Polls

I watched with debated breath for the polls results to tell me who won the leaders’ debate Tuesday night. Not that I planned to use the information in my scheme to vote strategically, but I wanted to be assured that my perceptions of the debate were not skewed by my personal bias. My wife was impressed by the quality of all of the leader’s French on Monday evening. I gather she took little stock in any of their promises but said any one of them was easier to listen to than the former Prime Minister.

Should we decide how to vote based on polls? If a party is reported to be leading in the polls, would you vote for that party’s candidate in belief that you will elect a member of the governing party? Are you afraid of a minority government and so plan to caste your vote to avoid that scenario?

If polls truly reflected the wishes of a percentage of the people any number of times out of a hundred, we wouldn’t need to vote. The problem is that too often the questions in the polling are skewed or biased towards the desired answer for the person or group who sponsors the poll. Good polls are supposed to be designed bias-free, but for instance, if the Liberal Party is surveying on how they are governing, you can be sure they are not going to have questions whose answer would put them in a bad light.

The people who design the polls have most certainly taken courses in political studies and statistics and English hyperbole. They know which buttons to push to get the results they, or their sponsor, wants. While integrity of the polling company is at stake, they have to get results or they will soon be out of business. Unfortunately, we only hear the results of the polls – we don’t often get to see the questions. The crux of the problem is in the question(s) asked to a person who has little time to reflect on the question, let alone the answer.

For instance, after watching the leader’s debate, you get a call from a pollster who asks, based on the debate, “Which person do you think would make the best Prime Minister -Jack Layton, Paul Martin or Stephen Harper?” Your answer might be the loquacious Layton – with the proviso that he change parties, but you only get to say a name. The results say that 4 out of 5 Canadians think that NDP Leader Layton would make the best Prime Minister, 99 times out of a 100. The question did not address policy or party but the individual characteristics of the three leaders. Moustachioed Jack is going to be disappointed on June 28 if he believed that poll.

Finding the answer to bias in poll results is the responsibility of the media. Before they report the news, the media is going to have to do a little background work. They are going to have to ask for the questions and they are going to have to ask to see the demographics of the polled people. The one person polled in northern Ontario may give a different answer about bears than the 50 who live in Toronto. Females may have a different perspective on a question than males, older people different than younger people, immigrants from native people, and university graduates from high school drop-outs.

Pollsters survey the demographic body in a statistical method, but lately the size of the poll has come into question by some media watchers. Can you accurately gauge the whole of Canada by asking a few questions of 1,600 people? Oddly enough, you can get a close approximation. In recent elections, the media can show who has won the election long before most of the polls have been returned. So mathematical models do work on a large scale, but you can be assured that in any close race, it will be the votes counted that decide the winner.

Should we, and our politicians, rely on poll results to make decisions? When Baytoday recently asked if people would support a buck-a-bag garbage fee to support at new recycling plant, most answered ‘yes’. The question should have been – a) do you support a new recycling plant; b) should it be funded by a buck-a-bag; or c) funded through a general levy. But now we have too many answers for a quick thumb-nail survey to determine a clear result. Obviously only people who read Baytoday would answer this survey, and while they may represent the best informed of our populace, I wouldn’t want to bet the farm on the answer. We know this was not a scientific poll, but this answer will be filed away in some councillor’s mind.

Angus Reid, the veteran pollster, in a CP interview, said that it is becoming more and more difficult to poll the public. Up to 80% of people answering a poll call will not participate in the poll. It would be most interesting to know why that 80% did not want to participate. Of the people who do agree to answer the poll, how many answer truthfully? Even though we may have studied the platforms, we still rely on polls to give us a sense of comfort that we are joining (or going against) the majority of our fellow citizens.

As far as the leader’s debate, I have my answers ready. Best debater? Gilles Duceppe, hands down – in both languages, both nights. Impressions: Harper is not as scary as the other parties are trying to make him appear; Martin has same old broken promises- not credible; Layton no chance of forming government but will enjoy seeing him in Commons; Duceppe will hold the balance of power in a minority government and this may be good for all provinces getting more from Ottawa.

Polls are today’s referendum and like it or not, polls are setting policy and direction for our government. If the pollster selects you (1 of 1,600) as the one respondent for Northern Ontario for the survey, think carefully about the question and if in doubt, become one of the ‘undecided’. With enough undecided responses all of us may get a chance to express our wishes as the polls then lose their influence.

They really do call 1,600 people, don’t they?