What fun we’ve had. Watching the polls like hawks. Seeing them go up, seeing them go down. Watching the media obsess about one poll here or one poll there. There are now only really two polls that count: John Curtice’s exit poll at 10 o’clock on election night followed by the actual result around eight hours later. And against those two results, all the polling of the last six weeks will be measured.
So first off, some stats for our readers. There have been 127 polls as we go to pixel during this election, 14 of them being the much fabled MRP polls. There were only two days when no polls were published, 15 and 22 June. Every other day, some pollsters somewhere hit us with some new stats.
So, what have we actually learned from six weeks of extreme polling?
1. None of the TV debates had any meaningful impact on the polls whatsoever. There are those occasions when the debate and its media write up has an impact; just look across the Atlantic at the disastrous and enormous impact of just one debate. But in this General Election, there was none. Which raises the question: are the debates worthwhile? They suck an enormous amount of time for the party leaders out of the campaign, who then trot through the same old, tired, predictable and endlessly rehearsed lines. They do give all the journalists lots of similar copy, but they have not helped the voting public in the least. They also tend to be watched mostly by the very politically engaged who usually have already made up their minds and just want to shout at their television either for or against their favourite or hate figure. So why bother? Previous party leaders have indeed opted out. Trump opted out of the entire primary season in the States. Are these dull shouting matches that very few people watch more for the TV channels and journalists than for anyone else?
2. The polls have had an enormous spread ranging from a lead for Labour as high as 27 points, down to as low as 11. As we have written numerous times before, it is the average, the so called ‘poll of polls’, not any individual poll that should be looked at. Whenever a media outlet trumpets one particular poll, utterly ignore them.
3. Farage arriving into the campaign did make a difference. He pushed Reform from the low teens to the high teens, sucking votes away from the Tories. But he peaked and then softened again. We shall see on election day what real impact his intervention had.
4. What about the hype around MRP polls? Truthfully, MRP polls delivered nothing wildly different to the rest of the polls as they tried to predict the election result constituency by constituency. We’ll see how well they did on election day. But it seems to the REC team these very large and very expensive polls are more a way of pollsters getting more extra and making a lot more money than anything really significant.
5. One of the big issues debated has been the difference between how the pollsters accounted for those who said ‘don't know’. There's been much commentary around the so-called ‘excluders’, pollsters which excluded the ‘don't knows’ and which tended to give a higher lead for Labour, or the so called ‘guesstimators’, those which through various means effectively guessed how the ‘don't knows’ might vote and which generally gave Labour a lower lead. We shall see on polling day if there's any real difference but as the polls unravelled over the six week period, from whichever camp all the pollsters moved towards the average, broadly speaking.
6. Did the pollsters ‘herd’? That is, did they play with their algorithms to get them to the average. It's very bad for pollster business if you're an outlier that gets it badly wrong on polling day, so they tend to fiddle with their models to get closer to the average. Certainly, the excluders started with a very high 20s leads for Labour but ended up on a 22 point lead. Whereas the ‘guesstimators’ started in the low teens and moved up to a consistent 18. This gave an average overall lead for Labour of 20 points, about where they started the campaign, indeed about where they have been for the last 18 months. Was this herding or was this simply voters, one by one, actually making their minds up hence getting the polls more accurate. We shall see.
So, what happens next in polling land?
Well, if the polls are significantly wrong, and there are those that think this might be the case, then there will be the inevitable inquiry. It is interesting that the recent past chairman of YouGov came out in The Sunday Times this weekend to say this was his view. The last two May local elections in 2023 and earlier this year produced real results for Labour of only around a 7-9 point lead, all while the polls showed the national lead to be around 20 points. So if for example, the rock solid 20 point Labour lead for the last 18 months turns out to be something like only a 15 point lead, then questions will be asked for sure. And in that scenario, bigger questions would lurk: did the persistent solid lead for Labour affect voters’ behaviour? Did the pollsters merely report public opinion or did they in fact help make public opinion?
If on the other hand, the pollsters have all herded to the safe average and the polls are near enough bang on, then standby for mailshots from every pollster in Christendom offering their services.
Comentarios