This post has been corrected and updated:
Although I am a fierce advocate of a free press, I am beginning to think there should be a ban on the reporting of political polls during election campaigns. Aside from providing easy headlines for the media, the only service they might perform for citizens is in helping some voters make strategic decisions on how to cast their ballots. But that's only when sample sizes are significantly large to produce useful results, the polls' timing and questions are right, and, last but far from least, the media interpret them correctly.
But journalists are notoriously innumerate. I know because I can't tell you how often my colleagues ask me to do percentages for them, as in if x dropped by y, what percentage did it go down?
This lack of familiarity with 'rithmetic sure showed up this past week, as we saw when the Globe and Mail ate crow -- actually a crow feather considering the error -- over a grossly misreported front page Strategic Counsel poll which apparently revealed a drastic drop in NDP support in Ontario.
Figures for party support in Ontario were incorrect in a story published Monday. Poll respondents were asked which party had the most momentum going into the Jan. 23 federal election as well as which party they would support. Those momentum numbers were wrongly cited as each party's level of support. The correct Ontario numbers were as follows: Liberals 46 per cent, Conservative 30 per cent, New Democratic Party 17 per cent and Green Party 7 per cent.
The NDP had been reported at 9 per cent, in a story with a sub-head that read: New Democrats' support level drops to single digits as Grits reach their highest mark since May, poll says
Maclean's blogster Paul Wells was all over this and another Globe/CTV polling muck-up
(Antonia sez: see Uppity Date below) like a TV news crew over a free breakfast buffet.
The jury's in. The most prestigious newspaper in Canada and the leading private television network can't read their own polls. Now: Has anyone read or seen a correction?
Attaboy had at 'em too.
It’s one thing to attribute any meaning to daily poll results; it’s another entirely to read the poll flat-out wrong and report it as fact. It’s embarrassing that reporters at a newspaper of the The Globe and Mail’s stature can’t read their own poll results properly.
Reader Dan Coleman, who complained to the Globe and tipped me to this earlier this week when I should have paid closer attention, suggested that perhaps the Strategic Counsel people doing the actual polling aren't communicating directly with the reporters who rely on chair Alan Gregg to do all the media interviews. Here's an example from last Monday's screw-up story:
Backing for the Tories is stable, whereas the number of those who say they would vote NDP has dropped from nearly 15 per cent in early December, into single digits.
"The NDP look like they're getting caught in a bit of a squeeze overall," Mr. Gregg said.
Anyway, what's done is done. The point is, it shouldn't happen at all. Ever.
All this to refer you to a comment -- referencing last night's instant poll on Global -- from York University's Bob Hanke, who posted in Commentapalooza below. (Scroll down to today at 12:55 p.m.)
In my view, citizens are the clear losers when instant polls are used to predict the outcome of debates or election campaigns. Internet-based polling technology speeds up the collection, collation and publication of polling data. These "surveys" appear to resolve the crisis of our increasingly unrepresentative democracy by allowing people to make their opinion known directly, uninfluenced by the news media's post-debate analysis. Yet, according to the Global National's website, only 2611 English-speaking voters who were "planning to watch the leaders’ debate" were surveyed--a self-selected population. Here we see one of the insidious effects of the "public opinion" produced by Ipsos-Reid and Global New--it exludes the opinion of nonviewing voters.
However, the problem with instant email or wireless polls goes deeper than their methods. They are techniques of forward communication where thought and time, democracy and duration, part company. At the very time when you would expect political opinion to be open and fluid, real-time media poll-itics tries to pin down and fix public opinions and emotions.
Unfortunately, reporters and news editors who may be concerned about the democractic deficit spend little time analyzing their own use of polls.
In a shamless, but welcome, bit of self-promotion, Hanke links to news of his upcoming study on all this. He argues that ''poll-itics'' actually distorts politics.
According to Hanke, when voter preference polls dominate political reporting and commentary, the “poll-driven” news media make an independent contribution to the political reality and to election campaign dynamics. He states that “Bell Globemedia and CanWest Global Communication Corporation commission leader or party preference polls rather than policy preference polls. The overload of pre-election, poll-driven news uses up space which could have been used for analysis and commentary.”
In Hanke’s view, the growth of instant polls represents a complete break with social science methodology. “Daily reporting of survey results is supplemented by up-to-the-second public opinion, accelerating the flow of information within the circuit between the TV screen, the newspaper headline, the Internet and the ballot box,” he says. “This accelerated public opinion has effects apart from any message.”
”By introducing seat projection technology and analysis, the news media aim to simulate the election in advance, pre-empting the time for deliberation and debate,” Hanke claims. “As a result, the public is inundated with political information but remains uninformed, uncertain and never quite sure of the differences in platform, policies or ideas among the contenders for political power.”
Of course, when that ''political information'' is misinformation, the distortions get further distorted, leaving the political landscaped warped.
UPPITY DATE: I have been reliably informed that the problem with the first Globe polling ''muck-up'' that Wells attacked was with the headline, not the story. The report itself was correct, as Wells later himself corrected without admitting he had erred in the first place. Which is why I followed his link.
One of the principles of blogging, I should think, especially when one is attacking others for making mistakes, is being open about making one yourself. I hope I followed that principle with this update. I hate making mistakes and, as my ex-husbands learned the hard way, I never make them.
But this time I did.
Recent Comments