UNASSIGNED: To provide a mathematically simpler framework, the current study introduces parsimonious pooled random parameters logit models. Then, the standard pooled random parameters logit models without considering temporal effects are also simulated for comparison. By comparing the goodness of fit measure and estimation results, the parsimonious pooled random parameters logit model is suitable for capturing the temporal instability. Then, the non-transferability among helmeted and non-helmeted overspeeding motorcycle crashes is illustrated by likelihood ratio tests and out-of-sample prediction, and two types of models provide robust results. The marginal effects are also calculated.
UNASSIGNED: Several variables, such as age, cloudy and weekday indicators illustrate temporal instability. Moreover, several variables are observed to only show significance in non-helmeted models, showing non-transferability across helmeted and non-helmeted models.
UNASSIGNED: More educational campaigns, regulation and enforcement, and management countermeasures should be organized for non-helmeted motorcyclists and overspeeding behavior. Such findings also provide research reference for the risk-compensating behavior and self-selected group issues under overspeeding riding considering the usage of helmets.
■为了提供一个数学上更简单的框架,当前的研究引入了简约的混合随机参数logit模型。然后,还模拟了不考虑时间效应的标准混合随机参数logit模型进行比较。通过比较拟合优度度量和估计结果,简约的混合随机参数logit模型适用于捕获时间不稳定性。然后,通过似然比检验和样本外预测说明了头盔和非头盔超速摩托车碰撞之间的不可转移性,和两种类型的模型提供了稳健的结果。还计算了边际效应。
■几个变量,比如年龄,多云和工作日指标说明时间不稳定。此外,观察到几个变量仅在非头盔模型中显示出显著性,在头盔模型和非头盔模型中显示不可转移性。
■更多教育活动,监管和执法,应对无头盔摩托车和超速行为组织管理对策。这些发现也为考虑头盔使用的超速驾驶下的风险补偿行为和自我选择群体问题提供了研究参考。