac.uk/depts/maths/histstat/lies.htm [Accessed February 6, 2017].

Rosenbaum, M. & Greenwood, G., 2017. Local voting figures shed new light on EU referendum – BBC News. BBC News. Available at: http://www.bbc.co.uk/news/uk-politics-38762034 [Accessed February 6, 2017].

World Population Review, 2016. United Kingdom population 2016. Available at: http://worldpopulationreview.com/countries/united-kingdom-population/ [Accessed January 9, 2017].

——————————-

[i] The phrase is popularly attributed to Disraeli, although versions of it existed prior to 1895 with the earliest recorded being Sir Charles Dilke in 1891 (Lee 2012)

[ii] 3708 comments by 23:30 on 6/2/2017