这是用户在 2024-3-25 16:32 为 https://snip.mathpix.com/haoyuw57/notes/43ff18be-9ab5-4152-91c0-8cbc5cd46245/edit 保存的双语快照页面,由 沉浸式翻译 提供双语支持。了解如何保存?

Juan de Oña ${ }^{a}$, Rocío de Oña ${ }^{a}$, Laura Eboli ${ }^{\text {b,*, Gabriella Mazzulla }{ }^{b}}$ \\ a TRYSE Research Group, University of Granada, Department of Civil Engineering, SeveroOcho s/n, Granada 18071, Spain \\ ${ }^{\mathrm{b}}$ University of Calabria, Department of Civil Engineering, Via Pietro Bucci, Rende 87036, Italy

\section*{A R T I C L E I N F O}

Available online 7 August 2013


Bus transit service

Overall Service Quality

Passengers' perceptions

Structural Equation Modelling

This paper proposes a methodology for evaluating the quality of service perceived by users of a bus transit service. A Structural Equation Model (SEM) approach is used to reveal the unobserved latent aspects describing the service and the relationships between these aspects with the Overall Service Quality. Data from a Customer Satisfaction Survey conducted by the Transport Consortium of Granada (Spain) are analyzed. A total of 1200 surveys were collected, and two passengers' statements about the Overall Service Quality were gathered: the first one when passengers have not reflected on the attributes describing the service, and the second one after they have thought about them. This is the first time that the Overall Service Quality of a public transport system has been jointly explained by these two overall evaluations when a SEM approach is adopted.

Some interesting results have been obtained. Three latent variables were identified representing the main characteristics of the service. The unobserved latent construct obtaining the highest weight on Overall Service Quality is Service, while Comfort and Personnel have little influence. The passengers' evaluation better explaining the Overall Service Quality is the evaluation made when passengers have reflected on the service.

The findings of this research can provide operating companies and transport managers valuable information for designing appropriate transport policies attracting new passengers and retaining the current ones.

(C) 2013 Elsevier Ltd. All rights reserved.

\section*{1. Introduction}

Nowadays the success of a public transport system depends on the number of passengers which the system is able to attract and retain. For this reason, the quality of a service becomes an issue of maximum importance because it is known that an improvement in the level of quality of the service leads to a higher satisfaction of the passengers and to an increase of the use of the system.

Service quality is related to a series of attributes describing the Public Transport (PT) service. Berry et al. $(1990)$ point out that "customers are the sole judges of service quality", and many authors have also supported this theory. Therefore, if service quality is measured from the customer's perspective, transit quality depends on the passengers' perceptions about each attribute characterizing the service.

In order to design appropriate transport strategies, operating companies monitor the perceptions of the users about the service every year or with a 6-month frequency. These perceptions are usually
* Corresponding author. Tel./fax: +39 0984496784.

E-mail address: laura.eboli@unical.it (L. Eboli).

measured by customer satisfaction surveys, and the data collected are used for developing indices providing useful information about the global quality of service and its evolution along the time. However, for determining these measures, they need not only to know the perceptions about the attributes of quality, but also to identify which attributes have the highest influence on the global assessment of the service. Asking customers to rate each attribute on an importance scale is the method mostly used by the operating companies.

However, previous studies showed that the factors affecting the global evaluation of the passengers about the service can vary when they are provoked into thinking about some attributes of the service which they did not consider before. dell'Olio et al. (2010) demonstrated that passengers may change his overall evaluation when they are made to reflect on the attributes characterizing the service, and de Oña et al. (2012) discovered that the key factors influencing the perception of the passengers about a bus transport service are different before and after their reflection.

So, asking customers to state the importance of each service attribute can lead to erroneous estimation, because some attributes can be rated as important even though they have little influence on overall quality, or they are important only in one of the moments of the assessment (before or after thinking).

For this reason, derived importance methods, which determine the importance of the attribute by statistically testing the strength of the relationship of individual attributes with overall satisfaction, are preferred by researches because of their numerous advantages (Weinstein, 2000), although they are not very used because of their high complexity.

In the field of public transportation and based on customer satisfaction surveys, the derived importance approaches mostly used for investigating on customer satisfaction and transit service quality have been: regression analysis (e.g. Aksoy et al., 2003; Dell'Olio et al., 2010; Huse and Evangelho, 2007; Kim and Lee, 2011; Tyrinopoulos and Aifadopoulou, 2008; Tyrinopoulos and Antoniou, 2008; Weinstein, 2000) and methods based on factor analysis, as Principal Component Analysis (PCA) (e.g. Ching-Chiao et al., 2009; Chin-Shan, 2007; Kolanovic, 2008; Lai and Chen, 2010; Pantouvakis, 2010; Rahaman and Rahaman, 2009; Sezhian et al., 2011), Confirmatory Factor Analysis (CFA) (e.g. Changa and Chen, 2007; Yu and Lee, 2011) or Structural Equation Models (SEM) (e.g. Andreassen, 1995; Eboli and Mazzulla, 2007, 2012; Irfan Syed et al., 2011; Karlaftis et al., 2001; Ngatia et al., 2010; Stuart et al., 2000).
\hline & & $x_{2}$ & Punctuality (Item2) & 1.459 & 0.121 & 0.000 & 0.591 \\



西班牙格拉纳达大学土木工程系 TRYSE 研究小组,地址:Severo Ocho s/n,格拉纳达 18071

卡拉布里亚大学,土木工程系,意大利雷恩德皮耶特罗布奇大街 87036 号


2013 年 8 月 7 日在线可用



本文提出了一种评估公交服务用户感知服务质量的方法。采用结构方程模型(SEM)方法揭示描述服务的未观察到的潜在因素以及这些因素与整体服务质量之间的关系。对西班牙格拉纳达交通联盟进行的客户满意度调查数据进行了分析。共收集了 1200 份调查问卷,并收集了两个乘客关于整体服务质量的陈述:第一个是在乘客没有考虑描述服务的属性时,第二个是在他们考虑了这些属性之后。这是首次采用 SEM 方法共同解释公共交通系统整体服务质量的这两个整体评估。



(C)2013 Elsevier Ltd. 版权所有。




为了设计合适的运输策略,运营公司每年或每 6 个月频率监测用户对服务的感知。这些感知通常是


然而,先前的研究表明,影响乘客对服务的整体评价的因素可能会因为他们被迫思考一些之前没有考虑过的服务属性而发生变化。dell'Olio 等人(2010)证明,当乘客被要求反思服务的特征属性时,他们可能会改变对整体评价的看法;而 de Oña 等人(2012)发现,在乘客反思之前和之后,影响他们对公交运输服务感知的关键因素是不同的。



在公共交通领域,基于客户满意度调查,用于调查客户满意度和交通服务质量的主要派生重要性方法有:回归分析(例如 Aksoy 等,2003 年;Dell'Olio 等,2010 年;Huse 和 Evangelho,2007 年;Kim 和 Lee,2011 年;Tyrinopoulos 和 Aifadopoulou,2008 年;Tyrinopoulos 和 Antoniou,2008 年;Weinstein,2000 年)和基于因子分析的方法,如主成分分析(例如 Ching-Chiao 等,2009 年;Chin-Shan,2007 年;Kolanovic,2008 年;Lai 和 Chen,2010 年;Pantouvakis,2010 年;Rahaman 和 Rahaman,2009 年;Sezhian 等,2011 年),验证性因子分析(例如 Changa 和 Chen,2007 年;Yu 和 Lee,2011 年)或结构方程模型(例如 Andreassen,1995 年;Eboli 和 Mazzulla,2007 年,2012 年;Irfan Syed 等,2011 年;Karlaftis 等,2001 年;Ngatia 等,2010 年;Stuart 等,2000 年)。

SEM 方法学已广泛应用于多个研究领域,近年来在公共交通服务质量领域中开始被最频繁地使用。这是因为服务质量是一个复杂、模糊和抽象的概念(Carman, 1990; Parasuraman et al., 1985),取决于一系列观察和未观察到的变量。这些未观察到的变量通常被称为维度。这些维度用于更好地理解客户如何感知各种服务属性,通过将它们分组成代表相似属性的因子。

当这些维度没有事先确定时,可以使用统计方法来确定它们。最流行的方法是因子分析,它分析大量属性是否与较少的未观察变量呈线性相关。各种作者在他们的研究中使用了这种方法,作为先前步骤中使用其他统计方法分析服务质量的方法,例如 Aksoy 等人(2003)在预测航空公司满意度之前应用了这种方法进行判别分析;Eboli 和 Mazzulla(2007)使用因子分析和结构方程模型评估公交运输方面对全球客户满意度的影响;Kim 和 Lee(2011)和 Weinstein(2000)在因子分析之后使用多元线性回归技术。Kim 和 Lee(2011)评估了韩国国内航空公司的质量,Weinstein(2000)研究了旧金山地区快速交通的服务因素对乘客整体满意度的相对重要性。

因此,本研究的主要目的是确定一系列描述公交运输服务质量的特征对整体服务质量(OSQ)的影响。本文的另一个目的是揭示哪些未观察到的潜在方面代表了服务的主要特征,这些特征由描述服务质量的属性所描述。在这项工作中,没有使用因子分析,而是采用结构方程模型(SEM)方法。提出了四种不同的模型,并找到了更好的拟合结构。这是首次使用 SEM 分析公共交通中的 SQ,使用两个不同的乘客对服务的整体评价(在反思服务属性之前和之后)作为解释 OSQ 的观察变量。考虑这两个关于 SQ 的评价可以帮助更好地理解 OSQ 的概念。在 Eboli 和 Mazzulla(2007)和 Eboli 和 Mazzulla(2012)中,OSQ 是通过唯一可用的满意度指标来衡量的,并且是通过一个重要性指标来不当地衡量的。拥有两个不同的整体满意度评价的可用性为更好地衡量 OSQ 的潜在构建提供了良好的机会。 此外,我们认为拥有这两种不同的判断是有优势的,因为它可以在面试的两个不同时刻对整体服务质量的评估进行有趣的研究,正如我们在实验环境部分所观察到的。



结构方程模型(SEM)方法是一种强大的多变量分析技术,可以建立观察变量和未观察变量之间的一组关系来对现象进行建模。尽管它是在 1970 年代开始的一个相对较新的方法(Fornell 和 Lacker,1981),但它已经广泛应用于各种研究领域,包括心理学、教育、社会科学、经济学、统计学等。SEM 方法是指一系列统计技术,如因子分析、路径分析和回归模型,用于分析数据。


结构模型的基本方程定义如下(Bollen, 1989):
η = B η + Γ ξ + ζ η = B η + Γ ξ + ζ eta=B eta+Gamma xi+zeta\eta=B \eta+\Gamma \xi+\zetaη=Bη+Γξ+ζ

其中 η η eta\etaη 是潜在内生变量的 m × 1 m × 1 m xx1m \times 1m×1 向量,