Tuesday, February 9, 2010

Quantifying Media Bias

When Parliament resumes at the beginning of March, I would like to conduct an experiment to quantify potential bias in the various Canadian media outlets. I have been considering ways to empirically chart bias in journalism and I would like to focus my attention to online content. Measure variables like font size for a specific story, where it is prioritized on the main page, positive or negative buzz words used, whether it is breaking news or opinion editorial, etc.

It is like having one of the largest newspapers printing the giant headline "BARBARIANS AT THE GATE" a week before a Federal Election when the Conservatives have taken the lead in the polls. Clearly that headline selection is opinion oriented and not fair and balanced journalism. You can then chart the bias of stories against newspaper circulation or television ratings against opinion polls to see if any individual story potentially impacted public opinion. For example the Star has been one of the most pointed publications against prorogation, and we now see that there has been a measurable shift in polling numbers in the greater Toronto area.

Below is a basic draft outline for my experiment. Ideas for variables to track and methods of assigning numerical values to headline bias are most welcome.

The Who

1) National Post
2) Globe and Mail
3) Toronto Star -should give the CBC a run for its money on the Bias-O-Meter
4) CBC.ca - considering the preference given to Kady O'Malley and Heather Mallick stories, I forecast this medium to be the most biased.
5) CTV.ca

The Where

1) Online Main News pages
2) Canadian Politics pages

The What

1) Priority on main site: what slot in the order does the story appear? Is it even listed on the main page? Is the font size bigger or smaller than standard titles?

2) Positive or negative spin: this will be tricky to ascribe a numerical value to, other than a simple yes or no on the critical or supportive meter.

3) Which stories are not being covered by specific outlets?

4) Is the story "breaking news", opinion editorial, or two in one?

The When

The month of March for what promises to be an emotional session of Parliament.

8 comments:

  1. There are several organizations in the US that have published articles on media bias. Some have been peer reviewed. You should look at their methodology so you don't have to recreate the wheel.

    However, the recent prorogue coverage is clear evidence of media bias, quite honestly.

    ReplyDelete
  2. So was the three-week feeding frenzy over Peter MacKay purportedly calling Belinda Stronach a dog. The Canadian media as a whole are juvenile.

    ReplyDelete
  3. Iceman, the graphics (photos, cartoons, etc.) are eye-catching and stick in the mind. You'll need to pay attention to those used.

    ReplyDelete
  4. Speaking of media bias, don't be shy to send your complaints to the Soloman Show politics@cbc.ca

    ReplyDelete
  5. Actually the most biased is the Quebec media. You need to track the Gazette and La Presse at a minimum.

    ReplyDelete
  6. Counting the number of stories and duration on covering a story. What percentage of the newscast is devoted to the political stories. I read a Poll the MSM employees indentify with the Liberals and NDP in much larger numbers than the electoral results. Will try to find that tidbit.

    Bias exists in everything, the problem is when the bias becomes glaring. The worship of OBAMA by several media outlets in the MSM including our own.

    ReplyDelete
  7. It is either the Liberals or the coalition and everything that comes with them that are promoted by the national media to run this country and no other particularly the Conservatives.

    The media has a duty to perform and if any reporter dare reveal or go his or her way to do what's right for their country against the liberals; he she will be 'fired' removed. They either follow the 'code of silence' or they are gone.

    ReplyDelete
  8. Perhaps you could use the current news coverage of McGuinty proroguing the Legislature as a control case.

    ReplyDelete