In Sunday’s Sun: lies, statistics and polls
Can public opinion polls suppress voter turnout? Of course they can.
Do they? Yes, a lot, a lot.
If a poll shows the electorate evenly split, turnout typically surges — because people feel energized by the notion they might help to cast a pivotal, deciding vote.
On this, studies abound.
In Israel, where tight races are common, a Hebrew University study concluded that “closeness in the division of preferences induces a significant increase in turnout.”
But, not surprisingly, the Israeli study also found that, if a particular political choice is way ahead, there will be “an important decrease in participation” by folks favouring that choice.
“Why bother,” to employ a less academic lexicon. “My vote won’t make a difference. He/she is going to win anyway.” Makes sense, right?
Closer to home, the experts find the same thing. One fairly recent study done for Elections Canada looked at vote rates in Canadian elections, and found that participation has been plummeting since about 1984.
One the biggest reasons for lack of turnout?
“A widespread feeling that political participation is meaningless,” the study found.
Which brings us, in a roundabout kind of way, to last week’s much-watched byelection in Brandon-Souris in Manitoba.
It was much watched because the Manitoba riding had been held by Conservatives for most of the last century. It was much watched because the Liberal party came fourth in the 2011 federal election, with a pitiful 5% of the vote.
And it was much watched because, the day before the vote, an outfit called Forum Research released a poll stating the Liberals had a nearly 30-point lead in the riding.
Thirty points! The resulting big headline in the Winnipeg Free Press: “Liberal candidate holds 29-point lead in Brandon-Souris byelection: Poll.”
Forum’s methods are “amazingly accurate,” said Forum’s media-friendly president, who became quite shy when this writer (a friend of the Liberal candidate, full disclosure) sent him a number of questions about his firm’s methods.
“Amazingly accurate.” That’s what he said.
Except, well, they weren’t. Not in Brandon-Souris; they were a joke. On election night, my Liberal pal lost by less than 400 votes — but, it should be noted, he lost.
About two percentage points separated the Liberal and Conservative candidates. Not 29.
Who paid for that poll, which landed like a bomb in Brandon-Souris, and mere hours before polls were to open? Forum didn’t say. Did the Conservatives? After all, Conservative backroomers are pretty smart, and they know all about the gist of the aforementioned academic studies.
They know that too many Canadians often regard their participation in elections as “meaningless,” and a poll that tells Liberal voters their guy has already won — well, such a poll might certainly have the effect of causing “an important decrease in participation,” to quote the Israelis.
But again, Forum — in other circumstances, always ready to provide a clip to the media — has become a bit less available, post-Brandon. They didn’t respond to my questions by deadline.
Is it important? Yes, actually, it is. Forum is the same polling firm that said the NDP would win a big majority in B.C., Wildrose would do likewise in Alberta and the Parti Quebecois would seize a majority in Quebec.
Wrong, wrong and wrong.
None of it would matter if people didn’t pay attention to media polls, and if their behaviour wasn’t affected by polls. But they do, and they are.
In the case of Brandon-Souris — and in the case of B.C., Alberta, Quebec and not a few other cases — I don’t actually blame Forum Research for this appalling situation.
I blame the media (including Sun Media, sometimes) who regurgitate bogus numbers without ever employing their critical faculties.
In Brandon-Souris, we will likely never know if a poll persuaded some folks to stay home. But, based on what the experts say, it seems likely.
If you care about democracy, that should worry you.
A lot.
