First thoughts from ComRes on why the Liberal Democrat vote share last week, although up for the third general election in a row, was much lower than the polls had been predicting:
We along with many others were surprised by the eventual Lib Dem figure; but that’s not the whole story….. our Conservative Party vote share was, in the words of the BBC’s analysis, ‘bang on the button’ while we understated Labour support to the tune of 1.14% (incidentally the first time the Labour vote share has been understated since 1983). Along with every other polling company we overstated Lib Dem support by some 4.5%.
We’re analysing what happened there but our initial view is that it is most closely related to one key factor – turnout. We were expecting 70% to 71% whereas it was 65.1% and much of the flakier, less certain vote was being indicated as Lib Dem support. As for the Labour understatement, there is some evidence of late swing, and it is indeed likely that some of that stated LD support ended up voting Labour while a lot more stayed away altogether on Thursday. In short we had Con and Labour well within the margin of error, and the LD figure was 1.5% outside of that.