Evaluating the Impact of Leading Pedestrian Intervals and Curb Extensions on Conflict Frequency and Severity

Leading Pedestrian Intervals (LPIs) and curb extensions are commonly implemented countermeasures intended to improve pedestrian safety at signalized intersections. While LPIs have a well-established crash reduction record, fewer studies have evaluated their impacts using surrogate safety measures or examined how curb extensions interact with LPIs to influence pedestrian-vehicle conflicts. This thesis investigates the effects of LPIs and curb extensions on pedestrian safety using detailed conflict data collected at eleven signalized intersections in the Portland, Oregon metropolitan area. Approximately 600 hours of video data were collected across four matched intersection pairs and three before–after sites. Pedestrian and vehicular volumes, turning speeds, and pedestrian–vehicle conflicts were manually extracted. Conflicts were quantified using Post Encroachment Time (PET) and categorized by severity based on PET and turning-vehicle speed thresholds. Analyses included descriptive statistics, statistical tests, conflict frequency modeling using count-data regression, and conflict severity modeling using discrete-choice methods. Results indicate that LPIs are consistently associated with lower pedestrian-vehicle conflict frequencies relative to untreated conditions, with more robust evidence observed in regression models that control for exposure and site characteristics. Intersections with both LPIs and curb extensions generally exhibited lower proportions of high-severity and high-speed conflicts, suggesting that geometric treatments may complement operational measures by influencing turning vehicle behavior. Sensitivity analyses show that severity-related findings depend on the choice of speed threshold, highlighting the importance of transparent severity definitions in conflict-based studies. Further, results from both the descriptive and regression analysis indicate that results vary based on the specific intersection and related characteristics. Overall, the findings suggest that LPIs are associated with meaningful reductions in pedestrian–vehicle conflicts, while curb extensions may further reduce conflict severity when implemented alongside LPIs, though evidence is weaker. The results also demonstrate the utility and limitations of surrogate safety measures for evaluating pedestrian countermeasures at a fine temporal and spatial scale.