The recent riot over at digg got me thinking again about the role of the media in our society.
The articles on digg.com only hit the front page if they’ve been submitted+voted there by the users. Digg presents quite a threatening, democratic model to ‘old’ media. This doesn’t mean there’s no editorial influence, it just means it takes a very different shape.
When the owners of digg buried a recent story for fear of litigation, the users revolted. It was a fantastic display of democracy in action (or not), watching the same story hitting the front page over and over again. I do wonder though, what has been buried in the past without the users noticing!
Users who submit to digg tend to use quite sensationalist headlines to grab attention (and diggs), but I’ve often found that they’re highlighting issues outside of the ‘media circle‘ (important issues that even the left wing press won’t draw attention to, such as corporate influence on the media itself). This was particularly noticeable when Tony Blair criticized the media recently: neither he nor the media even mentioned the underlying problem of media ownership. That, to me, is a far more important question than the ‘tone’ of public discourse, as it sets the agenda.
I digress. Since my impression of digg was that this (hopefully) democratic news channel might provide hidden secrets, I decided to do a quick test to see if it was worth further investigation. I asked a simple question:
Are popular articles of news on digg getting the same attention in the mainstream media?
Answering this question might give us a sense of the hunger that the public have for items that the mainstream won’t touch.
My chosen method for this test was to compare digg to Google News, where its easy to gauge the amount of attention an item of news is getting from ‘old’ media (leaving aside the issue of how Google does this and how reliable it may be!).
My methods were extremely unscientific, but again this was just a brief test to see if its worth taking further. I welcome criticism of and alternatives to these methods.
Here’s what I did:
- Took a snapshot of the top stories on a section of digg (politics - the a topic that gets plenty of attention at digg and in the mainstream)
- Recorded the popularity, the date/time of the story, and (rather subjectively) picked a few keywords from the story to search Google News with.
- Search on Google News, recording the number of media outlets reporting the same item.
- Chart the relationship between digg popularity and media exposure.
Here’s the chart. The initial results are are striking in their shape.

I’ve approximated some of the numbers here: the count on Google News was too small to chart in many cases, but for anyone interested, here’s the spreadsheet complete with links, stats, and more.
Step 3 is where my methods really fall down. There’s a big gap between the way something is reported on digg and the way its reported everywhere else. Some might say its not even the same item any more.
For example: there were two stories on digg highlighting the same issue: that a small group of US politicians are sponsoring a bill to impeach Dick Cheney. The digg articles highlighted the third and fourth senators to sponsor the bill. Searching for ‘cheney impeachment’ at Google News gave no clear results. Some articles in the media mentioned the bill, but mainly the results were were about efforts to impeach Bush.
Making it harder still was that occasionally Google News gave no grouping to the items (a helpful feature when it recognises that the media are reporting the same item), which could lead to several conclusions:
- there really wasn’t a central ’story’ - quite often the digg articles highlight a single fact/quote from a broader item (there’s that that sensationalism again)
- My keyword choices didn’t accurately connect with the media’s take on events
- Google News didn’t group the items (for whatever reason)
- There’s a potential gap between the time of the digg articles and the media exposure
My count of the Google News items was a bit of a compromise: if the grouping for what was obviously the same item, I took that number. If not, I went through the results counting the number of strongly related articles.
To conclude: this quick and dirty test indicates there might be a real desire for the media to cover a wider range of topics, and place a very different emphasis on some.
Moving forward: I’m going to spend more time on democratic news channels (e.g. ohmynews). I’d really love it if my digg test were a long term experiment (perhaps using the API’s of both sites to automatically record/chart the results), but I’m unsure of how to get over some of the problems with my method. It does seem like every item needs human input to collect/verify results.
I’d welcome ideas about how to get this going: for example, could you send a link to the Google News API to gauge media popularity instead of using clunky keywords?