Text as Data
By Matthew Gentzkow, Bryan T. Kelly, Matt Taddy. 2017
1,062 citations recorded by the source corpus.
An ever-increasing share of human interaction, communication, and culture is recorded as digital text. We provide an introduction to the use of text as an input to economic research. We discuss the features that make text different from other forms of data, offer a practical overview of relevant statistical methods, and survey a variety of applications. (JEL C38, C55, L82, Z13)
Subfields: Artificial Intelligence, General Social Sciences, Sociology and Political Science, Social Sciences & Humanities
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Susan Athey
- Guido W. Imbens
- Dario Caldara
- Caroline Chin
- George Skiadopoulos
- B. Seegmiller
- Rastin Matin
- Lawrence D. W. Schmidt
In General Social Sciences
- Guido W. Imbens
- Dario Caldara
- George Skiadopoulos
- B. Seegmiller
- Rastin Matin
- Lawrence D. W. Schmidt
- Matteo Iacoviello
- Renato Faccini
In Sociology and Political Science
- Susan Athey
- A. Salomons
- B. Seegmiller
- Lawrence D. W. Schmidt
- Caroline Chin
- Dario Caldara
- George Skiadopoulos
- Rastin Matin
In Social Sciences & Humanities
- David H. Autor
- Guido W. Imbens
- Dario Caldara
- Caroline Chin
- George Skiadopoulos
- B. Seegmiller
- Rastin Matin
- Susan Athey
Citations in the corpus
- Machine Learning Methods That Economists Should Know About (2019, background): “It is beyond the scope of this paper to fully describe this literature; Gentzkow et al. [2017] provides an excellent recent review.”
- New Frontiers: The Origins and Content of New Work, 1940–2018 (2022, background): “3Relative to conventional measures of text similarity (for example, the commonly used bag of words approached outlined by Gentzkow et al. (2019)), the key advantage of word embeddings is that they account for synonyms—which is crucial in our context since patent texts and occupational descriptions m”
- Technology, Vintage-Specific Human Capital, and Labor Displacement: Evidence from Linking Patents with Occupations∗ (2022, methodology): “Consequently, standard methods which require exact overlap in terms—such as the “bag-of-words" approach described in Gentzkow, Kelly, and Taddy (2019)—are likely to perform relatively poorly in measuring patent–occupation textual similarity.”
- Measuring Geopolitical Risk (2018, methodology): “23 See Appendix E.3 for details.”
- Dissecting Climate Risks: Are they Reflected in Stock Prices? (2021, methodology): “climate policy is priced in a pronounced way in the post-2012 period is consistent with previous findings that investors have become aware of climate change risks only in the most recent years (Painter (2020), Bolton and Kacperczyk (2021a), Goldsmith-Pinkham et al. (2021)).”