Tuesday, February 11, 2014

Wikipedia, News and the Murder of Meredith Kercher case

As some of you might know, I’ve become interested in the ways that the Wikipedian archives can be analysed to show how the reporting of particular events evolves over time. My essay which compared the English and Italian Wikipedia articles for the Murder of Meredith Kercher was published in Language and Literature last week. You can listen to a presentation of that research as a work-in-progress which is available on iTunesU.

One of the points I made in that essay is that Wikipedia makes it very easy to see how the knowledge presented in their articles is constructed. As Martin Poulter (the JISC ambassador for Wikimedia) put it at the EduWikimedia conference in Cardiff last year, looking at the archives of Wikipedia is like opening the bonnet of the car: it can help you understand how the car works.  In the case of the Murder of Meredith Kercher article, there are differences in how the events in this very controversial case were represented over time and which varied between the different language Wikipedias.

One of the ways in which the articles varied was in how they prioritised citations from different news sources as sources for verifiability. I did the research on the Murder of Meredith Kercher article over a year ago. Given that the verdicts from the retrial were announced 10 days ago, I wondered whether the recent news interest in the case would also influence how the article developed.

It’s too early for a substantial piece of research on this, but watching the article for the week following the verdict of the retrial, you can see several things.

1. There is a marked increase in the number of page views of the article as news interest in the verdict increased. The tool which measures the page views of Wikipedia articles shows the peak viewing for the English language version on 31 Jan with 281,167 page views and 81,445 views on the day preceding (30 Jan) and 88,165 views on the day after (1 Feb). This is much higher than in the preceding three months (by comparison, the most frequent views per day are only 6,071).

A similar pattern occurs in the page views for the article in the Italian  Wikipedia, though the peak viewing figures on 31 Jan is somewhat more modest at 19,821 page views.

2. There is an increase in editing activity for the week after the verdict of the retrial was announced.  Although the article has been edited regularly since it was first in November 2007 (see the Page History statistics for this article in the English Wikipedia and the Page History statistics for the article in the Italian Wikipedia), this has peaked at particular times: (1) in December 2009 (when Knox and Sollecito were first convicted) (2) in October 2011 (when Knox and Sollecito were acquitted) and now again when Knox and Sollecito have been reconvicted.

It’s a much smaller scale comparison, but here is the frequency of editing in the 10 days before and after the verdict of the retrial in the English and Italian Wikipedias.

3. In the English language Wikipedia, the editing doesn’t just include the addition of breaking news, but where it does, these are supported with citations from news sources. The verdict itself is added with a citation from the BBC News, reactions from the Knox family cited from the Daily Telegraph, Sollecito’s reaction cited from Sky News, and Knox and Sollecito’s plans for appeal cited from the Guardian Newspaper.

4. Other changes include substantially re-ordering the content of the material so that the first section which documents Prosecution process for each of the original suspects is no longer Amanda Knox, but is Rudy Guede.  I’ve argued in the past that the structure of the article had a particular focus on Knox (earlier this year the section documenting her treatment in the Prosecution was six times longer than that of Sollecito’s section or Guede’s section: this change seems to alter that perspective).

I should say at this point that my analysis of the Wikipedia article is not in any way making a judgement about the outcomes of the Meredith Kercher case: that’s not my remit. I’m a linguist, not a forensic specialist or a lawyer.

Nor does my analysis stereotype Wikipedia as a poor source of information. In fact, I think it is a very interesting source of information: information about how contemporary events are collectively documented in different cultural contexts and about where the editors of Wikipedia get their source material. So when you next type a search term into google and find yourself using Wikipedia as the first information source you find, don’t forget to ‘lift the bonnet’ and find out which resources the Wikipedia editors are using to support their points.