权益分类	功能权益	普通用户	{{item.name}}会员
{{category.name}}	{{benefitItem.name}}

A study on panel survey and panel analysis for transportation planning

交通规划面板调查与面板分析研究

基本信息

批准号：
08305019
负责人：
KITAMURA Ryuichi
金额：
$ 3.97万
依托单位：
KYOTO UNIVERSITY
依托单位国家：
日本
项目类别：
Grant-in-Aid for Scientific Research (B)
财政年份：
1996
资助国家：
日本
起止时间：
1996 至 1997
项目状态：
已结题

项目摘要

Models of attitude change or behavioral change require repeated measurements obtainable only in panels. If accurate measurements of travel behavior are required for the analysis, such data cannot be obtained from retrospective questions asked of a cross section, but must be obtained by repeated measurements from a panel survey. Analytically, "Longitudinal data often provide more statistical leverage in sorting out causal patterns because they enable the analyst to separate effects of persistent interpersonal differences from real inter-temporal relationships" (Duncan, et al., 1987). In particular, it is well established that consistent model estimation is not possible with cross-sectional data when unobserved heterogeneity is present which tends to be in variant over time (e.g., Davies & Pickles, 1985).With this recognition, our research group has developed several dynamic models for an individual's travel and activity behavior, household vehicle ownership and aggregated indicator for vehicle ownership. Travel behavior models that account for unobserved heterogeneity across individuals or bias of panel attrition are developed using panel data.Our research group held two special sessions for panel analysis and panel survey in annual meeting on infrastructure planning in 1996 and 1997. In these special sessions, models of individuals' dynamic behavior using panel data or methodology for dealing with biases generated with panel attrition, conditioning or fatigue are reported. And advantages and availability in actual transportation planning of panel are discussed.

态度改变或行为改变的模型需要重复测量，只能在小组中获得。如果分析需要对出行行为进行精确测量，则此类数据不能从对横截面的回顾性问题中获得，而必须通过小组调查的重复测量获得。从分析上讲，“纵向数据通常在整理因果模式时提供更多的统计杠杆，因为它们使分析师能够将持久的人际差异的影响与真实的跨时间关系分开”（邓肯等人，1987年）。特别是，已经明确的是，当存在未观察到的异质性时，不可能用横截面数据进行一致的模型估计，这种异质性往往随时间而变化（例如，Davies & Pickles，1985）。基于这一认识，我们的研究小组开发了几个动态模型，用于个人的旅行和活动行为、家庭车辆拥有量和车辆拥有量的综合指标。本课题组在1996年和1997年召开的基础设施规划年会上，分别进行了两次专题小组分析和小组调查。在这些特别会议上，报告了使用面板数据或方法处理面板损耗，条件反射或疲劳产生的偏见的个人动态行为模型。并探讨了该方法的优越性及其在实际交通规划中的有效性。