Did the EU Referendum Bring a Different Kind of Voter to the Polls?

Posted on 28 June 2017 by John Curtice

It has been a long time a-coming, but the publication today of the latest British Social Attitudes report means that a significant and unique source of evidence on who did and who did not vote in the EU referendum has now finally been unveiled.

Turnout was relatively high in the EU referendum. At 72%, it was higher than at any general election since 1992 – including at the ballot held most recently on June 8th.  This inevitably raises questions about whether the kind of people who voted last June were different too. In particular, is there any evidence that younger people and those with less interest in politics, amongst whom turnout consistently tends to be relatively low, were particularly more likely to have been engaged by the European ballot?

The difficulty we have had with answering this question up to now has been the absence of robust evidence. Opinion polls are not necessarily the most reliable source of evidence on differences in turnout – most polls are conducted over a very short period of time, many people will refuse to answer questions over the phone, while all internet polls are conducted amongst people who have previously agreed to take part in such exercises. As a result, polls often find it difficult to make contact with those who have little interest in politics and who thus fail to make it to the polls.

The annual British Social Attitudes (BSA) survey, in contrast, is undertaken over weeks and months. The people who are interviewed – face to face in their own homes – are all selected entirely at random. At the same time, the survey is not just about politics, so those with less interest in politics but with an interest in one of the many other subjects covered by the survey can sometimes still be persuaded by an interviewer to take part. It is thus more likely to be successful at securing an interview with those who do not vote.

Not only did the latest BSA survey ask people whether they voted in the EU referendum, but previous BSA surveys have asked about participation at other major ballots in the UK. The 2015 survey, for example, asked people whether they had voted in that year’s general election, allowing us to compare turnout in the EU referendum systematically with that at the general election two years ago when turnout was a more modest 66%. Moreover the 2011 BSA asked people whether they voted in the last UK-wide referendum, that is, the referendum held in May 2011 on whether the electoral system for the House of Commons should be changed to the Alternative Vote. That means we can also establish whether the pattern of voting in the EU referendum differed from that at a previous referendum in which the turnout, at just 42%, was much lower.

After weighting the sample to be demographically representative, 77% of those who participated in the 2016 BSA, conducted between July and October last year, said that they voted in the EU referendum. This is somewhat higher than the official turnout of 72%. However, the official figure will have been depressed somewhat by redundancy in the register of electors on account of deaths since the register was originally compiled the previous autumn and by the presence of people who are legitimately registered to vote at more than one address. Even so, the survey probably slightly over-represents those who voted, but would seem sufficiently close to the actual level of abstention for us to regard its estimates of who did and did not vote as likely to be relatively robust.

So were younger people, who some felt to have most at stake in the referendum, and who, indeed, were much more likely to support staying in the EU, particularly more likely to participate in the European referendum?  Today’s report suggests that to some extent they were. Table 1, which is taken from the report, shows the level of turnout broken down by age group for each of the ballots in 2011, 2015 and 2016. As can be seen, irrespective of whether we compare the turnout in 2016 with that in 2015 or the level in 2011, turnout did increase rather more amongst those aged 18 to 34 than it did amongst those aged 65 and over.



That said, two notes of caution are in order. First, given that turnout was already as high as 84% amongst older voters in 2015, there is something of an arithmetic limit to how much higher it could go. Second, although the difference in turnout between younger and older voters may have narrowed somewhat, it was still the case that younger voters were markedly less likely to have voted than their older counterparts.

But what of those with less of an interest in politics? After all, it has been suggested that those who were generally disenchanted with and disengaged from the political process regarded the referendum as an opportunity to express their feelings by turning out and voting to leave the EU. Indeed, as many 62% of those with not very much or no interest at all in politics who did make it to the polls voted to leave, compared with just 40% of those with quite a lot or a great deal of interest.

Table 2 shows the level of turnout in 2011, 2015 and 2016 broken down by people’s reported level of interest in politics. As can be seen, virtually all of the increase in turnout as compared with the 2015 general election occurred amongst those with not very much or no interest at all in politics. To that extent, the EU referendum does seem to have engaged some of those who often fail to make it to the polls – and given the way the less politically interested voted in the referendum this finding is consistent with previous NatCen research (based on its longitudinal mixed-mode panel) that suggests that those who voted in 2016 but did not do so in 2015 were disproportionately Leave voters.



Again, some caution is in order.  Given that nearly nine in ten of those with a great deal or quite a lot of interest in politics voted in the 2015 general election, there is clearly a limit as to how much higher turnout could go amongst this group. Indeed, if we compare the pattern of turnout in 2016 with that in the AV referendum in 2011, the increase in turnout is only a little higher amongst those with less interest in politics than it is amongst those with rather greater interest. The EU referendum was thus only marginally more successful at bringing less interested voters to the polls than was that much less high profile referendum ballot. Meanwhile, those with less interest in politics were still far less likely to participate – the gap simply narrowed somewhat, rather than disappeared.

However, perhaps the most interesting finding about political engagement in the wake of the referendum is not that those with less interest in politics (along with younger voters) were somewhat more likely to make it to the polls, but rather that more people expressed an interest in politics in the first place. No less than 42% said in the 2016 survey that they had a great deal or quite a lot of interest in politics, up from 36% in 2015, which itself was slightly above the previous all-time high of 35% recorded in thirty years of previous annual BSA readings. Of course, it may be that this greater level of interest in politics will not be sustained, but for the time being at least it looks as though perhaps the biggest impact of Brexit on patterns of engagement may have been to persuade some people that politics really does matter – either for good or ill.


John Curtice

By John Curtice

John Curtice is Senior Research Fellow at NatCen, Professor of Politics at Strathclyde University, and Chief Commentator on the What UK Thinks: EU website.

There is one comment

Your email address will not be published. Required fields are marked *