In many years of doing election polling and as a frequent client of research firms, one learns to be a bit sceptical. In fact, today, you have every reason to be downright suspicious. It appears that people are using selected poll results to try to stampede voters.
The voters have no reason to believe these pollsters who are trumpeting their polls. There is absolutely no way those polls can be accurate. Think about how you react to someone on the telephone asking you how you are going to vote. Do you graciously tell the caller everything they want to know? Are you kidding?
The problem today is that there is no way to get a truly random sample of Canadian voters. And if you could achieve that, it would still have to be a large enough sample to allow for breakouts of regional influences. The Bloc Quebecois in Quebec skew the figures there. If you include that 10 per cent of the vote with the rest of the country, it is not relevant. When you consider it can be 40 per cent of Quebec’s vote, you see why it can translate into more than 40 seats in the House of Commons. In the same way, the disproportionate Reform/Tory vote in the Prairies skews the national figures.
As the old chestnut goes, figures do not lie but liars can figure. To understand a poll, start with the bias of the person reporting it. Mind you, the person does not have to lie. All they have to do is be selective about the figures they give you.
With one-third of the House of Commons seats in the Province of Ontario, when have you heard in this election about polling in Ontario? That is because it is the toughest area in the country to predict. It is also the toughest area in which to get a statistically viable sample of voters. The pollsters do not want to expose themselves.
The current flurry of polls showing the NDP strengthened, are more a phenomenon of the news media than any real hope for an NDP breakthrough. This is not 1990 in Ontario. You have to be smoking some fairly strong stuff to imagine Jack Layton as Prime Minister. Jack is a small potatoes municipal politician who has not grown in his current role. Most of that supposed growth in NDP support is in Quebec where tired PQ voters are searching for some answers.
But what the pollsters fail to explain is that about 30 per cent of the potential voters are unlikely to go to the polls. Nobody can be sure who they are.
We used to. Back in our early days of political polling, we used to know how the voter would vote before asking them the specific question. It was done with qualifying questions. Their actual voting intent helped with the equation used to determine their probability of going out to vote. We used to do some amazingly accurate surveys.
The poll that interests us was where we lined up yesterday at the advance poll in our electoral district. There never used to be advance poll line-ups like that. It bodes well. There is going to be a good turnout of voters in this election. That is bad news for Mr. Harper. He wants sleeping voters, not voters eager to go to the polls. He wants a vote like 2008 when large numbers of Liberal votes did not make it to the polls. They have much more to vote for this time.
The most accurate election surveys are traditionally what are called exit polls. They are a quick survey done as voters are leaving the polling station. They are more likely to tell you the truth then.
– 30 –
Complaints, comments, criticisms and compliments can be sent to [email protected]