Guilty by Association: Using Word Embeddings to Measure Ethnic Stereotypes in News Coverage

Abstract

The current study provides a new level of empirical evidence for the nature of ethnic stereotypes in news content by drawing on a sample of more than 3 million Dutch news items. The study’s findings demonstrate that universally accepted dimensions of stereotype content (i.e., low-status and high-threat attributes) can be replicated in news media content across a diverse set of ingroup and outgroup categories. Representations of minorities in newspapers have become progressively remote from factual integration outcomes, and are therefore rather an artifact of news production processes than a true reflection of what is actually happening in society.