Self-directed learning is associated with knowledge and performance improvements, increased identification and amelioration of knowledge gaps, and heightened critical appraisal of available evidence. We developed and implemented a decision support system that could support self-directed learning for anesthesia residents by soliciting resident input in case selection. We hypothesized that residents would utilize this system to request complex cases, and that more advanced residents would request more complex cases. Prospective, observational study involving 101 anesthesiology residents. We used a web-based interface, RHINOS [Residents Helping in Navigating Operating Room (OR) Scheduling], which allowed residents to share their rank-ordered preferences for OR assignment. Number of cases per OR, anesthesia base units, time units, and proportion of inpatient cases were used as proxies for case complexity. Data were analyzed using a mixed linear model. Residents requested rooms with fewer cases [F(3,22,350) = 194.0; p < 0.001], more base units [F(3,19,158) = 291.4; p < 0.001], more time units [F(3,19,744) = 186.4; p < 0.001], and a greater proportion of cases requiring inpatient preoperative evaluation [F(3,51,929) = 11.3; p < 0.001]. In most cases, these differences were greater for more advanced residents. As hypothesized, residents requested ORs with higher case complexity, and these cases more often required inpatient preoperative evaluation. More advanced residents exhibited a stronger preference for more educational cases than junior residents.