Project description

In authoritarian countries news media are under government pressure, the media landscape is polarized and media tend to affiliate either with official propaganda or with independent coverage of events. In order to estimate censorship and to show bias in news coverage, I am developing methods based on the measurement and comparison of agenda from different Russian media, and on some NLP techniques. The method works with an annotated database with recognized named entities; the estimates of agenda difference use data on the share of texts mentioning particular persons.