​Seminar: Who Leads and Who Follows? Finding the Structure of Motions in the Swedish Riksdag 1971--2015

​Apart from the manifestos and documents parties create on their own, another valuable source of data to measure their interest in certain issues are the documents they produce and take part in while in parliament.

​

Here, we use such data - the motions of Swedish parties between 1971 and 2015 - to see if they can be used to: a. establish salient topics, b. observe and explain changes in saliency, and c. if one party's change in saliency can be used to ``predict'' the change in another's. To establish the topics, we use a Structural Topic Model (STM), as this allows us to include additional parameters such as year of publication and the party of the author. Apart from generating the topics, this model also allows us to study relations and dependencies between topics as well as identify if certain words and terms were more popular for some parties than for others. Doing so, we identify fifteen topics that are relatively stable over time, and whose popularity can be matched with various historical events. Also, we find that while some topics do predict changes, others do not show such effects. Given this, we conclude with a brief discussion on the promises and pitfalls of using large, free open data and the opportunities and challenges it generates for political science.

Information about the speaker

Bastiaan Bruinsma is currently a post-doc at Chalmers. His research interests include electoral behaviour, topic models, and automated text analysis. In addition, he previously published on topics related to Voting Advice Applications and dimensionality reduction and nationalism.

The event is hybrid:  zoom-link, password: mondays23

​​