Various methods of corpus analysis can be insightful in studying the ways newsworthiness of events is constructed, write Amanda Potts, of Lancaster University, Monika Bednarek, of University of Sydney, and Helen Caple, of University of New South Wales. They applied a number of analytical methods to a massive corpus of almost 42 000 news stories relating to Hurricane Katrina.
The authors implemented the research techniques of tagged lemma frequencies, collocation, key part-of-speech tags (POStags) and key semantic tags. The latter was found to be most useful, although under certain conditions all could be made use of in studying large corpora, the authors notice. POStag, for example, would require to be accompanied by qualitative methods.
The study was published as an online-first article of the journal Discourse & Communication. At the time of writing this (25.2.2015), the publisher is offering free access to the article here.