Even a poll-obsessive like me would have to admit that during election campaigns we often spend too much time focusing on the election ‘horse race’ – who is up and who is down in the polls – on an almost hourly basis.
Even in this election, where Labour has been well ahead of the Conservatives from the outset, the focus has been on how much.
Figures for the Labour lead span a huge range from 16 to 25 points. And now there’s talk of “crossover”.
Has Nigel Farage led Reform past the Tories into second place for the first time in a general election campaign?
There’s a 10-point difference between pollsters in the figures reported for Reform UK starting at nine points and reaching 19.
The unprecedented nature of what the top end of these results suggest means we could be heading for a seismic reordering of the political landscape on 4 July. Alternatively, many pollsters could end up with egg on their faces.
Read more
What are in the party manifestos?
Tory candidates are ditching party branding
Check our poll tracker here
What lessons can we learn from polling at previous British general elections?
Perhaps the most fundamental thing is that polls become more informative about the result as election day nears.
Time brings greater accuracy
Simply, the difference between how a party is polling and its actual vote falls by about two points on average over the final 100 days.
Of course, some campaigns see much bigger shifts in the polls while others hardly move at all.
Remember that many people – around one in five – will cast their vote by post, long before polling day. Even if the race tightens, it may be too late for some to change their minds.
The key question to keep in mind when considering the latest polls is whether and how much they move during the campaign and the possibility that they could be getting a particular party’s vote too high or too low. Recent elections offer some salutary lessons about whether we should trust the polls.
In 2019, both Boris Johnson’s Conservatives and Jeremy Corbyn’s Labour made substantial gains during the election campaign – peeling off support from the Brexit Party and Liberal Democrats respectively by firming up their appeals to Leave and Remain voters.
Polling anomaly in 2017
The final polls did a pretty good job of capturing voters’ intentions, being less than a point and a half out from the result across all four parties.
The 2017 election will forever be associated with Theresa May’s imploding campaign and the surge in support for Labour under Jeremy Corbyn.
Labour’s gains in the polls were exceptional, so shouldn’t be treated as a likely blueprint for future dramatic comebacks.
Unusually, the polls underestimated Labour’s vote by around five points – in part because of the very late momentum in public opinion but also as many pollsters had implemented methodological changes designed to correct the polling miss at the previous election.
In contrast, the 2015 campaign was notable for the relative stability of the polls – and the big miss on the Conservative vote in the final ones published.
Cameron surprise
David Cameron defied expectations and won a majority after five years of coalition government.
If you had put your faith in headline voting intention in 2015, the result would have been quite a shock to you. Those who looked at other indicators like ‘best prime minister’ or ‘best party on the economy’ would not have been so surprised by the outcome.
Historically pollsters have tended to be too low for the Conservatives and too high for Labour. The 1992 and 2015 elections were notable for the size of the polling miss – with each leading to inquiries.
There are exceptions to this – most recently 2017 as we have discussed, but also 2010 when support for both parties was underestimated.
Tories in trouble
Lots of comparisons have been drawn between now and the lead-up to Labour’s 1997 landslide. Yet Labour’s lead declined over that election campaign, while the final polls over-estimated the Labour vote.
While a few recent polls have shown a dip in Labour’s vote, overall Labour’s lead has been relatively stable and is now ahead at the same point in the campaign.
There is also no guarantee that the polls are overstating Labour’s lead as they did in 1997.
If current trends continue, the Conservatives could be on course for a far worse defeat than 1997.