Extracting meaning from newspaper archives
Title: Extracting meaning from newspaper archives
SNIC Project: SNIC 2022/22-344
Project Type: SNIC Small Compute
Principal Investigator: Miriam Hurtado Bodell <miriam.hurtado.bodell@liu.se>
Affiliation: Linköpings universitet
Duration: 2022-04-01 – 2023-04-01
Classification: 50401


When the world around us changes, they need to be understood by the individuals who experience them. This project aims to analyze a vast newspaper archive to extract the common interpretations that guided Sweden's understanding of immigration and sexual norms over the last 100 years. Our novel text-analytic approach, relying on seeded topic models, maps the importance of pictures painted in the media discourse and demarcates eras of meaning---periods of relative stability during which a distinct set of interpretations had dominated the discourse. We aim to identify eras where norms and ideas such as that coined Sweden's reputation as a haven for refugees and sexual liberalization. We ask how and when the media discourse turned. Our text-based measures provide a broader empirical foundation to the casing of timelines than the narrative accounts usually heralded in the historical social sciences. We are interested in comparing these empirical measures with the turning points found in qualitative studies. We are especially interested in investigating the relationship between "events" and how people discuss event-related concepts. In our study, we investigate the impact of both events clearly defined in time (e.g., terrorist attacks) and more extended "periods of rupture"/"unsettled ties" (e.g., the refugee crisis)---. By exploring how time-series data of topic co-occurrences, obtained by seeded topic models, changes after events or after long periods of uncertainty, we can add to the literature on if/how culture changes