Utilizing Beta Regression in Predicting the Underlying Factors of Motorcycle Rider Behavior
Abstract
Introduction: Motorcyclists have the highest proportion of casualty toll caused by street accidents in Iran, and they endanger themselves and others by those risky behaviors. Health and safety education will not be sufficient without knowing the causes of such behaviors. Since no studies have been carried out based on accurate statistical methods on bounded response variables for motorcyclists' high-risk behaviors in Iran, this study aimed to predict MRBQ by ADHD and the underlying predictors using the Beta Regression as an alternative strategy.
Methods: The present sectional study included 311 Motorcyclists randomly selected using a cluster sampling method in Bukan city to evaluating the relationship between the limited response MRBQ with ADHD and its subscales. We used an innovative beta regression method for the analysis and carried out unadjusted and adjusted modeling.
Results: Direct and significant relationships were observed between MRBQ score and ADHD score and its subscales, including (DSS score) (coefficients ranged over 0.01 to 0.6, All P<0.05). Additionally, the riding period (coefficients ranged over -0.32 to -0.48, P<0.05), hours of riding (coefficients ranged over: 0.31 to 0.34, P<0.05), using the helmet (coefficients: 0.26 to 0.31, P<0.05), and sub-accident (coefficients ranged over 0.35 to 0.37, P<0.05) also turned out to be significant predictors of MRBQ score.
Conclusion: ADHD score and riding parameters can be taken into account when contriving actions on the motorcycle rider behaviors as measured by MRBQ.