Converting the Texts of the U.S. State Constitutions into Quantifiable Data: A Text Analytics Project

Julia G. Clouse

Advisor: Alex T Tabarrok, PhD, Department of Economics

Committee Members: Garett B Jones, Joshua D Wright

Carow Hall, Conference Room
April 22, 2019, 10:00 AM to 12:00 PM


For this dissertation project, I converted the texts of two-hundred and three historical and current U.S. state constitutions into quantifiable data, using methods of categorization. In chapter one, I describe the methodology behind the creation of the “U.S. State Constitutions Dataset” (Clouse, 2018) and explain how the constitutions were gathered, standardized, and converted from text to numerical data. 

In the second chapter, I use the dataset to analyze the evolution of political ideas and priorities in the fifty states. I study the overall growth of the constitutional texts and highlight the different topics comprising that growth. I conclude the lengthening of these texts is primarily due to increases associated with local government, taxation & finance, and judicial.

In the third and final chapter, I analyze how the states approach businesses and corporations in their constitutions and how this has changed over time. I specifically focus on major shifts in constitutional language. My analysis reveals the first major shift in constitutional language occurred in the 1840s; when the state constitutions changed to limit both the legislature’s involvement in the establishment of corporations and the state’s ability to invest in these corporations. The second shift occurred in the 1870s, when states began regulating specific business industries through their constitutions. I conclude these shifts in language were driven primarily by economic and political conditions within the individual states, rather than by federal policy.