Workshop in digital methods for social sciences
- Actualité Sciences Po
Join a project class and learn a series of tools for exploration of a digital corpus. Learn how to collect, analyze and visualize large datasets of texts about the coverage of immigration.
Public: Master and PhD Students
Regular Instructors: Vincent Antonin Lépinay, Associate Professor in the Sociology Department and Medialab, Jean-Philippe Cointet, Associate Professor in Medialab, with contributions from Katharina Tittel and Dimitrij Muller.
Dates: From the 13th to the 17th January 2020
Credits: Masters students who will successfully complete the course will earn 4 credits.
The course is designed as a project class around which students will learn a series of tools for exploration of a digital corpus. Organized during an intensive week in between term 1 and 2 of their Master’s degree, the course will happen over 30 hours and will mostly consist in hands-on training for students to learn how to collect, analyze and visualize large datasets of texts about the coverage of immigration.
This year we work on immmigration and its press coverage since 2004. We look at 2 different kinds of presses, major newspapers in 3 different languages (English, French, German) and scientific (social sciences and humanities) .
Readings and discussions are meant to train students to articulate research design (with a battery of methods to use given the format of the data and the research question) in public. It is also serving the other purpose, important at the early stage of the research career of young scholars, of showing how to combine methods without losing the edge of any of them.
The exercise is to teach students to collectively discover the methods adapted to the case under study. The exercises will deploy a series of methods in an organic fashion: instead of listing them and stabilizing them across the years, they will be adjusted to the case at hand and the instructors will elicit the use of a variety of methods with the students. The important by-product of the exercise is also to teach students to prototype a research design, beyond what research proposals usually contain, by instructing students to explore their data before the dataset is fully defined.
The set of methods will cover data extraction, formatting and cleaning, text analysis and data visualization. One of the outcomes of the seminar is to teach students to document the steps that have been adopted, from the formulation of their research prototype to the stabilization of a research design. Work will be achieved in small groups.
We ask students to bring their laptop during the workshop and to read the following articles in preparation. These articles will both offer an overview of possible frameworks for analyzing textual material (topic analysis, sentiment analysis) and examples of empirical dataset coming from online, media and academic sources.
- Antilla L. Climate of scepticism: US newspaper coverage of the science of climate change. Global environmental change. 2005 Dec 1;15(4):338-52.
- Farrell J. Corporate funding and ideological polarization about climate change. Proceedings of the National Academy of Sciences. 2016 Jan 5;113(1):92-7.
- Flores RD. Do anti-immigrant laws shape public sentiment? A study of Arizona’s SB 1070 using Twitter data. American Journal of Sociology. 2017 Sep 1;123(2):333-84.
- Shwed U, Bearman PS. The temporal structure of scientific consensus formation. American sociological review. 2010 Dec;75(6):817-40.