Deciphering the Decline: A Computational Analysis of Two Decades of Canadian Newspaper Op-Eds on Freedom of Information

Abstract

Background: Newspaper op-eds are an underexplored mode of communication that frame social, cultural, and political issues. Analysis: This article uses an unsupervised machine-learning approach called structural topic modelling to map changes in the content of a corpus of Canadian newspaper op-eds on freedom of information (FOI) law spanning a 20-year period. This makes it possible to investigate changes in the content of newspaper op-eds over time and to decipher trends in the kinds of topics that national, regional, and local newspapers publish. Conclusion and implications: Computational approaches to analyzing news texts are used, and recommendations are offered for future research on FOI and political communications.