OPINION POLLS ARE JUST THAT — PEOPLE’S OPINIONS. AND THEY ARE RARELY ACCURATE. BUT SHOULDN’T GOVERNMENT POLICY BE DRIVEN BY MORE THAN THAT, ASKS ANDREW COYNE — LIKE ACTUAL DATA.
Trust the facts, instead of results from a survey
Jean-Yves Duclos was a well-regarded economist before he went into politics. As Minister of Families, Children and Social Development he has generally lived up to the advance billing: the government’s landmark reform of child benefits, for example, was his file.
And yet that reputation took a bit of a hit the other day. The minister had been sent out to soften up the media in advance of next week’s budget, specifically on that staple of Liberal rhetoric, the struggling middle class. It did not go well.
The centrepiece of his presentation was an incomprehensible chart with three boxes marked Middle Class, Growth and Trust, connected by arrows with labels like Feed, Support, Inclusive, and Engine. As reporters scratched their heads, the minister moved on to a chart that claimed to show how “inflation-adjusted wages have stagnated.” Alas, it in fact showed median real wages have been climbing steadily for the last 20-odd years.
If there’s any struggling going on, it’s the Liberals’ struggle to keep the stagnation thesis aloft in the face of the facts.
Median household incomes have risen fully 25 per cent after inflation since the early 1990s. The poverty rate, as measured by Statistics Canada’s Low Income Cut Off, is now at the lowest level since the agency began keeping statistics. The share of income going to the richest one per cent has been declining since 2006, and is now back to where it was in 1998.
All three of these indicators, it is true, showed very different trends in the 1980s and early 1990s — not coincidentally, a period bracketed by two of the deepest recessions in our history.
The question is why the public should be menaced with these long-ago events as if they were part of present reality. If the situation were the reverse — if things had been getting better then, but now were getting worse — would anyone be telling the public, “never mind how badly off you are, think how great things were in your parents’ time”?
If the minister had any new evidence to rebut this, he did not produce it. What he did have was survey data. One chart lamented that “many Canadians feel the next generation’s standard of living will be lower,” based on a Nanos poll taken last month. Another, based on an Ekos poll last year, showed a declining percentage of Canadians identify themselves as middle class. The label: “Canadians are increasingly feeling left out of the middle class.”
Liberals used to take a dim view of this sort of perception-based decision making. When the Harper government claimed it didn’t matter if the official statistics showed crime rates falling to their lowest levels in decades, because people felt as if crime was rising, Liberals rightly scoffed. Now a similar fact-free feeling — the middle class is getting nowhere — is the foundation of their whole economic platform.
Liberals are by no means the only ones playing this game. Rather than answer questions raised by her signature proposal to subject every refugee, immigrant or tourist to a quiz on their belief in “Canadian values” — questions such as why this is needed, what it would accomplish, and what it would cost — Kellie Leitch refers to polls showing sizable majorities of Canadians support the idea.
Likewise, those raising the alarm over Motion 103, unable to answer how a parliamentary motion with no legal force or effect could restrict free speech, have lately taken to citing polling data showing a majority of Canadians with varying concerns about the motion.
It’s easy enough to gin up a poll in support of just about anything, of course, depending on how you ask the question. The people waving them about today are in many cases the same ones who not long ago were railing ago about all the pollsters who failed to call Donald Trump’s victory (in fact, they called the vote to within a percentage point: Clinton beat him by two points, instead of the three points in the consensus forecast).
But let’s suppose these polls are genuine reflections of current public opinion. That’s a good answer to the question: what does the public think on these issues? It’s no answer at all to the question: are they right to think so? Yet that is how they are invoked: if that’s how the public feels, it must be true.
Skeptics are challenged, in tones of indignation: what, so you’re saying that millions of Canadians … are wrong? Well, yes. What of it? “Millions of people” are quite capable of believing things that aren’t true, particularly on matters to which they have given very little thought and with which they have little personal experience. The political science literature is filled with examples of people cheerfully offering their opinions to pollsters on entirely fictional events and people.
Climate skeptics rightly make the point that the overwhelming consensus of expert opinion on global warming is not enough, in itself, to prove it is right. Science is not a popularity contest: throughout history, individuals have stood against conventional opinion, and been vindicated, But let 1,340 randomly selected Canadians have their dinner interrupted to answer a question from a telemarketer about a subject they’ve barely heard of, and suddenly it’s gospel.
Experts, it is true, can sometimes be mistaken. But if experts can get it wrong, the public is at least as capable of it. And yet these days we are enjoined to reflexively reject the former, and just as reflexively to believe the latter. Perhaps we should rather trust the data.