Mediate Metrics FAQ #1

Thanks in large part to coverage initiated by Inside Cable News, interest in our media bias/slant rating system has increased dramatically. Rather than field all questions individually, we’ve decided to post some of the most popular ones below:

How does one measure “bias” in the media without introducing bias into the system?

We were diligent in trying to maintain objectivity by adhering to very strict social science/text analytics, guidelines, working with a partner who is very experienced in this area, and engaging multiple “coders” for the sake of system integrity . After many months of incrementally refining the system — and waiting until we achieved high levels of inter-coder correlation — we released our version 1.0 classifier, and have continued to refine it every day since. Systems like ours must be constantly refined to adapt to the changing political rhetoric of the day. Fortunately, our platform is designed to do just that.

Text classification systems use Natural Language Processing elements —- basically, a progression of statistical correlation techniques —- to mimic the results of expert human coders. That being the case, the human coding process is key, since that is where bias can most readily be introduced. Some of the provisions we included to minimize coder bias include:

• Defining VERY strict rules for identifying transcript statements which can be coded as either “Favoring Democrats/Critical of Republicans,” or “Favoring Republicans/Critical of Democrats.” For example, the experts can only code for slant if if the explicit terms or specific proxy labels for Democrats or Republicans are contained in the text.
• Randomizing transcript statements for the human coding process so that “slant inertia” is drastically reduced. Even expert coders tend to bring outside context into their evaluations, especially when reading a narrative which has a repetitive theme. Randomizing statements helps the “man” component of this man-machine partnership to be more clinical, and enhances objectivity.
• Regular adjudication sessions, in which the team members review their mismatches and recommend rule refinements to improve coding clarity. Having done this innumerable times, and operating under the proviso of, “When in doubt, code NEUTRAL,” I can tell you that bias is controlled rather effectively this way.
• Partitioning statements related to the Republican Presidential primaries. This was critical to making the ratings fair and reasonable. News coverage about the Republican primaries is decidedly negative, and is often about Republican candidates bashing other Republican candidates, while we specifically target inter-party comparisons. Once again, we have VERY strict guidelines for how we treat this situation.
• Following slant assessment templates which involve identifying the speaker, determining the object of his/her discussion, assessing inter-party comparison(s),uncovering embedded judgments, and noting factual references that clearly reflect positively-or-negatively towards a particular party.

Hopefully, you get the idea. We’ve gone to great pains to make our ratings objective, but I am not so bold as to represent that it is perfect. Even the best text analytics systems have limitations. This one is no exception.

What is the business model for this service?

Beyond the high-level slant metrics we have initially provided free-of-charge, there is additional business value to be reaped from:

• Networks, news analysts, and interest groups, through secondary slant studies on specific topics such as health care, labor/union issues, military spending, right-to-life, tax reform, regulatory measures, etc.?
• Watchdog agencies, via insight reports on the political views of prominent news anchors, correspondents, and contributors?
• Various political groups desiring a deeper understanding of each network’s Republican Primary coverage and slant.
• Commercial, governmental, and educational bodies desiring to analyze the resonance of TV News slant through social media platforms like Twitter, Facebook, and the blogosphere.
• Media outlets, who want to certify that their content meets a specific political/informational criteria, for the purpose of differentiation

Say the President has a bad news day…something bad happens…bad job numbers, court case goes against the administration, scandal in the West Wing…whatever. How does your system handle that scenario?

A bad (or good) day by the President will influence our ratings. Slant ratings effectively “move with the market.” Therefore, our ratings are best viewed as relative measure. Said another way, you would find that certain networks and programs are more slanted than others during a “bad” news week, for Democrats or Republicans, but all will be effected by a dominant political news theme.

How does one evaluate “bias” in content that is, by design, supposed to be opinionated?

From our perspective, Op-Ed news content is absolutely valid, as long as viewers are aware that the content they are watching is indeed that. Frankly, we think that boundary between opinion pieces and straight news is often blurry for the general public. News wonks know the difference intuitively, but we have all experienced instances in which an uninitiated viewer proudly states that, “The only news program I watch is {INSERT YOUR OP-ED PROGRAM OF CHOICE}.” Furthermore, straight news programming often contains a subtle-but-consistent political tilt, despite claims to the contrary.

The fact is that TV news programs, regardless of type, often frame the political discourse of the day, which then translates into voting behavior and government policies that dramatically affect our daily lives. That being the case, don’t you think an object entity should “watch the watchers” in order to serve the greater good?

That may sound pretentious, but I don’t know how else to say it.

If Mediate Metrics had been through a rigorous process of development, which can take several months of hard work, they’d be telling us about it, because it would be a big step forward. The biggest trouble is that the initial degree of inter-annotator agreement, depending on how you define it and measure it, is likely to be spectacularly low, say around 30%.

Actually, our inter-coder reliability reached a peak of over 80% before the 1.0 classifier was released.

Our system had been in development for many months, and the supporting the code book is substantial. Still, there are many different outlets for this service, many of which are not staffed with linguistic/text analysis experts.  Knowing that, and in consideration of our limited resources, we did not publicize all of our details, but they are available with certain concessions to confidentiality.

 

LIBERAL MEDIA BIAS?: Impact Report #2 Provides the Answer(s)

Is there a liberal bias (or as we prefer to call it, political news slant) in the U.S media? It depends largely on which news content you think matters most.

Based on the data we currently have under analysis, early indications are that the leading broadcast and cable new outlets lean towards the Democratic party. With that said, the recently released Pew Research Center News Interest Index indicates that more people are now turning to cable news for information about the 2012 presidential campaign than to other forms of media. Adjusting for their findings yields an entirely different result.

For Impact Report #2,  we have quantified the amount of network slant (aggregate speaker bias plus editorial influence) and combined it with viewership data in order to assess the total “Impact.” In this installment, we have computed the weighted-average Impact Rating for each network during the weeks of January 16th and January 23rd, focused on their evening/weekday news programming. The results are depicted below:

Aggregate Slant Impact: Weekdays - January 16 to 27

As in all of our charts, content favoring the Republican party is represented in red (greater than zero), while that which leans towards the Democratic Party is shown in blue (less than zero). We have indicated depth-of-coverage by lightening the base colors as required. For example, the bases representing NBC,CBS, and to a lesser degree Fox News, were purposely made lighter to reflect the limited transcript coverage for those particular networks. Regardless, the basic message depicted in chart above is that the slant rating base is amplified by the number of people viewing the content. The table below shows the numerical analysis supporting the chart:

The Cumulative Slant Ratings  in column 2 were derived from data previously discussed and shown at this site on Wednesday, February 8th. We have also factored in viewership data for the networks and programs under review. As you can see, the cumulative rating for this report — across all networks and programs we are studying —  slightly favors Democrats/Liberals, as indicated by the numbers highlighted in light blue.

If, however, we account for the Pew Research data, which states that 36% of people regularly go to cable networks to learn about Presidential campaigns (and, we presume, national political news in general) versus only 26% turning to national nightly news programs, our cumulative results shift in favor of Republican/conservative interests (as highlighted in light red).

Liberal or conservative media bias? It all depends on whose news one views.

TV News Political Bias Impact Ratings: January 16 – 20, 2012 (Revised)

Much debate has been devoted to assessing whether there is a liberal or conservative media bias (or as we prefer to call it, political news slant) in the U.S. Most tend to focus on the source of the bias, but our view is somewhat different. At Mediate Metrics, we prefer to quantify the amount of slant (bias plus editorial influence), combined with the quantity of slant recipients, to assess the full impact of media bias.

As presented in our November 15 post, total objectivity and fairness in the news is a virtual impossibility. Still, our hypothesis is that networks will be less biased when their reputation is built upon informing viewers and being an objective resource. In contrast, news outlets which rely on affirming the political preferences of their loyal viewers will have a natural tendency to be more slanted.

Building on our previous 2 posts, we have added a Political Slant “Impact” Rating comparison for January 16th to the 20th, depicted in Chart 3 below:

CHART 3: Slant Impact Rating - January 16 to 20

As in our other charts, content favoring the Republican party is represented in red (numerically positive), while that which leans towards the Democratic Party is shown in blue (numerically negative). We have graphically indicated depth-of-coverage, or lack thereof, by lightening those colors. For example, the bars representing NBC,CBS, and to a lesser degree Fox News and ABC, were purposely made lighter to reflect the relatively small transcript coverage for those particular networks. Regardless of coverage, the basic message depicted in Chart 3 is that the slant “foundation”, depicted at the base of each pyramid, is amplified by the number of people viewing the content. Table 3 below shows the numerical analysis supporting the chart:

TABLE 3: Political Slant Impact Measures - January 16 to 20, 2012

The Composite Weekday Slant Ratings in column 2, along with the number of statements classified in column 3, were derived from data previously discussed and shown here on Tuesday, January 31st. Statement coverage and “Confidence Factors” relate directly to the color shades in Chart 3. Most importantly, we have factored in viewership data for the networks and programs under review. This is where things get interesting, given that:

• The nightly news programs from the major broadcast networks achieve the highest ratings per program by far, but their impact is mitigated by the fact that they are only broadcasted for 30 minutes a night;
• By definition, the 3 top cable networks broadcast a continuous line-up of news shows, each of which is 60 minutes long, and representing as much as 420 minutes of programming/network for the nightly time period (5 PM to 11 PM) under consideration.
• Public service programming, such as the Republican Presidential debates, was purposely omitted from our calculations since it does not reflect the editorial views (slant) of the network-or-program they were broadcast on.

Some compelling notions, however preliminary, can be drawn from this analysis. While the aggregate slant of content delivered during this time period appears to favor Democrats (as depicted in light blue in the “Totals” row), the aggregate impact tilts towards the GOP (as shown in the light red cell, same row).

Admittedly, our classifier and database need further refinement, but we think these initial results are rather intriguing. Still, we’d love to know what you think. Don’t hesitate to leave a comment below, or to send one directly to: barry@mediatemetrics.com.

TV News Political Slant Report by Show: 1/16 – 1/20

Building on our previous post, today we our publishing a separate version of our TV news measurement metrics which focuses on the political slant of individual programs aired by the 3 major broadcast networks (ABC, CBS, and NBC) and the top 3 cable news channels (CNN, Fox News, and MSNBC), for shows aired from 5 PM until 11 PM eastern time, Monday through Friday. As highlighted yesterday, our analytical coverage varies by network, program, and date, but our intention is to augment it over time.

CHART 2: Slant Rating by Program - January 16 to 20, 2012

Content favoring the Republican party in Chart 2 is portrayed in red (numerically positive), while content that slants towards the Democratic Party is shown in blue (numerically negative). Those interested in the underpinnings of the Mediate Metrics slant rating system should review our January 31st post, or see our primer on Text Analytics Basics at: http://wp.me/p1MQsU-at.

Since our analytical coverage varies by network, program, and date, so does the associated confidence factor in our slant ratings. The exact amount of coverage per network is shown in the Table 2 below, but we have graphically indicated depth-of-coverage by way of color shading in Chart 1. For example, the cones representing The Five, Hannity, and On The Record were purposely made lighter to reflect the relatively small transcript coverage for those particular networks. Low transcript coverage likely accounts for certain results that may seem counter-intuitive; we expect those metrics to adapt with volume and time.

TABLE 2: Slant by Program - January 16 to 20, 2012

As mentioned yesterday, we have partitioned statements about the Republican Presidential primaries, since they tend to be disproportionately negative and often lack inter-party comparison, and have largely excluded them from these slant ratings. Similarly, the Republican Presidential debates and other such dedicated program segments have been omitted in their entirety since they do not reflect the political positions of the networks, programs, or contributors under a consideration.

We’ll publish an “impact rating” for the same January 16 – 20 time period tomorrow.

TV News Political Slant Report by Network: 1/16 – 1/20

Welcome to the Mediate Metrics inaugural TV news political slant measurement report, based on our version 1.0 text classifier.

To our knowledge, this is the first objective TV news slant rating service ever published. Slant, by our definition, is news containing an embedded statement of bias (opinion) OR an element of editorial influence (factual content that reflects positively or negatively on a particular U.S. political party). This initial report focuses specifically on evaluating  slant contained in the weekday transcripts of  the national nightly news programs on the 3 major broadcast networks (ABC, CBS, and NBC) , as well as programming aired from 5 PM until 11 PM eastern time on top 3 cable news channels (CNN, Fox, and MSNBC). At this stage, analytical coverage varies by network, program, and date, but our intention is to fill in the blanks over time.

CHART 1: Slant by Network - January 16 to January 20, 2012

In keeping with U.S. political tradition, content favoring the Republican party in Chart 1 is portrayed in red (positive numbers), while content that tilts towards the Democratic Party is shown in blue (negative numbers)

To grossly over-simplify, the numerical slant ratings supporting the Chart 1 emanate from a custom text analysis “classifier,” built to extract statements of political slant from TV news transcripts. (For more on the underlying technology, see our post on Text Analytics Basics at http://wp.me/p1MQsU-at.) We have trained our classifier to interpret slant quite conservatively, conforming to strict guidelines for the sake of consistency and objectivity. As such, the ratings we present may be perceived as under-reporting the absolute slant of the actual content under review, but the appropriate way to view our ratings is as relative to similar programming.

As mentioned, our analytical coverage varies by network, program, and date. Correspondingly, our rating confidence is directly proportional to the amount of transcript text available for classification.The exact amount of coverage per network is shown in the table to the right, but we have graphically indicated depth-of-coverage in Chart 1 by way of color shading. For example, the bars representing the slant ratings for both NBC and CBS were purposely made lighter to reflect the relatively small transcript coverage for those particular networks.

During development, we determined that the Republican presidential primaries are an enterprise for which scrutiny is a normal-and-valuable part of the vetting process.  Related news content, however, tends to be disproportionately negative, and often times does not contain a clear inter-party comparison — an element we view as a crucial condition for the evaluation of political slant. With those factors mind, we have partitioned statements about the Republican Presidential primaries, and have excluded them from most slant ratings at this juncture. Similarly, the Republican Presidential debates and other such dedicated program segments have been excluded in their entirety from classification since they do not reflect the political positions of the networks, programs, or contributors under a consideration.

We’ll publish slant ratings by program for the same January 16 – 20 time period tomorrow.

Understanding The News

I’m about halfway through Blur by Tom Rosenstiel and Bill Kovach, and I can’t recommend this book enough. Classes on its content should be made an educational requirement.

Here are some factoids I gleaned from my readings:

• Newspaper staffs are down by roughly 30 percent from ten years earlier. For network news, the cuts have been even steeper.
• Traditional online news sites have gotten bigger, not smaller, despite the proliferation of online news outlets.
• When changes in communicating to the masses have occurred in the past, existing power elites have tried to exploit the transition in almost every case.
• The argument culture limits the information we get, rather than broadening it.

The fundamental premise of the book is that the objective-and-disciplined mediation function has largely disappeared from news reporting. In parallel, business models have arisen that focus more on assertion (goal: disseminate information quickly) and affirmation (goal: maintain audience loyalty) than on verification. As such, the burden of editing and analyzing the news falls to individuals. In essence, we must become our own investigative journalists.

The authors point out that the clash between fact-and-faith has occurred numerous times in history, providing both risks and opportunities in every case. In this instance, the challenge of “skeptically knowing” the truth about the news is indeed burdensome, but we can now interact with the media, political leaders, and others in an unprecedented manner.

Powerful stuff, if you think about it.

Related articles

• Blur: Who is Who in Journalism? (teesdiary.wordpress.com)

Open Season on Journalists: Trump versus Todd

It seems like every time I turn around, some political figure is taking on a journalist. Here’s the latest installment, featuring a very edgy Donald Trump chastising Chuck Todd of MSNBC during his Daily Rundown Show this morning.

http://www.youtube.com/watch?v=MyVq-Vj9IlY

Wouldn’t you love to see focus group ratings during one of these exchanges?

My sense is that many political figures already have.

Editorial Selection: Fox and MSNBC

Building on the theme of editorial selection and the news, I decided to once again use my “tag cloud” (most popular words) tool on evening and prime time broadcasts from Fox News and MSNBC on November 14th and 15th. As I highlighted yesterday, media outlets can broadcast but a tiny portion of the available news, so I decided to see what these 2 competitors decided to emphasize.

DISCLAIMER #1: I could not wait to get this out, so I’m sure I will be making additional edits and refinements.

DISCLAIMER #2: Tag clouds are not surgical instruments. That fact, combined with the knowledge that I manually culled words that did not directly relate to specific topics and messaging themes should tell the reader to view the following with a critical eye…. as you should with all interpretative journalism.

Which virtually all political news is.

Disclaimers aside, examining the content selection of Fox and MSNBC is like having box seats at a gun fight. It’s clear that MSNBC is putting Republican Presidential candidates under a microscope, taking pot shots at local Republican candidates whenever possible, and positioning themselves as the mouth-piece for the middle class. Similarly, Fox has President Obama and the 2012 election in the cross hairs, featuring topics that cast him or his administration in a negative light, with specific emphasis on job creation (or a lack thereof).

Those are the highlights — or low-lights, depending on your point of view — but there is more information in the clouds if you are willing to stare at them briefly …

**********************************************************************************************************************

MSNBC “TOP 25” TAG CLOUD:

• Substantial Republican Primary/Candidate focus, with Herman CAIN (236 occurrences) still drawing the most attention, ROMNEY (82 occurrences) a distant second, and Perry (52 occurrences) in third.
• Occupy Wall Street is a significant topic, as evidenced by the occurrence of the related tag words MOVEMENT, OCCUPY, and STREET. Why WALL did not make the top 25, I have no idea.
• SCOTT is in the top 25 primarily due to parallel references to Republican governors Scott Walker (Wisconsin) and Scott Brown (Florida). Similarly, JOHN was also mentioned frequently in relation to Ohio governor John Kasich, but I removed that name because several other JOHNs were intermingled in the word count.
• Frequent references to AMERICANS (and AMERICANS by default, since my tag cloud tool intermittently extracts root words in parallel) and the middle CLASS seems to represent a positioning theme for MSNBC
• JUDGE generally shows up in 2 different contexts: 1.) The judge who let Penn State coach Sandusky out on reduced bail and; 2.) The impartiality Judges Scalia and Thomas related to the Supreme Court case on health care.
• CASE shows up in several different contexts, again related to the tag cloud tools penchant to extract root words — ObamaCARE, HealthCARE, MediCARE, and are “they” sCAREd?

FOX “TOP 25” TAG CLOUD:

• No references to the Republican Primary candidates by name in the Top 25 tag words. In contrast, PRESIDENT (65 occurrences) and OBAMA (42 occurrences) are the top 2 most popular tag words in the cloud. When viewed in relation to the MSNBC tag cloud, one cannot help but conclude that negative politics extends to these 2 networks.
• Similar, but not exactly the same, thematic positioning around AMERICA, but not so much on CLASS.
• BOOK was an area of focus mostly because of controversies surrounding Bill O’Reilly’s new book (“Killing Lincoln”), and Peter Schweizer’s book about alleged congressional insider trading.
• A greater focus on activities in the SUPER COMMITTEE, and with question on whether a satisfactory DEAL can be made.
• DEAL was also used in the context of favorable (and ethically questionable) deals made on IPOs and land, leveraging the insider trading immunity afforded to congressman.
• CONGRESS was primarily used in 2 contexts: 1.) There were several CONGRESS persons on the prime time Fox News programs I analyzed, and; 2.) Numerable references were made along the lines of our “Do-nothing CONGRESS. ..”
• ELECTION appeared primarily as part of 2 topics: 1.) Forward-looking statements related to the 2012 Presidential election, and; 2.) The fact that negative news related to Solyndra was allegedly throttled by administration officials.
• FLORIDA made the top 25 based on Florida government officials on the shows whose transcripts I analyzed.
• JOB and JOBS are in the top group because of a focus on the subject of job creation.
• LEGAL is attached to either the constitutional rights that should or should not be afforded terrorists, as well as related to immigration issues.
• The term SPEAKER rose to the top because of references and sound bites from House Speaker John Boehmer, as well as an interview with FORMER SPEAKER of the House Newt Gingrich.

**********************************************************************************************************************

If you would like to know more about the specific details of my process or the specific programs I included in this analysis, just email me at: barry@mediatemetrics.com.

White Paper Excerpt: “Bias and Objectivity in the News Media”

I remain convinced that one can measure media bias electronically, at least to some extent, by examining the text of news broadcasts and objectively identifying the speaker’s personal value judgments. With that said, it is far more difficult to extract bias based on that content that is chosen to be aired. The following excerpt, taken from a 2004 white paper published by The Foundation for Critical Thinking titled, “How to Detect Media Bias and Propaganda” by Dr. Richard Paul and Dr. Linda Elder, explains this far more eloquently than I ever could.

Enjoy.

********************************************************************************************************************

The logic of constructing news stories is parallel to the logic of writing history. In both cases, for events covered, there is both a massive background of facts and a highly restricted amount of space to devote to those facts. The result in both cases is the same: 99.99999% of the “facts” are never mentioned at all (see Figure 1).

FIGURE 1

If objectivity or fairness in the construction of news stories is thought of as equivalent to presenting all the facts and only the facts (“All the news that’s fit to print”), objectivity and fairness is an illusion. No human knows more than a small percentage of the facts and it is not possible to present all the facts (even if one did know them). It isn’t even possible to present all the important facts, for many criteria compete for determining what is “important.” We must therefore always ask, “What has been left out of this article?” “What would I think if different facts had been highlighted here?” “What if this article had been written by those who hold a point of view opposite to the one embedded in the story as told?”

*****************************************************************************************************************
Amen.

Who’s News? YOU Decide.

The more I study media bias, the more I realize that TV coverage flows (and often overflows) in certain directions because viewers vote with their eyeballs.

The blogosphere is crackling today with reports on the CBS internal memo which directed their debate moderators to devote fewer questions to Michelle Bachmann. The issue certainly has ignited the fanaterati. Don’t get me wrong; editorial selection bias is a very real phenomenon. Still, a thinking person should consider other possibilities.

So here is one: Perhaps we get a disproportionate amount of coverage on certain issues and people because they drive viewership. Combined with the extensive amount of news capacity that needs to be filled, media outlets are motivated to keep popular stories alive because lots of people are following them. As an unfortunate by-product, reporters and commentators fan the  flames over time by digging up all kinds of corner-cases, then sensationalizing them as “New Developments!” And that’s when we enter the realm of the absurd.

Circling back to the issue du jour, giving Michelle Bachmann more debate time does not make sense for the network in that context. It’s an inexact science, but it is a network executive’s job is to promote viewership … which drives ad revenue …which increases company profits, equity value, and personal paychecks.

It’s tempting to see a conspiracy here, and maybe there is one, but I think it is equally possible that this is just capitalism in action.