BACKGROUND: Artemisinin combination therapy (ACT) is first-line treatment for malaria in most endemic countries and is increasingly available in the private sector. Most studies on ACT adherence have been conducted in the public sector, with minimal data from private retailers. METHODS: Parallel studies were conducted in Tanzania, in which patients obtaining artemether-lumefantrine (AL) at 40 randomly selected public health facilities and 37 accredited drug dispensing outlets (ADDOs) were visited at home and questioned about doses taken. The effect of sector on adherence, controlling for potential confounders was assessed using logistic regression with a random effect for outlet. RESULTS: Of 572 health facility patients and 450 ADDO patients, 74.5% (95% CI: 69.8, 78.8) and 69.8% (95% CI: 64.6, 74.5), respectively, completed treatment and 46.0% (95% CI: 40.9, 51.2) and 34.8% (95% CI: 30.1, 39.8) took each dose at the correct time ('timely completion'). ADDO patients were wealthier, more educated, older, sought care later in the day, and were less likely to test positive for malaria than health facility patients. Controlling for patient characteristics, the adjusted odds of completed treatment and of timely completion for ADDO patients were 0.65 (95% CI: 0.43, 1.00) and 0.69 (95% CI: 0.47, 1.01) times that of health facility patients. Higher socio-economic status was associated with both adherence measures. Higher education was associated with completed treatment (adjusted OR = 1.68, 95% CI: 1.20, 2.36); obtaining AL in the evening was associated with timely completion (adjusted OR = 0.35, 95% CI: 0.19, 0.64). Factors associated with adherence in each sector were examined separately. In both sectors, recalling correct instructions was positively associated with both adherence measures. In health facility patients, but not ADDO patients, taking the first dose of AL at the outlet was associated with timely completion (adjusted OR = 2.11, 95% CI: 1.46, 3.04). CONCLUSION: When controlling for patient characteristics, there was some evidence that the adjusted odds of adherence for ADDO patients was lower than that for public health facility patients. Better understanding is needed of which patient care aspects are most important for adherence, including the role of effective provision of advice.