Abstract
There is a considerable disagreement about, the effectiveness of the use of traffic training gardens within the frame-work of traffic education. It is therefore useful to review the existing experimental data. Four variables are important (1) the training condition or situation (where), (2) the learning objectives or tasks (what); (3) the learning method (how); and (4) the age (who). The results of the experiments are mostly opposite, but it is shown that the real traffic situation is better fitted for the learning of pedestrian behaviour than a traffic training garden.