The mid-week polls so far are below:
SavantaComRes (25th-26th) – CON 41%(-1), LAB 34%(+2), LDEM 13%(-2), BREX 5%(nc)
YouGov/Sky/Times (25th-26th) – CON 43%(-1), LAB 32%(+2), LDEM 13%(-3), BREX 4%(+1)
ICM/Reuters (22th-25th) – CON 41%(-1), LAB 34%(+2), LDEM 13%(nc), BREX 4%(-1)
Kantar (21st-25th) – CON 43%(-2), LAB 32%(+5), LDEM 14%(-2), BREX 3%(+1)
Survation/GMB (20th-23rd) – CON 41%(-1), LAB 30%(+2), LDEM 15%(+2), BREX 5%(nc)
Taken individually, almost all the changes in these polls are within the margin of error (Kantar is the only exception). However, looking at them as a group there is a clear trend, with every poll showing a slight drop in Tory support and a slight increase for Labour. Taken together it’s clear there’s been a slight narrowing of the race though, of course, that still leaves a Conservative lead between 7 and 11 points. As usual, it is almost impossible to ascribe specific causes to this.
As well as the standard polls this week, YouGov published their MRP model. MRP is a method of using a large national sample to project shares at smaller geographical areas – in this case Parliamentary constituencies. By modelling how different demographics vote in seats with different characteristics, and then applying that model to each constituency, the MRP model produces vote shares for each individual constituency and, via that, projects seat totals for each party. Famously the YouGov MRP model projected a hung Parliament in 2017 when most people expected a Conservative majority.
The model this time is less surprising – it projected national vote shares of CON 43%, LAB 32%, LDEM 14%, BREX 3% (so very much in line with YouGov’s traditional polling), and seat numbers of Conservative 359, Labour 211, SNP 43, Liberal Democrat 13. This represents a Conservative majority of 68, much what we would expect to find on those shares of the vote (though the detailed projection is interesting, with the Conservative gains coming largely in the North and the urban West Midlands, with notable gains in West Bromwich, Wolverhampton and Stoke). Full details of the MRP model are here.
Finally this week, we’ve seen what is only the second Scottish poll of the campaign, this time from Ipsos MORI. Topline figures with changes from the 2017 election are CON 26%(-3), LAB 16%(-11), LDEM 11%(+4), SNP 44%(+7). Tabs for that are here.