I work on music-theoretic and cognitive applications of symbolic music modelling. These symbolic music models are based on statistical learning, and derive knowledge of musical structure on the basis of large corpora of music from a variety of genres. I leverage music-theoretic and cognitive research to develop improved techniques for predicting musical structure, and use the resulting models to provide insight into compositional practice and into music cognition.
I also work on modern psychometric approaches to testing musical abilities. My collaborators and I have developed a number of computerised adaptive tests of musical abilities, including the ability to discriminate between melodies, to perceive the beat in a piece of music, and to detect mistuning in vocal music. These tests use adaptive procedures to home in on the test-taker's ability during the test, administering hard items to able candidates and easy items to less able candidates. They also use automatic item generation to produce large banks of items with predetermined psychometric characteristics. The resulting tests have been used in a variety of applications, including an ongoing longitudinal study of musical and academic development through adolescence.