This manual is one of the products of a study "Development of a Methodology for Accident Causation Research." The objective of this study was to fully develop and apply a methodology to study accident causation, which was outlined in a previous study. This methodology is based on a stochastic concept of causation: factors which influence the probability that an accident occurs in a given pre-crash situation are "causal" factors. Accident probabilities are estimated by accident rates. To calculate accident rates, exposure measures matched to the pre-crash situations are needed. Such exposure measures have to be obtained by observation of pre-crash situations. Data collection techniques to collect certain basic data by roadside observations were developed. A data collection design was developed, indicating when and where to collect data in Ulster and Schenectady Counties, New York. Data were collected, and exposure estimates for several pre-crash situations derived. Using police accident reports, accident rates were calculated for these pre-crash situations. The relations of these rates to various pre-crash factors and their interactions were studied. The methodology described in this manual is based on a previous study, it was further developed in this study, and parts of it were applied to and revised on the basis of the study in Ulster and Schenectady Counties.
Abstract