Categorizing music by genres.

Abstract

In arts and music, exemplars are categorized into genres, but those are not static and change dynamically. In musical taste research, there are strong differences in number and width of genre categories, and it is unclear how category width (broad genres vs. narrow subgenres) affects liking evaluations. The use of broad genre labels has been strongly criticized. We address this issue by a quantitative approach, comparing liking evaluations of 15 musical genres with 100 embedded and two nonexistent subgenres (N = 804). We applied a wide range of analyses (correlations, regressions, random forest modeling, factor analyses). Results converge, showing that evaluations of the majority of nested subgenres were highly similar to the related genres. The nonexistent subgenres revealed evaluation heuristics, enhancing consistency. For some genres, liking of subgenres grouped based on historical (traditional-modern), cultural (German subcultures), or functional (danceability) reasons. The genre labels pop and rock were less appropriate. We provide a list of 24 genres to assess musical taste sufficiently for general applications. Importantly, the concept of genre for taste studies is still useful.