This contribution gives a report of some practical experiences that have been gained in two statistical investigations. The subject of these investigations was the effect of one-way traffic on road safety in residential areas. The numerical data for both investigations consisted of accident statistics covering several years. A fairly conventional method was used, which cannot exactly be considered as "short-term". The short-term character was obtained in this case by using exclusively historical data. If the design of the investigation has been adequately worked out and the input data are at the disposal of the researchers, such a "post-hoc" study should in principle demand very little time. (in practice, however, this expectation is unfortunately found to be over-optimistic in many cases). It will shortly become clear from the brief descriptions that the two investigations can be regarded as "area-wide". One-way traffic is in most cases introduced for reasons other than the improvement of road safety. It does not, therefore, represent a specific safety measure. However, the effect of this measure on road safety was the central factor in the investigations. In this respect too, the contribution falls within the scope of the seminar. The first investigation is described briefly in section 2. A reasonably well known statistical technique was used here. For this reason, the experiences described in section 3 refer primarily to the problems related to the subject to which the technique was applied in this case. Some of the remarks made refer equally to the second investigation, which is described briefly in section 4. In this investigation, an attempt has been made to obtain more specified results from the same data by means of a different statistical technique. Since this technique is less well known, the experiences recorded in section 5 apply more to the method used in this second investigation. (a) for the covering abstract of the conference see IRRD abstract no 264967.
Samenvatting