Development of crash modification factors for roadway illuminance: A matched case-control study

Roadway lighting is used to ensure nighttime safety and security for multimodal road users. However, the absence of reliable quantitative analyses of the safety effects of roadway lighting photometric characteristics prevents accurate assessment of street lighting maintenance and retrofitting projects. This study aimed to investigate the relationship between nighttime crash risk and two critical photometric criteria, i.e., average lighting level and uniformity. To achieve this goal, high-resolution horizontal illuminance data were collected in Florida for 300 + center-miles from 2011 to 2014. Based on the data, a matched case-control study was conducted to address two major issues existing in previous studies: (1) the confounding effects of illuminance standard deviation on illuminance mean and (2) spatially-unrelated extreme values for ratio-based uniformity. By eliminating the confounding effects through a random matching strategy (one case, a segment with nighttime crashes, to one control, a segment without nighttime crashes at 1,046 strata), this study successfully isolated the negative effects of the standard deviation and developed more significant crash modification factors (CMFs) for average lighting levels: 0.679 for increasing the average lighting level from [0 fc, 0.5 fc] to (0.5 fc, 1.0 fc] and 0.581 for increasing the average lighting from [0 fc, 0.5 fc] to higher than 1.0 fc. Additionally, a CMF of 1.391 for a max-min ratio greater than 10 was identified by controlling the segment length at a short uniform unit (1,200 ft). The developed CMFs overcame the underestimation issue in previous studies and are implementable in current street lighting design and safety management.