Advancing Text Analysis for Nonprofit Research: Using Semantic Role Labeling to Automate Institutional Grammar Coding of Nonprofit Laws and Policies

---

# Description

The Institutional Grammar is a rigorous tool for examining the laws and policies that govern nonprofit organizations, yet hand-coding has traditionally limited its reach because of the time required.

We introduce a natural language processing advance: a semantic role labeling classifier that reliably codes the rules guiding nonprofit organizations. The paper details how to hand-code with the Institutional Grammar, prepare text for machine learning, and apply the classifier to automate the coding pipeline.

By comparing manual and automated coding we demonstrate strong accuracy, opening new opportunities to explore inter-organizational collaboration, government contracting, federated compliance, and nonprofit governance across languages and countries. Throughout, we provide practical examples showing how researchers can pair the Institutional Grammar with the SRL classifier to pursue questions that once demanded prohibitive effort.