Climategate study - simpler methods into the mix

Image from a HuffPo article by Shan Wells

Climategate was a controversy unfolded in November 2009 after thousands of emails and files from the Climatic Research Unit (CRU) at the University of East Anglia (UEA) were published online without the owners' consent. The climate change opponents used the content of the emails to argue that scientists manipulated data to prove their argument for human responsibility of climate change. Several investigations didn't find any scientific misconduct at the CRU, but the reports called for opening up access to research data and more transparency in methods and communication of results (see Climactic Research Unit email controversy in Wikipedia).

Controversies are always hard to sort through, but they present an interesting research case for those like me who are interested in discourse, language, and media. A recent study "The creation of the climategate hype in blogs and newspapers: mixed methods approach" (paywalled) looked at the Climategate controversy and compared discussions in blogs and newspapers.

Newspaper and blog data were collected from the LexisNexis Academic database using the search term ‘climategate’. Two methods were used to analyze the data: a) ARIMA (Auto Regressive Integrated Moving Average) modeling to create a model of the daily frequencies of postings and to examine the mutual influence of newspapers and blogs and b) semantic co-word maps of blogs and newspaper headlines to compare framings of climategate.

The results of the modeling seemed a bit confusing as they showed a significant link between a high number of blogs and a high change in newspapers articles (either increase or decrease) on the same day. (I'd really like to see simple descriptive statistics of posts per day, etc. Also, a pre-print where all the images and tables are at the end of the article is very hard to read). At the same time an increase in newspaper articles on one day had no effect on the number of blog postings on the next day. The conclusion of the article is that blogs influenced newspapers, but not the other way around. The semantic maps showed (predictably) that the blogs used a more informal language and framed the topics more negatively, while the newspapers were more formal and stayed more neutral. Both blogs and newspapers picked up similar sub-topics, such as climate change, scientists, and so on, although the word "climategate" occurred more in blogs.

Several thoughts / questions upon reading this interesting, although a bit too methodologically complicated for such simple variables and questions, study:

• How different are "traditional" and "new" media nowadays? They may be still different in their language style, but what about the speed of publication, audiences, contributors, and so on? The headlines don't get to the differences in main posts and comments either.
• The word "climagate" did originate in a blog, but it was a journalist who picked it up and popularized it via a newspaper-hosted blog (see Climategate: how the 'greatest scientific scandal of our generation' got its name). Does it change the conclusion that "blogs were independent of the attention in newspapers" (p. 20) if journalists write for both media?
• It would've been helpful to establish the actual sequence of events via an additional documentary analysis. The paper argues that the word "climategate" originated in blogs, which promoted the hype. But according to the Wikipedia article, news about emails release were published almost simultaneously in blogs and newspapers - on November 20, 2009. So is the hype about the word or other, more nuanced exchanges and actions as well?
• Three blogs received links to leaked documents. It seems that it was intentional - the blogs were skeptical of climate change. Did it matter for how the hype have originated and developed? Again, what is the connection between the actual controversy and its naming as climategate?
• How can the link between the large number of blog posts and the decrease in newspapers articles be explained? More quotes and examples of interactions and influences between blogs and newspapers could be very helpful in illustrating all the findings.

Overall, it seems that the studies of controversies benefit from careful tracings of words and actor connections rather than from complicated modeling that is rather confusing and not so eye-opening.