Replication Data for: Anchoring in the past, tweeting from the present: Cognitive bias in journalists’ word choices

Drawing upon insights from theories of the mind, memory, and language, this study explores how cognitive biases are embodied in journalistic work across different media. We built a large-scale dataset of text corpora that consisted of more than 220,000 news articles, broadcast transcripts, and tweets generated over a year by 73 campaign reporters in the 2016 U.S. presidential election. Leveraging this unique dataset of journalistic outputs from a campaign season, we conducted automated text analyses. Results suggest that heuristics and intuitive thinking played a significant role in the generation of content on Twitter.