NLP for understanding the constants and changes in concepts of scientific (mis)conduct

Skip to first unread message

Marco Spruit

Oct 1, 2021, 4:06:10 AM10/1/21
to LIACS thesis projects
This is a new collaboration between the faculties of Law, Medicine and Science.

Scientists in the Netherlands who work at institutes who underwrite the Dutch code of conduct for research integrity are expected to work according to the five guiding principles of honesty, scrupulousness, transparency, independence, and responsibility. When accusations of scientific misconduct arise, it is key that these accusations are investigated and that the results of this investigation are acted upon. Recently, a new step in the professionalization has been initiated by us through the collection of all previous and future rulings (i.e. free texts) of the committees for scientific integrity (CWIs) in a single, open, database.

You are provided with the opportunity to explore this unique unstructured dataset by performing 1) analyses of frequency over time and whether certain events like landmark cases or changes in policy lead to pattern deviations 2) sentiment analyses in which we investigate whether the content/tone of the rulings is different between fields and universities, and 3) clustering analyses in which we investigate whether external factors can explain the differences and similarities between rulings. These factors include landmark cases and changes in policy such as the 2018 renewal of the integrity code in the Netherlands. 4) And more...

For this project it is important that you are proficient in Dutch.

LIACS thesis supervisor is prof.dr. Marco Spruit
Reply all
Reply to author
0 new messages