Proceedings of the Annual Meeting of the Cognitive Science Society

Scientific communication includes primary scientificliterature written by and for scientists, as well as pressreleases written about these scientific articles that are used toinform the popular press. By the time new scientific findingsare reported by the press, the reporting can often reflect 'spin',or reporting that minimizes uncertainties and exaggeratesimpact, as compared to the original study. In this work, weexamine the role that the press release may play incommunicative change, in particular with respect todifferences in portrayed confidence between abstracts ofscientific articles and press releases. We examine a largecorpus of over 15,000 documents collected from onlinedatabases covering a range of scientific topics, leveragingautomated analysis tools from natural language processing toexamine how the readability, sentiment, subjectivity, andportrayed confidence varies between primary literature andpress releases. We find that press releases are often easier toread, portray more positive sentiment, use language thatimplies greater objectivity, and demonstrate higher confidencein the findings. Future work should focus on examining ifthese differences between press releases and primary articlesdo indeed engender different perceptions in readers.