OBJECTIVE When performing middle ear operations, such as ossiculoplasty or stapes surgery, patients and surgeons expect an improvement in air conduction (AC) hearing, but generally not in bone conduction (BC). However, BC improvement has often been observed after surgery, and the present study investigated this phenomenon. METHODS We reviewed the preoperative and postoperative surgical outcomes of 583 patients who underwent middle ear surgery. BC improvement was defined as a BC threshold decrease of >15 dB at two or more frequencies. Subjects in group A underwent staged ossiculoplasty after canal wall up mastoidectomy (CWUM), group B underwent staged ossiculoplasty after canal wall down mastoidectomy (CWDM), group C underwent ossiculoplasty only (thus, they had no prior history of CWUM or CWDM), and group D received stapes surgery. We created a hypothetical circuit model to explain this phenomenon. RESULTS BC improvement was detected in 12.8% of group A, 9.1% of group B, and 8.5% of group C. The improvement was more pronounced in group D (27.0%). A larger gain in AC hearing was weakly correlated with greater BC improvement (Pearson's r=0.395 in group A, P<0.001; r=0.375 in group B, P<0.001; r=0.296 in group C, P<0.001; r=0.422 in group D, P=0.009). Notably, patients with otosclerosis even experienced postoperative BC improvements as large as 10.0 dB, from a mean value of 30.3 dB (standard error [SE], 3.2) preoperatively to 20.3 dB (SE, 3.2) postoperatively, at 1,000 Hz, as well as an improvement of 9.2 dB at 2,000 Hz, from 37.8 dB (SE, 2.6) to 28.6 dB (SE, 3.1). CONCLUSIONS BC improvement may be explained by a hypothetical circuit model applying the third window theory. Surgeons should keep in mind the possibility of BC improvement when making a management plan.