The referendum on EU membership took place on 23 June 2016. Opinion polling for the United Kingdom European Union membership referendum was ongoing in the months between the announcement of a referendum and the referendum polling day. Polls on the general principle of the UK's membership of the European Union were carried out for a number of years prior to the referendum. Opinion polls of voters in general tended to show roughly equal proportions in favour of remaining and leaving. Polls of business leaders, scientists, and lawyers showed majorities in favour of remaining. Among non-British citizens in other EU member states, polling suggested that a majority were in favour of the UK remaining in the EU in principle, but that a similarly sized majority believed that if the UK were only able to remain in the EU on renegotiated terms then it should leave.
Analysis
Demographics Younger voters tended to support remaining in the EU (but are generally less likely to vote) whereas older people tended to support leaving. There was no significant difference in attitudes between the genders. According to two out of three pollsters, managerial, professional, and administrative workers were most likely to favour staying in the EU, while semi-skilled and unskilled workers, plus those reliant on benefits, were the largest demographic supporting leave. University graduates are generally more likely to vote remain compared to those with no qualifications. White voters were evenly split, and all ethnic minority groups leant towards backing Remain, but registration is lower and turnout can be up to 25% lower in this demographic. Support for remaining in the EU was known to be significantly higher in Scotland than it is in the United Kingdom as a whole.
Polling methods The way voters are polled is known to affect the outcome. Telephone polls have consistently found more support for remaining in the EU than online polls. YouGov, which uses online polling, has criticised telephone polls because they "have too high a percentage of graduates", skewing the results. Ipsos MORI and ComRes, and Peter Kellner, the former president of YouGov, have said telephone polls are more reliable. ICM has said "as good a guess as any is that the right answer lies somewhere in between". A joint study by Populus and Number Cruncher Politics in March 2016 concluded that telephone polls were likely to better reflect the state of public opinion on the issue.
The results of the Referendum, as with the results of the 2015 General Election, show that there is still a problem with the polling methodology. Overall, however, online polls seem to have had a better performance than phone polls. Online surveys, on average, predicted a "leave" win with a 1.2% margin, whereas those with a phone methodology had "remain" win with a 2.6% margin. All in all, 63% of online polls predicted a Leave victory, while 78% of phone polls predicted that Remain would win. Kantar TNS and Opinium, both pollsters with online methodologies, were the two groups that forecast a Leave victory just ahead of the vote.
Polls of polls Several different groups have calculated polls of polls, which collect and average the results of opinion polls across different companies. They have different methodologies; for example, some give more weight to recent polls than others, some deal with undecided voters differently, and some attempt to adjust for the consistent gap between telephone and online polling. As a result, the polls of polls give a spread of results.
Conducted by Date Remain Leave Undecided Lead Notes What UK Thinks: EU 52% 48% N/A 4% Six most recent polls. Elections Etc. 50.6% 49.4% N/A 1.2% Twelve most recent polls. Telephone polls are adjusted in favour of Leave and online polls in favour of Remain. HuffPost Pollster 45.8% 45.3% 9% 0.5% Number Cruncher Politics Alt URL 46% 44% 10% 2% Equal weighting to phone and online polls. Financial Times 48% 46% 6% 2% Five most recent polls. The Telegraph 51% 49% N/A 2% Six most recent polls. The Economist 44% 44% 9% 0% Excludes polls with fewer than 900 participants.
Standard polling on EU membership
The tables show polling on whether the UK should be in or out of the EU. Polling generally weights the sample to be nationally representative. Polls were usually conducted within Great Britain, with Northern Ireland and Gibraltar normally omitted from the sample. This has historically been the case in British opinion polling because Northern Ireland has a different set of political parties from the rest of the UK, reflecting the political divide between unionism and nationalism or republicanism. Similarly, Gibraltar was not included in standard polls because it has its own local legislature and does not take part in British parliamentary elections, although Gibraltar does take part in elections to the European Parliament and took part in the referendum.
Most of the polls shown here were carried out by members of the British Polling Council (BPC) who fully disclose their findings, methodology and the client who commissioned the poll. As non-members, Qriously (Qriously has since become a member), Greenberg Quinlan Rosner Research, Pew Research Center and Lord Ashcroft Polls are not bound by the standards of the BPC, and their polls should be treated with caution.
The percentages who "would not vote" or who refused to answer are not shown below, although some pollsters have excluded these in any case.
2016