Abstract
This article describes a model by which it is possible to get insight into the factors influencing the occurrence of accidents. By its multiplicative form and the input of observations on situations where accidents occurred, a linear transformation was impossible without biasing the estimates considerably. By introducing the likelihood theory a model was developed which also analyses situations where no accidents occur. This method is applied successfully in the analysis of accidents on polder roads interurban provisions for bicycles and the influence of lane width on accidents.