fact-based news analysis?

[varwoche]

This post is addressed to news junkies, anyone with an interest in fairness in the media, and/or database nerds.

Suppose one wanted to quantify biases / tendencies on TV news by boiling down news and news-like programs to detailed hard facts. Suppose there is a detailed database that catalogs every program (world news on the major networks, all programs on the cable news networks), daily detail, stories covered, time spent on each story, hosts and guests, political leanings of hosts/guests when known, etc.

Does this seem interesting or valuable? What data would enhance the value? Any database types who would be interested in critiquing a schema?

TIA for any feedback.

 

[T'ai Chi]

This post is addressed to news junkies, anyone with an interest in fairness in the media, and/or database nerds.

Suppose one wanted to quantify biases / tendencies on TV news by boiling down news and news-like programs to detailed hard facts. Suppose there is a detailed database that catalogs every program (world news on the major networks, all programs on the cable news networks), daily detail, stories covered, time spent on each story, hosts and guests, political leanings of hosts/guests when known, etc.

Does this seem interesting or valuable? What data would enhance the value? Any database types who would be interested in critiquing a schema?

TIA for any feedback.

Hi, I'm not necessarily interested in the news per say, but I am interested in statistical theory and methods.

I think something like this would be very valuable, especially to counter or prove claims of people who are always saying so-and-so channel is so left or so right, etc.

 

[Bjorn]

Suppose one wanted to quantify biases / tendencies on TV news by boiling down news and news-like programs to detailed hard facts. Suppose there is a detailed database that catalogs every program (world news on the major networks, all programs on the cable news networks), daily detail, stories covered, time spent on each story, hosts and guests, political leanings of hosts/guests when known, etc.

How/who would rate statements such as this:

President GWB showed evidence that Iraq has WMD
President GWB made it clear that Iraq has WMD
President GWB stated that Iraq has WMD
President GWB insisted that Iraq has WMD
President GWB defended himself by saying that Iraq has WMD
President GWB still insists, against evidence, that Iraq had WMD

'Detailed hard facts' wouldn't do you much good when it comes to bias - 'made it clear' and 'insisted' have very different meanings, even if the 'hard facts' in the rest of the statements are exactly the same.

 

[Suggestologist]

This post is addressed to news junkies, anyone with an interest in fairness in the media, and/or database nerds.

Suppose one wanted to quantify biases / tendencies on TV news by boiling down news and news-like programs to detailed hard facts. Suppose there is a detailed database that catalogs every program (world news on the major networks, all programs on the cable news networks), daily detail, stories covered, time spent on each story, hosts and guests, political leanings of hosts/guests when known, etc.

Does this seem interesting or valuable? What data would enhance the value? Any database types who would be interested in critiquing a schema?

TIA for any feedback.

Are you looking for a list of qualities to quantify? Remember that things like tone of voice and (obvious forms of) body language can turn what looks like a fair story in the transcript into a biased story as-seen-on-TV.

When you have facts to consider, you have to have people to fact check.

When you have general story types to consider, you have to determine if and for which political leaning that story helps; and if it was fair anyway due to the facts of the matter.

You might also consider looking into common storylines that may seem, at first, not to lean the way they actually do. Consider that a few years ago, FOX ran a special during which they stated as a passing fact that the Sierra Club was against windpower because it kills birds. In actuality, the Sierra Club only had a problem with a particular situating of some windturbines that were in the migratory path of one particular threatened bird species or something. But in general, the Sierra Club was and is in favor of the spread of windpower. The misinforming story that the Sierra Club was against windpower had been spread on many right wing radio shows (as documented by FAIR.org) and not surprisingly had turned up in a FOX special.

You also have to be careful about political leanings. Chris Matthews' commentary and leading questions lean toward the republicans (IMO); though he seems to like to be considered a democrat.

The press often says that they report what people want to know and are talking about, and therefore aren't responsible for the spread of misinformation and rumor. And they often conveniently forget that they are the one's who provide the menu of what will be talked about and spread by the people.

 

[Suggestologist]

Re: Re: fact-based news analysis?

How/who would rate statements such as this:

President GWB showed evidence that Iraq has WMD
President GWB made it clear that Iraq has WMD
President GWB stated that Iraq has WMD
President GWB insisted that Iraq has WMD
President GWB defended himself by saying that Iraq has WMD
President GWB still insists, against evidence, that Iraq had WMD

'Detailed hard facts' wouldn't do you much good when it comes to bias - 'made it clear' and 'insisted' have very different meanings, even if the 'hard facts' in the rest of the statements are exactly the same.

Yeah, and what if they called him Mr. Bush, instead of "President"? What if they called him Mr. Monkey?

In any case, any of the options you listed could mean different things depending on how they are said. The most neutral of your examples is "President GWB stated that Iraq had WMD." - But what if this was said in a humorous or sarcastic or quizical (or etc.) tone of voice? What if the newsperson's eyebrows were raised or their mouth was frowning as they said this?

 

[varwoche]

Re: Re: Re: fact-based news analysis?

The intent is for there to be little or no subjectivity to the data. Simply a list of topics covered, time for each, hosts, and guests.

Only in the case of talk/debate programs would opinions be cataloged, and even then only when unambiguous. (Program X, segment 1, topic: xxxx, Time: xx, Guest1: Yes, Guest2: No, etc.)

The political leaning of all hosts/guests would be part of the model (when known).

Thanks for feedback.

 

[Bjorn]

Re: Re: Re: fact-based news analysis?

Yeah, and what if they called him Mr. Bush, instead of "President"? What if they called him Mr. Monkey?

In any case, any of the options you listed could mean different things depending on how they are said. The most neutral of your examples is "President GWB stated that Iraq had WMD." - But what if this was said in a humorous or sarcastic or quizical (or etc.) tone of voice? What if the newsperson's eyebrows were raised or their mouth was frowning as they said this?

True.

And that's one of the reasons I don't think we can quantify biases by counting 'facts' on news channels.

Larry King discussing talking to dead people proves me right.

 

[T'ai Chi]

This could be generalized, of course, but take two news channels, X and Y. Make an X-Y plot and draw a 45 degree line up and out to the right, starting at the origin.

A datapoint (x,y) could represent the ratings or scores regarding a similar topic (say how many times X and Y, respectively, mentioned WMD in a given time period). One could plot (x,y) for a variety of topics or timeframes, and plot all (x,y) points on the graph.

If X and Y produce the exactly same results, the points should lie exactly on the 45 degree line (the 45 degree line is the line such that x = y). If they don't, the degree can be measured by something called the 'concordance correlation coefficient', which is somewhat like the Pearson correlation coefficient, but measures agreement, and not just correlation (which is a component of agreement (the other component being precision)).

There are many other measures of agreement that could be worth exploring.

 

[varwoche]

Re: Re: Re: Re: fact-based news analysis?

True.
And that's one of the reasons I don't think we can quantify biases by counting 'facts' on news channels.
Larry King discussing talking to dead people proves me right.

Not at all. It's worth knowing afterall which news network/channel covers the most hard news, which covers the most woo woo, etc.