A Test of a Configurational Model of Agency Performance in the United States Federal Government Using Machine Learning Methodology

ABSTRACT

This paper takes PA research on organizational performance in a new direction by testing a configurational model using self-organizing maps, a machine learning methodology. The model was built and tested using six performance dimensions from 2017 Federal Employee Viewpoint Survey (FEVS). Four distinct performance profiles or groups were identified: very low performers, average performers, transitional performers, and high performers. Implications for theory development and practice of configurational models of public organizational performance were discussed.

KEYWORDS:

• configurational models
• agency performance
• US Government
• machine learning

Disclosure statement

No potential conflict of interest was reported by the author(s).