Understanding Polling: A General Election 2019 Guide from Ipsos MORI


[Music] So here are some simple lessons about understanding polls as we go into a very exciting and potentially unpredictable 2019 General Election. One thing I think is worth getting straight is that actually if you look at the history of opinion polling – and people like Professor Will Jennings have done so – polling has actually remained pretty accurate for the last seventy years and in fact if you look at it Ipsos MORI’s poll in the EU election of May this year we were accurate to each party share of the vote to about 0.75% and pollsters are always trying to improve what they do so polls generally are accurate however polls don’t predict the future. Very important to remember that pollsters sometimes do but polls don’t they just tell us what people are thinking at the time they’re taken so the fact that Boris Johnson goes into this campaign with a 10-point lead tells us very little about what might be happening in six weeks time when the election actually takes place. Even perfect polls also have a margin of error so all polls that you are going to see in this campaign will probably have a margin of error of about plus or minus three to four percentage points and that means that if one says that a party is on 36% well that could be as high as 40% it could be as low as 32% it’s probably somewhere in the middle but that margin of error is there even when a poll has been perfectly conducted. That’s simple laws of statistics. Because of that margin of error when you’re looking at polls try not to look at the lead – how far ahead one party is over another – but rather look at the vote share because you could have one poll showing a four point lead another poll showing a seven point lead. Nothing has changed. It’s simply the margin of error that’s at play they’re they’re effectively the same. The other thing to remember in all of this is that because of our first past the post system and because of the fact that most constituencies will not change hands in this election as in all previous elections – even in the most turbulent of elections in 1997 only about 28% of seats actually switched control – the national picture doesn’t always predict very well what’s going to happen in terms of seat share in the House of Commons so we can measure accurately what percentage of vote each party will get but in terms of the seats, that really depends on what’s happening in the marginals and most national polls are not focused on those marginals so if you really want to know exactly what the seats will be, look for the marginal polls or an exit poll. So do always look at the fieldwork dates and the longer-term trends, do look at the question wording and question order. If you have any doubts about a survey all the reputable companies will publish them on their websites. And finally of course polls are much more than just the headline who’s ahead in the race you want to look at what are the issues that determine how people are making up their minds what are they going to be voting on what do they think about the party leaders and all of the other tools that we have like focus groups and neuroscience are important in really getting beyond just this headline voting intention question to really understand the dynamics of the election. And finally remember our dear politicians they don’t actually lie anywhere near as much as people think they do but one lie they do say is that they don’t pay any attention to opinion polls. They’re paying just as much attention as you or I. Have a great election. Thank you. [Music]

About the author

Leave a Reply

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