Many standard shoulder outcome measures do not adequately cover the range of problems and issues specifically encountered in glenohumeral joint instability and have been shown not to be sensitive enough to detect clinical change with intervention adequately. The purpose of this report is to present a prospective evaluation of a new self-administered patient questionnaire specifically designed to assess glenohumeral joint instability. The evaluation involved test-retest reliability and comparison with the Shoulder Rating Questionnaire (SRQ). Sixty-four patients with confirmed glenohumeral joint instability were assessed with both the Melbourne Instability Shoulder Scale (MISS) and SRQ 12 preoperatively and at 6 months after shoulder reconstructive surgery. Twenty-two patients were recruited into a reliability study of the MISS questionnaire. The test-retest reliability of the MISS was found to be 0.98 (interclass correlation coefficient, mixed-model analysis of variance, absolute agreement). Assessment of agreement between the MISS and SRQ questionnaires indicated very poor pretest agreement (0.33) and moderate agreement at 6 months (0.66). The differences between the MISS and SRQ were statistically significant both before surgery (paired t = 13.2, degrees of freedom [ df ] = 63, P < .001) and at 6 months' follow-up (paired t = 7.9, df = 63, P = .001). Change in the questionnaire scores measured from surgery to 6 months' follow-up was significantly greater in the MISS (mean, 30; SD, 19.1; median, 30.8) than in the SRQ (mean, 16.6; SD, 12.8; median, 14.3) (Wilcoxon test: z = -5.8, P = .0001). The results of this study show that the MISS questionnaire is a reliable outcome questionnaire and has a greater range to detect changes in shoulder instability than more global outcome questionnaires such as the SRQ. The higher scores encountered on the SRQ may mean that it underestimates the severity of a patient's instability problem.