Representation in Corpus Studies of Music: Commentary on Shea's (2022) "A Demographic Sampling Model and Database for Addressing Racial, Ethnic, and Gender Bias in Popular-music Empirical Research"

Abstract

This commentary responds to the article by Nicholas Shea in Vol. 17(1) of this journal, which offers a method to encode the demographic information of the artists in a corpus of popular music and then describes a resampling procedure to create a new corpus based on specific demographic targets. I argue that attempts to diversify the demographic distribution of popular music corpora should occur not at the sampling stage but instead at a prior stage when the researcher is determining the musical style or era to study. Once this musical style or era has been determined, empirical principles oblige the researcher to create a corpus that faithfully represents the statistical population.