Why Adrian Dix and the NDP shouldn't freak out about the latest Forum Research poll
Today's big provincial political story is a new poll, which suggests the B.C. Liberals are narrowing the gap against the NDP.
Forum Research has reported that the governing party has the support of 35 percent of decided voters. The NDP stands at 39 percent.
As I read about these results in this morning's National Post, a few things jumped to mind.
First of all, there's the margin of error, which is at plus or minus three percent, 19 times out of 20.
The ThreeHundredEight.com site indicates that Forum is reporting significantly higher support for the B.C. Liberals than all of the other polls.
The next closest was 33 percent by Abacus, which polls for the right-wing Sun News Network.
Secondly, many polling companies have fallen into disrepute in recent elections for missing the mark.
For instance, Forum Research predicted shortly before the September 2012 Quebec election that the Parti Québécois was "headed to a comfortable majority". In fact, the PQ barely won a minority.
Two days before the Alberta election last April, Forum Research called for a "bare majority" government for the Wildrose Alliance. The Progressive Conservatives won a majority.
That's not all. Before the 2011 Vancouver municipal election, Forum Research head Lorne Bozinoff claimed that the NPA's Suzanne Anton had a chance of winning. She lost by nearly 20,000 votes.
With a track record like this, perhaps the B.C. Liberals might want to look upon this recent poll with some skepticism.
Thirdly, Forum relies on what it calls "an interactive voice response telephone survey".
It doesn't publish the rate of refusal in its news releases.
Anyone who ever took the time to read Margin of Error, which was journalist Claire Hoy's 1990 exposé on pollsters, knows that the rate of refusal is the industry's dirty little secret. Most people refuse to participate in telephone surveys.
Another problem with telephone surveys is that they often under-report the preferences of young people, who are more likely to vote NDP.
That's because young people live on their cellphones and often don't have access to landlines.
Angus Reid Public Opinon relies on online surveys, which are better suited for capturing the views of young people.
The last Angus Reid poll had the NDP at 45 percent, compared to just 31 percent for the B.C. Liberals.
Fourthly, there's the issue of motivation. Was the Forum poll driven by the B.C. Liberals or financed by its supporters with the goal of boosting support and rallying the base? It will certainly help the B.C. Liberals raise money in the final push toward election day on May 14.
My final concern about the Forum Research poll regards its overall conclusion.
The president, Bozinoff, determined that the NDP would win a one-seat majority with a four-percent lead in the poll.
That misses how the New Democrats have traditionally fared in B.C. elections.
Because the B.C. Liberal vote has been heavily concentrated in its strongholds—such as Richmond, the North Shore, and wealthy ridings like Vancouver–Quilchena and Vancouver-Langara—the B.C. Liberals sometimes require a higher percentage of the popular vote to achieve as many seats as the NDP.
This was most apparent in 1996 when the NDP won a majority of seats even though the B.C. Liberals captured a greater percentage of the popular vote.
The NDP vote is more efficient in that it's distributed in a way to maximize the number of seats.
Bozinoff is based in Ontario. For him to suggest that the NDP would win a one-seat majority with a four-percent lead in the popular vote tells me that he's not a scholar of B.C. politics. If the two parties were tied, the NDP would still likely win a majority—and it would be greater than by one seat.
Then again, Bozinoff didn't exactly hit the mark last year in Alberta or Quebec, either.
I would rather hear about a few more polls before I'll place a great deal of stock in his most recent prognostication.