Service Quality Assessment and Optimization of High-Speed Railway Waiting Halls Using a Kano Model and Multidimensional Questionnaire Analysis

Abstract

With the rapid development of high-speed railways, the quality of service in the waiting halls of high-speed railway stations has become a subject of great concern. In order to clarify the impact of various service elements on the overall satisfaction associated with high-speed railway passenger stations, this study offers an in-depth exploration of the service quality of the waiting halls of high-speed railway stations by considering the physical environment (such as thermal environment, acoustic environment, light environment, and air quality), environmental design (including architectural design, route design, and hygiene situations), and service facilities (such as rest facilities, information facilities, safety features, commercial facilities, and ticketing facilities). The study uses a combination of an online questionnaire and an on-site questionnaire to collect data, and we ensured the reliability and validity of the research results through reliability and validity analyses. The Kano model was used to accurately identify the demand attributes of passengers for various service elements in the departure hall. Linear regression analysis was used to conduct a detailed study of the quantitative relationship between the influencing factors and overall satisfaction, and the satisfaction level of each dimension was systematically calculated to accurately quantify the impact of different factors on the overall satisfaction. Pearson correlation analysis was used to carefully explore the correlations among the factors and reveal the potential relationships. The study clearly depicts the performance of each service element. According to the demand classification of the Kano model, Must-Have Quality (M) elements include air quality, thermal environment, route design, the hygiene situation, and information facilities; Attractive Quality (A) elements include the acoustic environment, light environment, and architectural design; rest facilities, commercial facilities, and ticketing facilities are classified as One-Dimensional Quality (O); and safety facilities are of Indifferent Quality (I). Combined with regression analysis and correlation analysis, these results were used to further determine the focus of service element optimization. By clarifying the attributes of different service elements and their degree of impact on overall satisfaction, the corresponding optimization direction is proposed.

1. Introduction

1.1. Background

High-speed rail brings convenience to people’s lives and is generally regarded as one of the most sustainable development projects for surface transportation [1,2]. With the rapid development of the global economy, high-speed rail travel has become an important option for people, with its advantages of high efficiency and convenience, and the number of high-speed railway stations is increasing day by day. However, at present, there are deficiencies in the quality of high-speed rail passenger service [3]. As waiting halls are a key place where passengers wait for trains, the quality of service in the waiting halls of high-speed rail stations directly affects the travel experience of passengers [4]. Passenger satisfaction increases with improvements in the perceived service quality [5]. However, the relationship between various aspects of service quality and customer satisfaction may not follow a linear trend, which indicates that better service quality does not necessarily lead to higher customer satisfaction [6]. Therefore, it is necessary to classify the components of service quality in terms of their impact on customer satisfaction [7]. High-speed railway stations need to fully understand customer expectations and needs in order to reasonably allocate resources to improve service quality [8].

As people’s living standards continue to rise and their travel experiences accumulate, passengers’ demands and expectations for high-speed railway station services are escalating. Passengers not only expect the basic facilities to be complete, the environment to be clean, and the waiting space to be comfortable but also increasingly demand more personalized and distinctive services. Unfortunately, the current waiting-area services in high-speed railway stations still face numerous challenges.

From the perspective of the indoor environment, the indoor conditions in high-speed railway stations are more complex than those in traditional railway stations. Research by Du, X.H. (2020) [9] has shown that, during winter, the air temperature in high-speed railway stations often drops well below the comfort threshold. In summer, illuminance uniformity in some stations is unsatisfactory. Although the overall acoustic and air-quality conditions are generally acceptable, the acoustic environment has a relatively negative impact on passenger comfort. Regarding public facilities, using the universal design theory, Zhang, Jia-Rui (2024) [10] evaluated the public facilities in the waiting halls of Chinese urban high-speed railway stations and found that many facilities only meet the basic barrier-free design requirements. The application of universal design principles is insufficient, making it difficult to meet the diverse needs of all passengers. There are issues in multiple areas, such as toilet facilities and the placement of guidance signs.

These results indicate that there is considerable scope for enhancement in the overall service quality of high-speed railway stations. Research on the relationship between service quality and passenger satisfaction serves as the cornerstone for the rational allocation of resources and the effective enhancement of satisfaction levels. Geng, Y. (2017) [11] conducted a study of the relationship between indoor environmental quality (IEQ) factors and passenger satisfaction, finding that factors such as “thermal comfort” and “spatial layout” have distinct impacts on passenger satisfaction. From the perspective of comprehensive service quality evaluations, different studies have approached the topic from various angles. For instance, Cheng, X. Y. et al. (2018) [12] focused on bus-transfer service quality. However, these studies on individual service links lack comprehensiveness and are not systematic, and there is an urgent need for more in-depth research to accurately analyze the impact of different service factors on passenger satisfaction. Therefore, researching the factors that influence the improvement of high-speed railway station service quality is not only essential for meeting passengers’ needs and enhancing their travel experiences but also crucial for high-speed railway stations to maintain their competitive edge and achieve sustainable development in the highly competitive transportation market.

1.2. Objectives of the Study

This study aims to construct a comprehensive theoretical framework for analyzing the service quality of high-speed railway station waiting halls, elucidating the relationship between service elements and passenger satisfaction; we also propose targeted optimization strategies. These strategies are intended to enhance passenger satisfaction and travel experiences, thereby contributing new academic insights to the field of transportation hub service management. The specific research objectives are as follows:

• In terms of theoretical contributions, this study conducts multi-dimensional in-depth research on high-speed railway station waiting halls and performs a cluster analysis on the components of service quality. By moving beyond the linear hypothesis of traditional service quality research, we evaluate the service quality requirements of high-speed railway station waiting halls using the Kano model and an importance–satisfaction analysis, exploring the differentiated impact mechanisms of various service elements.
• Regarding methodological innovation, this study develops a dual data source fusion method combining “professional evaluation + passenger perception” to conduct an in-depth investigation of the service satisfaction status of different samples of high-speed rail station waiting halls. By conducting a questionnaire analysis, the satisfaction levels for sub-factors within each dimension are statistically analyzed, their importance ranked, and the influence of different factors on sub-project satisfaction and overall satisfaction clarified.
• In terms of practical value, based on the attribute classification results, importance rankings, linear regression equations, and relevant analysis outcomes, this study deduces optimization strategies and establishes an evaluation index system for the service quality of high-speed railway waiting halls. This system prioritizes resource allocation attributes, providing evidence-based decision support for operations.

The structure of this study is organized as follows: Section 2 presents a comprehensive literature review; Section 3 details the research methods, conceptual framework, and the design of research tools; Section 4 and Section 5, respectively, focus on the research process and analysis of results; and Section 6 draws conclusions, summarizes service-optimization strategies, reflects on research deficiencies, and looks ahead to future research directions.

2. Literature Review

2.1. Research on High-Speed Railway Waiting Hall Services

A comprehensive review of the relevant literature reveals that research on high-speed railway waiting hall services encompasses multiple key aspects, including the comfort of the physical environment, satisfaction with the building space environment, satisfaction with service facilities, and the relationship between service quality and passenger satisfaction.

Regarding the comfort of the indoor physical environment, numerous scholars have conducted in-depth investigations. Liu, G. (2015) [13] focused on summertime thermal comfort issues in high-speed railway stations in cold regions of China. By conducting questionnaire surveys and physical measurements in two typical high-speed railway stations, data on passengers’ thermal preferences and the in-station environmental conditions were collected. Du, X. (2020) [9] initially measured and surveyed high-speed railway stations and found that the air temperature in the winter did not meet standards and the illumination uniformity in the summer was unsatisfactory, but the overall acoustic and air environment was good. Moreover, different groups of passengers showed varying sensitivities, tolerances, and satisfaction levels towards the environment. Subsequently, Du, X. (2020) [14] focused on the interactive effects on human comfort of the thermal, lighting, and acoustic environments in high-speed railway waiting halls. With the development of China’s high-speed rail construction, the indoor physical environment of high-speed railway waiting halls has become more complex, with some areas deviating from the design expectations and affecting passengers’ environmental satisfaction. Jia, X. (2021) [15] undertook field research and simulation experiments, demonstrating that passengers’ adaptability to the indoor thermal environment was stronger than predicted by models, and their thermal comfort requirements varied depending on the length of stay and travel itinerary. Liu, G. (2016) [16] emphasized that in large-scale and high-occupancy-density buildings such as high-speed railway stations, the thermal comfort of waiting passengers in non-air-conditioned areas deserved attention. Yuan Y. (2023) [17] studied the changes in passengers’ thermal comfort during the space-transition process in high-speed railway stations, explored their dynamic thermal responses throughout the departure process, and investigated the thermal comfort requirements in different functional areas. Later, from the perspective of energy consumption, Yang, L. [18] (2015) took a medium-sized high-speed railway station in southern China as an example; using DeST software to simulate the load characteristics of the HVAC system, the study indicated that reducing air infiltration had significant energy-saving potential. Yuan, Y. (2024) [19] studied the thermal comfort of passengers during the multi-space transition process in high-speed railway stations in cold regions of China. Through on-site investigations and laboratory experiments, the optimal thermal environment sequence was determined.

In terms of service facilities, Zhang, Jia-Rui (2024) [10] used universal design theory to evaluate the public facilities in the waiting halls of Chinese urban high-speed railway stations through applicability assessments and user satisfaction analysis. The existing problems were pointed out, and suggestions for improvement were put forward, such as optimizing toilet facilities and rationally setting guidance signs. Wu, X. et al. (2015) [20] proposed a fuzzy quality function deployment method based on the evidence theory for the evaluation of high-speed rail catering service stations. Jung, B.-d. (2018) [21] applied the Importance–Satisfaction Analysis (ISA) method to evaluate the parking lot services of high-speed railway stations. Zhou, H. (2024) [22] aimed to optimize the utilization of the existing railway network capacity, focusing on the carrying capacity of high-speed rail hub stations. Brumercikova, E. [23] used the analytic hierarchy process to study the common problem of insufficient infrastructure capacity, quantitatively evaluating the performance of carriers.

Regarding the built space environment, Castaldo, A. G. (2022) [24] introduced a method for evaluating the current level of attractiveness of Italian high-speed railway stations in a GIS environment. This method comprehensively considered station services and passenger flow parameters. Kim, H.-S. (2022) [25] empirically analyzed the relationship between railway station characteristics and the number of passengers. Tetiranont, S. (2024) [26] analyzed the sustainable design of Thai railway stations in tropical climates, pointed out the existing problems in high-speed railway station design, and proposed future design directions. Zeng, Z. (2024) [8] used social media data to analyze passengers’ satisfaction and travel behaviors in major railway hubs in China. It was found that, in addition to transfer efficiency, factors such as sign quality had a significant impact on satisfaction. Monsuur, F. (2017) [4] analyzed the demand for passenger waiting areas based on urban development characteristics, passenger travel characteristics, and station departure passenger flow. A prediction model for the number of people staying in the waiting hall considering passenger flow and train departure schedules was constructed, and predictions were made using Beijing South Station, Xi’an North Station, Hefei South Station, and Zhoukou East Station as examples. Xie, F. (2023) [27] developed a Gradient-Boosting Regression Tree (GBRT) model to explore the contributions of relevant design parameters of the waiting hall to indoor lighting environment indicators. It was concluded that, for high-speed railway waiting halls in cold regions, factors such as the skylight ratio and the cross-section length-to-width ratio had a significant impact on the lighting environment, providing a basis for early-stage lighting design. The study proposed improving the service quality of waiting halls by optimizing the layout design and enhancing the intelligence level of facilities and equipment.

Based on this literature review, it is evident that the existing research on high-speed rail waiting halls predominantly centers on the physical environment or individual facilities. There is a notable absence of a comprehensive service quality evaluation framework grounded in user needs. Moreover, much of this research remains at a theoretical level, lacking integration with management practices and operations. Additionally, the unique characteristics of China’s high-speed rail system, such as its large scale and high passenger density, have not been adequately considered. Thus, a more in-depth demand analysis aligned with national conditions is warranted.

2.2. Research on Public Transport Service Quality

In the field of airport services, Batouei, A. (2020) [28] explored the impact of the airport experience on passengers’ satisfaction and behavioral intentions from the dimensions of sociology, psychology, and service marketing. Based on data collected from 377 passengers and a partial least-squares analysis, it was found that service fairness, the service environment, service contact, and self-service technology had a significant impact on passengers’ satisfaction, and passengers’ satisfaction was related to their intention to revisit and word-of-mouth dissemination. This provides strategies for airport management and marketing. Antwi, C. O. (2020) [29] evaluated the applicability of the Airport Indicator Passenger Experience (AIPEX) model to Shanghai Pudong International Airport and constructed a theoretical model to explore the relationship between airport service quality, passengers’ emotional image, and satisfaction, as well as the moderating role of passengers’ travel purposes. The Kano model was used to analyze the importance of different service attributes. Hong, S.-J. (2020) [30] constructed a conceptual model including interaction quality, outcome quality, and physical environment quality to analyze the differences in service quality perception between passengers and service providers at Incheon International Airport. The Kano model was helpful for classifying service quality attributes to better understand user needs. Go, M. (2018) [31] integrated the Kano model and a service blueprint to explore the management elements of negative customer interactions (NCCI) during flights. Based on passenger surveys, the Kano classification of NCCI management attributes and the differences among passengers with different flight frequencies were analyzed, and a service blueprint was drawn. Tseng, C. C. (2020) [7] proposed the IPA–Kano model for the classification and diagnosis of airport service attributes and verified it using Taoyuan International Airport as an example. By combining the Skytrax scale, the Kano model, and an importance–performance analysis, airport service attributes were classified and targeted service improvement strategies were proposed based on the results, providing a decision-making basis for airport managers. Lippitt, P. (2023) [32] used cluster analysis, the Kano model, and Importance–Satisfaction Analysis (ISA) to classify business travelers and investigate their expectations and satisfaction with 14 Service Quality Attributes (SQAs). The results revealed the importance and satisfaction levels of different SQAs and provided resource allocation suggestions for airlines. Choi, S. (2024) [33] adopted a mixed method combining the Delphi method and the Kano model to determine the key information service attributes of intelligent airports and clarify the impact of technical services on passengers’ satisfaction. Bezerra (2021) [34] analyzed the relationship between passengers’ expectations and the dimensions of airport service quality through a structural equation model analysis of the survey data of passengers at a large airport in São Paulo, Brazil. The study compared the differences among different passenger groups, and the results showed that the expectations of non-frequent travelers had a significant impact on all dimensions of airport service quality, while the expectations of frequent travelers only had a significant impact on the “process” dimension.

Regarding rail transit satisfaction, Ding, Y. (2022) [35] tested the indoor environment of four representative urban rail transit stations in Chongqing, including the thermal environment, air quality, acoustic environment, and lighting environment. Based on questionnaire surveys and on-site tests, the characteristics of passengers’ satisfaction and environmental conditions were analyzed, and existing problems and improvement directions were identified. Huang, W. (2021) [36] proposed a method for evaluating urban rail transit passenger service quality based on the Kano–Entropy–TOPSIS model. Taking the Chengdu Metro in China as a case study, an evaluation index system was constructed based on MOPES 2.0. The Kano model was used to screen service quality indicators, calculate their sensitivities, and rank them. Then, the entropy–weight method and TOPSIS method were used to calculate and rank passengers’ satisfaction, and the indicators that needed to be improved were determined based on the differences between the two. It was found that the Chengdu Metro should prioritize meeting passengers’ basic needs and improving indicators such as transfer service, staff service quality, and station cleanliness. Li, X. (2023) [37] pointed out that accurately predicting short-term subway passenger flow was highly significant for improving operation efficiency and passenger satisfaction. However, the nonlinearity and non-stationarity of the passenger-flow time series presented challenges. An improved variational mode decomposition (IVMD) and multi-model combination prediction model, IVMD–SE–MSSA, was proposed. Sumaedi, S. (2016) [38] took Jakarta public transportation passengers as the research object and empirically explored the simultaneous impact of perceived value, image, perceived ease of use, and perceived usefulness on passengers’ satisfaction. It was found that perceived value, image, and perceived usefulness had a positive impact on passengers’ satisfaction, while perceived ease of use had no significant impact.

Regarding bus satisfaction, Pi, X. (2018) [39] pointed out that the bus load factor (i.e., bus congestion level) is one of the key indicators for measuring public transportation service quality and consumer satisfaction. Kökalan, Ö. and Tutan, A. (2021) [40] developed a scale to measure the satisfaction levels of public transportation passengers. Via an exploratory factor analysis (EFA), they obtained a structure consisting of 22 items and four factors, namely, “technical satisfaction”, “service satisfaction”, “comfort satisfaction”, and “cleanliness satisfaction”. The study examined the relationships between these four sub-dimensions and found a significant positive correlation. The Cronbach’s Alpha value of the scale was 0.88. Cheng, X. et al. (2018) [12] constructed a structural equation model (SEM) to explore the relationships among bus transfer services, passengers’ perceived value, and passengers’ satisfaction. A scientific and reasonable evaluation of current bus transfer services is of great significance for improving the operation efficiency and system utilization rate of high-speed rail.

In another study of transportation satisfaction, Liu, X. (2024) [41] focused on the functional optimization of electric vehicle charging stations. The aim was to enhance user satisfaction and promote the development of the electric vehicle market by analyzing user requirements. The study used the KJ method to collect and summarize 23 user requirements. It employed the Kano model to analyze the types of requirements, then utilized the analytic hierarchy process (AHP) to determine the importance weights of each requirement. Chen, M.-C. (2021) [42] evaluated transportation service quality during large-scale events. The study used the Kano model and the service quality model proposed by Parasurarman, Zeithaml, and Berry (PZB) to investigate passengers’ satisfaction with various elements of transportation services.

Based on the above research review, we can deduce that the models employed have evolved from a single model to the integration of multiple methods, from static attributes to dynamic demand responses; research methods are in a state of constant innovation. In the study of public transport service quality, airports focus more on service satisfaction, while rail transit and bus systems pay more attention to passenger flow and transfer, gradually forming a differentiated research system. The existing research is still limited and lacks a systematic framework for evaluating the service quality of public transport hubs.

2.3. Research Methods for the Comprehensive Evaluation of Service Quality and Passenger Satisfaction

By reviewing the research methods used by scholars to study travel satisfaction, we found that Wang, J. (2011) [2] used the SERVQUAL model to construct an evaluation index system for railway passenger satisfaction and calculate passenger satisfaction. Bezerra, G. C. L. (2021) [34] and Cheng, X. Y. (2018) [12], respectively, employed structural equation models to explore the relationships between passengers’ expectations and quality dimensions in airport services and bus transfers. Zhou, Z. (2022) [43] combined questionnaire surveys with the Rasch model to study the intermodal transportation services between hubs and urban transportation. Tsoi, K. H. et al. (2023) [44] investigated the relationship between stressors and perceived traffic stress via a factor analysis and multiple linear regression. Rodriguez-Valencia (2024) [45] further explored the origins and qualities of four common service indicators (satisfaction, likability, service quality, and experience) in airport terminals by comparing regression variables. With the development of information technology, Zeng, Z. et al. (2024) [8] analyzed passengers’ satisfaction and travel behavior using social media data. It can be concluded that factor analysis and multiple linear regression remain classic research methods for the issues addressed in this study.

Regarding the comprehensive evaluation of high-speed railway station service quality and passenger satisfaction, numerous studies have analyzed the relevant influencing factors in some depth. Monsuur, F. (2017) [5] used the partially constrained proportional odds model to study the impact of train and station types on the perception of railway service quality. Cheng, X. (2018) [12] constructed a structural equation model and found that economy and convenience were key indicators. Lee, K.-D. (2016) [46] utilized exploratory factor analysis (EFA) and then employed the structural equation model (SEM) to analyze the important factors that enhance customer satisfaction. Güner, S. (2023) [47] adopted a SERVQUAL-based social media analysis method, using techniques such as topic modeling and sentiment analysis to analyze the dimensions of conventional and high-speed railway service quality, passenger satisfaction, and their change trends. Zeng, Z. (2024) [8] analyzed passengers’ satisfaction and travel behavior using social media data. By means of keyword frequency analysis, semantic classification, and network visualization, the factors influencing passenger satisfaction were explored. Niu, H. (2019) [3] constructed an evaluation system based on the SERVQUAL model. Kim, H.-S. (2022) [25] used a multiple regression model, taking station characteristics as independent variables and the number of passengers as the dependent variable, to explore the relationship between railway station characteristics and the number of passengers. Pan, J. Y. (2019) [48] suggested that there were deficiencies in segmenting high-speed rail passengers based on demographic characteristics, and the results of using cluster analysis and multivariate analysis of variance (MANOVA) were proven to be effective.

Based on a comprehensive review of the research methods used to evaluate service quality and passenger satisfaction, it is evident that these methods can be categorized into two primary approaches: one being the classical method based on static attribute analysis, and the other comprising innovative techniques leveraging advancements in big data technology. While both approaches possess distinct advantages and limitations, the traditional method tends to lack insight into dynamic behavior patterns, whereas the newer approach, despite handling vast amounts of data, predominantly remains at the level of descriptive analysis. Consequently, there is a need to develop an integrated methodology that combines traditional models with contemporary multi-dimensional analytical techniques.

2.4. Using the Kano Model in Service Quality Evaluation

In traditional studies, customer satisfaction is mostly viewed in a one-dimensional way, with a higher perceived quality of products or services leading to higher customer satisfaction and vice versa. However, due to the asymmetric and nonlinear relationship between product or service performance and customer satisfaction, meeting the requirements of individual products or services does not necessarily result in higher customer satisfaction. The Kano model proposed by Professor Kano (1984) [49] has unique advantages in solving this problem; it designs dysfunctional and functional questionnaires to evaluate the characteristics of product or service demand, so as to assess customer satisfaction with product or service quality attributes and classify demand attributes, in order to formulate appropriate measures to improve satisfaction. Another feature of the Kano model is that it is easy to operate and does not require advanced statistical knowledge.

The Kano model is widely applied in service quality evaluations to analyze the relationship between service attributes and customer satisfaction. Sunil Kumar, C. V. et al. (2015) [50] utilized the Kano model to help suppliers become the preferred choice of manufacturers, clarifying the direction for improvement. Yin, J. et al. (2016) [51] employed the IPA–Kano model to analyze the environment-related factors influencing the residential satisfaction of Xi’an residents. They classified community characteristics, providing a basis for local government resource allocation. Witte, J.-J. (2024) [52] conducted surveys and analyses of Dutch car-sharing users, demonstrating that the Kano model can be used to analyze users’ demands for car-sharing services.

However, although the Kano model is used to classify service attributes and customer requirements, it ignores the performance and importance of these attributes. Therefore, scholars use the combination of the Kano model and other research methods to make up for this deficiency. For example, in the study of passenger satisfaction in the field of transportation, Go, M. et al. (2018) [31] integrated the Kano model and a service blueprint to explore the management elements of negative customer interactions during flights. Huang, W. et al. (2022) [36] proposed an evaluation method for urban rail transit passenger service quality based on the Kano–Entropy–TOPSIS model. They used the Kano model to screen indicators and determine directions for improvement. Choi, S., Moon, C., Lee, K., Su, X., Hwang, J., and Kim, I. (2024) [38] adopted the Delphi method and the Kano model to determine the key information service attributes of intelligent airports. Liu, X. (2024) [46] used the Kano model to analyze the types of requirements and then determined the importance weights of each requirement using the Analytic Hierarchy Process (AHP). Tseng, C. C. (2020) [7] combined the Skytrax scale, the Kano model, and importance–performance analysis to classify airport service attributes and proposed targeted service improvement strategies based on the results.

Based on the above research review, we can conclude that the application of the Kano model has developed from a single tool to a multi-dimensional integrated method. Some scholars have combined the Delphi method, analytic hierarchy processes, materiality performance analysis, and other methods. This study innovates to combine Kano model with linear regression and correlation analysis. On the basis of dividing attributes, this study analyzes the internal correlation and importance ranking of each service element and applies it to the service scenario of high-speed railway waiting halls. This approach allows us to put forward appropriate optimization strategies.

2.5. Literature Summary

Based on the above literature review, it can be seen that, on the one hand, there has been no comprehensive analysis of how spatial and environmental factors and service facility factors work together to create passenger experience. The existing service quality assessment methods need to be improved in terms of accuracy and suitability, and it is difficult to accurately distinguish the degree of influence of different service attributes on passenger satisfaction. On the other hand, although the Kano model and its evolutions are widely used in many fields, its application in the service scenario of the waiting halls of high-speed rail stations is still relatively limited. In-depth research on how to comprehensively and accurately classify the service attributes of the waiting halls of high-speed rail stations by combining the Kano model with a variety of tools is of practical significance for formulating feasible optimization strategies.

Based on a multidimensional questionnaire analysis, this study intends to comprehensively use the Kano model, reliability and validity analyses, linear regression analysis, Pearson correlation analysis, and other methods to conduct in-depth research into the physical environment, environmental design, service facilities, and other aspects of the service quality of the waiting halls of high-speed rail stations. The research includes the Kano attribute classification results, linear regression equations, and correlation analysis results of various factors. It produces satisfaction statistics and importance rankings for each dimension to determine the impact of different factors on overall satisfaction, and the optimization strategy based on these analyses provides a systematic theoretical basis and practical guidance for the service optimization of the waiting halls of high-speed railway stations.

3. Research Methodology

In summary, the current waiting hall services at high-speed railway stations exhibit deficiencies in multiple areas, and the existing service quality evaluation methods are inadequate. Consequently, conducting research on the service quality evaluation and optimization of high-speed railway station waiting halls based on multidimensional questionnaires and the Kano model holds significant practical value. Scholars such as Choi, S. (2024) [33], Liu, X. (2024) [46], and Tseng, C. C. (2020) [7] have developed online questionnaires to analyze demand types using the Kano model. This study integrates factor analysis and multiple linear regression from Tsoi, K. H.’s (2023) [44] offline survey questionnaire to investigate the importance rankings of impact factors and their relationship with overall satisfaction, thereby conducting a multidimensional analysis through importance–satisfaction ranking and other methodologies. The research scheme process is illustrated in Figure 1.

3.1. Questionnaire Design and Development

3.1.1. Online Questionnaire

The online questionnaire was administered in the preliminary stage of the research. It focused on the attention and importance that all potential passenger groups attach to various service elements in the waiting hall. The results can be used for the attribute analysis of the Kano model. The questionnaire was designed with three aspects in mind: the physical environment, the spatial environment, and service facilities. The physical environment includes the thermal environment, light environment, acoustic environment, and air quality. The spatial environment covers architectural design, route design, and hygiene situation. Service facilities involve facilities for rest, information, safety, commerce, and ticket-checking. The questionnaire consists of two types of questions: positive-oriented and negative-oriented. The Kano type of each factor is determined based on the degree of importance that respondents attach to the factor under different scenarios. The scoring scale ranges from 1 to 7. In positive-oriented questions, a score of 1 indicates “extremely important”, 2 means “important”, 3 represents “relatively important”, 4 is “average”, 5 is “relatively unimportant”, 6 is “unimportant”, and 7 is “extremely unimportant”. In negative-oriented questions, a score of 1 means “not mind at all”, 2 is “don’t mind”, 3 is “relatively don’t mind”, 4 is “average”, 5 is “relatively mind”, 6 is “mind”, and 7 is “extremely mind”. The questionnaire can be found in Supplementary File S1.

3.1.2. On-Site Questionnaire

The on-site questionnaire was designed based on the online questionnaire and follows the same classification system. It mainly focuses on the satisfaction levels of passengers, who are the users in the waiting hall, in relation to various elements. Respondents subjectively rate their item-by-item satisfaction and overall satisfaction according to the physical environment, spatial environment, and service facilities of the high-speed railway station they are in. The rating ranges from 1 to 7. A score of 1 means “very dissatisfied (inappropriate)”, a score of 2 means “unsatisfactory (inappropriate)”, a score of 3 means “relatively dissatisfied”, a score of 4 means “average”, a score of 5 means “relatively satisfied (appropriate)”, a score of 6 means “satisfactory (appropriate)”, and a score of 7 means “very satisfied (appropriate)”. The questionnaire can be found in Supplementary File S2.

In the design of the on-site questionnaire, the variables in the physical environment, spatial environment, and service facilities are divided in a clearer and more detailed manner.

In the research on indoor physical environment quality and satisfaction, Kim, J. and de Dear, R. (2012) [53,54] classified the indoor environmental quality factors in office buildings and educational buildings, and Yin, J. (2016) [51] undertook the same task for residential buildings. Both used the Kano model in combination with other research methods to study the indoor physical environment. Based on these studies, Buratti, C., Belloni, E., Merli, F., and Ricciardi, P. (2018) [55], as well as Mewomo, M. C. (2023) [56], found that the key factors characterizing a building’s indoor physical environment are generally recognized as the indoor air quality, indoor thermal comfort, visual comfort, and acoustic comfort, which are used to comprehensively evaluate indoor comfort. Similarly, in the research on high-speed railway waiting halls, Lee, K.-D. (2016) [46] found that the impact differences in different service factors on customer satisfaction are similar to those of different types of factors in the Kano model, providing a practical reference for the application of the Kano model in the service quality assessment of high-speed railway waiting halls. This study draws on the above-mentioned research and organizes the sub-factors affecting the satisfaction of the physical environment into the thermal environment, lighting environment, acoustic environment, and air quality.

In the research on the building space of transportation hubs, referring to the basic factors of the spatial layout of airport terminals mentioned by Geng, Y. (2017) [11], the future sustainable design directions of high-speed railway stations emphasized by Tetiranont, S. (2024) [26], and the research basis of optimizing the layout design of waiting halls to improve their service quality by Xie, F. (2023) [27], this study draws on these studies and organizes the sub-factors affecting the satisfaction of the spatial environment into greening design, artistic style design, color-matching display, and detail decoration.

In the research on the service facilities of waiting halls and terminals, Zhang, Jia-Rui (2024) [10] focused on the public facilities in high-speed railway station waiting halls, such as optimizing toilet facilities and guidance signs. Wu, X., Nie, L. (2015) [20] focused on high-speed rail catering service stations. Kim, H. -S. (2022) [27] focused on the characteristics of railway stations and the number of passengers. Cao, Y. L (2022) [4] studied the travel characteristics of passengers and the demand for passenger waiting areas influenced by the departure passenger flow of stations. Based on this previous research, this study organizes the sub-factors affecting the satisfaction of service facilities into rest facilities, information facilities, safety facilities, commercial facilities, and ticket-checking facilities.

3.2. Data Collection and Sampling

The online questionnaire was disseminated via the Wenjuanxing platform, which offers convenient sharing and dissemination functions. It generated QR codes and links for the questionnaire, which were then spread via various channels. The platform’s built-in sample-service function was utilized to target and send the questionnaire to potential user groups of high-speed railway station waiting halls, distributing it to users from different regions and age groups. The distribution and collection of the online questionnaire lasted for two weeks. After the initial screening, cleaning, removing incomplete questionnaires to ensure data integrity, and passing logical checks, 510 valid questionnaire samples were finally obtained.

For the field-questionnaire survey, the sample collection scheme of “standardized evaluation by professional researchers” was adopted; the research involved multi-dimensional professional indicators (such as thermal environment parameters, spatial layout rationality, etc.), and the evaluation standards of different sites should be unified. The six selected testers have backgrounds in architecture, management, traffic engineering, and other related disciplines. Compared to passenger samples, professionals can make use of interdisciplinary advantages, transform researchers into “professional sensors”, obtain more accurate spatial environment data, and ensure the objectivity and professionalism of the evaluation process. The six testers who participated in the study went to the waiting halls of 11 high-speed railway stations to conduct experiments and questionnaires, and a total of 66 valid questionnaires were recovered. Regarding the waiting hall design type, there are many differences in spatial form. The waiting hall in this study refers to the centralized public space in the high-speed railway station for passengers to wait and transfer. Based on the Code for Design of Railway Passenger Station (TB 10100-2018) [57], the 11 stations are classified according to the maximum number of people gathered by design. It covers small (Shangqiu East Station), medium (Chongqing Shapingba Station and Yangzhou East Station), large (Chongqing North Station, Suzhou Station, and Shangqiu Station), and super large (Shanghai Hongqiao Station, Beijing South Station, Hangzhou East Station, Zhengzhou East Station, and Chongqing West Station). The geographical areas of the samples include Southwest China, East China, and Central China. The survey sample covers high-speed railway stations with different geographical locations, sizes, and operational characteristics, aiming to maximize the diversity and representativeness of the sample.

The testers underwent rigorous training to ensure that they could accurately understand all of the questions in the questionnaire and collect data based on objective evaluation criteria. To ensure that the testers could represent a wide range of passenger groups, healthy adults aged 20–40 were selected. This age group encompasses the majority of high-speed rail passengers. The male-to-female ratio among the testers was maintained at a relatively equal level to balance the perspectives and experiences of different genders. Prior to the official start of the survey, the testers received comprehensive and meticulous training. The training content covered the meaning of each question in the questionnaire, the evaluation criteria, and how to score in different scenarios, with the aim of avoiding data errors caused by biases in individual understanding. In the testing scenarios at each station, the testers engaged in activities typical of passengers, such as walking, dragging luggage, and sitting and waiting, to simulate the real waiting experience of passengers. During the data collection process, the testers scored different dimensions of the high-speed railway station waiting hall based on their on-site observations and experiences. For instance, when evaluating the physical environment, they took into account factors such as on-site temperature, humidity, ventilation, noise, and illumination. For the environmental design dimension, they considered aspects such as architectural design, the greening layout, route design, and the hygiene situation. For service facilities, they evaluated the user experience and satisfaction levels regarding waiting seats, consulting platform services, safety facilities, and commercial stores.

3.3. Analytical Methods

In this study, SPSS 26.0 was selected as the primary statistical analysis tool. The main analytical methods included reliability analysis, validity analysis, Kano model analysis, Pearson correlation analysis, and linear regression analysis. Reliability analysis was employed to evaluate the stability and reliability of both the online and offline questionnaires. The internal consistency of the questionnaires was assessed by calculating the Cronbach’s Alpha coefficient. Validity analysis was used to evaluate the accuracy and effectiveness of the questionnaires. The structural validity of the questionnaires was determined by indicators such as the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity. Pearson correlation analysis was applied to measure the degree of linear correlation between two variables. This method explores the relationships among different elements in the offline questionnaires and their relationships with overall satisfaction. This further reveals the complex relationship network among the service elements in high-speed railway station waiting halls, helping to comprehensively analyze the internal connections among different elements and providing a basis for considering the synergistic effects and mutual influences among factors when optimizing services. Kano model analysis is one of the core methods used in this study. For the online questionnaire, based on the basic principles of the Kano model, the Kano attribute analysis of various factors such as the physical environment, spatial environment, and service facilities was conducted. The results help us to understand the position of different factors in the hierarchy of passenger needs, thus providing clear guidance for formulating service optimization strategies. Linear regression analysis is used to study the linear relationships between two or more variables. In this study, it was applied to determine the impact degree and contribution of different factors to overall satisfaction. By establishing linear regression equations, the impact relationships between independent variables from different dimensions, such as the physical environment, environmental design, and service facilities, as well as the dependent variable of overall satisfaction, are explored. This helps to identify the key influencing factors in different dimensions and provides a quantitative basis for decision making.

4. Kano Model Analysis of Online Questionnaires

The Kano model has played an important role in many studies, providing an effective method for the classification of service quality. For example, Wencheng Huang (2021) [36] combined the Kano model and the Entropy–TOPSIS method to study the passenger service quality of urban rail transit. This study combines this model with other methods. For example, referring to Mu-Chen Chen (2021) [42], the Kano model is combined with quantitative analysis tools, and a matrix is constructed by calculating the coefficient of dissatisfaction and the coefficient of customer satisfaction, so as to more accurately define the key service elements and determine the strategic priority of service optimization. In the scenario of high-speed railway stations, the applicability of the Kano model is improved by combining Kano, model regression analysis, and correlation tests.

4.1. Data Preprocessing

After the online questionnaire was released, the reliability of the questionnaire was improved through the verification function of the “Questionnaire Star” platform. After software deletion and the manual screening of the questionnaire, a total of 510 groups of data were collected. A total of 32 items, including both positive- and negative-oriented questions, were extracted. The reliability of the data was then assessed using SPSS 26.0 software. The directly tested Alpha value was −0.382. Since the Alpha coefficient represents reliability and there were items that were negatively scored in this questionnaire, the data related to the reverse-worded questions were reverse-coded. After the data adjustment, the overall Cronbach’s Alpha coefficient of the questionnaire was found to be greater than 0.75, indicating good reliability. This suggests that the questionnaire has a high level of consistency and reliability in investigating the demand satisfaction of high-speed railway station passengers. The detailed results are presented in

Table 1. Reliability analysis of the overall questionnaire.

Dimensions	Cronbach’s Alpha	Number of Terms
Whole	0.753	32

.

Simultaneously, a validity analysis was performed on the data. The results showed that the KMO (Kaiser–Meyer–Olkin) value under the overall dimension was 0.876. Since this value is greater than 0.85, it indicates that the questionnaire has excellent validity. Additionally, the significance level was less than 0.01, suggesting the high degree of accuracy of the questionnaire. The detailed results are presented in

Table 2. Validity analysis of the overall questionnaire.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Entirety	0.876	4928.793	496	0.000

.

4.2. Kano Model Classification of Factors

Based on the fundamental principles of the Kano model, a matrix–vector table was employed to conduct positive and negative inquiries into three aspects: the physical environment, environmental design, and service facilities. Each aspect encompasses multiple sub-factors, and each sub-factor has seven options. In positive questions, the values from 1 to 7 represent the following seven dimensions: “extremely important”, “important”, “relatively important”, “average”, “relatively unimportant”, “unimportant”, and “extremely unimportant”. In negative questions, the values from 1 to 7 represent “not mind at all”, “don’t mind”, “relatively don’t mind”, “average”, “relatively mind”, “mind”, and “extremely mind”. Thus, each respondent has 7 × 7 response options.

According to the nature of each factor in the Kano model principles, the evaluation results are divided into six categories: Attractive Quality (A), One-Dimensional Quality (O), Must-Have Quality (M), Reverse Quality (R), Indifferent Quality (I), and Questionable Results (Q). Taking the thermal environment in the physical environment as an example, the Kano model attribute evaluation table is shown in

Table 3. Kano model attribute evaluation (physical environment–thermal environment).

Physical Environment–Thermal Environment		The Degree to Which the Condition Is Not Appropriate
		1	2	3	4	5	6	7
The degree of importance to which the condition is appropriate	1	Q (a11)	A (a12)	A (a13)	A (a14)	A (a15)	A (a16)	O (a17)
	2	R (a21)	I (a22)	I (a23)	I (a24)	I (a25)	I (a26)	M (a27)
	3	R (a31)	I (a32)	I (a33)	I (a34)	I (a35)	I (a36)	M (a37)
	4	R (a41)	I (a42)	I (a43)	I (a44)	I (a45)	I (a46)	M (a47)
	5	R (a51)	I (a52)	I (a53)	I (a54)	I (a55)	I (a56)	M (a57)
	6	R (a61)	I (a62)	I (a63)	I (a64)	I (a65)	I (a66)	M (a67)
	7	R (a71)	R (a72)	R (a73)	R (a74)	R (a75)	R (a76)	Q (a77)

.

The calculation formulas for the proportion of each factor type in the questionnaire statistics are as follows:

$$A=\sum _{j=2}^6{\displaystyle \frac{\left(a_{1j}\right)}{510}}$$

(1)

$$R=\left[\sum _{i=2}^6\left(a_{i1}\right)+\sum _{j=1}^6\left(a_{7j}\right)\right]/510$$

(2)

$$I=\sum _{\begin{array}{c}i=2\\ j=2\end{array}}^6{\displaystyle \frac{\left(a_{ij}\right)}{510}}$$

(3)

$$M=\sum _{i=2}^6{\displaystyle \frac{\left(a_{i7}\right)}{510}}$$

(4)

$$Q={\displaystyle \frac{a_{11}+a_{77}}{510}}$$

(5)

$$O={\displaystyle \frac{a_{17}}{510}}$$

(6)

where a_ij is the corresponding attribute in the attribute evaluation table of the Kano model; A is the Attractive Quality, R is the Reverse Quality, I is the Indifferent Quality, M is the Must-Have Quality, Q is the Questionable Results, and O is the One-Dimensional Quality.

The results of the Kano attribute analysis are presented in

Table 4. Kano attribute result analysis.

Requirements	Factors	M	A	O	R	I	Q	Stats
Physical environment	Thermal environment	46.7	12	12.2	8.4	19.8	1	M
	Acoustic environment	0	47.1	37.1	0	14.5	1.4	A
	Light environment	5.1	39.4	31	1.8	21.6	1.2	A
	Air quality	45.5	1	40	1.2	12	0.4	M
Environmental design	Architectural design	3.5	48	0.8	9	37.5	1.2	A
	Route design	51.6	12.4	12	6.1	17.3	0.8	M
	Hygiene situation	68.8	2.5	2	3.6	21.6	1.4	M
Service facilities	Resting facilities	24.5	14.5	38	7.8	12	3.1	O
	Information facilities	76.5	2.4	1.6	3.7	13.5	2.4	M
	Safety features	29	2.5	24.5	7.8	35.1	1	I
	Commercial facilities	27.6	32.4	34.9	0.4	4.5	0.2	O
	Ticketing facilities	25.1	14.3	42.9	4.9	12	0.8	O

. It can be observed that, among the 16 factors, there are 7 Must-Have Quality (M) factors, 3 Attractive Quality (A) factors, 5 One-Dimensional Quality (O) factors, and 1 Indifferent Quality (I) factor. There are no Reverse Quality (R) factors or Questionable Results (Q). The Must-Have Quality (M) factors include the thermal environment and air quality in the physical environment, the route design in spatial design, and the information facilities and hygiene conditions in basic services. The Attractive Quality (A) factors mainly include the acoustic and lighting environments in the physical environment and the architectural design in spatial design. The safety facilities in basic services belong to the Indifferent Quality (I) category. The One-Dimensional Quality (O) factors are the resting facilities, commercial services, and ticket-checking facilities in basic services.

4.3. Better–Worse Coefficient and Demand Prioritization

The Better–Worse coefficient indicates the degree to which a certain function can increase satisfaction or eliminate dissatisfaction. It was developed by Matzler and Hinterhuner through further research on the Kano model, and it consists of two indicators for improving customer service quality. Specifically, “Better” is the satisfaction coefficient after addition, with a value ranging between 0 and 1. It indicates that, if a product provides a certain function or service, the user satisfaction will increase. The closer the value is to 1, the greater the increase in passenger satisfaction. “Worse” is the dissatisfaction coefficient after elimination, with a value ranging between −1 and 0. It means that, if a product does not provide a certain function or service, the user satisfaction will decrease. The closer the value is to −1, the faster the passenger satisfaction will decline.

The calculation formulas are as follows.

$$Better\left(\mathit{SI}\right)={\displaystyle \frac{\left(A+O\right)}{\left(A+O+M+I\right)}}$$

(7)

$$Worse\left(\mathit{DSI}\right)=-{\displaystyle \frac{\left(M+O\right)}{\left(A+O+M+I\right)}}$$

(8)

where Better (SI) is the satisfaction coefficient, Worse (DSI) is the degree of dissatisfaction, A is the Attractive Quality, R is the Reverse Quality, I is the Indifferent Quality, M is the Must-Have Quality, Q is Questionable Results, and O is the One-Dimensional Quality.

According to the Better–Worse coefficients of 16 demands, we took SI as the horizontal coordinate, |DSI| as the vertical coordinate, and the mean values of SI and |DSI| as the critical line to divide the demand quadrants of the Kano model. Then, we established the demand coordinate system and substituted various demand indicators to create the demand scatter diagram of the Kano model, as shown in Figure 2.

According to the scatter diagram, the importance ranking of passenger demand is determined. S represents the satisfaction degree of a single demand, and the formula is as follows:

$$S=\sqrt{{\mathit{SI}}^2+{\mathit{DSI}}^2}$$

(9)

where S is the satisfaction degree of a single demand, SI is the satisfaction coefficient, and DSI is the degree of dissatisfaction.

In the above formula, S takes SI and |DSI| as the distance from the point with coordinate values to the origin. According to the characteristics of SI and |DSI|, the larger the value of S, the greater the impact of satisfying the demand on user satisfaction: that is, the higher the importance of the demand. Therefore, the importance of the same demand categories is ranked according to the sensitivity S. According to the SI value and DSI value, the importance of waiting hall demand is determined, and the results are shown in

Table 5. Importance of passenger demand.

Demand	Factors	Better (SI)	Worse (|DSI|)	Satisfaction Sensitivity S	Stats	Importance Ranking
Physical environment	Thermal environment	0.27	0.65	0.49	M	10
	Acoustic environment	0.85	0.38	0.87	A	3
	Light environment	0.73	0.37	0.66	A	7
	Air quality	0.42	0.87	0.93	M	1
Environmental design	Architectural design	0.54	0.05	0.30	A	12
	Route design	0.26	0.68	0.53	M	9
	Hygiene situation	0.05	0.75	0.56	M	8
Service facilities	Resting facilities	0.59	0.70	0.84	O	5
	Information facilities	0.04	0.83	0.69	M	6
	Safety features	0.30	0.59	0.43	I	11
	Commercial facilities	0.68	0.63	0.85	O	4
	Ticketing facilities	0.61	0.72	0.89	O	2

.

According to the relevant Kano theory, when ranking importance based on attributes, among the Must-Have Quality (M) attributes, air quality has the highest importance, followed by information facilities, the hygiene condition of the waiting hall, route design, and the thermal environment. Among the One-Dimensional Quality (O) attributes, the order is ticket-checking facilities, commercial facilities, and rest facilities. Among the Attractive Quality (A) attributes, the acoustic environment ranks the highest, followed by the lighting environment, and then the architectural design. The safety facilities belong to the Indifferent Quality attribute and are ranked last. The demand hierarchy of the high-speed railway station waiting hall is shown in the following Figure 3. Based on this demand hierarchy, improvement suggestions can be put forward according to the actual situation.

5. On-Site Questionnaire Results and Analysis

5.1. Physical Environment Analysis

Among the physical-environment-related factors, the thermal environment includes the temperature, humidity, uniformity of temperature distribution, and ventilation. The acoustic environment includes the basic factors of noise and ambient vibration, as well as the factors related to the sound information and atmosphere of the waiting hall, such as the sound color, the clarity of the broadcast, the appeal of the broadcast, and the music atmosphere. The light environment includes indoor lighting, natural light, and glare.

The thermal environment includes temperature, humidity, uniformity of temperature distribution, and ventilation. Temperature measures the perception and satisfaction of passengers with the temperature in the waiting hall and directly affects the comfort of passengers. It may be affected by factors such as the air conditioning system, the season, and the building structure of the waiting hall (such as whether there are good insulation measures). Humidity affects the comfort and perception of the human body. High humidity may make people feel stuffy, while low humidity may cause discomfort such as dry skin, which is related to the dehumidification and humidification function of the air conditioning system. The uniformity of temperature distribution reflects the spatial uniformity of the thermal environment, and an uneven temperature distribution may make passengers feel a large temperature difference in different locations, affecting their experience. Good ventilation, which can improve air quality, regulate temperature and humidity, and reduce odors, is essential for the waiting halls of high-speed rail stations with large enclosed spaces, and it may be affected by the performance and layout of the ventilation equipment.

The sound environment includes the basic factors of noise and ambient vibration, as well as factors related to the sound information and atmosphere of the waiting hall, such as the sound color, the clarity of the broadcast, the appeal of the broadcast, and the musical components of the atmosphere. The waiting hall is usually a densely populated place, and the noise comes from various sources, such as speech, radio, the sound of baggage being dragged, etc. Noise levels that are too high affect the comfort of passengers and the information obtained by the radio. Environmental vibrations may come from the nearby rail transit or the operation of equipment in the station, affecting passengers’ feelings and evaluation of environmental quality. Factors such as timbre, clarity, interest, and music in the waiting hall are related to the information service and creation of an atmosphere in the waiting hall. High-quality timbre and clear information transmission can improve the service experience, while the interest and music atmosphere of the waiting hall may affect the mood of passengers and the enjoyability of the waiting experience.

The light environment includes indoor lighting, natural light, and the level of glare. Reasonable lighting ensures that travelers can clearly see their surroundings and read information, but overly strong or weak lighting may cause discomfort. The introduction of natural light can improve the comfort and visual effect in the space and can also save energy to a certain extent, but this may be affected by the orientation of the building, the design of the windows, and the shading facilities. Excessive glare can cause visual discomfort and affect the visual experience of travelers, which may be affected by the arrangement of lamps and light reflections.

Air quality includes air freshness, air pollutants, and air fragrance. Air freshness reflects the oxygen content of the air and the degree of removal of pollutants, which directly affects the respiratory comfort and health of passengers; it is related to the filtration function of ventilation and air conditioning systems and the density of people in the area. Air pollutants may include dust, odors, harmful gases, etc., which reduce passengers’ satisfaction and perceptions of health. While the aroma of the air is not the key factor, pleasant scents can improve travelers’ psychological feelings to some extent, but overly strong or pungent odors should be avoided.

5.1.1. Physical Environment Regression Analysis

A reliability test was conducted according to the questionnaire data obtained from the field investigation, and each factor was summarized and averaged into four parts: thermal environment, acoustic environment, light environment, and air quality. The obtained data were tested against the overall satisfaction with the physical environment, as shown in

Table 6. Reliability analysis of the overall physical environment.

Dimensions	Cronbach’s Alpha	Number of Terms
Overall physical environment	0.902	5

. The obtained α reliability coefficient is 0.902, indicating excellent reliability.

The validity analysis of the data shows that the KMO value of the overall dimension of the result is 0.848, greater than 0.80, with excellent validity and significance lower than 0.01; this indicates the high accuracy of the questionnaire. The results are shown in

Table 7. Validity analysis of the overall physical environment.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degree of Freedom	Salience
Physical environment as a whole	0.848	211.929	10	0.000

.

The four parts of the physical environment (thermal environment, sound environment, light environment, air quality) and the total satisfaction of the physical environment were subjected to linear regression using the following formula:

$$Y_1={\alpha }_1{X}_1+{\alpha }_2{X}_2+{\alpha }_3{X}_3+{\alpha }_4{X}_4$$

(10)

where Y₁ is the total satisfaction of the physical environment, X₁ is the thermal environment, X₂ is the acoustic environment, X₃ is the light environment, X₄ is the air quality, a₁ is the coefficient of the thermal environment, a₂ is the coefficient of the acoustic environment, a₃ is the coefficient of the light environment, and a₄ is the coefficient of air quality.

As shown in the above

Table 8. A summary of the models.

Models	R	R2 b	Adjusted R	Errors in Standard Estimates
1	0.993 a	0.985	0.984	0.65014

Note: (1) Predictive variables: air quality, light environment, thermal environment, acoustic environment. (2) For over-origin regression (no-intercept model), the R square is used to measure the variable proportion of the dependent variable with respect to the origin explained by this regression. This R square cannot be compared to R squares for models that include intercepts.

, R² is 0.985, indicating that the independent variable can explain 98.5% of the change in the dependent variable, and the model fits well. At the same time,

Table 9. ANOVA (dependent variable: physical environment satisfaction).

Models		Sum of Squares	Degrees of Freedom	Mean Square	F	Significance (p)
1	Regression	1750.794	4	437.699	1035.539	0.000 c
	Residual	26.206	62	0.423
	Total	1777.000 d	66

Note: (1) Dependent variable: overall satisfaction—physical environment (satisfaction). (2) Linear regression through the origin. (3) Predictive variables: air quality, light environment, heat environment, sound environment. (4) Because the constant is zero for cross-origin regression, this total sum of squares is not correct for the constant.

shows that the statistic F is 1035.539 and the corresponding p-value is less than 0.05. The multiple linear regression passes the overall significance test, and the regression model is meaningful.

Finally, the obtained linear regression table shows the relationship between the four aspects of physical environment and the total satisfaction with the physical environment, and the specific data are shown in

Table 10. Coefficients between physical environment satisfaction and the four aspects.

		Unstandardized Coefficients		Standardized Coefficient
Model		B	Standard Error	Beta	t	Significance (p)
1	Thermal environment	0.555	0.096	0.519	5.777	0.000
	Acoustic environment	0.223	0.135	0.203	1.658	0.102
	Light environment	0.039	0.109	0.037	0.354	0.725
	Air quality	0.264	0.101	0.239	2.625	0.011

Note: (1) Dependent variable: overall satisfaction—physical environment (satisfaction). (2) Linear regression through origin.

.

The p-values of thermal environment and air quality are less than 0.05, showing a significant feature. This indicates that these two variables have a significant impact on overall satisfaction, but the impact of the acoustic environment and light environment is not significant. At the same time, the comparison of Beta values shows that the thermal environment has the greatest influence, followed by air quality, then sound environment, and, finally, light environment. The four influencing factors all positively affect the total satisfaction with the physical environment Y. The formula can be derived as follows:

$$Y_1=0.555X_1+0.223X_2+0.039X_3+0.264X_4$$

(11)

where Y₁ is the total satisfaction with the physical environment, X₁ is the thermal environment, X₂ is the acoustic environment, X₃ is the light environment, and X₄ is the air quality.

5.1.2. Physical Environment Factor Correlation Analysis

The single factor of physical environment in the field questionnaire is subjected to a reliability test, as shown in

Table 11. Reliability analysis of the physical environment factors.

Dimensionality	Cronbach’s Alpha	Number of Terms
Factors of physical environment	0.950	17

, and the obtained α reliability coefficient is 0.950, which is greater than 0.9, indicating excellent reliability.

Then, the data were analyzed for validity, and the KMO value in the overall dimension of the result was 0.891, greater than 0.85. The validity was very good, and the significance was also lower than 0.01, indicating the high accuracy of the questionnaire, as shown in

Table 12. Validity analysis of the physical environment factors.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Factors related to physical environment	0.891	1034.195	136	0.000

.

After we conducted a basic data analysis, a Pearson correlation analysis was conducted between physical environmental factors and total satisfaction with the physical environment. The results are shown in

Table 13. Pearson correlation analysis of total satisfaction with the physical environment and each factor.

Thermal Environment		Acoustic Environment		Light Environment		Air Quality
Temperature	0.654 **	Noise	0.593 **	Illumination	0.524 **	Air freshness	0.737 **
Humidity	0.762 **	Ambient vibration	0.567 **	Natural light	0.312 **	Air pollutants	0.731 **
Uniformity of temperature distribution	0.748 **	Broadcast timbre	0.524 **	Glare	0.525 **	Air retention of fragrance	0.628 **
Ventilation	0.776 **	Broadcast clarity	0.543 **
		Broadcast interest	0.507 **
		Musical atmosphere	0.530 **

Note: ** At level 0.01 (two-tailed), the correlation is significant.

below.

The results of the regression analysis show that, among the physical-environment-related factors, the improvement in satisfaction with the thermal environment, light environment, acoustic environment, and air quality is positively correlated with total satisfaction. The values for thermal environment and air quality are significantly higher than those for the acoustic environment and light environment, which are the core factors of passenger satisfaction. When the satisfaction with the thermal environment increases by one unit, the total satisfaction with the physical environment increases by 0.555. The air-quality satisfaction increased by one unit, and the total physical environmental satisfaction increased by 0.264. According to the correlation analysis, in the thermal environment, ventilation (0.776) and humidity (0.762) are strongly correlated with the overall satisfaction with the physical environment, followed by the uniformity of temperature distribution (0.748). Regarding air quality, the correlation between air pollutants (0.731) and freshness (0.737) was strong. In terms of the acoustic environment, the noise factor (0.593) is strongly correlated with the ambient vibration (0.567). Regarding the light environment, glare (0.525) and illuminance (0.545) have higher values.

In terms of the combination of requirement types mentioned in the Kano model analysis, the thermal environment and air quality are basic demands; it is necessary to make the strong correlation between the factor conditions as perfect as possible. The acoustic environment and the light environment are attractive qualities, and we can try to identify more appealing ways to attract passengers and to improve their satisfaction with the sub-factors. Therefore, the broadcast timbre, broadcast appeal, music atmosphere, and natural lighting in the acoustic environment and the light environment can be considered feature services in need of improvement.

5.2. Spatial Environmental Analysis

The categories of spatial environment satisfaction and architectural design include green design, art style design, color matching display, and detailed decoration. The traffic route design includes the pedestrian flow line design, the width of the walkway, the length of the pedestrian flow line, and the congestion of the passageway. The environmental hygiene condition includes the hygiene status of the waiting hall and the odor.

Green design can beautify the environment, improve the air quality, adjust the indoor microclimate, and induce more comfortable visual and psychological feelings in passengers, reflecting the ecological and humanistic concerns of the station. A unique artistic style can enhance the cultural atmosphere and attraction of the waiting hall and give passengers an aesthetic experience. The right color combinations can create a different spatial atmosphere and affect the psychological visual experiences of passengers. The detailed decoration reflects the sophistication and sense of quality of the waiting hall, and small decorative details may bring unexpected surprises and pleasure to passengers.

The design of traffic routes includes the design of the pedestrian flow line, the width of walkways, the length of the pedestrian flow line, and the congestion of passageways. Reasonable pedestrian flow line design can ensure the efficient passage and transfer of passengers, reduce crowd congestion, and prevent the crossing and conflict of flow lines. The width of the walkway affects the convenience and comfort of passengers, and the capacity of different passenger flows should be considered. Overly long flow lines may lead to passenger fatigue, so the distance between key nodes should be shortened as far as possible. The congestion of the channel reflects whether the design is reasonable, and a crowded channel will affect the passenger’s transport efficiency and psychological feelings.

The environmental health condition includes the health condition of the waiting hall and the odor. The environmental hygiene of the waiting hall, including the cleanliness of the floor, seats, toilets, etc., represents a basic passenger need. The odor of the waiting hall may originate from a variety of sources, such as garbage, toilets, food, etc., which will seriously affect passenger satisfaction.

5.2.1. Spatial Environmental Regression Analysis

First, the questions about environmental design requirements in the 66 collected questionnaires were extracted, and the factors were classified. The average values of green design, artistic style, color matching, and detailed decoration were selected as the architectural design part, and pedestrian flow line design, corridor width, pedestrian flow line length, and congestion were selected as the route design part. The sanitation and odor conditions were taken to represent environmental hygiene. We first carried out a reliability analysis of the two parts and the overall satisfaction with environmental design, and the specific data are shown in

Table 14. Reliability analysis of the overall environmental design.

Dimensions	Cronbach’s Alpha	Number of Terms
Overall environmental design	0.837	4

.

The table shows that the α reliability coefficient is 0.837, indicating excellent reliability. Meanwhile, the validity analysis shows that the KMO value is 0.672, greater than 0.6, meaning that the validity is acceptable. See

Table 15. Validity analysis of the overall environmental design.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Significance
Environmental design as a whole	0.672	127.455	6	0.000

for details.

A regression analysis was conducted between the three parts (architectural design, route design, and environmental hygiene) and the total satisfaction with environmental design. The specific formula is as follows.

$$Y_2={\beta }_1{X}_5+{\beta }_2{X}_6+{\beta }_3{X}_7$$

(12)

where Y₂ is the total satisfaction with environmental design, X₅ is the architectural design, X₆ is the route design, X₇ is the hygiene situation, β₁ represents the coefficients of the architectural design, β₂ is the coefficient of the route design, and β₃ is the coefficient of the hygiene situation.

SPSS 26.0 software was used for data collation and regression analysis, generating the results shown in

Table 16. A summary of the models.

Models	R	R2 b	Adjusted R	Errors in Standard Estimates
1	0.996 a	0.991	0.991	0.50143

Note: (a) Predictors: sanitary environment, route design, architectural design. (b) For over-origin regression (no-intercept model), R square is used to measure the variable proportion of the dependent variable with respect to the origin explained by this regression. This R square cannot be compared to R squares for models that include intercepts.

. The table shows that R² is 0.991, indicating that the independent variable can explain 99.1% of the variation in the dependent variable, and the model is well fitted.

At the same time, according to the ANOVA table generated by the regression process,

Table 17. ANOVA (dependent variable: environmental design satisfaction).

Models		Sum of Squares	Degrees of Freedom	Mean Square	F	Significance (p)
1	Regression	1754.160	3	584.720	2325.527	0.000 c
	Residual	15.840	63	0.251
	Total	1770.000 d	66

Note: (1) Dependent variable: environmental design (total satisfaction). (2) Linear regression through the origin. (3) Predictors: sanitary environment, route design, architectural design. (4) Because the constant is zero for cross-origin regression, this total sum of squares is incorrect for the constant.

shows that the statistic F is 2325.527 and the corresponding p-value is less than 0.05. Multiple linear regression passes the overall significance test, and the regression model is significant.

Regression is established according to the obtained data, which are detailed in

Table 18. Coefficients between environmental design satisfaction and the three aspects.

		Unstandardized Coefficients		Standardized Coefficient
Model		B	Standard Error	Beta	t	Significance (p)
1	Architectural design	0.282	0.077	0.260	3.669	0.001
	Route Design	0.746	0.074	0.716	10.113	0.000
	Hygienic environment	0.024	0.079	0.024	0.300	0.765

Note: (1) Dependent variable: environmental design (total satisfaction). (2) Linear regression through the origin.

. Among them, the p-values of architectural design and route design are less than 0.05, indicating that the influence of these two items on the total satisfaction with the dependent variable of environmental design is significant, while the influence of the hygiene situation is relatively low. A comparison of Beta values shows that the coefficient of route design is greater than that of the architectural design and sanitary environment, which indicates that route design has a greater impact.

According to the regression analysis, the following formula is obtained:

$$Y_2=0.282X_5+0.746X_6+0.024X_7$$

(13)

where Y₂ is the total satisfaction with environmental design, X₅ is the architectural design, X₆ is the route design, and X₇ is the hygiene situation.

5.2.2. Spatial Environmental Factor Correlation Analysis

There are 10 factor options for each factor in environmental design, which are analyzed in accordance with the total satisfaction with environmental design. First, a reliability analysis is carried out, and

Table 19. Reliability analysis of the environmental design factors.

Dimensions	Cronbach’s Alpha	Number of Terms
Factors of environmental design	0.901	11

is obtained. The α value is 0.901 (a value greater than 0.9 indicates that the reliability is high).

After conducting the validity analysis, it can be concluded that the KMO value is 0.819, which is greater than 0.8; this means that the validity is excellent and the accuracy of the questionnaire is good. Detailed data are shown in

Table 20. Validity analysis of the environmental design factors.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Environmental design factors	0.819	498.413	55	0.000

.

On this basis, a Pearson correlation analysis was conducted for each factor in environmental design and for the total satisfaction with environmental design, and the data in

Table 21. Pearson correlation analysis of total satisfaction with environmental design and each factor.

Architectural Design		Route Design		Hygiene Situation
Green design	0.462 **	Pedestrian flow line design	0.619 **	Sanitary conditions	0.547 **
Art style design	0.448 **	Walkway width	0.598 **	Odor conditions	0.466 **
Color-matching display	0.506 **	Walk streamline length	0.776 **
Detail trim	0.546 **	Aisle congestion	0.679 **

Note: ** At level 0.01 (two-tailed), the correlation is significant.

were obtained.

The results of the regression analysis show that, among the environmental design factors, the improvement in satisfaction with architectural design, route design, and the hygiene situation is positively correlated with total satisfaction. Route design has a significantly higher value than architectural design and the hygiene situation and is the core factor in passenger satisfaction. When the satisfaction with route design increases by one unit, the total satisfaction with environmental design increases by 0.746. When the satisfaction with architectural design increases by one unit, the total satisfaction with the environmental design increases by 0.282. The coefficient of hygiene was the lowest (0.024) but also showed a positive correlation. According to the correlation analysis, in route design, the walk streamline length (0.776) is the factor with the highest correlation, and the correlations between aisle congestion (0.679), pedestrian flow line design (0.619), and walkway width (0.598) and total satisfaction are high. In architectural design, detailed decoration (0.546) has a strong correlation with color matching display (0.506). In terms of environmental hygiene, the correlation level between sanitary conditions (0.547) and odor conditions (0.466) was also strong.

According to the above analysis of the demand types of the Kano model, route design and hygiene are both basic needs. Although the regression coefficients of the two are quite different, each should receive basic improvements. As architectural design is an appealing demand, color matching display and detail, two sub-factors with a high correlation, can be selected to enhance the satisfaction with the overall architectural design factor, and a characteristic architectural design space can be created to stimulate the excitement of users, so as to achieve a multiplier effect of improving satisfaction.

5.3. Service Facility Analysis

In the third category of service facility satisfaction, the resting facilities include waiting seats, massage seats, accessibility settings, mother and baby rooms, toilets, water heaters, and other public facilities. Information facilities include consulting service platforms, network signal, and the charging facility layout. Security facilities include the configuration of police stations and security guards. Commercial facilities are represented by the number of commercial shops. The check-in facilities mainly include the problem of overcrowding, the sensitivity of the gate machine, the efficiency of checking ID cards at the station, and the efficiency of processing security checks.

Rest facilities include waiting seats, massage seats, accessibility features, mother and baby rooms, toilets, water heaters, and other public facilities. As a value-added service, the number, price and convenience of massage seats affects the frequency of use and passenger satisfaction. Barrier-free design reflects the inclusiveness of the station and facilitates the use of special groups, such as the disabled, the elderly, families with small children, etc. It involves facilities such as ramps, elevators, barrier-free toilets, and so on. The mother and baby rooms are set up to provide convenience for passengers with infants and should be equipped with the necessary facilities, such as baby nursing tables and nursing chairs. Public facilities such as toilets and water heaters meet the basic physiological needs of passengers, and their cleanliness, completeness, and convenience of use are key.

The information-related facilities include consulting service platforms, network signal, and the charging facility layout. The consulting service platform provides information consulting services, and the professionalism of service personnel and the accuracy and timeliness of information are the key factors affecting satisfaction. Regarding network signal and the layout of charging facilities, to meet the needs of passengers in the information age, the strength of network signal and the quantity and convenience of charging facilities are very important.

Security facilities include the configuration of police points and security guards. The police station configuration ensures the safety of passengers, and its location and personnel configuration impact passengers’ sense of security. The security guard allocation plays a role in personnel guidance and emergency handling, and the professionalism and number of guards affect the safety guarantee effect.

Commercial facilities are represented by the number of commercial shops. Commercial stores meet the shopping and dining needs of passengers, but the presence of too many or too few may have a negative impact on the passenger experience, and the type and distribution of stores should be considered.

Check-in facilities are mainly discussed in terms of the problem of overcrowding, the sensitivity of the gate machines, the efficiency of checking ID cards at the station, and the efficiency of the security check at the station. Solving the problem of personnel congestion reflects the management and dredging ability of the passenger flow of the station; it is related to the number, layout, and personnel organization of the ticket gates. The sensitivity of the gate affects the efficiency with which passengers pass through, and low sensitivity may lead to queuing and congestion. The efficiency of ID card inspection affects the speed at which passengers enter the station and is related to the performance of equipment and personnel operations. The efficiency of inbound security checks is an important link to ensure the safety of the station, and there needs to be a balance between its efficiency and passenger experience.

5.3.1. Service Facility Regression Analysis

The options for service facility requirements in the 66 questionnaires collected were classified. The average values of waiting seat settings, massage seat settings, accessibility settings, mother and baby rooms, and toilet water heaters were set as leisure facilities. The average values for the consulting service platform, network signal, and charging facilities were taken as information facilities. The average police station configuration and security guide configuration were used as security facilities. The number of commercial shops represents commercial facilities, while the ticketing facilities give the average values of the degree of personnel congestion resolution, the sensitivity of the gate machine, the inbound ID card inspection, and the inbound security inspection efficiency. A reliability analysis of the five parts and the total satisfaction of the service facilities was conducted, and the results are shown in

Table 22. Reliability analysis of the overall service facilities.

Dimensions	Cronbach’s Alpha	Number of Terms
Overall service facilities	0.784	6

. Where the α value is 0.784, greater than 0.75, there is good reliability.

On this basis, a validity analysis was conducted, and detailed data are shown in

Table 23. Validity analysis of the overall service facilities.

Dimensions	KMO Sampling Appropriateness	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Overall service facilities	0.770	155.486	15	0.000

. Among them, the KMO value is 0.770, which is greater than 0.7, indicating good validity. Meanwhile, the significant coefficient of these data is lower than 0.01, indicating that the accuracy of the questionnaire is relatively high.

In the same way, the linear relationship between the five aspects of service facilities and the overall satisfaction with service facilities can be obtained via a regression calculation. The specific formula is as follows:

$$Y_3={\epsilon }_1{X}_8+{\epsilon }_2{X}_9+{\epsilon }_3{X}_{10}+{\epsilon }_4{X}_{11}+{\epsilon }_5{X}_{12}$$

(14)

where Y₃ is the total satisfaction with service facilities, X₈ is the rest facilities, X₉ is the information facilities, X₁₀ is the safety features, X₁₁ is the commercial facilities, X₁₂ is the ticketing facilities, ε₁ is the coefficient of the rest facilities, ε₂ is the coefficient of the information facilities, ε₃ is the coefficient of the security facilities, ε₄ is the coefficient of the commercial facilities, and ε₅ is the coefficient of the ticketing facilities.

The summary of the model shown in

Table 24. A summary of the models.

Models	R	R2 b	Adjusted R	Errors in Standard Estimates
1	0.995 a	0.989	0.988	0.57457

Note: (a) Predictive variables: ticketing facilities, commercial facilities, leisure facilities, security facilities, information facilities. (b) For over-the-origin regression (no-intercept model), R square is used to measure the variable proportion of the dependent variable with respect to the origin explained by this regression. This R square cannot be compared to R squares for models that include intercepts.

shows that R² is 0.989, indicating that the independent variable can explain 98.9% of the change in the dependent variable, and the model fits well.

According to the ANOVA shown in

Table 25. ANOVA (dependent variable: service facilities satisfaction).

Model		Sum of Squares	Degrees of Freedom	Mean Square	F	Significance (p)
1	Regression	1862.862	5	372.572	1128.565	0.000 c
	Residual	20.138	61	0.330
	Total	1883.000 d	66

Note: (1) Dependent variable: service facilities (total satisfaction). (2) Linear regression through the origin. (3) Predictive variables: ticketing facilities, commercial facilities, rest facilities, safety facilities, information facilities. (4) Because the constant is zero for cross-origin regression, this total sum of squares is incorrect for the constant.

, the statistic F is 1128.565, and the corresponding p-value is less than 0.05. Multiple linear regression passes the global significance test, and the regression model is meaningful. Therefore, the data from the linear regression model are obtained and are shown in

Table 26. Coefficients between service facilities satisfaction and the five aspects.

		Unstandardized Coefficients		Standardization Coefficient
Model		B	Standard Error	Beta	t	Significance (p)
1	Rest facilities	0.566	0.128	0.511	4.423	0.000
	Information facilities	0.028	0.110	0.026	0.257	0.798
	Safety facilities	0.118	0.081	0.121	1.469	0.147
	Commercial facilities	0.004	0.066	0.004	0.066	0.948
	Ticketing facilities	0.352	0.088	0.338	3.977	0.000

Note: (1) Dependent variable: service facilities (total satisfaction). (2) Linear regression through origin.

. The table shows that the p-values of rest facilities and ticketing facilities are less than 0.05, showing significant characteristics. This indicates that these two variables have a significant impact on the total satisfaction with the dependent variable service facilities, while the impact of information facilities and commercial facilities is not significant. In terms of its Beta value, the rest facility is the highest, followed by the ticketing facility, indicating that it has the greatest impact on the overall satisfaction with service facilities. The Beta value of security facilities is lower, and lowest of all are information facilities and commercial facilities.

The regression equation can be obtained through regression, and the specific equation is expressed as follows:

$$Y_3=0.566X_8+0.028X_9+0.118X_{10}+0.004X_{11}+0.352X_{12}$$

(15)

where Y₃ is the total satisfaction with service facilities, X₈ is the rest facilities, X₉ is the information facilities, X₁₀ is the safety features, X₁₁ is the commercial facilities, and X₁₂ is the ticketing facilities.

5.3.2. Service Facility Factor Correlation Analysis

The correlation between the 11 factors related to service facility demand and the total satisfaction with service facilities was analyzed. First, the data belonging to the service facility dimension in the 66 datasets were identified, and the reliability analysis was carried out using SPSS 26.0 software. The results are shown in

Table 27. Reliability analysis of the service facilities factors.

Dimensions	Cronbach’s Alpha	Number of Terms
Service facilities by factor	0.892	15

. The table shows that the α value is 0.892, greater than 0.8, indicating excellent reliability. This suggests that the questionnaire evaluation has a high level of consistency and reliability in investigating the demand for service facilities in terms of various factors.

On this basis, a validity analysis was conducted and

Table 28. Validity analysis of the service facilities factors.

Dimensions	KMO Sampling Appropriateness Measure	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Service facilities (all factors)	0.770	600.731	105	0.000

was obtained. The KMO value is 0.770, greater than 0.7, indicating good validity, and the p value is also less than 0.01, indicating the relatively high accuracy of the questionnaire.

The correlation between each factor in the demand for service facilities and the total satisfaction with service facilities can be obtained by conducting a Pearson correlation test, and the detailed data are shown in

Table 29. Pearson correlation analysis of total satisfaction with service facilities and each factor.

Rest Facilities		Information Facilities		Safety Features		Commercial Facilities		Ticketing Facilities
Setup of waiting seats	0.551 **	Consulting platform services	0.467 **	Police point configuration	0.278 *	Number of commercial stores	0.176	Overcrowding to solve the situation	0.602 **
Setup of massage seats	0.634 **	Network signal with charging pile	0.465 **	Safety guard configuration	0.303 *	Types of commercial stores	0.123	Pass-through sensitivity	0.577 **
Accessibility settings	0.637 **							Efficiency of ID card checks at station	0.537 **
Mother and baby room	0.524 **							Inbound security efficiency	0.522 **
Toilets, water heaters, and other public facilities	0.587 **

Note: ** At a level of 0.01 (two-tailed), the correlation is significant. * At a level of 0.05 (two-tailed), the correlation is significant.

.

The results of the regression analysis show that, among the service facilities, the improvement in satisfaction with the rest facilities, information facilities, safety features, commercial facilities, and ticketing facilities is positively correlated with the total satisfaction. Among them, the correlation of rest facilities (0.566) is the strongest in the overall value, followed by ticketing facilities (0.352) and safety facilities (0.118). The coefficients for information facilities (0.028) and commercial facilities (0.004) were lower. Rest facilities (0.566) were significantly higher than other factors: their satisfaction increased by one unit, and the total satisfaction of environmental design increased by 0.566. From the perspective of the correlation analysis, among rest facilities, accessibility settings (0.637) and the massage seat setup (0.634) have a high correlation, followed by public facilities such as toilets and water heaters (0.587), the waiting seat setup (0.551), and mother and baby rooms (0.524). Among ticketing facilities, the highest correlation is between overcrowding to solve the situation (0.602) and pass-through sensitivity (0.577), which are directly related to passenger efficiency.

According to the above analysis of the demand types of the Kano model, information facilities are basic requirements. Although the correlation between the regression coefficients and the consulting platform, charging setup, and network is less than that of the rest facilities and ticketing facilities, a certain amount of attention should be paid to them. Rest facilities and ticketing facilities are expected needs in the classification of the Kano model questionnaire, and their regression coefficients and correlations are high, so they should be given priority in order to respond to users’ expectations. Although the regression coefficient and correlation of commercial facilities are relatively low, they still represent an expected demand, and the number and type of commercial stores can be appropriately improved according to the real-world conditions of the site. Finally, it is classified as a safety facility with no different needs and a low correlation. Therefore, the existing security facility configuration can be maintained when the basic safety conditions have been thoroughly guaranteed.

5.4. Overall Satisfaction Analysis

A total of 66 questionnaires were collected. They covered the four dimensions of overall satisfaction, total satisfaction with the physical environment, total satisfaction with the environmental design, and total satisfaction with service facilities. This study explores the linear relationship between the comprehensive overall satisfaction and the three levels of satisfaction. First, the data are sorted out, and then the reliability test is carried out. The results are shown in

Table 30. Reliability analysis of the overall satisfaction.

Dimensions	Cronbach’s Alpha	Number of Terms
Overall satisfaction	0.881	4

.

As shown in the table, the overall α value of the four datasets is 0.881, which is greater than 0.8, indicating excellent reliability. Meanwhile, a validity analysis was carried out, as shown in

Table 31. Validity analysis of the overall satisfaction.

Dimensions	KMO Sampling Appropriateness	Bartlett Sphericity Test
		Approximate Chi-Square	Degrees of Freedom	Salience
Totality	0.792	163.024	6	0.000

. It can be concluded that the KMO value is 0.792, greater than 0.75, indicating good validity and the relatively high accuracy of the data.

On this basis, we developed a linear regression equation with overall satisfaction as the dependent variable and total satisfaction with the physical environment, environmental design, service facilities, and ticketing facilities as the independent variables. The specific equation model is as follows:

$$Y={\zeta }_1{Y}_1+{\zeta }_2{Y}_2+{\zeta }_3{Y}_3$$

(16)

where Y is comprehensive overall satisfaction, Y₁ is the total satisfaction with the physical environment, Y₂ is the total satisfaction with environmental design, Y₃ is the total satisfaction with the service facilities, ζ₁ is the regression coefficient of Y₁, ζ₂ is the regression coefficient of Y₂, and ζ₃ is the regression coefficient of Y₃.

As shown in

Table 32. Model summary.

	R	R2 b	Adjusted R	Errors in Standard Estimates
1	0.996 a	0.993	0.992	0.46349

Note: (a) Predictive variables: total satisfaction with service facilities, total satisfaction with environmental design, total satisfaction with physical environment. (b) For over-the-origin regression (no intercept model), R square is used to measure the variable proportion of the dependent variable with respect to the origin explained by this regression. This R square cannot be compared to R squares for models that include intercepts.

. the summary of the model shows that R² is 0.993, indicating that the independent variable can explain 99.3% of the variation in the dependent variable, and the model fits well. At the same time, by analyzing the ANOVA in

Table 33. ANOVA (dependent variable: overall satisfaction ).

Models		Sum of Squares	Degrees of Freedom	Mean Square	F	Significance (p)
1	Regression	1811.466	3	603.822	2810.789	0.000 c
	Residual	13.534	63	0.215
	Total	1825.000 d	66

Note: (1) Dependent variable: overall satisfaction. (2) Linear regression through the origin. (3) Predictive variables: total satisfaction with service facilities, total satisfaction with environmental design, and total satisfaction with physical environment. (4) Because the constant is zero for the cross-origin regression, this total sum of squares is incorrect for the constant.

, we determine that the statistic F is 2810.789, and the corresponding p value is less than 0.05. The multiple linear regression passes the overall significance test, and the regression model is significant.

Based on the B value, the following formula can be obtained:

$$Y=0.243Y_1+0.276Y_2+0.482Y_3$$

(17)

where Y is comprehensive overall satisfaction, Y₁ is the total satisfaction with the physical environment, Y₂ is the total satisfaction with environmental design, and Y₃ is the total satisfaction with the service facilities.

As shown in the coefficient table (

Table 34. Coefficients of overall satisfaction.

		Unstandardized Coefficients		Standardized Coefficient
Model		B	Standard Error	Beta	t	Significance (p)
1	Total satisfaction with physical environment	0.243	0.061	0.240	3.954	0.000
	Total satisfaction with environmental design	0.276	0.073	0.272	3.800	0.000
	Total satisfaction with service facilities	0.482	0.085	0.490	5.698	0.000

Note: (1) Dependent variable: overall comprehensive satisfaction. (2) Linear regression through origin.

), the p values of the coefficients obtained from the three dimensions are all less than 0.05, indicating that the linear relationship is significant. At the same time, the analysis of its Beta value shows that the total satisfaction with service facilities (0.482) has the greatest impact on overall comprehensive satisfaction, followed by the total satisfaction with environmental design (0.276), and, finally, the total satisfaction with the physical environment (0.243).

5.5. Discussion of Results

(1)

First, the limitations of the Kano model are discussed. The first limitation is that static classification characteristics may not be able to fully capture the impact of dynamic factors on passenger demand. The second limitation is that the classification results of the Kano model depend on the subjective judgment of passengers to a certain extent. Different passengers may have different needs and expectations for the same service quality indicator, which may lead to uncertainty in the classification results. In view of this limitation and the limitations of the Kano model, future research can further improve the model, improve its adaptability and reliability, and combine big data analysis, sentiment analysis, and other technologies to automatically identify passenger needs and reduce the influence of subjective judgment.

(2)

Regarding the classification results of the Kano model, although we have classified air quality, thermal environment, and other indicators as basic demands and the acoustic environment, light environment, and other indicators as exciting demands, the demand attributes of these indicators may change in different regions and across different seasons. For example, in the northern winter, the demand attributes of the thermal environment may change from basic requirements to more critical requirements and even affect passengers’ satisfaction with the overall service of the high-speed rail station. This point suggests that, in the follow-up research, it will be necessary to consider the influence of regional and seasonal factors on the attributes of service demand, further improving the applicability of the Kano model.

(3)

According to the results of the regression analysis, although we found that variables such as the thermal environment and air quality in the physical environment, route design and architectural design in the spatial environment, and rest facilities and ticketing facilities in service facilities have significant effects on the corresponding satisfaction, there may be complex interactions between these factors in real-world scenarios. For example, good building design may compensate for some of the decline in satisfaction caused by poor thermal conditions. This suggests that, in future studies, we should further explore the interaction between various factors and build a more perfect model to explain the relationship between service quality and satisfaction.

(4)

It is also necessary to discuss the choice of regression model. A linear regression model was used to describe the relationship between total satisfaction and the physical environment, spatial design, and service facilities, as well as the relationship between the three factors and their respective sub-factors. A “cross-origin regression model” is adopted because a “value-free” linear relationship is assumed between service quality attributes (such as facility convenience, environmental comfort, etc.) and passenger satisfaction. That is, when the score of all service attributes is zero (i.e., the complete absence of service), passenger satisfaction should theoretically be zero, and the intercept term has no practical significance in this scenario. In similar studies [43,44], scholars adopted the intercept constraint model to enhance the explanatory power of independent variables.

(5)

The prediction ability of the model is also discussed. For both the sub-regression model and the overall satisfaction regression model, a high R square indicates that the model explains more variance, while a low one may indicate that there are other important factors that are not considered. If the R square value is low, we objectively analyze possible causes such as potential variables or interactions. In the existing model, the R square values are 0.985, 0.991, 0.989, and 0.993, indicating that the independent variable can explain the change in the dependent variable and that the model fits well. In addition, based on the discussion of the p-value and KMO value, the model passes the overall significance test with excellent validity.

6. Conclusions and Future Work

6.1. Main Research Conclusions

• The Kano model was used to accurately identify passengers’ demand attributes for various service elements in waiting halls. The research results showed that the air quality, thermal environment, information facilities, hygiene situation, and route design were classified as Must-Have qualities. The three indexes of sound environment, light environment, and architectural design are categorized as Attractive Qualities.
• The linear regression model shows that thermal environment and air quality (p < 0.05), two variables in the physical environment, have a significant impact on overall satisfaction, but the impacts of the acoustic environment and light environment are not significant. Among the environmental design factors, route design and architectural design (p < 0.05) have significant effects on the total satisfaction with the environmental design, while the impact of the hygiene situation is relatively low. Among the factors of service facilities, the two variables of rest facilities and ticketing facilities (p < 0.05) had a significant impact on the total satisfaction with the service facilities, while the influence of information facilities and commercial facilities was not significant.
• The results of the Pearson correlation analysis showed that air freshness and air quality, ventilation in the thermal environment, noise control in the acoustic environment, detail in architectural design, the walk streamline length in route design, accessibility settings and massage seats in rest facilities, consultation platforms in information facilities, etc., have a significant positive effect on overall satisfaction, which confirms the decisive role of basic service quality.

6.2. Recommendations for Optimization

Using a Kano model and multidimensional questionnaire analysis, this study reveals the optimization pathway for service quality in high-speed railway waiting halls. We present the following recommendations for optimization.

• For the Must-Have qualities (M), the optimization of service elements is a high priority. Must-Have qualities such as air quality, sanitation, route design, and information facilities are the basis of passenger satisfaction, and, according to regression models and correlation analysis, appropriate temperature, good ventilation, and air quality should be guaranteed. Increasing the fresh air volume during peak hours is an effective improvement measure. Sanitation and information facilities should receive continuous investment to meet the basic needs of users.
• For the One-Dimensional qualities (O), the efficiency of ticketing facilities and the degree of congestion directly affect the travel efficiency of users. Therefore, improving the hardware facilities and the sensitivity of ticketing facilities, combined with reasonable route design, can effectively reduce congestion and improve travel satisfaction. Similarly, as a One-Dimensional quality (O), rest facilities are also a high priority in terms of the regression coefficient and correlation. Decision makers should consider increasing mobile rest facilities during peak hours to meet user demand. Although the regression coefficient and correlation of commercial facilities are low, they are still categorized as One-Dimensional qualities (O), and the number and type of commercial shops can be appropriately increased according to the real-world conditions of the site.
• For Attractive qualities (A), we can pay attention to their differentiation and characteristics to improve the service experience design and appropriately weaken the attention to regression coefficient and correlation, mainly in terms of the light environment, sound environment, and architectural design. Combining the two elements of architectural design and light environment, natural lighting can be considered in the design. In addition, special designs such as decorative details and color matching can also produce twice the result with half the effort. In terms of the acoustic environment sub-factors, noise and environmental vibrations are highly correlated but difficult to improve. Therefore, interesting broadcasts and a musical atmosphere can be realized at a lower cost, and the characteristic acoustic environment can play a significant role in improving passenger satisfaction.
• The safety facility, as an Indifference quality (I), can be left as is without additional optimization-oriented investment.

6.3. Limitations of the Study and Future Research Directions

This study focuses on the physical environment, environmental design, and service facilities of waiting halls; however, the importance of broader perspectives such as the connectivity between the station and the urban environment and the psychosocial experience of passengers for service optimization should also be considered. Moreover, the sample coverage of this study is insufficient, which may lead to the adaptive bias of the model towards extreme climate scenarios. In future research, the sample size should be increased, covering different regions, different high-speed railway station sizes, and more passenger types. Attention should be paid to the needs of special groups, as detailed by Mandhani, Jian, Nayak, J. K., and Parida, M. (2023) [58]; particular attention should be given to elderly and disabled passengers. Finally, this study is based on a static model, which inherently has certain limitations. In future research, the complex nonlinear relationships among influencing factors, neural networks, decision trees, and other models can be considered to accurately capture the relationship between variables. The introduction of big data to capture dynamic characteristics over time improves our ability to understand and predict satisfaction-influencing factors.