Road Infrastructure Characteristics Associated with Jaywalking-Related Versus Legal-Crossing Pedestrian Injury Crashes in Abu Dhabi and Al Ain

Pedestrian crossing safety is influenced by the accessibility and configuration of road infrastructure; however, crash-based evidence on jaywalking-related events remains limited in Gulf countries. This retrospective case–case study compared police-recorded jaywalking-related and legal-crossing pedestrian injury crashes in Abu Dhabi and Al Ain. The dataset comprised 1548 crossing-pedestrian injury crashes recorded during 2012–2020: 1201 (77.58%) jaywalking-related and 347 (22.42%) legal-crossing crashes. Conventional binary logistic regression with forward selection was used after excluding outcome-linked variables, with Firth regression as an estimation-method sensitivity analysis. The final model identified adjusted associations with pedestrian crossing green time, numbers of crossing facilities within 200 m and 500 m, road type, road width, number of lanes, median width, posted speed limit, urban/rural setting, and lighting condition. Wide versus narrow roads had an adjusted odds ratio of 6.23 (95% confidence interval: 3.17–12.24). The conventional model had an apparent area under the receiver operating characteristic curve (AUC) of 0.86 and Nagelkerke’s pseudo-R2 of 0.409; a 1000-resample Firth-based sensitivity analysis yielded an optimism-corrected AUC of 0.84. These findings may inform infrastructure audits and pedestrian-safety planning, although estimates for sparse categories warrant caution. The associations are conditional on a police-recorded injury crash and should not be interpreted as causal effects or estimates of crash risk or jaywalking prevalence among the general pedestrian population.