The August Poll

August 19, 2004

August Poll as Expected

As expected the Mustel Group's August poll came back into line with what Ipsos-Reid found in July and what Mustel found in its February, April and May polls, namely that the Campbell Liberals are in a dead heat with the NDP, 40% Liberal, 42% NDP. Rather than admitting that their July results were likely wrong, skewed by being too close to the federal election or simply in error, the Mustel Group issued a news release with the headline "Support for the BC Liberals on the Rebound".

Polls Sept 2001 - Aug 2004 The only thing the BC Liberals' support rebounded from was a poll that was likely bad in July. Nevertheless, Evi Mustel claimed that "The recent positive economic news and the prospect of an NDP government may also have driven some voters back to the BC Liberals" before adding that recent polls are volatile. That sounds straight off the Campbell song sheet. Recent economic news consists primarily of slightly improved forecasts which are an improvement relative to last years' forecasts but not as good as the actual economic experience in the 1990s. Payroll employment is growing more slowly than in did in the three years before the election, 1.37% compared to 1.77%. As for the prospect of an NDP government, look at the chart of polling results since the last election. The NDP has shown steady growth. Since the July Mustel poll the Campbell government and BC Ferries were in the news almost daily taking heat over their "blatantly stupid" decision to build three super C class ferries offshore. It is inconceivable that the Liberals rebounded 7 points during that time, but it is conceivable that they never really fell to 33%.

August 9, 2004

The August Poll

With just nine months to go until the May 17, 2005, election public opinion polls will generate more interest. The two major polling companies, Mustel Group and Ipsos-Reid, have shown roughly similar results over the past three years, at least they did until last month. Provincial polls 2001-04

That is when Ipsos-Reid placed the NDP and the Campbell Liberals in a dead head, but the Mustel Group showed a 12 point lead for the NDP, 45 to 33.

The Mustel Group was "in the field" again last week and their results will be made public in the next few days. Their August poll should be considered relative to their May results, not relative to what may have been an abnormality in July. In May they showed the Campbell Liberals slightly behind at 39 to 42. If the Liberals fall and the NDP gain relative to the May results, then New Democrats can take satisfaction that the right course is being followed for a competitive election nine months from now. If the July results are confirmed, the knives will be out to dump Campbell as leader before then. If the August results are somewhere between the May and July numbers, the Liberal spin machine will talk about a recovery and how their message is being heard, even though the only validation will be measurement relative to a rogue July poll. The following table shows the actual percentage of the popular vote in the past 8 elections. The size of the legislature changed over the last 32 years, from 55 seats in 1972 to 79 today, but the NDP's percentage of the popular vote was reasonably constant until 2001. The best the NDP did was 46% in 1979; that is another reason why it is hard to believe last month's poll which put the NDP at 45% and more reason why the next poll should be judged relative to May's poll and past elections.

Of course, the only poll that counts is the one on May 17, 2005. Between then and now attempts will be made to project seats based on estimates of the split in the popular vote. It may be reliable 19 times out of 20 to forecast the split in the popular vote with a margin of error of 3.5%, based on a sample between 500 and 800 potential voters, but taking those results and projecting the distribution of seats is not scientific. Very large sample size polls and large individual constituency polls are conducted by political parties, and those results are seen by very few people, and are never published.

As interesting as polls may be, the results of the last federal election should reinforce the point that they are an amusement that should not distract from a thorough discussion of the issues. August is a slow news time so the soon to be released poll may receive a disproportional amount of attention. Time might be better spent looking at the costs and mistakes of the Olympic games in Athens in order to avoid the same in 2010. During Premier Campbell's stay in Athens, he should reflect on the blank cheque he signed for all cost overruns in 2010, something he wouldn't offer for BC's shipyards. That may have more impact on the vote in 2005 than any of the spin around interpreting what a few hundred people say when they answer a pollster.