Defining population problems: Using topic models for cross-national comparison of disciplinary development

• Correlated topic modeling (CTM) allows inductive identification of themes in texts.
• CTM facilitates cross-national comparisons of disciplinary development over time.
• CTM shows how French and British demography evolved differently in the 20th century.
• High fertility dominates British research agenda, French agenda focuses more on low.
• Agendas reflect cultural, institutional differences that shape ideas on fertility.

The content of academic journals provides insight into disciplinary boundaries and priorities. This paper uses correlated topic modeling (CTM), an innovative approach to textual analysis, for a cross-national comparison of the development of research agendas in the discipline of demography. Using articles from leading demographic journals from 1946 to 2005, CTM shows how the set of concepts relevant to the study of fertility was defined differently in France and Great Britain. Results indicate that demographic research agendas reflected both cultural and institutional differences that shaped different understandings of fertility decline. While British demography focused on high-fertility contexts, French demography focused on lower-fertility contexts. This difference reflects national intellectual traditions shaped by larger cultural discourses: the dominance of demographic transition theory and fears of overpopulation in Britain versus the co-existence in France of a second salient model, a theory of demographic “revolution” with sustained low fertility leading to depopulation. Relationships between expert concerns and broader public concerns are then examined in the British case by comparing journal publications to mass-media coverage of fertility and population issues. This comparison shows that British academic demography passed over some policy-relevant population issues, such as discussions of immigrant fertility, that were featured in the popular press.