The latest Angus Reid poll released today, though conducted a week ago, has the NDP at 50% support. I think this is a first for the NDP in BC in a poll, though they may have been at the level in 1991 while Bill Vander Zalm was still premier.
I have my concerns about the methodology of Angus Reid. The took 802 people from their pool of online panelists. The problem is how can the company be certain the respondents reflect public opinion?
In three elections in the last year, federal, Ontario and Alberta, there have been some significant discrepancies between the polls and the election results and it seems that the online panels are skewing low on support for the government.
In the 2011 federal election Angus Reid seemed to be high on NDP support and low on Conservative support. In comparision, EKOS tended to be lowest of all the pollsters on the Conservatives but high for the Greens and "other" and just a bit high on the Liberals.
In the Alberta election Angus Reid only released one poll, so it is hard to known if they have a trend going on. They did not seem to be far off what the other pollsters were coming up with, but with only one data point it is hard to know. Their one poll in Manitoba was close to the final result
Their biggest faux pas was during the Ontario election. They released a poll on October 4th that was significantly lower than all the other polls released for Liberal support. They then released a poll the next day that showed a reversal of the support for the PCs and the Liberals which was closer to final result. In general the polls using online panels (Angus Reid, Abacus Data, and Leger) seemed to have lower numbers for Liberals and higher numbers for the PCs.
If I had the time I would do more rigorous data crunching on different polling methods, but on the surface of it I think there is a tendency of the online panels to underestimate the support for government.
All that said, the NDP still has a commanding lead in public support in BC even if Angus Reid is showing higher NDP support and lower Liberal support than is actually the case. If the NDP is leading by 15-20 percentage points on election day next year, there will be an NDP landslide. There have only been a few elections with more than a 15 percentage point gap between the winners and losers