Minimising visitor impacts is a critical component of protected-area management. Teaching visitors to reduce impacts is preferred to enforcing regulations. But does education actually reduce impacts? Marion and Reid (2007) recently reviewed this question in this journal. They found only two studies that measured changes in site conditions following minimal-impact education, both in the USA. In fact, the effectiveness of minimal-impact education in reducing actual impacts of park visitors has been tested in Australia using controlled experiments (Littlefair, 2003, 2004). The education programme tested involved verbal interpretation by skilled guides working with ecotourists in World Heritage rainforest. Education for different groups taking the same guided walk was modified experimentally to include five levels: a control group with no environmental information; general environmental education but no minimal-impact information; verbal appeal for specific behaviours; role model demonstration by the guide, and both appeal and role model.