Abstract

Traffic signs are one of the oldest safety and traffic control equipments. Drivers’ reaction to installed signs is an important issue that could be studied using statistical models developed for target groups. There are 527 questionnaires have been filled up randomly in 45 days, some by drivers passing two northern cities of Iran and some by e-mail. Therefore, minimum sample size of 384 is fulfilled. In addition, Cronbach Alpha of more than 90 % verifies the questionnaire’s validity. Ordinal logistic regression is used for 5-level answer variables. This relatively novel method predicts probability of dif-ferent cases’ considering other effective independent vari-ables. There are 18 parameters of factor, man, vehicle, and environment are assessed and 5 parameters of number of accidents in last 5 years, occupation, driving time, number of accidents per day, and driving speed are eventually found as the most important ones. Age and gender, that are considered as key factors in other safety and accident studies, are not recognized as effective ones in this paper. The results could be useful for safety planning programs.