This week, the Daily Telegraph newspaper published a poll predicting that the Conservative Party will face annihilation at the forthcoming election. The Telegraph said that ministers including the Chancellor, Jeremy Hunt and leadership contender, Penny Mourdant would lose their seats, alongside a further 194 Conservative MPs. Polls like this grab the attention but do they really have any impact on the election results?

The simple answer is no. The more difficult answer is that we probably can’t ever know what their impact is.  This is partly because the methodology used by polling  firms can never ensure certainty and also because of the different motivation it provides to different groups of voters. Take the EU referendum of 2016. The majority of polls predicted that remain would win, most suggesting a narrow victory. However, most of the polls asked whether you would vote leave, remain or you were undecided. What the polls failed to pick up was that the majority of the undecided would vote to leave when they arrived at the polls. Herein lies a problem with polls, they can’t tell us what undecided voters will do. Moreover, they can never be 100% reliable with voters occasionally changing their mind or concealing their true feelings. The phenomenon of ‘shy Tories’ whereby voters refuse to admit they will vote Conservative to opinion pollsters is well known and polling companies do try to make allowances for it, but this helps to show us how difficult it is for the polling firms to get accurate data.

The 2017 General Election was also marked out for widely inaccurate polling. All but one poll in the six months prior to the election predicted that the Conservative Party would win the election, most predicted by more than 10%. In the end, the Conservatives only won by 2.5% and Theresa May lost her majority. How had the polls got it so wrong? Ironically, most of the polling companies had originally predicted this outcome but actually changed their voting methodology because they themselves couldn’t quite believe what they were seeing. Burnt by previous predictions on youth turnout, the polling companies were sceptical that young people would turn out to vote and so adjusted their figures to take account of this. Here is another problem with opinion polls, they find it hard to predict turnout, and particularly which groups will turnout to vote, usually over predicting the youth vote. This caused a problem in 2017 because in making allowances for this, they actually underpredicted youth turnout.

We should also consider what polling does to people’s voting intention. It is highly conceivable that opinions polls predicting a comfortable victory for Remain in the EU referendum or a comfortable victory for the Conservative Party in 2017 might encourage those who might vote in line with these results to stay at home, believing that their vote is not as important. Whilst this is most probably true in some cases it is hard for us to be certain of such. Nonetheless, it is unlikely that this would have a big enough impact on the final result.

A further point that needs to be considered is why the poll has been published in the first place. Is the poll being conducted by an independent polling firm or has it been commissioned on behalf of a political party or a cause? In the case of this week’s mega poll, it had been commissioned by a group of Conservative donors called the Conservative Britain Alliance. This group is on the right of the Conservative Party and have been sitting on the poll for a number of weeks. Why publish it now? The website Conservative Home, amongst others, have suggested that the publication of the poll in the week of a vote on the Rwanda Bill should be no surprise. Perhaps it was the intention of Conservative Britain Alliance to publish the poll in hope of jolting Conservative MPs to vote against what they saw as a bill that needed toughening up.

In summary, it is difficult to know the impact of opinion polls. One thing for sure is that polls move all the time and the Labour Party, who currently hold a sizeable lead in run up to the election, would be well advised to ignore them and campaign as if they were behind.