June 25, 2016 by AK
The Financial Times’ latest poll of polls had a Remain-Leave-Undecided ratio of 48%-46%-6%. The Economist’s tracker had 44%-44%-9% (not sure where the missing 3% went). The actual vote went 52%-48% for Leave. Were the polls misleading or misinterpreted?
First, the polls may have accurately estimated the breakdown among the public in general but not among those most likely to vote. Second, the undecided may have made all the difference on the voting day. In any case, the six percentage point discrepancy between the Leave-Remain difference in the FT’s aggregate poll and in reality – between -2% and nearly +4% – cannot be waved away as a statistical blip.