We assessed the differential complications and mortality rates of teaching versus nonteaching hospitals in July against other month-to-month differences in a cohort of 324,988 elderly patients hospitalized for a femoral neck or intertrochanteric fracture (data taken from the 1998-2003 National Inpatient Sample). Demographics were similar between teaching and nonteaching hospitals and across admission months. The overall mortality rate was 3.64% and was slightly higher in teaching hospitals compared with nonteaching hospitals (3.69% vs. 3.61%, relative risk [RR] = 1.0062, 95% CI 0.99-1.02). The adjusted relative risk (RR) for mortality in July/August was significantly higher than the overall adjusted RR and compared with all other month pairs, indicating higher in-hospital mortality rates in teaching hospitals compared with nonteaching hospitals. Intraoperative complications and length of stay were statistically significantly greater in teaching hospitals but did not demonstrate a "July effect." Teaching hospitals had lower perioperative complication rates. Elderly hip fracture patients treated at teaching hospitals had 12% greater relative risk of mortality in July/August (ie, experience a "July effect") compared with nonteaching hospitals during that time period (1998-2003). Although various methods exist for exploring the "July effect," it is critical to take into account inherent month-to-month variation in outcomes and to use nonteaching hospitals as a control group.