Analyzing open-responses in a post-exam survey using natural language processing methods

• In Event: Surveys of/for the Measurement Community

Abstract

This study compares two Natural Language Processing (NLP) methods and human coding for analyzing open-response survey data in a high-stakes medical examination. Initial findings revealed that NLP categorized open-responses in a similar way as human coding. The possibility and future directions of utilizing NLP to grasp candidates’ feedback are discussed.

Authors

• Yunyi Long, National Board of Osteopathic Medical Examiners

• Xia Mao, NBOME