Text as data in economic analysis

Advancements in information technology, coupled with the rapid growth of digitally accessible text, have greatly enhanced economists' ability to use text as data in their research (Baker, Bloom, and Davis 2016; Gentzkow, Kelly, and Taddy 2019). Today, the range of such text sources is virtually limitless, including books, contracts, transcripts, news articles, social media feeds, and many more. Of course, economists have long engaged in analyzing text, as evidenced, for example, by Romer and Romer (1989, 2023), who analyzed historical transcripts of Federal Reserve policymaking meetings. This article focuses on the technological advancements that have transformed this practice. We explore how these innovations enable faster, larger-scale, and more consistent processing of text, while also facilitating the measurement of multiple textual dimensions.