To investigate whether and how use of an online remediation system requiring reflective review of performance and self-assessment influenced students' performance on objective structured clinical examination (OSCE) station repeats (subsequent to failure on the first attempt) and their self-assessments of their performance (between the first and second attempts). Fourth-year medical students' performances on seven OSCE stations were videotaped at University of Michigan Medical School in 2006. Failing students took the exam again; remediation included self-assessment and review, plus faculty guidance for failures that were greater than one standard error of measurement of the distribution. A total of 1,171 possible observations of students' actual performance and performance self-assessments were analyzed using independent and dependent t tests and within-subjects ANOVA. Results indicate statistically significant changes in students' performance between first and second attempts and statistically significant improvements in self-assessment between first and second attempts. No significant changes were found between self-assessed and faculty-guided remediation. This study provides evidence that OSCE remediation combining review, reflection, and self-assessment has a salutary effect on (subsequent) performance and self-assessment of performance.