This research aims at building a multivariate statistical model for assessing users' perceptions of acceptance of ride-sharing services in Dhaka City. A structured questionnaire is developed based on the users' reported attitudes and perceived risks. A total of 350 normally distributed responses are collected from ride-sharing service users and stakeholders of Dhaka City. Respondents are interviewed to express their experience and opinions on ride-sharing services through the stated preference questionnaire. Structural Equation Modeling (SEM) is used to validate the research hypotheses. Statistical parameters and several trials are used to choose the best SEM. The responses are also analyzed using the Relative Importance Index (RII) method, validating the chosen SEM. Inside SEM, the quality of ride-sharing services is measured by two latent and eighteen observed variables. The latent variable 'safety & security' is more influential than 'service performance' on the overall quality of service index. Under 'safety & security' the other two variables, i.e., 'account information' and 'personal information' are found to be the most significant that impact the decision to share rides with others. In addition, 'risk of conflict' and 'possibility of accident' are identified using the perception model as the lowest contributing variables. Factor analysis reveals the suitability and reliability of the proposed SEM. Identifying the influential parameters in this will help the service providers understand and improve the quality of ride-sharing service for users.

翻译：本研究旨在构建一个多变量统计模型，以评估达卡市用户对拼车服务接受程度的感知。基于用户报告的态度和感知风险，开发了一份结构化问卷。从达卡市的拼车服务用户及相关利益者中收集了350份正态分布的响应数据。受访者通过陈述偏好问卷，表达其对拼车服务的体验和意见。采用结构方程模型（SEM）对研究假设进行验证。通过统计参数及多次试验筛选出最优SEM模型。同时，使用相对重要性指数（RII）方法对响应数据进行分析，以验证所选SEM模型的可靠性。在SEM模型中，拼车服务质量由两个潜变量和十八个观测变量衡量。潜变量“安全与保障”对整体服务质量指数的影响大于“服务绩效”。在“安全与保障”维度下，“账户信息”和“个人信息”这两个变量被证实是对决定与他人拼车行为影响最为显著的因素。此外，感知模型识别出“冲突风险”和“事故可能性”为贡献度最低的变量。因子分析结果证明了所提SEM模型的适宜性与可靠性。识别这些关键影响参数将有助于服务提供商理解并提升用户的拼车服务质量。