News

The Harper Conservatives will enter a federal election campaign with a polling lead that puts them within striking distance of winning a majority government, according to a new survey taken on the eve of an expected vote.

The new poll for The Globe and Mail-CTV News finds Canadian voters satisfied with the direction of the country and significantly more confident in the leadership abilities of the Tories and Prime Minister Stephen Harper than they are in those of his main rival, Stéphane Dion and the Liberals.

According to the poll, conducted by the Strategic Counsel, 37 per cent of Canadians would opt to vote for the Tories were an election to be held today, compared with 29 per cent for the Liberals, 17 per cent for the NDP and 9 per cent for the Green Party.

What questions do you have about the latest numbers? What do you want to know about polling? Are you curious about how voter preferences today will impact the next election, which could be called any day now?

We are pleased that Tim Woolstencroft, managing partner of the Strategic Counsel, the firm that did the poll for The Globe and CTV, joined us online to answer your questions.

Questions and answers appear at the bottom of this page.

As one of the founding partners of The Strategic Counsel, Mr. Woolstencroft's research experience spans both public and private sectors.

His expertise regularly involves him in a broad spectrum of research projects, particularly in the financial services, government, and telecommunications sectors.

Mr. Woolstencroft has undertaken long-term comprehensive brand positioning and tracking for clients such as Bank of Montreal, CN North America, Microsoft and the Government of Canada.

He has been in the market research field for close to 14 years.

After earning an MBA from the Ivey Business School at University of Western Ontario, he joined Decima Research as vice-president in the late 1980s, moving to Toronto from Ottawa, where he had been executive assistant to the federal Minister of External Relations.

Editor's Note: globeandmail.com editors will read and allow or reject each question/comment. Comments/questions may be edited for length or clarity. We will not publish questions/comments that include personal attacks on participants in these discussions, that make false or unsubstantiated allegations, that purport to quote people or reports where the purported quote or fact cannot be easily verified, or questions/comments that include vulgar language or libellous statements. Preference will be given to readers who submit questions/comments using their full name and home town, rather than a pseudonym.

Brodie Fenlon, globeandmail.com: Thank you for joining us Mr. Woolstencroft. There has been so much interest in today's discussion that I'm going to turn it over right away to our readers. We'll try to get to as many questions as we can.

Dave Roberts from GTA Canada writes: Hello Mr. Woolstencroft. Could you comment on how likely a voter is to actually go to the polls and vote relative to each party? For instance, is someone who identifies himself as an NDP voter more likely to actually vote than someone who identifies as a Liberal voter etc.?

Tim Woolstencroft: Dave, we don't have the specific answer to your question. But we do know that younger people are less likely to actually vote than older people. The Green/NDP/Liberal parties tend to be more reliant on the voter support of younger people, while the Conservatives are more dependent on older voters.

Greg King from Ottawa writes: I'm always keenly interested in how language is used in politics and especially opinion polls. How a question is phrased is extremely important to the answers that you will receive and the results of a poll. Can you let us know the exact phrasing of the questions that were posed to individuals in this poll?

Tim Woolstencroft: Greg: you are absolutely right. This is a critical consideration. We spend a tremendous amount of time in developing questions that are phrased in a neutral manner. We also use questions that we have developed and tested over many years. We suggest that you visit our website where you will see the specific questions that we used in this latest poll.

John Northey from Georgetown writes: Have you studied where the Green Party vote is coming from? As a former PC/Reform voter I know I have been turned off by the current Conservative party (broken promises cause that) and since the NDP doesn't fit my views and I have no trust in the Liberals I have shifted to the Greens as have many other former PC voters that I know. This is the opposite of what I hear via the media so I was curious if polls matched my personal experience.

Tim Woolstencroft: John, the Green Party appears to be drawing support from all of the major parties. If we look at second choices of Green party voters, it is clear that most of them would go to either the Liberal or NDP parties. A smaller proportion of the Green voters indicate that their second choice would be the Conservative party — many of whom (I suspect) are Red Tories. Our polling suggests that one of the challenges facing the Liberal party in this election is that liberal/progressive/left-wing voters are more fractured across three parties, while the conservative voters are more consolidated under the Conservative banner.

Bob Onyschuk from Toronto writes: What kind of data do you have on the 416 area from this poll?

Tim Woolstencroft: Unfortunately, given the relative proportion of the Toronto population to the national population, our sample size for 416 is too small to report on and publish. But, when we do look at the data for Toronto, we do find that 416 continues to be a Liberal fortress with little change since the last election. The dynamics of the election campaign may cause a shift in these initial findings.

Michael Sharp from Victoria writes: The recent CROP poll in Quebec has the CPC ahead there. Do you believe declining support for the Bloc will translate into more support for the CPC in Quebec?

Tim Woolstencroft: Michael, we are seeing a fairly significant drop in support for the Bloc in Quebec. As we noted over a year ago, the Conservatives have emerged as the dominant federalist party in Quebec outside Montreal. Our latest poll shows that the Conservatives currently have the best opportunity to capitalize on declining Bloc support outside Montreal. Interestingly, we are also finding Dion's Liberals have an opportunity to exploit the shift away from the Bloc in Montreal where the Liberals have traditionally been very strong. In addition, our latest poll also shows that francophone voters are starting to be more open to Mr. Dion suggesting he may be able to attract additional support for the Liberals because he is a Quebec native son.

Geoff Ondercin-Bourne from Hamilton writes: Although your poll appears to be good news for the Conservatives, I think it could be more of a blessing for the Liberals, because they could argue more convincingly for strategic voting. Whatever their weaknesses, how successful do you think the Liberals would be at stopping Stephen Harper from achieving a majority, given your numbers? A concerned New Democrat

Tim Woolstencroft: Geoff, it really depends on the effectiveness of the Liberal campaign. The Liberals need to aggressively increase the risk of voting Conservative. In effect, the Liberals need to make trusting Harper with a majority "the" issue of the campaign. This will mean the Liberals using strong and aggressive negative advertising to drive down the Conservatives's favoruable ratings. In this way, the Liberals will be able to consolidate progresssive/left-wing voters to support them. If the Liberals don't do this, there could be an opportunity for the NDP because many progressive/left-wing voters will be more comfortable voting for their first choice, the NDP.

Darren Moore from Belleville writes: The Canadian voting public seems to be very good at electing the government we want; during the last election, popular support had fallen for the Liberals, but we were still uncertain about Harper's agenda, so we elected a minority Conservative government. Aside from perhaps the first elected Green MP this time around, what are the chances that we end up with something very, very similar to what we have now?

Tim Woolstencroft: Darren, there is a high probability that we are going to end up with the results that are comparable to 2006. However, we do know that campaigns matter. Things happen. Unforeseen events can impact the dynamics of an election campaign. Just look south of the border and how the McCain VP choice is causing a significant upheaval in the Presidential campaign. In the last two election campaigns in this country, we have seen that events (for example the RCMP income trust letter) have had a huge impact on the election outcomes. So, we suggest that this forthcoming election campaign will probably be no different. So, let's hang on to our hats.

Chris Borst from Hamilton writes: Thanks, Tim, for taking our questions. Recent polls have certainly raised some! This summer's polls have been consistently showing the Conservatives and Liberals in a dead heat, and recent releases from Harris-Decima and Nanos continue to do so. The Strategic Counsel's polls showed a much larger gap back in the spring and this newest release seems to be returning to that pattern - though the most recent Ipsos-Reid and Angus Reid releases seem to point in the same direction. How would you explain the divergent results from recent polls and particularly the Conservatives' consistently greater strength in your own polling? As well, the Globe has headlined your latest as 'the brink of majority', yet the numbers seem to be almost exactly those we saw in the last election. Are there actually signs that something has changed or is the Globe's headline simply positive (from a Tory perspective) spin? Thanks! Likewise, William Friesen from Winnipeg writes: Poll after poll, including several last week by other reputable firms have shown that the Conservatives and Liberals are in a statistical tie. Your poll is the only one that shows a gap between the parties. How can we know if your poll is just an anomaly or if in fact it really is telling us that the Conservatives have a significant lead over the Liberals?

Tim Woolstencroft: Thank you for your questions. Chris and William both asked about why some polls point to a statistical tie, while our poll and others indicate a Conservative lead.

There are essentially two reasons for this. The first is essentially a symptom of the margin of error, which in most credible polls is usually in the ballpark of +/- 3 per cent in 95 per cent of a theoretically infinite cluster of samples. This means that a poll showing the Tories at 37 per cent and the Liberals at 29 per cent is statistically no different from a poll showing the Tories at 34 per cent and the Liberals at 32 per cent. It's important to remember that a statistical tie is just that — it is statistical.

The second reason for differences is a result of when the data was collected. It is important to note when the data was collected when looking at a poll. One of the polls that showed a much closer fight between Liberals and Tories was collected earlier than ours. We were in field collecting data over 6 days. We did see a shift over those six days in the numbers favouring the Conservatives after their pre-election advertising kicked in.

So, how should we interpret these polls on an eve of an election call? Our poll suggests that a majority government is within reach of the Conservatives, but yet to be within their grasp. The other polls suggest that that race between the Liberals and Conservatives is modestly tighter. We are sure that we are going to see additional movement in the numbers over the next several days and weeks as the election campaign gets started.

R. Carriere from Maritimes writes: Good Day, While polls are always interesting, I do have problems with 'national' polls where provincial and regional sampling is small, and where the margins of error are extremely high. That said, a competing polling firm in Quebec, polls voting intensions of Quebec residents ONLY- broken down into specific voting regions that represent and better reflect potential seat distributions. Why don't we see this type of survey for Ontario. This would better reflect voting intensions by region-perhaps breaking Ontario into 4-5 specific regions.

Tim Woolstencroft: You are right — we live in a country where there are significant differences in the political cultures between regions and within provinces like Ontario. More regional polling would be interesting and revealing. However, the reason this does not occur very often is due to the limited resources/budgets of organizations who commission these polls.

Don Portz from Trochu, Alberta writes: I always question the accuracy of polls in any given situation, especially that of politics. When conducting polls at various times do they follow the same format. i.e. do they cover the entire country based on per cent of population, province, territory, number of seats etc. and is it consistent? Given the wide differences in people and economies today, could it not be inconsistent if either say most of the polls came from Toronto or alternatively from Alberta? Just interested in the procedure.

Tim Woolstencroft: Don, this is a question that is often asked. We use a rigorous process to ensure that our national samples representative of the country and each of the different regions of the country. This methodology is consistently adhered to by all of the reputable national polling firms in this country

W M from Canada writes: Tim, inasmuch as the popularity of sitting governments often increases during recesses, the Conservatives have had most of the media attention for the past few weeks and have been getting their side of the story out pretty effectively, and they have also been the only ones running campaign ads for the past 2 years, including intensive coverage for the past few weeks, is 37 per cent really good news?

Tim Woolstencroft: The tranquillity of the summer has been good for the Conservatives after a turbulent spring when the Government was under constant attack from the Opposition. It is also clear that Harper did not want to face the House of Commons which, in effect, gives the Opposition a platform to hit at the government. This explains why we are having a snap election after the summer. So, 37 per cent is probably pretty good news for the Conservatives, but the more revealing numbers will be in a few weeks once the dynamics of the election campaign have kicked in.

D B from Canada writes: Hello, I was hoping you could comment a bit more on the method of your survey. The Globe article only says that 1,000 people were polled. How did you select these people and what steps did you take to ensure that they truly constituted a standard random sample of all Canadians? How did you reach the conclusion that the Conservatives may win a majority? The numbers reported were 37 per cent Conservative, 29 per cent Liberal, 17 per cent NDP, and 9 per cent Green. The reported error is 3.1 per cent at a confidence level of 95 per cent. I presume that the unreported 8 per cent is a combination of Bloc Québécois and undecided. How does 37 per cent with an error of 3.1 per cent lead to a Conservative majority? Even with the error, the data presented suggest no more than 40.1 per cent for the Conservatives. Is a Majority assured with 40.1 per cent of the popular vote? This conclusion is not apparent. If the error is assumed to be in the other direction, the Conservatives will get 33.9 per cent and the Liberals 32.1 per cent, which could lead to a minority government for either party. The only conclusion that I can draw from the data presented is that the Conservatives will likely have a minority government. Could you please share with us how you interpreted the same data to conclude that the Conservatives may win a majority?

Tim Woolstencroft: D.B. you make some really great points. I'll start with your first question about how we collected the sample. First, respondents were selected randomly. This means that if single women living in houses represent 5 per cent of the population, then a single woman living in a house will have a 5 per cent chance of being selected for our sample. This is the basis of random probability theory. A proportional number of respondents from each province were selected.

With respect to your second question ask how we can infer that the Conservatives are on the brink of a majority. Given the existence of five parties, 41 per cent of the popular vote would certainly be sufficient for the Conservatives to win a majority. Don't forgot that Mr. Chretien won a majority with 38 per cent in 1997 We also factor into our analysis where the Conservatives are drawing their support. If their support is increasing in battleground areas, we can assume that they are looking pretty strong. In the last election, 15,000,000 people voted, but 10,000 people in battleground ridings held the keys to government.

Brodie Fenlon, globeandmail.com: Thank you again to Mr. Woolstencroft for his answers and to our readers for all of your questions. You can continue the discussion now on our comment board .