A comparative analysis of dangerous locations on the public roads in Serbia Krsto Lipovac a , Miladin Nešic ´ b , Milan Tešic ´ c,⇑ , Alan Ross e , Vladan Tubic ´ a , Bojan Maric ´ d a The Faculty of Transport and Traffic Engineering, Vojvode Stepe 305, 11 000 Belgrade, Serbia b The Academy of Criminal and Police Studies, Cara Dušana Street 196, 11080 Belgrade-Zemun, Serbia c The Paneuropean University Apeiron, College of Traffic Engineering, Vojvode Pere Krece 13, 78000 Banja Luka, (Republic of Srpska) Bosnia and Hercegovina d The Faculty of Transport and Traffic Engineering, Vojvode Mišic ´ 52, 74000 Doboj, (Republic of Srpska) Bosnia and Herzegovina e Road Safety Management Adviser, 25 Sefton Avenue, Newcastle upon Tyne, England, United Kingdom article info Article history: Received 1 November 2014 Received in revised form 19 December 2015 Accepted 21 December 2015 Keywords: Dangerous locations Objective and subjective risk Comparison abstract The proper identification of dangerous locations is one of the most significant phases of the Black Spot Management process. Identifying dangerous locations is a complex procedure, as they are not always located only in hazardous road sections. In fact, dangerous locations can be identified in safe sections (those having a low level of objective risk). This comparative analysis includes the results of two studies focusing on identifying hazardous road sections on the state roads in Serbia. The first study is based on field research and interviews of experts according to territorial principles (subjective method). The second study relies on analysis of data related to the history and spatial distribution of road accidents (objective method). The subject of this analysis is focused on results obtained in the analysis of the sub- jective and objective methods used for identifying dangerous locations. The objective of the comparative analysis is to identify differences between the results of the subjective and objective analyses, as well as to more precisely prioritize dangerous locations on Serbian roads. The comparative analysis has shown that sections with low risk according to the subjective method may include 1 km sections with very high risks, which as a consequence, may lead to incorrect selection of dangerous locations. Additionally, there are sections with a small number of recorded road accidents and a very high subjective risk. Therefore, a reliable list of prioritized dangerous locations is obtained using both methods for identification of dangerous locations and the comparative analysis of their results. Ó 2015 Elsevier Ltd. All rights reserved. 1. Introduction A review of the professional literature demonstrates the wide range of methods used for measuring the safety of road sections and identification of dangerous locations. Apart from the long- term efforts, the principles and techniques for identification of dangerous locations have not yet been harmonized, and the approaches used for that purpose differ from country to country (Elvik, 2008b). Methodologies range from a simple marking of a location with a large number of road accidents to the sophisticated techniques used to estimate the expected number of road crashes and determine potentials for improving road safety (Geurts and Wets, 2003). The methods for identifying dangerous locations can be divided into methods based on data concerning the struc- ture, number and severity of road crashes, their spatial distribution per road sections (objective methods), and methods based on sur- vey questionnaires offered to respondents and experts during field research of the road network (subjective methods). Subjective methods are less popular in academic circles because the output results depend on a range of factors, including professionalism and commitment of interviewed experts, content of the survey questionnaires, etc. When applying objective methods for identifi- cation of dangerous locations, it is obvious that there are safe road sections. However, some road sections may have 1 km road sec- tions with an outstanding number of road crashes, i.e., locations where the subjective risk of a road accident (subjective risk repre- sents the individual level of psychological comprehension of hazards on road sections, which leads to the change in driving behavior) is less obvious. For that reason, when putting up a list of dangerous locations for the rehabilitation of the road network in question, it will be necessary to simultaneously use both subjec- tive and objective methods. For example, in this case, application of a comparative analysis of results obtained in the subjective and http://dx.doi.org/10.1016/j.ssci.2015.12.019 0925-7535/Ó 2015 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +387 66633458. E-mail address: milan.te.sicm@gmail.com (M. Tešic ´). Safety Science 84 (2016) 190–200 Contents lists available at ScienceDirect Safety Science journal homepage: www.elsevier.com/locate/ssci