Measuring Media Criticism with ALC Word Embeddings

Abstract

The ability of news media to criticize government is a core pillar of media freedom and is often taken as evidence of meaningful democratization. Existing indices typically use scoring criteria or expert surveys to develop over-time measures of media freedom. In this article, we use the largest existing dataset of Arabic-language news to evaluate how political reporting about the government changes over the course of successful and failed democratic transitions in Egypt and Tunisia. Using entirely unsupervised ALC word-embedding techniques, we demonstrate how to generate temporally granular measurements of media criticism that closely correlate with measurements derived from expert surveys for both countries. Crucially, the technique we propose is computationally inexpensive, open-source, and cost-free––making it eminently scalable. Our work therefore points to new possibilities in the monitoring and measurement of media capture within authoritarian and transitional settings.

How Syrian Refugees Engage with Online Information

“…an "affluence dimension" in their books corpus by subtracting the vector for the word "poor" from the vector for the word "rich," "poverty" from "affluence," and "worthless" from "priceless" before summing these together. Rozado and al Gharbi (2021) use this technique to derive an axis of "economic development" in their news corpus, and Barrie et al (2023) use a similar approach to measure criticality of the regime in Egyptian and Tunisian state media. We can then project a target word of interest onto these dimensions in order to determine any temporal change (or other document-or covariate-level unit) in the relationship between words (or concepts) in vector space.…”