There’s a great analysis by John Curtice this morning on the latest Scottish polls particularly the one out today from ICM for the Guardian.
The whole article is well worth reading but the following is a key part when it comes to making Scottish Westminster seat projections:-
“..What, however, this poll brings anew to the evidence is an estimate of how well the parties are doing in different types of seats in Scotland. Inevitably, these estimates are based on relatively small sample sizes and thus have to be treated with caution. But what they do suggest is that there is no reason to believe that the swing to the SNP does not extend to what are supposedly the safest Labour seats in Scotland. If anything, the swing appears to be even greater in such seats. In seats where Labour is defending a majority of more than 25 points the swing in the poll from Labour to the SNP since 2010 is 24 points, rather higher than the 19.5 point swing for Scotland as a whole.
That means that, if anything, estimates of how many seats the SNP might win that are derived by assuming that the Scotland-wide movement uncovered by a poll would be replicated in each and every constituency in Scotland could actually underestimate the scale of SNP gains. In the case of this poll, projecting the Scotland-wide movement across the country as a whole produces an estimated seats outcome of SNP 45, Labour 26, Liberal Democrats (on 6% of the vote), 3 and the Conservatives (on 13%), 1. But if we take into account the difference in the movement in different types of seats then the estimate becomes SNP 53, Labour 3, and Liberal Democrat 3 (while the Conservatives emerge empty handed). In short, pretty much every Labour seat in Scotland has to be regarded as currently at risk of being lost to the SNP.“
A bleak picture for new Scottish LAB leader, Jim Murphy and Ed Miliband. Securing 3 Scottish seats rather than the 41 currently held completely changes everything for the party at the coming election.