Validation of a video-based tool to assess risk perception both objectively and subjectively in pedestrians

David Herrero-Fernández; Carla Álvarez Ferradas; Smaranda R. Bogdan-Ganea; Juan Luis Martín Ayala

Resumen:

Introduction: Road crashes involving pedestrians are still a relevant cause of death and injury in Spain. Risk perception in pedestrians has been proposed as one of the main predictors of risky behaviors and crash-related events. The current research aimed to validate a video-based tool to assess risk perception in pedestrians, considering both the subjective way (self-report) and the objective way (skin conductance level and response). Method: A set of eight main videos (four involving a high-risk situation, and the other four involving low-risk situations) were presented to 403 participants (Mage = 22.41, 61.5% female) who rated the amount of risk perceived subjectively, following a 5-point Likert scale. Out of which, 90 participants taken incidentally were also assessed psychophysiologically, analyzing electrodermal activity while watching the videos. Results: The results showed a good fit of the bifactor SEM carried out with the self-report, supporting the existence of two uncorrelated factors (Risk Perception at High-Risk Situations, ω = 0.84; Risk Perception in Low Risk Situations, ω = 0.67) and a global factor (Risk Perception). Significant differences were obtained in the comparison of risk perception in high-risk situations and low-risk situations, both in the self-report measure (η2 = 0.93) and in skin conductance response (η2 = 0.20). Discussion: It is concluded that the tool is valid for assessing risk perception both subjectively and objectively. This could have several practical implications, particularly as a measure to validate intervention programs aimed at improving risk perception ability. Moreover, given the tool’s non-verbal nature, it may be suitable for individuals with low reading skills or cognitive disabilities. However, its validity should be further examined in these specific populations.