Research lab "Online Political Polarization"[go to overview]
Winter Term 2020 / 2021
- This research lab will be advertised soon via the student mailing list. It will have a small list of questions to answer.
- There is no need to write to me in advance in order to register.
See the past slides from the first meeting here
This Research Lab (Forschungspraktikum/Projektpraktikum) organizes participants into groups that address various areas of the multi-faced phenomenon of political polarization online. Each group will be provided with datasets, ready tools, and tasks that are specific to several thematic areas. (please see kickoff slides above)
Social media data are full of controversies and debates. These are important tools for understanding the issues of our society, but also tell us a lot about how people process information and why. In a nutshell, online opinions reflect political views, which are a type of bias. Hence, competing truths clash with each other but are can be “equally true” with no objective solution. We must find ways to understand if, how, and when this is a problem for society.
This lab is a chance for applying your ideas to real world problems. We provide starting points by melding Web Science and Social Science, and leave it to you to develop an idea. Several lab meetings will be held where you present your progress. This semester will offer a chance to use argumentation logic, for example identifying conflicting information in a set of social media posts. Examples from past semesters include the classification of hate speech and fake news, or the association of negative emotions with politicial images in social media.
- Intro meeting: November 2, 2020 (10am)
- First lab meeting: November 16, 2020 (10am)
- Second lab meeting: December 7, 2020 (10am)
- Third lab meeting: January 25, 2020 (10am)
- Deadline term paper: End of semester
What we expect (and how we grade)
This lab places focus on not just “throwing algorithms at the problem” but how and why to combine analytical elements in an efficient manner. Example: Sentiment analysis is cheap to do, but does not do much unless combined with other elements. Think about the problems you see online and where you see a potential that can amplify negative sentiments instead of constructive debate. This can be text, images, news, and more.
Places are given first come first serve. Please respond to the questionnaire via the student mailing list, soon.