Re‐Imagining the Epistemic Possibilities of GPT for Public Administration Research in Competitive Settings

ABSTRACT Innovation is desirable for the public sector. Yet understanding what and how some innovation projects survive and thrive in a competitive landscape—or public sector innovation—is often challenging. The challenges not only rest in the invisibility of the features of an innovation to human eyes but also in the lack of their accessibility for analysis. This study showcases a methodological framework using a generative pre‐trained transformer (GPT) for scale development and synthetic data generation to measure, predict, retrodict, and calibrate innovation outcomes using real‐world and synthetic data and a human‐in‐the‐loop process. This study demonstrates the epistemic gains of the framework in predicting and manipulating competitive texts to simulate the past, present, and possibly the future. The approach offers avenues for future research on a wide range of competitive phenomena using large‐scale text analysis across the social sciences.