INTRODUCTION Over-prescription of antibiotics for common infective conditions is an important health issue. Infective conjunctivitis represents one of the most common eye-related complaints in general practice. Despite its self-limiting nature, there is evidence of frequent general practitioner (GP) antibiotic prescribing for this condition, which is inconsistent with evidence-based guidelines. AIM To investigate the prevalence and associations of GP registrars' (trainees') prescription of antibiotics for infective conjunctivitis. METHODS We performed a cross-sectional analysis of the Registrar Encounters in Clinical Training (ReCEnT) ongoing prospective cohort study, which documents GP registrars' clinical consultations (involving collection of information from 60 consecutive consultations, at three points during registrar training). The outcome of the analyses was antibiotic prescription for a new diagnosis of conjunctivitis. Patient, registrar, practice and consultation variables were included in uni- and multivariable logistic regression analyses to test associations of these prescriptions. RESULTS In total, 2333 registrars participated in 18 data collection rounds from 2010 to 2018. There were 1580 new cases of infective conjunctivitis (0.31% of all problems). Antibiotics (mainly topical) were prescribed in 1170 (74%) of these cases. Variables associated with antibiotic prescription included patients' Aboriginal or Torres Strait Islander status, registrar organisation of a follow up (both registrar and other GP follow up), and earlier registrar training term (more junior status). DISCUSSION GP registrars, like established GPs, prescribe antibiotics for conjunctivitis in excess of guideline recommendations, but prescribing rates are lower in later training. These prescribing patterns have educational, social and economic consequences. Further educational strategies may enhance attenuation of registrars' prescribing during training.