Global Attention Profile FAQ

A Global Attention Profile looks at how a media outlet pays attention to 180 different nations. Instead of comparing how many stories Google has on the Green Bay Packers versus the Dallas Cowboys, we compare Google stories on Sudan versus Libya versus Lebanon. The resulting data gives us a picture of how an individual media source, at a given moment in time, pays attention to different nations.

What do you mean by "news media outlet"?
News media outlet, in this case, is a broad term referring to any publisher or aggregator of news stories. The first crop of Global Attention Profiles includes profiles of two news aggregators (news.google.com and altavista.com/news), two television news channels (CNN, BBC), three newspapers (New York Times, New York Post, Washington Post) and two news wires (AP, Reuters). More specifically, "media outlet", in the context of GAP, refers to the websites maintained by each of these news outlets. If the BBC reports something on the radio, but it doesn't make the BBC website, it does not get considered in a BBC GAP.

Why would I be interested in the information in a Global Attention Profile? Why should I care what a media outlet pays attention to?
A GAP is analagous to the ingredients and health information on food packaging. When you eat potato chips, it's worth knowing that you're not ingesting much Vitamin C. Eat nothing but potato chips and you're likely to end up catching scurvy.

Similarly, a GAP can tell you whether a media outlet is giving you information on Africa, or whether using this media source is going to lead to an Africa deficiency. While your gums won't bleed and your teeth won't fall out, an Africa deficiency will mean you're less likely to understand future media stories that involve Africa, less likely to understand Africa's role in global politics and more likely to be surprised by events triggered by events in Africa.

Immediately post 9/11, much of the world discovered that it was suffering from an information deficiency concerning Central Asia. For a brief interval, there was a great deal of media interest in Afghanistan and surrounding nations where Al Qaeda operated training camps. Now that US military involvement has wound down in Afghanistan and wound up in Iraq, Afghanistan still receives media attention, but media attention is radically reduced in the other countries in Central Asia. This raises the disturbing question: "Is our media paying enough attention to Central Asia?"

One of the goals of the GAP project is to help media consumers identify nations and regions they may be underinformed about. Armed with a news media outlet's GAP, one can urge an outlet to cover stories more thoroughly... or choose a better balanced media diet.

How do I read the GAP maps?
GAP scipts produce two kinds of maps, hit maps and correlation maps. The hit maps (this one from the AP, for example) track what proportion of stories detected by GAP scripts involve a particular country. Countries colored deep red are mentioned in at least 3.2% of the stories the GAP script found. As the colors shade from red to blue, attention wanes - countries colored the deepest shade of blue were mentioned in less than 0.032% of the stories GAP scripts found. And countries colored grey received no mention at all (or the GAP script failed to retrieve data.)

Correlation maps show a relationship between story counts and either population or national gross domestic product, a rough measure of the size of a nation's economy. GAP scripts ask the question: "How would we expect media attention to be distributed if it were directly correlated to population (or GDP)?" Then GAP scripts look at actual story counts and look for differences between the predictions and reality. Countries colored in red have more stories than GAP scripts estimated, while countries in blue have fewer. Countries in white have about as many stories as we'd expect.

For instance, in this map from the AP, Iraq and Liberia are colored dark red, suggesting that they're receiving more media attention than we would expect, given their population. We know, of course, that both are receiving a great deal of media attention because of military conflicts.

Much more information about the methodology behind GAP and GAP maps is available within a recent working paper.

Why do certain countries consistenly get a lot of media attention? Why are others consistently ignored?

Those are the types of questions the GAP project is most interested in answering. While research is incomplete, we're starting to have some theories. Early research suggests that nations with a large GDP get more attention than nations with small GDPs - in other words, if you live in a poor country, you're less likely to be newsworthy than someone in a large country, even if your nation is experiencing a war. But the GAP project also raises new questions: no single factor explains, for instance, why the Middle East commands so much attention and Eastern Europe commands so little. A thorough answer to the question of why some nations are ignored and others focused on would require a comprehensive anthropological analysis of newsrooms around the world. How do editors decide where to deploy reporters? How do reporters decide what is and isn't a story? The GAP project isn't asking those sorts of questions at this point - instead, it's trying to gather as much information as possible from raw story data and draw conclusions statistically.

Why these nine news media outlets? What about my favorite paper/TV station/website/blog?
The GAP project is in its early stages. One of the main avenues of future research is to apply the GAP methodology to hundreds or thousands of media outlets from different nations. The nine considered in initial research were chosen because they were relatively easy to "scrape" and because they were fairly prominent. If you feel like there's a media outlet that calls out for inclusion, let me know and I'll try to ensure it gets included in the next version of the scripts.

I have a problem with your methodology. I think you should do it this way instead.
Good. You've got two options. You can email me with your suggestions and if they make sense to me, I'll include them in the next version of the tools. Alternatively, you're welcome to alter the methodology and try it your way. I'll even send you the existing scripts. (They're not online yet, as I haven't had time to make a pretty Makefile for them, but if you want them, I'll send them off.) You just have to agree to let me know what results you get...

I have a question not answered here.
Email me - if it's a good question, I'll include it in a subsequent version of the FAQ.