The study of emotions in music is a popular topic in the field of music psychology, but methods used to investigate this phenomenon rather focus on how emotions in music are perceived rather than how they are experienced or felt. This problematic leads to the two main objectives of this thesis: In the first part, a review of methods used for studying emotions in music will be given and it will further be evaluated to what extent the results obtained with these methods give insight about the two distinct phenomena referred to as perceived emotions in music and music-evoked emotion. The methods that have mainly been used in this area are self-report, and physiological measurements or a combination of the two. Self-report is based on introspection and can be classified as first-person description of musical experience, whereas methods based on repeatable measurements fall into the category of third-person descriptions. Both these standard methods are highly based on interpretation and retrospective evaluation of an experience, which makes it questionable if a subjective and partly subconscious phenomenon like an emotion can reliably be inferred this way. In the second part of this thesis a methodological framework for studying music-evoked emotion will be developed that takes a second-person description as its point of departure. The main characteristic of a second-person description is that methods following this framework can give insight to a subjective experience via articulation and expressive behaviour, without requiring introspection and or interpretation. As facial expressions for a specific set of so-called basic emotions are suggested to be universal across cultures and provide a reliable indicator of an emotional experience, it will be discussed to what extent these expressions could be used as a form of second-person description to study music-evoked emotions. A pilot-study using this framework has been conducted and the results will be presented.