Media bias consists of two parts: choosing what topics to cover, and what language is used to write about them. We are hoping to disentangle these two components using text matching. By pairing documents that are about the same topic and examining how their language differs, we can figure out whether topic selection or language matters more for media bias.
Getting pairs of documents, however, is difficult. We have developed new statistical tools to do so, but we need your help figuring out if the tools we've designed work well.
Interested in learning more about the researcher conducting this experiment?
Click this link to be directed to this researcher's profile.