This study investigated ways to improve the quality of music performance evaluation in an effort to address the accountability imperative in tertiary music education. Severe difficulties with meeting the increasingly market-driven demand in Australian tertiary education for the demonstration of accountability, fairness and improvement in teaching and learning outcomes, particularly in music performance assessment contexts, have resulted from an absence of a commonly accepted set of criteria and standards to measure performance. Significant theoretical, empirical and measurement limitations and resistances in the evaluation of music performance have contributed to these difficulties. In response, the study investigated ways to improve the quality of music performance assessments that could address the accountability imperative and offer directions for improvements in teaching and learning outcomes. The study outlined a new theoretical framework encompassing an ecological model to account for the multi-layered, interactive musical and nonmusical variables that can influence music performance evaluations. An enhanced scientific methodology was employed incorporating ecological validity and a design employing recognized qualitative and quantitative methods. This approach enabled the development and refinement of an instrument-specific, criterion-referenced rating scale for empirically measuring music performance outcomes that demonstrated levels of standards in music performance, and enabled the influence of intrapersonal and interpersonal characteristics involving the performers' flow state, self-evaluation, performance experience, gender and instrument type, and examiner fairness to be empirically examined. The dimensions of music performance were determined by analyzing 655 written examination reports. The examiners' individual sorting of constructs from this analysis enabled a consensus of intersubjective objectivity to be distilled among 36 examiners within the classical music genre of an Australian music institution. From these data, an instrument-specific, criterion-referenced rating scale was developed using consensus validation with the Heads of Department. The Performance Evaluation Report (PER) was constructed individually for strings, piano, brass, woodwind and voice, and contained from 15 to 17 instrument-specific criteria and two levels of standards. The PER was completed by examiners in live student performance examinations across several semesters from which 829 PERs were produced. Using these data, structural equation modelling was conducted to statistically refine the PERs, resulting in a rating scale for each instrument family that provided high internal reliability and concurrent validity. This scale provided an instrument-specific way of measuring and demonstrating teaching and learning outcomes in music performance that addressed the accountability imperative, and fulfilled a University requirement for standards-based assessments that had not previously been achieved. It also enabled the demonstration of fairness. Results showed acceptable levels of interrater agreement and consistency, and the performers' gender was not found to substantively bias the examiners' rating. Results also showed that the interpretation of the standards measured on the PER required caution, and the implications for the application of standards in music performance assessment were discussed. The study measured the influence of performer characteristics in music performance evaluations involving a sample of 373 students, and offered direction for improving teaching and learning outcomes. A unique investigation was undertaken of the impact of the optimal performance state of flow proposed by Csikszentmihalyi (1975) in music performance outcomes. This state has received considerable and productive research attention in performance domains such as sport, but there has been an absence of empirical investigation in live music contexts. The findings provided the first empirical confirmation of the validity and reliability of the flow model in live music performance and was consistent with research in other performance domains such as sport activities. Results suggested that an abrupt shift from suboptimal to optimal flow was likely to occur at a particular threshold of high flow before performance outcome improved, and that this relationship was likely to be instrument-specific. The findings showed that most students in the sample did not believe they were sufficiently skilled to meet the challenge of the performance and most did not find the experience of the performance examination absorbing or enjoyable. This result, together with the finding that a better performance outcome was achieved by students with high flow levels, suggested that the assessment performance experience was unlikely of itself to promote substantive teaching and learning benefits for most students, and that the fostering of an optimal psychological state such as flow was likely to enhance teaching and learning benefits through an improvement in the quality of the students' performance outcome and experience in the assessment context. Findings also showed that overall, students did not hold accurate evaluations of the quality of their performances when compared with those of their examiners, and that a high majority were likely to have underestimated their ratings. The implications of this finding for improving teaching and learning outcomes were discussed in regard to the use of the PER and the frequency of summative forms of assessment. The results of the study addressed the dearth of empirical research on music performance experience by measuring a range of performance experience variables involving solo and ensemble performance experience before and since students began tertiary music studies, as well as their assessment performance experience. The findings indicated that overall, the students had not been exposed to high levels of these forms of performance experience, and that the effect of performance experience on performance outcome was not strongly linear. Rather, a threshold effect was found in which at least monthly solo and ensemble performance experiences and at least frequent levels of assessment performance experience were required before an improvement in performance quality was evident. The limitations of the study are discussed and directions for future research are examined in offering further improvements in the evaluation of music performance in the continuing effort to address the accountability imperative and improve teaching and learning outcomes in tertiary music education.