There is concern that graduating surgery residents are not prepared for independent practice. This study aimed to identify predictors of performance, autonomy, and readiness for independence ratings of trainees by attendings for thyroidectomy and parathyroidectomy with respect to repeated resident-attending exposure. We hypothesized that increased exposure with a particular attending increases resident autonomy.All residents and faculty at a single institution performing parathyroidectomy or thyroidectomy were invited to complete an operative performance evaluation at case competition using the Zwisch scale to measure performance and autonomy for individual operative steps. In addition, each survey evaluated the trainee's readiness for practice in a straightforward procedure as a binary variable. Categorical variables were evaluated via Chi-squared or Fisher's exact tests and ordinal variables were evaluated with Wilcoxon or Kruskal-Wallis tests. Multivariable analysis was conducted with random effects logistic regression, and learning curves were generated for each procedure.Operative performance evaluations were obtained from 36 individual learners and 6 faculty members, with a total of 145 evaluations for parathyroidectomy and 116 for thyroidectomy. On bivariate analysis, readiness for practice ratings was significantly associated with increasing chronologic procedure number, but not resident gender or case difficulty. The multivariable model demonstrated that increasing chronologic procedure number, while a significant predictor without accounting for exposure, did not remain a significant predictor of practice-readiness for parathyroidectomy when accounting for resident-attending exposure. Bivariate analysis comparing resident and attending ratings showed no difference between the 2, but there were significant differences in autonomy and performance scores for both groups of raters. Trainees rated by attendings as independence ready completed a median of 7 parathyroidectomies and 5 thyroidectomies. Descriptive learning curves generated serve as a model of the multistate nature that residents undergo when moving from novice to proficiency.Not surprisingly, the more operations residents perform with a single attending, the higher their ratings for performance and autonomy from that individual, with increased exposure allowing improved performance with less attending autonomy. By contrast, our data also show that repeated exposure between resident and attending may confound the use of procedural numbers alone when predicting resident ability in the operating room.