Parsing “regret”: A computational linguistics approach to the standardized rejection letter for PhD applications

The rejection letter is a unique genre of academic communication designed to deliver bad news while maintaining a façade of institutional benevolence. This study employs Natural Language Processing (NLP) to analyze a corpus of 1,000 rejection emails received by the authors over a decade. We identify a consistent “Sympathy Sandwich” structure and a near-100% textual similarity across institutions. Our findings suggest that the “difficult decision” mentioned in letters is often a “copy-paste decision.” We conclude by proposing an open-source automated rejection generator to restore honesty to the academic rejection process.