PURPOSE. To evaluate the efficacy of a physician-targeted website to improve knowledge and self-reported behavior relevant to strabismus and amblyopia ("vision") in primary care settings. METHODS. Eligible providers (filing Medicaid claims for at least eight well-child checks at ages 3 or 4 years, 1 year before study enrollment), randomly assigned to control (chlamydia and blood pressure) or vision groups, accessed four web-based educational modules, programmed to present interactive case vignettes with embedded questions and feedback. Each correct response, assigned a value of +1 to a maximum of +7, was used to calculate a summary score per provider. Responses from intervention providers (IPs) at baseline and two follow-up points were compared to responses to vision questions, taken at the end of the study, from control providers (CPs). RESULTS. Most IPs (57/65) responded at baseline and after the short delay (within 1 hour after baseline for 38 IPs). A subgroup (27 IPs and 42 CPs) completed all vision questions after a long delay averaging 1.8 years. Scores from IPs improved after the short delay (median score, 3 vs. 6; P = 0.0065). Compared to CPs, scores from IPs were similar at baseline (P = 0.6473) and higher after the short-term (P < 0.0001) and long-term (P < 0.05) delay. CONCLUSIONS. Significant improvements after the short delay demonstrate the efficacy of the website and the potential for accessible, standardized vision education. Although improvements subsided over time, the IPs' scores did not return to baseline levels and were significantly better compared to CPs tested 1 to 3 years later. © 2011 The Association for Research in Vision and Ophthalmology, Inc.