Music researchers regularly test the hypothesis that participants will respond differently based upon their levels of musical sophistication. Circumstances may not permit any extensive pre-testing and researchers often select or group participants according to their answers to simple survey-type questions related to their musical background. Although some indicators have been criticized for their inadequacy, their validity has remained untested.
The goal of this dissertation was to test the validity of selected indicators, using expert ratings as the criterion variable, and to compile a short questionnaire of the optimum combination of indicators that researchers may use to classify research participants. To that end, three survey projects were conducted.
The first project sampled 743 published music studies to obtain information about how researchers were measuring musical sophistication. Two of the most commonly used indicators were “Formal musical training” (e.g., years of private lessons) and “Year or level in a formal music program” (e.g., music or non-music major). However, 38 different categories of indicators were identified in the sample and were used in 173 unique ways—either singly or in combination.
The second project sampled 27 experienced musicians using a web-based questionnaire to discover their suggestions for gauging musical sophistication. The musicians’ suggestions added four new categories to the 38 established from the first survey.
For the third project, 29 indicators from the previous two studies were selected and developed into a 36-item questionnaire. It was administered to a sample of 633 adults who belonged to various types of groups involved in music-related behaviors, ranging from introductory university music classes for non-majors to professional orchestras. The leader of each group (a music expert) provided a rating of every member’s level of musical sophistication using a 10-point scale. When the expert ratings were blocked into two categories labeled less and more musically sophisticated, and the data were analyzed using logistic regression, a significant model emerged made up of nine indicators (model chi-square = 296.133, df = 32, p < .001). The model classified 79.5% of the sample accurately. On the basis of this research, the Ollen Musical Sophistication Index is presented.