The CBC has an explicit policy on the reporting of public opinion polls during an election campaign. It was formulated by both the English and French services of CBC/Radio-Canada before the 2004 election campaign. The reason for it is that, as a news organization, we have concerns that the use of polls during election campaigns skews the direction of coverage.
In our view, it has the effect of sucking the "oxygen" out of campaign debate by focusing on "who's ahead?" and obscuring debate about genuine policy differences among the contending parties. This is made even more absurd by the fact that headlines are often made of tiny changes in public opinion as charted by polls of widely different composition and sample size.
We believe that an election is the opportunity for citizens to decide on the direction and leadership of the country. Elections are about much more than who is ahead in what poll at what moment.
All polling is difficult at the best of times: getting the questions right, finding the most representative population sample and conducting the poll at the most appropriate time are all difficult issues. In recent times, the reliability of polls during Canadian election campaigns has been suspect, even disgraceful.
Given all of this, CBC News - in conjunction with its French-language Radio-Canada counterparts - has adopted the following guidelines:
Guidelines: CBC/Radio-Canada and Election Campaign Polling
CBC/Radio-Canada believes that election campaigns are about much more than simply trying to predict who will win. Our coverage during the campaign should lay before the electorate the different party platforms and positions, and reflect the debate among the parties and the public about these issues. We accept that too much media focus on the party "horse race" - reflected in the constant stream of "voter-preference" polls - can crowd out more meaningful campaign news and information.
Therefore, our guidelines about campaign polling for the federal election will be the following:
CBC/Radio-Canada will not commission any polls during the campaign that focus on "voter preference" or try to suggest the eventual outcome.
We will place limits on the systematic reporting of polls conducted by other media organizations. We will report primarily on polls whose results constitute a major campaign story. The preferred strategy will be to present a weekly wrap-up of poll results to illustrate a trend.
We will be similarly restrained in the discussion of such poll results on our news and current affairs programs, and in the careful way these results are promoted and placed in our programs.
We will not base headline stories for newscasts solely on poll results.
As for regional polls conducted by recognized polling firms, we will generally be able to report their results. These polls are less frequent than national ones and, with a sufficient sample size, can illustrate a meaningful regional trend.
The intended effect of these guidelines will be to ensure that more coverage and attention during the campaign will be devoted to the actual issues in front of the electorate - leaving the determination of actual "voter preference" to the voters on election day.
Editor in Chief
[an error occurred while processing this directive]