Statistical Errors in “How Graduate Students With Vocal Fry Are Perceived by Speech-Language Pathologists”

Abstract

Purpose: Perspectives of the ASHA Special Interest Groups recently published a study (Gallena & Pinto, 2021) that investigated how graduate students with vocal fry are perceived by speech-language pathologists. The statistical analysis of the data presented in this study contains errors. The purpose of this letter is to describe these errors, outline practices that may facilitate error detection prior to publication, and position these errors in the broader context of vocal quality biases that may influence hiring and supervision decisions among speech-language pathologists. Conclusions: The statistical errors in Gallena and Pinto (2021) reflected a violation of the assumption of independence required by the employed statistical tests and inaccurate calculation of effect size measures. The severity of these errors makes it difficult to determine whether the reported results would hold if the correct statistical analyses were applied to the data and, thus, challenge an understanding of the degree to which vocal quality may influence perceived speaker traits and hiring or supervision decisions. Open science practices, including the Registered Report model of publication and open access to data and analysis pipelines during the review process, provide mechanisms to support error detection prior to publication.