Over-prescription of antibiotics for non-pneumonia respiratory tract infections (RTIs) is a major concern in general practice. Australian general practice registrars (trainees) have inappropriately high rates of prescription of antibiotics for RTIs. The 'apprenticeship' educational model and the trainee- trainer relationship are drivers of this inappropriate prescribing. We aimed to reduce registrars' non-pneumonia RTI antibiotic prescribing via an educational intervention (a 90-min face-To-face workshop supported by online modules), complemented by delivery of the same intervention, separately, to their trainers. We conducted a pre-and post-intervention comparison of the registrars' intention to prescribe antibiotics for common RTIs using McNemar's test. We similarly tested changes in supervisors' intended prescribing. Prescribing intentions were elicited by responses to six written clinical vignettes (upper respiratory tract infection, otitis media, sore throat and three acute bronchitis vignettes). We found that, for registrars, there were statistically significant reductions in antibiotic prescribing for the sore throat (24.0% absolute reduction), otitis media (17.5% absolute reduction) and two of the three acute bronchitis (12.0% and 18.0% absolute reduction) vignettes. There were significant reductions in supervisors' antibiotic prescribing intentions for the same four vignettes. We conclude that our intervention produced a significant change in registrars' intention to prescribe antibiotics for non-pneumonia RTIs.