Abstract
Every means heretofore identified for evaluation full-scale delineation systems has notable strengths and weaknesses. Before and after accident studies require lengthy time periods and may not able to distinguish extraneous effects. An increasingly viable analytic alternative is to search for more indirect symptoms of hazardous driver performance. An Attempt is made to bring into better forms the rather difficult matter of systematically defining and interpreting in a practical manner the monitored behaviour of the driver.