The Mercatus Center at George Mason University is pleased to announce the creation of the Policy Analytics Society as a part of its QuantGov initiative. Policy analytics is, at base, the measurement of government policy for the purpose of studying its causes and consequences.

The new Society will serve the needs of scholars, researchers, and practitioners who use advanced analytical methods such as machine learning and natural language processing to identify and measure latent variables within public policy text. It aims to foster discussion of effective and legitimate ways of creating new data-driven research using human- and AI-powered algorithms. Because most public policy is written in unstructured text, using policy analytics to turn such text into structured data will lead to more and better research about the causes and effects of government policy.

To launch this effort, the Policy Analytics Society has released a series of working papers that showcases the development of policy analytics through novel datasets and methodologies. The following papers advance the study of best practices for the measurement of government policy and seek to empirically identify innovations for data-driven policy research. Topics covered by the symposium include a theoretical framework for policy analytics, methods for data validation, the use of machine learning to analyze trade agreements, data analytics related to occupational licensing, and a computational analysis of contracts.

Concept and Methodology

Towards a Formalization of Policy Analytics – Dustin Chambers

How to Improve Data Validation in Five Steps – Danilo Friere

Democratizing Policy Analytics with AutoML – Danilo Friere

Specific Applications

Using Machine Learning to Capture Heterogeneity in Trade Agreements – Scott Baier and Narendra Regmi

Validating Readability and Complexity Metrics: A New Dataset of Before-and-After Laws – Wolfgang Alschner

Measuring a Contract's Breadth: A Text Analysis – Joshua C. Hall, Bryan McCannon, and Yang Zhou

Man vs. Machine: A Novel Evaluation of Data Analytics Using Occupational Licensing as a Case Study – Edward J. Timmons and Conor Norris