Causation Modeling And Intervention Techniques Of Risky Riding Behavior Based On Behavior Change Theory

In the urban road traffic environment,electric two-wheelers are widely used because of their convenience.With the rapid growth of their number,the proportion of accidents involving electric two-wheelers in urban road casualties has gradually increased,posing a serious threat to people’s lives and property safety.Based on a review of the research progress on the characteristics of electric two-wheeler accidents,accident influencing factors,and rider risk behaviors and behavioral interventions at home and abroad,this study takes electric twowheeler accidents and rider risk behaviors as the research objects,with the objectives of improving the safety of electric two-wheeler travel and enhancing the intervention effects of rider risk riding behaviors.We systematically describe the main patterns of accidents and the characteristics of pre-crash risk behaviors,focus on the mechanism of the influence of pre-crash behaviors on accident severity,and construct a quantitative prediction model to identify the risk behaviors with significant influence.The risk behavior change questionnaire is developed and validated to reveal the main theoretical domains that influence the occurrence of behavior change and explore the mechanisms of risk behavior change.To develop intervention protocols and conduct intervention control experiments through intervention strategy modeling to evaluate the intervention effects and provide scientific basis for developing effective intervention strategies for cyclists’ risky behaviors and changing individual cyclists’ inherent risky behaviors.Firstly,based on the literature review,the basic theories and methods applied to carry out this study are systematically sorted out,and the three aspects of traffic accident depth survey theory,discrete choice model and behavior change theoretical are elaborated to clarify the research methods used in this study,such as statistical modeling regression analysis,questionnaire survey method and behavior intervention experimental method.Secondly,based on the CIDAS project database,the vehicle-electric two-wheeler collisions from 2011-2020 were screened according to the conditions to construct an accident dataset under urban road environment.The statistical description of the accident number distribution under different road and environment characteristics,accident patterns,accident occurrence time and other conditions,and the deep excavation of rider behavior characteristics and injury characteristics further support the necessity of conducting this study.Further,for the typical scenario of electric two-wheeler accidents,based on the refined accident videos provided by Xi’an traffic police department,four factors including the behavior of traffic participants before the accident affecting the accident severity,such as people,vehicles,roads and environment,were extracted,and the traditional fixed-parameter discrete selection model was improved by considering data heterogeneity,endogeneity and orderliness,and a random-parameter Logit and a partial proportional odds models for accident severity prediction.Using random utility theory,the mechanism of the influence of road traffic constituents on accident severity is quantitatively analyzed by interpreting the results of model parameters and elasticity analysis and marginal effect analysis.It was found that the parameters of risky behaviors such as speeding and red light running of cyclists were positive,and the risky behaviors significantly increased the probability of injury as well as fatal accidents.Meanwhile,targeted countermeasures to prevent the generation of risky behaviors are proposed.Videobased data mining achieves a focus on pre-crash behaviors and expands the range of accessible data on factors influencing accident severity.Again,the behavioral research method based on the theoretical domain framework(TDF)was elaborated based on the frontier theory of social psychology,and the phased development process and cross-validation strategy of the behavior change questionnaire were determined,taking the red light running behavior of online delivery workers in an occupational setting as the research object.The test entries of each theoretical domain were designed,the pretest questionnaire was improved by a small sample survey,and a measurement model containing39 entries and 13 potential factors was developed.The valid samples of the formal survey were subjected to descriptive statistics,and the validity of the questionnaire was verified by reliability and validity analysis,and the behavior change questionnaire was found to have good reliability and validity.The quantitative prediction model of behavioral intention of online delivery workers waiting for green lights was established by using the ability,opportunity,motivationbehavior(COM-B)model in the field of social psychology,and the quantitative prediction model of behavioral intention of online delivery workers waiting for green lights was established by using a hierarchical regression analysis method to determine the degree of influence of factors in each theoretical domain on behavioral intention,and to identify the barriers and facilitators that affect the behavioral change of delivery workers waiting for green lights.The behavioral intention to wait for green lights was found to be higher among the welleducated,less experienced delivery workers,and those who had no traffic police penalty experience in the past year,while those with low scores in the domain of "emotional regulation" and "social/professional role identity","social influence",and "knowledge" were found to have higher behavioral intention to wait for green lights.The final model explained 75% of the variance in behavioral intention.The TDF has shown promise in understanding the predictors of traffic safety behavioral intentions,and the questionnaire developed can be used by safety managers and distributors to provide a basis for developing targeted intervention programs.Finally,aiming at the potential barriers to the safe riding behavior of online delivery workers,based on the Behavior Change Wheel(BCW)theory,a three-stage eight-step method was used to model a nonlinear and systematic intervention strategy,and 7 intervention functions,4 policy categories and 26 key behavioral intervention techniques.The carrier forms corresponding to various technologies are clarified,behavioral intervention programs are developed,and intervention effect evaluation indicators including the possibility of occurrence of risky behaviors and the ability to identify risky behaviors are determined.Through the implementation of the intervention plan,carry out the intervention control experiment,compare and analyze the key parameters before and after the implementation of the intervention,and evaluate the effect of the proposed intervention technology based on BCW theory.It was found that there were differences in the influence of two different intervention modes on the risky cycling behavior of delivery personnel.The probability of risky cycling behavior in the experimental group and the control group decreased,but the decline in the experimental group was more obvious.Compared with the first intervention,the probability of risky behavior in the second intervention was significantly reduced,and the accuracy of risky behavior identification was significantly improved.The effectiveness and persistence of the behavioral intervention technology based on BCW is relatively good,confirming the possibility of sustained and effective behavioral change.It has practical guiding significance for traffic safety managers and enterprise intervention strategy makers to take intervention measures,and provides a theoretical basis for solving the risky riding behavior of delivery personnel.The findings provide new ideas and methods for the prevention of accidents and injuries of electric two-wheelers,and confirm their effectiveness,enriching the existing theoretical system of accident prevention.Additional evidence is provided for the importance of understanding the underlying factors influencing behaviour,with implications for the practice of risk assessment of cyclist behaviour,confirming the need for an enhanced understanding of people’s intrinsic risk.