Measuring media freedom: An item response theory analysis of existing indicators

Abstract

The media's ability to freely gather and disseminate information remains a critical aspect of democracy. Studies link media freedom to other concepts including human rights, corruption, democratic peace and conflict, natural resource wealth, political knowledge and foreign aid. However, media freedom's many facets make it difficult for any single index to fully capture. To develop a more robust measure, this article treats media freedom as a latent variable and analyzes ten extant indicators by fitting an item response theory model. Utilizing a Bayesian approach, the model generates time-series, cross-sectional data on a bounded, unidimensional scale from 0 to 1 that measures media freedom in 197 countries from 1948 to 2017. After numerous validity checks, the authors utilize their new Media System Freedom data to replicate Egorov, Guriev and Sonin's (2009) analysis of media freedom and natural resource wealth. The findings indicate that the published results do not hold once the more robust measure is included.