Political polling gets a bad press, but that's because we rarely look at the bigger picture and tend to treat each poll as a prediction. It's polling trends we should be watching closely and the trends right now look familiar.
The latest polling from Opinium (sample size 2,001) showed the gap between Labour and the Tories narrowing - the Tories down four points (41%) and Labour up three (29%) since their previous poll. It’s interesting that the Guardian chose to report the poll that way, whilst George Osborne’s Evening Standard spun the result as “Boris Johnson’s Tories tipped to be biggest party after 12 December.” I suppose that’s not that surprising. But it is way too early to make any kind of prediction - we are still in the first few days of the campaigns.
On the eve of the 2017 general election, the aggregated polling company data gave the Tories an 8 point lead. On election day, they managed a 2.4 point lead. Most (but not all) pollsters got it wrong. This is fairly universally considered to be because they weighted their data wrongly for the voter turnout for the different parties.
2019 will not necessarily be the same. The pollsters are all being really careful to say that this is probably the hardest to forecast election ever and doubtless they will be adjusting their weighting for turnout and other factors this time.
My image shows a couple of poll trackers (crudely lined up with each other on the same horizontal scale) - one from 2019 and one from 2017. I couldn't find two from the same source for both periods, but they both used a similar panel of polling companies. The 2019 graph (Britain Elects) excludes the SNP and Plaid Cymru, but this does not affect the headline findings.
What's interesting is that, in the lead up to the election being announced and the period following, the trends for both elections are thus far strikingly similar. The two biggest parties have seen their polling rise, with Labour closing the gap on the Conservatives. The smaller parties have seen their polling numbers drop. For the LibDems and the Brexit Party, this fall has been underway for several weeks, contrary to what their leadership would have us believe. If you look at the poll trackers from a variety of sources (sse at the bottom of this very long post for some links), you will see an identical pattern.
There are some specific issues, other than how well the party campaigns do, which I think we should be mindful of:
Because voter turnout is generally significantly lower amongst people on lower incomes, traditionally we have seen Labour penalised when turnout is low. In this election, I think it is reasonable to expect turnout to fall compared to 2017 (when it was high compared to the previous 20 years). There is a feeling over voter fatigue - this is our fifth national poll in four and a half years (three general elections, the EU referendum and the European Parliament election). But I think it would now be fair to assume that it will not just be Labour that takes the hit from a lower turnout. Labour's traditional working class voter base is already fragmented, with UKIP (and now the Brexit Party) and the Tories having picked up significant numbers of voters in the Labour heartlands.
In addition, the increase in young voter turnout has been and is likely to continue to be important. Labour's radical agenda in 2017 undoubtedly helped bring about a higher turnout amongst under 25s. But, as post-election polling showed, it was the Labour vote for 25-45 year olds that was most important to Labour's relative success in 2017.
This time, the LibDems, the SNP and potentially the Greens and Plyd will likely eat into Labour's vote in some areas because of the Brexit-Remain effect. But it's reasonable to expect Labour's strong offer on climate change and the environment will help sustain or increase turnout amongst the under 25s.
For me, there are two big lessons from polling here that I think are important:
Firstly, polls can be very misleading but that’s not surprising. Most are working with a fairly large margin of error and voting intentions are perhaps more dynamic than we have ever witnessed. It's a moving target.
Secondly, there is everything to fight for and helping get the vote out in the day is absolutely critical. If you do one thing to help your party do well, it should be helping to get people to the polling stations on the day, in whatever way. And if you want to be really targeted, help get young voters out especially.
One final observation: it's widely believed that the Brexit Party will disproportionately hit the Labour vote, particularly in the Labour heartlands in Wales and the Midlands and North of England. However, polling researcher professor John Curtice recently stated that he believes that the Brexit Party is likely to impact on the Tory vote more - by a factor of around two to one.
Here are a few useful sources:
Britain Elects poll tracker:
The Guardian's poll tracker:
The BBC's poll tracker:
Democratic Dashboard - where you can look up historic voting results in different constituencies:
Joseph Rowntree Foundation research on political engagement of engagement amongst people on low incomes in leave-leaving seats (which perhaps contradicts my argument about voter fatigue and turnout):