Investigating the Emergence of Implicit Attitudes to Immigration through Text Analysis

While newspapers like the Guardian and Daily Mail have different editorial focuses (the former being more left-leaning politically, the latter, more right-wing), it is unclear whether such biases are detectable purely through statistical analysis of the words they use. It is also unclear how the level of bias present in a newspaper’s language is related to the level of bias shown by a newspaper’s readers. For example, if a newspaper takes a biased position on a particular issue, is this bias reflected in the attitudes of the reader? This project makes a first step in examining this link by combining statistical analysis of word co-occurrences in a corpus of UK newspapers, and applying commonly used measures of people’s attitudes (e.g., Implicit Association Test), to quantify the relationship between a reader’s attitudes and the news content they consume. In the first phase of the project we focus on the topic of immigration, but hope to apply this approach to a broad range of concepts in the future.

Team:

Dermot Lynott

Tony McEnery

Liam Cross