Despite its exaltation as the preeminent gatekeeper to the epistemic body of scientific knowledge, the scholarly peer review process has been criticized for its many shortcomings—not the least of which is its propensity for bias, and more recently, racial bias. Given its considerable downstream influence (e.g., the establishment of programs of research and scholarly career advancement), and the dearth of empirical studies examining how racial bias may operate within this context, this study explored the presence of racial bias in the peer review process using natural language processing techniques (i.e., sentiment analysis). Two freely and openly accessible sentiment analysis tools were used to quantify reviewer sentiment: (1) VADER (Valence Aware Dictionary and sEntiment Reasoner; Hutto & Gilbert, 2014), and (2) SEANCE (SEntiment ANalysis and Cognition Engine; Crossley et al., 2017). Matched pairs (N = 85) were used to detect systematic variations in linguistic patterns of reviewers across two manuscript conditions (diversity-oriented v. nondiversity-oriented). Statistically significant findings revealed a nuanced picture of how racial bias may be manifesting in the peer review process—reflecting modern trends in racism in the U.S. Limitations and future research directions are discussed.
