Election polls blurred independence of news media

In April 2012, just after the provincial election in Alberta, Tomorrow declared a position on Facebook that requires repeating after the 78-day Canadian federal election.

Between the start of the campaign on August 7 and election day on October 19, there were a staggering 347 polls, or an average of more than four a day.

This included:

  • 113 national polls
  • 167 individual constituency polls
  • 67 polls on leadership preference and perception1

There were even more done privately by political parties but never released to the public.

Some, such as polls from Nanos Research,2 have been daily and some have been sponsored or organised with major media firms. They became the basis for much of the narrative of the election.

CBC’s poll aggregator, for example, always flagged up who was leading and which party had the biggest change, up or down. So the first thing a reader saw was the top party and who was losing. And that created a trend. Then opinion writers questioned parties who don’t accept the “reality” of polls.

Media can gauge public opinion in a number of ways. The most direct is the “vox pop”, quizzing around four to six people for their opinions. Sometimes we identify them by name, age and occupation, sometimes just as faces on a TV screen. This is usually somewhat balanced and expresses viewpoints with occasionally nuanced positions.

The other way is to commission a poll from a wider sample of hundreds or even thousands of people with set questions and potential responses.
These approaches are used to both fill space and time for a media outlet but also to appear relevant: “We know what ‘the people’ are thinking – here it is.”

But they also allow the media to dictate narratives. As has been better described by Jon Stewart, former host of The Daily Show, Fox News uses comment in the morning to drive news to drive comment to drive news to drive comment throughout its daily cycle.

“Polls say the party is sliding.”
“People say the party is sliding.”
“Reports have the party sliding.”
“A new poll says the party is sliding.”3

The polls drive comment to drive news to drive further polls. And in this most recent election, they became a self-fulfilling prophecy.

Even worse was unethical abuse of “polling” public opinion on election night when a CBC reporter announced he had told voters waiting in line to vote in British Columbia that they, as a media outlet, had declared the winner based on the rest of the country’s early results. That used incomplete numbers to sway further results.

And it is a breach of the third principle, “independence and accountability”. Even though our ninth principle is “observe and engage”, attempting to direct results through the power of the media is a step beyond engaging.

In 2012, we stated:

The Alberta election results, wildly different from pre-election polls, prove the flaw of polling and the dangerous media addiction to said polls. To meet our own commitments to accuracy, Tomorrow will not commission, use, or comment on political polling as a sole basis for news. Polls may be referenced in features but should not be taken as fact. That a poll was conducted is a fact. Percentages of small samples for polling do not guarantee the facts of eventual electoral results.

Without core principles, there can be no core journalism strength.
Tomorrow reasserts this position. It is the correct one for maintaining independence and it ensures our focus remains on asking important questions and presenting original and exclusive news coverage, particular of politics, as we did during the 78-day election without needing a single poll.

  1. Wikipedia lists public released polls here and here.
  2. Nik Nanos, who was involved in the Charlottetown Accord referendum in 1992 was interviewed by Tomorrow in 2012 and his website can be found here.
  3. These are not direct quotes from any news source, just impressions of the tone.

Follow Us

Comments are closed.