Full-text search across 30,000+ papers from top economics journals and NBER working papers
It turns out that tracking how empirical methods spread through economics is actually pretty important — but also kind of a pain to do manually.
This tool lets you search the full text of papers from the American Economic Review, American Economic Journal series, and over 30,000 NBER working papers. Not just titles and abstracts -- I've actually parsed the full-text of the underlying pdfs, so you can look for text directly in the paper (this follows the strategy from Currie, Kleven and Zwiers (2020) and Goldsmith-Pinkham (2024)). Want to know when "difference-in-differences" started taking over? Or whether anyone is actually using that new method you just read about? You can do that here.
The short version: I want to make it easier for applied econometricians to know what methods practitioners actually use.
The slightly longer version: As I discuss in Goldsmith-Pinkham (2024), understanding how causal inference methods diffuse through economics matters for how we think about empirical research. I built this tool partly to extend that work, but more practically, to give people an easy way to document method adoption.
Many applied econometrics papers include sections like this one from Mogstad, Torgovitsky, and Walters (2021):
Empirical researchers often combine multiple IVs using 2SLS. To document this practice, we searched the Web of Science Database for articles published between January 2000 and October 2018 containing the words "instrument" or "instrumental variable" in the abstract, title, or topic words. We restricted the search to the following five journals: Journal of Political Economy, American Economic Review, Quarterly Journal of Economics, Review of Economic Studies, and Econometrica. In total, 266 articles matched our search criteria.
Think of this tool as making that kind of analysis much easier. Search for "double/debiased machine learning" or "AKM" and immediately see how widely these methods have been adopted. Use this to motivate why we should care about your methodological contribution!
The search tool is particularly useful for understanding how empirical methods spread through economics. Here are two examples that show the kinds of patterns you can uncover:
The Abowd-Kramarz-Margolis (AKM) model for estimating worker and firm fixed effects has become increasingly popular in labor economics since the Card, Heining, and Kline (2013) paper. You can see the adoption afterward here:
Try searching for AKM or "worker fixed effects" yourself.
Probit models were everywhere in the early 2000s. But their usage in NBER working papers has declined steadily since peaking around 2006-2007. This likely reflects the influence of Angrist and Pischke (2009) and the shift toward linear probability models for estimating causal effects.
Search for probit, or compare it with "machine learning" or
logit.
The tool includes both published papers and NBER working papers, so you can compare trends across published research and the working paper pipeline. Here's "difference-in-differences" in both:
Search the full text of papers, not just titles and abstracts. Find every mention of a method, theory, or concept — even if it's buried in an appendix.
See how research topics and methods have changed over time. Track adoption curves for empirical methods and watch research agendas shift.
Download a CSV file with DOIs for all papers matching your search. Makes it easy to build literature reviews or cite the papers you find.
What's in here now: Papers from 2010-2024 for most journals, and NBER working papers from 1999-2024.
What's coming: This is an ongoing project. I'm working on:
Contributions and suggestions welcome!
Questions? Suggestions? Want to help out?
This tool is provided as-is for research purposes. Paper coverage and text extraction quality may vary.