What’s going on with opinion polls?

Part of engaging as Christians with public issues is simply understanding, as far as possible what is going on.  This is true when it comes to things like General Elections.  It is good to have a level of understanding about what is happening in the US with Donald Trump and Joe Biden, it helps us to pray for that situation in an informed manner.  I hope also that we have some awareness of the political climate in countries where we have mission partners.  For example, our church supports people involved in mission in Egypt and we have members who used to serve there,  I would encourage our church members to find out a little about Egyptian current affairs. This means understanding a little about the constitution, about recent history meaning about people like Presidents Mubarak and Sisi, organisations like the Muslim Brotherhood and the events of the Arab Spring in 2011 and what followed.

Here in the UK it means having an understanding of what is going on in the General Election. Much of that might be bewildering.  Why are politicians pursuing specific strategies? Why is the election happening now? Who are all the different parties and personalities etc.  One question of course will be “what is going on with the opinion polls?”  This is not least because the polls are so shocking.  There may be plenty of people around who want to see the Conservatives completely wiped out, just as there would have been plenty rubbing their hands in glee when the situation seemed to be reversed in 2017.  However, is it really possible that the party who only recently one a landslide majority of more than 80 seats could be reduced to a rump of  potentially as low as 60 seats? Such a scenario seemed inconceivable as late as 2022.

There’s also a lot of mythology around polls and that doesn’t help.  You will frequently here politicians say “the only poll that matters is the one on election day” which of course is true in that election day is the real deal but can give the impression that the actual poll and the opinion poll.  There’s the presumption that the polls always get it wrong.  That’s based on 1992 when all the leading pollsters were completely out in their projections, though whilst partly that seemed to be about a failure to correctly identify Conservative voters, the so called “shy Tories”, there may also have been a late shifting of opinion.  Hold on to that thought because there is a possibility that both of those factors could come into play this time around.

Other myths include “they cannot be right. Because I know lots of people who vote x and nobody who votes y.”  We tend to know about people who are like us. That’s because of three reasons.  First, we are more likely to spend time with people who share our outlook. Second because people tend to adjust what they say in line with the expected reaction.  Thirdly, we tend to hear what we want to hear or at least what we assume of others.  

Finally, there is the myth that this or that pollster is the best/only reliable one.  Back in 2019 the common theme was “I’ll wait for Survation.”  That’s because it was only really Survation along with the You Gov MRP methodology that had identified the likelihood of Corbyn running May close.  However in 2019, Survation failed to deliver the same level of optimism for Corbyn and there were plenty of other pollsters that were just as, if not more accurate.  You are only as good as your last performance.  It’s possible one pollster to be seen as the gold standard one year and then crash and burn four years later.

So, what are we seeing at the moment?  Well first of all, there is general agreement.  Labour are polling well, they are in the mid 40s.  That’s General Election winning territory and in modern times likely to guarantee a sizeable majority due to the First Past the Post system in a multi-party democracy.  The only time this didn’t prove true was in 2017 when although Theresa May got about 42-43% of the popular vote, Corbyn also polled in at about 40-41%.  In effect we reverted to a two party system. 

Secondly, the Conservatives are doing extremely poorly and look to suffer a landslide defeat.  The question is “exactly how badly are they doing?”  Polls are ranging from 19% to 28%.  That’s another question.  Why is there such a broad range?  There are a couple of reasons for this.  First of all, you will get outliers, that can be due to anything from odd timing to unique methodology.  Secondly, you expect a level of error so a party could be plus or minus 2 or 3 percent on their poll score.  That’s why people tend to look at an average of the polls hoping to iron out variations.  So, it’s assumed that if one poll shows the Tories at 21% and another at 27% then it looks possible that the first has understated by 3% and the second overstated.  On the balance of probabilities, if the second poll were also understating, you would expect others to occasionally come in above 27%.  So, whilst it is possible that the Tories are as high as 30% or as low as 16% both scenarios are unlikely. 

This still doesn’t completely explain the wide range and so we need to factor in something else.  The reported figures are not the raw data. In other words, if Savanta poll 1000 people and report 44% to be voting Labour, that doesn’t mean they have found 440 Labour supporters.  You see, I you do  a random survey, then this doesn’t mean you have a genuinely representative cross sample.  You may struggle to find certain sectors. This was notable back a few decades ago when pollsters were either relying on in person conversations but struggling to find working people at home or landline surveys when younger people might not have phones.  Similarly, in the early days of internet polling it was easier to find younger voters than older ones.  So, the numbers are weighted to account for discrepancies.  This means that pollsters use models based on what they know about previous voting habits.  They also take into account hoe the people themselves say they have previously voted.  Oh and then you have the “don’t knows”.  Now often, it’s assumed that these people really have pretty much made their minds up and will break in line with the do knows.  However, sometimes it is possible that people are genuinely uncertain and in those cases, you may see a surprise late shift.  I think we saw this in the EU and Scottish Independence referendums.

Now, why are the Tories polling s badly? How have they got here from 42-44% in the last two elections.  I think the answer is simple. The Conservative’s support in those elections was wide but shallow.  Here’s why.  The Conservatives had lost most of their support from the Thatcher years and never really recovered it.  They’d inched back to the mid 30s between 2005-2010 but never really built on it.  Then in 2016, the decision to go all out as the pro Brexit party meant they lost a lot of natural Conservatives who were remainers.  I don’t tnink they have won them back and so I suspect the Conservative’s core support had already shrunk from about 32% to about 25% by 2019. 

In the 2017 and 2019 elections, Theresa May and Boris Johnson decided to pursue another group of voters, those who were disaffected with politics who probably were traditional Laouur supporters but didn’t trust Cobyn and were pro-Brexit.  They seemed to assume that they had achieved a great realignment, that these were now fully signed up Conservative supporters.  However, I think they miscalculated.  A lot of those people were not switching sides but were lending their support, holding their noses and voting Tory to get Brexit through and keep Corbyn out.  With Boris Johnson as an insurgent gone and Corbyn too, that lent support has also gone too.

Now, there are a couple of unknowns.  First, how will Farage’s return to the fray affect things. It’s assumed that it will harm Conservative chances still further but it may not.  Farage is a populist and popular enough to harm the Tories but he doesn’t have magical properties and his indecision and gamesmanship may have damaged his own brand.  Time will tell.  Secondly there may be more surprises

A third factor is to what extent it is possible that a lot of people are still very undecided, conflicted even.  Note that there may well be shy Tories who are not willing to say out loud to pollsters that they are going to vote Conservative. Even in 1997 and 2001 Labour were overstated and the Conservatives understated.  Then there may be people who are unsure.  They may be disillusioned with the Conservatives and they may not want them back in power but they may still be unsure about Labour.  Even if they are happy with Starmer having a majority, they may think that he doesn’t deserve a landslide and they may see a need to still have a healthy opposition.  So, it is possible that the opinion polls may narrow.