Interview: Peter Kellner


The influence of opinion polls in our society is hard to overstate. From arguments between friends to changes in governmental policy, they are constantly invoked. Perhaps the most notable recent instance of the power of polling is the Scottish referendum: the shock YouGov poll of the 7th September 2014, putting Yes in the lead, caused serious panic amongst Westminster politicians. Indeed, it was ultimately responsible for the infamous pledge from the three Westminster party leaders, guaranteeing further devolution and a continuation of the Barnett formula. It is fair to say, therefore, that pollsters marshal a very potent force and so it was with much anticipation that I looked forward to interviewing Peter Kellner, President of YouGov and one of the foremost in their ranks.

As one might expect from a pollster, Kellner studied Economics and Statistics at Cambridge. But for the first 30 years of his career, he worked as a political journalist, both in print and on the BBC before joining YouGov when it was founded in 2000. Interestingly, “90% or more of YouGov’s work is market research, not the polling we are known for”, according to Kellner, and when it was set up, polling was only one of many activities it was meant to undertake. Nonetheless, it is their polling that they are known for and it is their polling which Kellner continues to write on in the commentaries he produces. He goes so far as to describe himself as a “recovering journalist”. Yet, he now writes “just on what the numbers say” and tries to maintain a neutral standpoint in his journalism. 


This issue of neutrality led onto the question of potential abuses of polling. It is, after all, fairly easy to skew greatly the results of a poll through the phrasing of the questions or the order in which they are asked and so I wondered what safeguards YouGov had in place to ensure that their polls were as objective and neutrally conducted as possible. The answer to this was simple: “Transparency”. All the majoring polling companies are members of the British Polling Council, which requires them to publish their findings in full, including the exact details of any questions asked. Even a cursory look over the full findings would reveal if a “push poll” (i.e. a politically loaded one) had occurred and, as Kellner points out, “No one wants to be shown up to be asking loaded questions.”

Another hot ethical issue is the question of whether polls, in fact, have too much influence, causing politicians too often to change direction and not stay the course. Indeed, it could be suggested that it would be better for our political system if we only had one poll, namely that held on Election Day itself. Kellner pours cold water over this idea, arguing that,”Polls are merely a mechanism for transmitting the public voice” and therefore only serve to “deepen and improve … the dialogue in a democracy.” Indeed, in his eyes, the problem, if problem there be, lies not with an overabundance of polling data but with the weakness of politicians, “If you have got a politician who is so feeble that they have got no principles or views of their own and simply want to do what the polls tell them, then they are a pretty terrible politician.”

Yet, Kellner also says “in my experience, that is not what happens.” Rather, polls and pollsters help to inform politicians on “how to present the argument, not on what the policy should be,” a perfectly legitimate role. By showing politicians what ordinary people are thinking, they equip those politicians with the means of selling their message most effectively. According to Kellner, “when you look at politics, it is incredibly hard to change voters’ minds through an advertising campaign or a slogan or a speech. The effective way of doing things is to take what people already think and go with that.” This is where the pollsters come in, as they can provide them with that vital information. A good example of this is the Conservatives’ infamous 1979 advertisement, adorned with a picture of a long dole queue and the caption, ‘Labour isn’t working’. According to Kellner, that poster was so effective because it “underpinned what voters already thought, that the economy was in a mess under Labour, and expressed it succinctly and vividly.”

This issue of whether polls have too much influence over politicians has most recently been keenly felt over the issue of Scottish independence. The Daily Mail slated YouGov, in particular, for its polling on the issue, arguing that the final result suggests that inaccurate polling panicked party leaders into making an unnecessary, dangerous pledge on devolution. Faced with this charge, Kellner is unrepentant, arguing that the poll was a true representation of the state of affairs as it was ten days to two weeks before referendum day. The proof, he argues, lies in the fact that after the initial YouGov polling showing a lead for yes, “Each of the four companies in the next week produced polls which were 50-50 +/- 1.” The discrepancy between these polls and the actual result, he would contend, was not caused by faulty polling but by a genuine change in people’s attitudes between when the polls were conducted and Referendum Day. Moreover, if you look at the YouGov poll conducted on the day of the election, it was only 1% off the actual result and Kellner actually cites the Scottish referendum as a YouGov success story, saying it shows a “very good record of accuracy.”

On this subject of accuracy, the big debate in polling circles recently has been on whether polls are more accurate when conducted via telephone or online. YouGov have placed themselves squarely in the latter camp, conducting their polls online, where they have a panel of 400,000 British people to choose from. Whilst Kellner accepts that online polling is not the method for every poll, he sees it as preferable in most cases since “in commercial terms, online is a faster and cheaper thing to do. Then the question is, ‘Can you get as good a sample?’ to which the answer is yes. More than 80% of people now have access to the Internet and to become part of the panel, they complete a detailed questionnaire about themselves. Therefore, when we select people to take part in a survey, we are able to do it in a much more sophisticated and fine-grained way to represent the population as a whole.” He also points out that response rates to telephone polls have been dismal in recent years, reaching into the single digits in the USA.

Whilst the Scottish referendum is behind us, one of the most interesting things about interviewing one the UK’s top pollsters is finding out what he thinks lies ahead. He has already published widely on his thoughts on the 2015 General Election, arguing that the uncertainty surrounding UKIP and the Liberal Democrats make the election impossible to predict, but that a dead heat between Labour and the Conservatives in a hung parliament is likely. However, I picked his brain on some of the less well-covered upcoming elections.

Luckily for him, he hit the first prediction I asked for on the nail, namely the results of the fiercely anticipated Clacton by-election. He said that he “would be astonished if Douglas Carswell did not win Clacton quite comfortably,” a view proven by Carswell’s whooping great victory there, gaining 59.7% of the vote. Yet, the by-election is but the beginning of an intriguing narrative arc in British politics, namely the trajectory of UKIP and, in particular, the Conservative MPs who have defected to it. Indeed, the next by-election, the one to be held in Rochester and Strood on the 6th of November, is by far the more interesting one. For whilst Carswell has an immensely strong personal following and thus Clacton can be seen as an outlier, Mark Reckless does not, thereby making the race much closer and a potential UKIP victory there all the more significant. Hearteningly for UKIP, Kellner is predicting a UKIP win, albeit a closer one, “As for Rochester, I would expect Reckless to hold it, but it is quite interesting that the Tories seem to have their gander up on Rochester and are going to make more of a fist of it.” But that is not the end of the story. For Kellner also predicts that whilst Reckless might win Rochester in the by-election, the battle there in May 2015 could be “very close indeed”.

It seemed natural, after so much talk of UKIP, to broach the topic of the potential European referendum in 2017. On this, Kellner proved to be intriguing. For him, David Cameron seemed to be the key, as such a referendum would only happen in a world where the Conservatives won the next election and thus Cameron remained as Prime Minister. In such a world, the result would depend on the outcomes of Cameron’s attempts at renegotiation: “As long as Cameron comes back and says ‘I have protected vital interests and I recommend a vote to stay in,’ the UK will stay in. If Cameron comes back and says, ‘I failed to get a suitable deal and we will leave,’ then I think it is going to be very tight.” According to Kellner, the polling evidences the importance he places on David Cameron. The key demographic in determining the election is a particular type of Tory voter who, although sceptical about Europe, would “toe the line if their leader said so.” And indeed, his overall prediction is that it will be a vote to stay in the EU,on the back of Cameron returning from his negotiations with a good deal he can sell to the electorate.

I came away from my conversation with Kellner conscious of a man who saw his profession more as a vocation. For him, the provision of accurate representations of public opinion, for the information of politicians and normal people alike, is an essential element and one that he is determined to provide. In that sense, he is a man on a mission. 


Please enter your comment!
Please enter your name here