Conclusions
Most predictions of election results make the assumption (implicitly, perhaps) that polls are unbiased. But the implications of of this are far from being merely psephological, they are also political. They drive the narrative and set the tone. Parties have ousted their leaders based on poll ratings.
Pollsters have clearly improved their methods, but as they do so, their task has progressively become harder. While 1992 stands out as the worst opinion polling failure in British electoral history, it is far from unique. It may never be repeated on the same scale, but it doesn’t need to be to cause a significant upset.
The models all suggest Conservative leads of over 6 points, pretty much consistent with intelligence from Atul Hatwal and Stephen Bush (close contests in seats that were close last time, when the Conservatives won by 7 points nationally). And while I would urge caution with respect to this analysis, it looks a lot like the Tories are set to outperform the polls once again.
