Welcome to Union Election Stats, a user-friendly hub for data on National Labor Relations Board (NLRB) representation elections. Here you can explore the history of union organizing through interactive visualizations, learn about recent trends in the labor movement, and download data for your own bespoke analysis. The database was constructed by a team of academic economists/political scientists, Zachary Schaller, Sammy Young, and Jonne Kamphorst, and is based on NLRB records. It contains the known universe of representation elections from 1962 to 2024. Soon it will also track the most recent elections as cases close. Use the Downloads tab to access the database, check out the Featured Analysis tab to dive deeper into union scholarship, and use the Contact Us tab with any inquiries or bug reports. Use of data requires citation

This website is about union election data (the flow of organizing activity), it is not about union membership data (the stock of unionized workers). For membership data, please visit unionstats.com. For info on strikes and work stoppages, use the labor action tracker at striketracker.ilr.cornell.edu. And for additional election analysis (for years after 2000) and some excellent real-time case tracking, check out unionelections.org.

We stand on the shoulders of giants to bring you this resource. Many thanks to Henry Farber, Bruce Western, J.P. Ferguson, Thomas Holmes, David Lee, Alexandre Mas, and Jack Fiorito for their seminal data collection and research on union elections. We also thank the Russell Sage Foundation for their support (grant # 2307-44744).
Custom Visualization
Responsible Data Use
As the saying goes, “figures never lie, but liars sure figure.” Please use these data responsibly and authentically. Be transparent about your methods and avoid temptations to graph hack and p-hack. Additionally, real people are behind the data generating process, so please be respectful of the blood, sweat, and tears that have been shed on all sides of industrial relations.
Research Purpose

In a time of high income/wealth inequality and widespread wage suppression and wage theft, the labor movement is enjoying renewed interest. But we must learn from history if we are to ensure that modern labor relations fulfills the needs of both capital and labor. To that end, quality data is essential. This team brings over 15 years of combined expertise on union organizing records to provide the richest and most complete dataset on NLRB elections ever compiled. We do not say this to brag, but to highlight that this project is the outcome of a multi-year effort to gather, harmonize, and merge many fragments of data collected from previous scholars and various FOIA requests. We also painstakingly geocoded the data whenever possible, chasing down location info by hand when necessary. The database is thus ambitious in its scope and novel in its comprehensiveness and accessibility.


In a time of high income/wealth inequality and widespread wage suppression and wage theft, the labor movement is enjoying renewed interest. But we must learn from history if we are to ensure that modern labor relations fulfills the needs of both capital and labor. To that end, quality data is essential. This team brings over 15 years of combined expertise on union organizing records to provide the richest and most complete dataset on NLRB elections ever compiled. We do not say this to brag, but to highlight that this project is the outcome of a multi-year effort to gather, harmonize, and merge many fragments of data collected from previous scholars and various FOIA requests. We also painstakingly geocoded the data whenever possible, chasing down location info by hand when necessary. The database is thus ambitious in its scope and novel in its comprehensiveness and accessibility.

Principal Investigators
coolest

Jonne Kamphorst

I am an Assistant Professor in Political Science at Sciences Po in Paris. Previously, I was a postdoctoral scholar at Stanford University's Sociology and Computer Science departments. I received my PhD from the European University Institute in 2023.

coolest

Zachary Schaller

I am an applied microeconomist specializing in industrial relations, regional economic development, construction IO, and economic history. I am particularly interested in labor market institutions, with my current research focused on unions and how deunionization in the US has affected local labor markets.

coolest

Samuel Young

I am currently an assistant professor of economics at Arizona State University. Previously, I was a postdoctoral fellow at the U.S. Census Bureau. I received my Ph.D. in economics from MIT in 2022.

Site Developers
coolest

Lucy Lewark

I am an undergraduate student in Art and Computer Science at Colorado State University. I am particularly interested in topics at the intersection of Computer Science and visual design, whether that's game design or general optimization for human-computer interaction.

coolest

Nathan Crane

I am an undergraduate student studying Software Engineering and Business at Colorado State University.
Clean dataset

Below are the links to download the cleaned NLRB data. The full dataset is at the election level (unique case_number, unitID combinations). For convenience, we also offer aggregated versions at the state, county, annual, and industry levels

Raw dataset

FOR ADVANCED RESEARCHERS ONLY. In merging the fragments of data from many sources, we had to make decisions about how to reconcile differences and prevent duplication. We did our best, but we’re not offended if you want to do it differently. This raw dataset is simply the fragments we have stacked on top of each other. Variable definitions were harmonized, but the fragments were not deduplicated/reconciled. Only use this raw version if you are A) interested in a particular election and want to see the full set of info available on it, even if some of it is wrong, messy, or unintelligible; or B) Very comfortable with data wrangling and well-versed in the NLRB election process/procedures.

Documentation

Below are the variable definitions for both the cleaned and raw dataset, as well as a history of version changes. It is automated through the site’s github which can be found here.

Version History
v1.0.0
Contact Us