milo
Jack L. Jones
Just a quick glance away from Brexit...
The most accurate pollster for the last 2 General Elections was Survation. Their latest polling:
http://survation.com/labour-extends-polling-lead-8-points-conservatives/
New Polling
We have a new political poll, conducted online Wednesday to Thursday on behalf of Mail on Sunday indicating that Labour’s support is back to its post election high of 45%. With the Conservative party share drifting lower in the poll at 37% the net effect of these factors is a polling lead for Labour of 8 points.
State of the parties – December 3rd (changes vs Survation polling 4th-5th October)
LAB 45% (+1) CON 37% (-1) LD 6% (-1) UKIP 4% (NC) SNP 3% (NC) GRE 1% (NC) AP 3% (NC)
We’ve not seen such a lead for Labour in a Survation poll since late 2013. An 8 point lead would put the Labour party into overall majority territory if such vote share totals were reflected at the ballot box.
Are Survation the outliers again? unchanged methodology gives us confidence
Much of the polling being conducted currently in the industry at large has some form of amended methodology since before the General Election. The industry’s problems with probabilistic turnout assumptions (deciding turnout based on historical voting patterns) served to incorrectly suppress change in opinion – for Labour support in particular – which contributed to the overall overestimation of the Conservative lead.
As Survation’s final online and telephone polling for GE 2017 had the “correct answer” – a small Conservative lead over Labour and a hung parliament picture we have not needed to change our methodology for either mode as both produced accurate vote share estimates, with the final poll being the most accurate in the industry.
Our final pre-election online poll (June 3rd) produced a Labour figure of 39% while our final polling of the general election was by telephone and had a GB (ex NI) figure for Labour of 40.4 – both methods slightly underestimated Labour’s actual election performance of 41%. We also underestimated the Conservatives who took 43.5% of the actual vote vs our final poll of 41.3%.
Still not experimenting
Having an unchanged methodology does simplify things greatly if an avid poll watcher wanted to contextualise this current polling with that conducted ahead of the General Election – which is shown in detail below.
It also gives us confidence that Labour is indeed enjoying a decent lead over the Conservatives at the current time. There’s no experimental methodology at work here, the methodology is that which proved accurate at the election.
50% to 34% in favour of a second referendum too. I could be wrong but I think that is a big swing.