Classification of scientific tweets (KlawiT)

Project description

Nowadays, social media are frequently employed in academia to communicate new findings or to discuss them. However, misgivings that scientists who use social media actively in their work act unprofessionally are also still widely held. Nonetheless, many researchers are active on Twitter, and lists naming these scientists are available on the internet. A recent study shows that researchers from social and information sciences are over-represented. The contents of these tweets are mostly of a personal nature, such as comments on political events or news. Recent online research makes little effort to analyse whether the rest of these tweets, which have no personal content, contain scientific information. If we continue the thought, we arrive at the question: when is a tweet scientific in nature or has at least a scientific context? This leads on to the research question to be addressed in this research project: what is a scientific tweet?

A database of heterogeneous tweet collections serves as the research object. Relevant medical terms will be used as search terms to compile the tweet collections. Medical domains are considered particularly suited to this project since they attract researchers, but also patients, interested laypeople, politicians etc., which should result in a heterogeneous data pool. Based on this data pool, concrete differences between scientificity and non-scientificity can be identified. The characteristics of scientific tweets in these collections will be evaluated exploratively by user opinions; at a later stage of the research project confirmatory evaluation will follow. Based on these evaluations, a classification framework will determine what characterises a scientific tweet.

As a by-product of this research project, we obtain additional indications for a possible classification of scientists and guidelines for creating scientific tweets. Such guidelines are available already, but they mostly look at using Twitter as a medium in certain domains and not so much at the scientificity aspect. Within the framework of this research project, recommendations will be developed that focus on the aspect of scientificity. Usually, the focus is on the person who tweets and not so much on the actual content of tweets as an object of classification.