National Differences and Transnational Communalities - Studying Homologies between Fields Using Topic Modeling and MCA

Kristoffer Kropp, Anton Grau Larsen, Thierry Rossier, Pierre Benz

Publikation: KonferencebidragPaperForskningpeer review


Social studies of economics have been dominated by case-oriented approaches. With this paper we suggest and show the advance in adding a quantitative field theoretical approach. More concretely we outline a perspective studying
economics as a social field by focusing on the homology between research topics and the resources and characteristics of their authors. We focus on the Swiss case, where the discipline is highly internationalised and integrated into
national elite networks. Our study relies on the combination of two data sources: The Swiss National Science Foundation (SNSF), which provides abstracts of funded projects since 2008, and the Swiss Elite Database, which contains
extensive prosopographical data on all tenured economics professors in the Swiss universities. In a first analytical part, we build the space of topics based on n=637 abstracts through a topic modelling technique (Latent Dirichlet Allocation. We identify a set of recurring topics, which are then projected into a geometric space through multiple correspondence analysis. We identify three main factors of differentiation between topics: 1) financial market vs. research in labour and behaviour economics topics; 2) macroeconomics vs. microeconomics topics; and 3) public economics vs. labour economics topics. In a second part, we map the most frequent used terms in relation to the profiles of the n=647
applicants (including n=156 economics professors). Our findings highlight a homology between the space of topics and the space of individual positions. Research in macroeconomics is linked to scientific and academic prestige, unlike
microeconomics topics. Other individual properties and resources, such as related to public expertise, corporate net-works, or gender, are linked, respectively, to the study of state and public concerns and market surveillance, corporate governance, and gendered inequalities at the workplace. This article provides an original and reproducible quantitative and computational approach that will help us expand social studies of economics and history of economic thought to new and promising research avenues.
Publikationsdato17 sep. 2022
StatusUdgivet - 17 sep. 2022
BegivenhedRelational Spatial Methods: Methods Conference - TU Berlin, Berlin, Tyskland
Varighed: 16 sep. 202217 sep. 2022


WorkshopRelational Spatial Methods
LokationTU Berlin

Citer dette