The goal of the anonymous source tracker is to make the media's use of anonymous sources more transparent. It's an experiment, and as such it's imperfect and subject to change.

While it finds many examples of the use of anonymous sources, it doesn't find all anonymous sources used by newspapers, magazines, TV stations, wire services or other news outlets online.

It gets its examples from the English version of Google News. Phrases commonly used to identify anonymous sources are fed to Google News, which produces an Atom feed for each phrase. Those feeds are then combined under a single label, "anonymous," in Google Reader. That feed is public. Every hour a PHP script grabs the Google Reader feed, extracts the summary text, highlights the anonymous source phrasing, and puts it in a database to display on the anonymous source tracker.

Some examples are rejected, even though the articles they point to used anonymous sources, because the anonymous source phrasing isn't in the summary.

Some examples are duplicates. If a URL is already in the database, those examples are rejected. But sometimes the same story can have different URLs, so the same story can appear more than once. The same wire story may also be run by multiple outlets.

The news outlets scanned are the same outlets scanned by Google News. I don't know what criteria Google News uses to decide whether to include a Web site.

Typically Google returns a search result for a phrase giving a summary for only one outlet, with an "and more" link pointing to other matches for stories on the same subject. The anonymous source tracker doesn't grab those "and more" results, so many examples are undoubtedly missed.

I don't know how Google does what it does or why, or why one outlet is given prominence for a given search while another isn't, so I don't know if all outlets are being treated equally by the anonymous source tracker.

The count for each news outlet doesn't include every anonymously sourced story produced by that outlet. The counts shouldn't be considered valid rankings.

To quote Donald Rumsfeld, "there are known unknowns."

"That is to say we know there are some things we do not know. But there are also unknown unknowns — the ones we don't know we don't know."