How much should you trust polls for the general election?
By Joe Harker
Back in 2017 most opinion polls were predicting a significant Conservative victory for Theresa May, but YouGov called the election as a hung parliament and were proven right once all the votes had been counted.
Their MRP has become an event for this general election, something seen as an indicator of the way the vote will go in less than two weeks time.
The poll said the Tories would have a parliamentary majority with 359 seats while Labour would drop to 211 seats. That would be just what Boris Johnson wanted and a disaster for Jeremy Corbyn, but how trustworthy are polls?
YouGov's MRP got it right last time when everyone else got it wrong, so there might be something to their methodology that bears paying attention to.
They try to apply their polling to specific areas rather than just tracking how vote share ebbs and flows, as a surge in overall vote share doesn't mean a party is boosted by that amount everywhere.
MRP stands for Multilevel Regression and Post-stratification, which sounds very jargon-y but basically means the estimates get broken down into smaller areas to better measure how support might translate into seats.
It would mean something close to a wipeout for the independent MPs who jumped ship from their parties and are attempting to go it alone.
A lot of work goes into creating and refining a model that works for predicting general election results, the polls aren't perfect but they are worth paying attention to.
The Counter Claim:
It should be pointed out that there's still almost two weeks to go in the general election campaign. The MRP is a prediction of the result if the vote was taken now, a week is a long time in politics and anything can happen between now and then.
Polls are also just estimations. They can hit close to the correct result but one which calls everything right is exceedingly rare and almost a matter of chance.
When the data was being collected the manifestos weren't released and plenty of debates hadn't yet happened.
Swing voters often don't make their minds up until the last moments, so there's still plenty of time for the figures to change and the result to shift.
It wouldn't take all that much of a change for a Tory majority to become a hung parliament. As is always the case with general elections, the only numbers that matter are the ones that make up the final result.
If the MRP is to be believed Labour's defeat would be the second worst the party has experienced since the Second World War, only Michael Foot's loss to Margaret Thatcher in the aftermath of the Falklands War in 1983 would have been worse.
In 2017 the poll called 93 per cent of the seats correctly, which adds credence to the model.
The poll also says if the Tory lead falls from the around 11 point gap to about seven points then the chances of a hung parliament get much higher.
You can learn more about the MRP results here.