Research project Media Slant through Images Print Share on Facebook Share on X Share on LinkedIn Send by email Images have a predominant role in modern mass communication. However, the current state-of-the-art methodology for measuring media slant of newspapers focuses only on language used in published articles. A systematic study of potential “image bias” in newspaper outlets has not been done yet. This project is a first step towards filling this gap. Using machine learning techniques we empirically investigate whether media slant is embedded in the images news outlets choose, if such “image bias” substitutes (or complements) the bias of the article content and if it is affected by how politically controversial the covered topic is.