November PollsPoliticians are fond of saying that the only poll that matters is the one on election day, but in truth they can live and die by the polls. The consequences for party strategy and the future of a leader can be severe if polling numbers go bad. That is why political advisors must find it extremely frustrating that there is a substantial difference between what is found by Angus Reid (Vision Critical) and what is found by the Mustel Group.
In a telephone survey of 502 BC adults taken between November 4th and 15th, the Mustel Group found 42% for the NDP and 37% for the Liberals, a gap of just 5% - down from a 9% gap they reported with their September poll. In a web based survey of 807 BC adults taken between November 2nd and 3rd (completed before Campbell announced his resignation at 11:30 AM on Nov. 3rd), Angus Reid found 47% support for the NDP and 26% support for the Liberals, a gap of 21% - down from 25% in October.
For the 2009 provincial election, Angus Reid had the best electoral prediction out of five polling companies. Its final survey before the election put the NDP at 42%, exactly what the party received on election night. The May 2009 Mustel Group poll underestimated the NDP support, giving it only 38%; both firms were within 2 points of the BC Liberals' actual vote.
Both polling firms reported a disapproval rating of 45% for NDP Leader Carole James, but the disapproval rating for Campbell was 75% in the Angus Reid poll and "only" 60% in the Mustel Group's. It probably is not surprising that supporters of a party disapprove of the leader of an opposing party. What both polling firms failed to report is what they found with respect to disapproval ratings from those potential voters who support the political party of the leader they were rating. Until the BC Liberals pick a new leader, all the polling numbers should probably be taken with a grain of salt.