Better later than ever: comparative analysis of feedback strategies in a dynamic intelligent virtual reality training environment for child pedestrians

Children require practical roadside training to learn safe pedestrian behaviour.
However, problems associated with exercising on real roads greatly restrict
the opportunities to provide such training. This paper presents the results of
a study on an alternative approach for practical safety training using a combination
of Intelligent Tutoring and Virtual Reality. In a classroom experiment, children
of second and third grades worked on virtual road crossing exercises. They
received instructions and feedback according to several different strategies. We
have observed a high general acceptance for this form of training and compared
effects of different feedback strategies on children's’ performance. The delayed
feedback strategy has been the most successful; its impact has been especially
notable on more advanced pedestrian safety skills that are the most challenging
for the children of the target age.