The traditional method, `product' evaluation, applied to traffic safety work in industrialised countries cannot be transferred successfully to developing countries. In the latter, motorisation has occurred suddenly and rapidly; infrastructure and education have not kept pace; new ways of living, working and moving due to the above coexist with old ways; individual and social values as determinants of traffic behaviour may not have evolved signficantly. In developing countries, pedestrian accidents occur outside crossings because: there are few crossings; even adults have little experience in judging vehicle speeds; sight distances are short because of commercial activities at roadsides; public lighting is insufficient. Drivers' reactions are also inadequate. Safety work should encompass: (1) safety measures in line with causation processes; (2) evaluation of safety measures, mainly by `process' evaluation; (3) collation of information from different evaluation studies and production of training material for safety professionals and researchers. `Product' evaluation is difficult because of basic problems associated with lack of criteria to define controls and scarcity of accident data; some measures need a long time to show tangible results, and changes in developing countries occur so fast that the effects are not identifiable. Although `process' evaluation needs qualified staff, it must be combined with `product' evaluation for rational safety policies to be adopted.
Abstract