Analysis of the Transfer Time and Influencing Factors of Air-Rail Integration Passengers: A Case Study of Shijiazhuang Zhengding International Airport

Abstract

As China’s high-speed rail and civil aviation gradually move from a competitive relationship to a cooperative relationship, air-rail integration services (ARISs) have begun to be promoted, and an increasing number of passengers are choosing the combined travel mode of high-speed rail and aviation. As two long-distance transportation modes are involved, transfer time also affects transfer service satisfaction, and the factors affecting the transfer time of ARIS need to be further clarified. In this study, a stated preference (SP) survey on the transfer reservation time of ARIS passengers was conducted, the perception accuracy of the transfer reservation time was analyzed, and the influencing factors of the transfer reservation time in each link were analyzed by using a multivariate ordered logistic regression model. The results show that the average transfer time of air-rail combined transportation passengers is 102 to 128 min; more than half of passengers’ reservation time is reasonable; and passengers’ genders, educations, occupations, incomes, travel modes, costs, and distances have different degrees of impact on the transfer reservation time. This study holds a certain guiding significance for improving the service quality of air-rail intermodal transportation and improving transfer efficiency and provides a theoretical basis for personalized and customized services for air-rail intermodal passengers.

1. Introduction

The appearance of high-speed railways (HSR) has brought both competition and cooperation opportunities with air transport, e.g., improving passengers’ travel experiences and expanding the service coverage of the airport [1,2]. Several air-rail integration services (ARISs) have been operated successfully for years in Europe, such as AiRail in Germany, TGV Air in France, Rail & Fly in Germany, and FlugZug Basel and Fly Rail Baggage in Switzerland. To attract more passengers from Beijing and other cities along the Beijing–Shijiazhuang high-speed railway, Shijiazhuang Zhengding International Airport (SJW), in cooperation with Beijing Railway Administration and Spring Airlines, proposed ARISs on 26 December 2012, the same day that the Beijing–Shijiazhuang high-speed railway was opened [3]. However, as the first ARIS in China, SJW paid more than two million RMB in subsidies to only 75,000 ARIS passengers from December 2012 to July 2013. This indicates that the market share of ARIS-SJW is far lower than expected [4].

Although China’s ARIS started late and its popularity is low, due to the rapid development of high-speed railway construction, the optimization and upgrading of civil aviation services, and the increasing travel demand of passengers, increasing attention has been paid to the development of ARISs. China’s high-speed rail and civil aviation services have gradually moved from a traditional competitive relationship to win-win cooperation [5]. To provide passengers with more convenient and personalized travel services and to meet market demand, ARISs have gradually been promoted. Several major cities in China have built air-rail intermodal airports, but compared with developed countries, China’s ARISs are still in the primary stage of development [6], and each transfer link needs to be further optimized and improved. The transfer time is the quantification of the transfer process, which reflects the bottleneck of the transfer process to a certain extent, and the factors affecting the transfer time of ARIS need to be further clarified. Based on this background, investigating the transfer time of ARIS passengers, analyzing the influencing factors of passengers’ reservation time in each link of air-rail intermodal transfer, and understanding the bottleneck in the transfer process can provide optimization suggestions for ARIS transfer services. The research on transfer time is helpful to understand the passenger’s satisfaction with transfer services and provide passengers with more targeted and personalized travel services.

The remainder of the paper is organized as follows. Section 2 reviews the research on ARIS and travel reservation times. Section 3 explains how data was collected for this study and describes the distribution of reservation time. Section 4 presents the classification of ARIS passenger reservation times. Section 5 presents the ordered logistic regression model that we used to analyze the factors that influence ARIS passengers’ transfer time. The paper then concludes in Section 6 with the main findings.

2. Literature Review

At present, research on ARISs is mostly based on the macro-perspective, such as the relationships of competition and cooperation between high-speed rail and civil aviation and the economic benefits of air-rail intermodal transportation. Givoni [7] divided the relationship between railways and air transportation into competition, (simple) cooperation, and integration (full cooperation).

The impact of the emergence of HSR on the civil aviation market has also been a concern of researchers. The European experience has shown that civil aviation passenger transportation services increase with increased HSR operation times [8]. At the same time, the expansion of HSR networks has also led to a decrease in the number of flights on some routes (e.g., the London–Paris Route) [9,10,11]. Similarly, in China, the entry of HSR has had a strong negative impact on air transportation demand. After the introduction of parallel HSR services, air transportation demand has become more flexible [12]. Under certain conditions, HSR has competitive advantages; researchers have confirmed the conventional theory that HSR attracts more passengers than air transportation within distances of 300–500 km [13], and HSR would be competitive against air transportation in terms of network connectivity, total travel time, and cost efficiency in the short–medium-haul market [14]. As mentioned above, HSR’s competitive advantage is not absolute; it decreases with increasing passenger travel costs [15]. This relationship between competition and complementarity has also contributed to cooperation and integration; e.g., airlines use railway services as additional spokes in the networks of hub airports to complement their services, and the cooperation between air transportation and HSR also leads to increased demand for railways [7]. Improving the connectivity between air and HSR at the airport will increase the number of international passengers and lead to the gathering of more inter/intra-transit passengers [16].

However, only a few researchers [17,18,19,20] have paid attention to the transfer time of ARIS passengers considering the perception of transfer services. From the perspective of travelers, travel time and distance affect their choice of services; en route travel time is the most important factor affecting the market share of ARIS services, and when the intercity travel distance exceeds a threshold, passengers become less sensitive to the connection time of AH services [17]. The facilities and service level of multimodal transportation hubs affect the transfer experience of passengers. A hub with good information integration can greatly shorten the transfer time of passengers and improve transfer efficiency. Some researchers pay attention to passenger satisfaction with the transfer process to put forward improvement suggestions for ARISs in developing countries. Taking the Shanghai Hongqiao hub as an example, although the Hongqiao hub gathers a variety of transportation modes, it achieved the lowest level of information integration. Further improvement is expected with respect to multimodal services and ticketing, such as time coordination and luggage delivery facilities, through ticketing and interchange discounts [18]. Passengers who buy ARISs are concerned not only about price but also about travel time, that is, time reliability. By analyzing the willingness of passengers to pay for air-rail intermodal transportation-related services, it was found that passengers pay more attention to the connection and transfer time and have a higher willingness to pay for ticket integration [19]. Diversification seekers tend to choose new air-rail intermodal products, but an excessively long transit time seriously affects the attractiveness of intermodal services [20].

Research on travelers’ travel reservation times mainly focuses on the urban commuter travel reservation time and analyzes path selection behavior based on travel time. Generally, the literature concludes that travel time reliability is one of the most important factors affecting travel planning, mode, route, and departure time [21,22]. Generally, travelers are more likely to adjust the departure time according to their own experience, especially when there is a small deviation within the acceptable arrival time range, and a large deviation may cause changes in both departure time and route [23]. The analysis of the influencing factors of urban commuter travel reservation time shows that the value of travel time reliability differs significantly for different income levels, time constraint levels, and transportation modes [24]. A statistical analysis of the route and departure time choice behavior of Japanese commuters under the influence of travel time variability and real-time information showed that travel time variability and the availability of real-time information have a significant impact on departure time and route choice behavior [25]. In research on the influencing factors of actual travel times and ideal travel times, it was found that the mode, trip purpose, activities during trips, and companionship were positively associated with the ideal travel time [26].

Different from a single or combined travel mode of intercity commuting, the ARIS transfer process is more complex, with more influencing factors and more diversified types of travelers. Additionally, the reservation time can reflect travelers’ perception of the complexity of the travel process [27,28,29].

In summary, this literature review suggests that previous studies on air-rail integration services are mostly based on the macro-perspective, such as the relationships of competition and cooperation between high-speed rail and civil aviation and the economic benefits of air-rail intermodal transportation. However, only a few studies have been conducted on the transfer process, and there is a lack of research on the influencing factors of transfer reservation time. Therefore, it is necessary to analyze each transfer link from a micro-perspective and classify the reservation time of ARIS passengers, by focusing on the factors that affect the transfer process, we can propose targeted optimization measures.

3. Reservation Time for the ARIS Transfer Process

3.1. Data Collection

In this study, Shijiazhuang Zhengding International Airport was selected as the survey point to conduct a four-day SP survey on ARIS transfer passengers in August 2019. Zhengding Airport HSR Station is about 3.5 km away from the airport (Figure 1). Eight voluntary interviewers were recruited to conduct the survey. Before issuing the survey, we conducted professional training for interviewers according to the suggestions of experts in the relevant fields. The interviewers distributed questionnaires (shown in Appendix A) at the shuttle bus pickup point of the HSR station and the airport departure lounge and randomly sampled the ARIS passengers. After the survey, the interviewers entered the questionnaire into the WJX platform to calculate the questionnaire results. A total of 786 questionnaires were collected, of which 697 were valid. The effective response rate of the questionnaire was 88.7%. The questionnaire was divided into two parts:

• The personal and travel information of passengers was collected. The personal information included the passenger’s gender, age, education, occupation, and income, and the travel information included the number of trips, the trip purpose, the transportation mode, seats, distance, and costs.
• The ARIS reservation time included the shuttle bus, check-in, and baggage check; the security check; the waiting time; and the interval between high-speed rail arrival and flight departure (regardless of delays).

3.2. Reliability and Validity Test

Reliability analysis was used to analyze and verify the reliability of questionnaire questions and to study whether the answers of the respondents were contradictory and reliable. Cronbach’s α was used to evaluate the reliability of the questionnaire. George and Mallery [30] provided the judgment standard of the Cronbach coefficient: if it is higher than 0.8, the reliability is high; if it is between 0.7 and 0.8, the reliability is good; if it is between 0.6 and 0.7, it is basically acceptable; and if it is less than 0.6, the reliability is poor, and the survey scale should be reconsidered and revised. The questionnaire data were input into SPSS for validity and reliability tests, and the standardized Cronbach coefficient was 0.842, slightly higher than 0.8, indicating that the questionnaire had high internal consistency.

The purpose of a structural validity analysis is to verify the consistency between the questionnaire questions and the research purpose. The results of a validity analysis mainly refer to the KMO value and significance. If KMO > 0.7, it indicates that there is a certain relationship between the independent variables designed in the questionnaire, and the questionnaire is effective. If sig. < 0.001, it indicates that the questionnaire meets the requirements of factor analysis. After the test, the KMO value was 0.941, and the Bartlett sphere test’s p < 0.001, which indicates that the questionnaire had good structural validity.

3.3. Descriptive Statistical Analysis

According to the statistics from the valid questionnaires, male passengers accounted for 54.2%, which was slightly higher than female passengers. Passengers under the age of 30 accounted for 70.7% of the total number. Students, employees, and self-employed individuals accounted for 39.9% and 35.5%. Civil servants, workers, and others accounted for 13% and 11.6%. The distribution of different income levels was relatively average, as shown in

Table 1. Personal information statistics of ARIS passengers.

Classification and Proportion
Gender		Occupation
Male	54.2	Student	39.9
Female	45.8	Civil servant	13.0
Age		Employee or self-employed	35.5
≤29	70.7	Worker and others	11.6
30–49	25.3	Income
≥50	4.0	<2000 RMB	25.7
Education		2000–6000 RMB	29.9
Technical secondary school or below	9.9	6000–10,000 RMB	25.0
College or undergraduate	43.9	>10,000 RMB	19.4
Master degree or above	46.9

.

In terms of travel information, in the majority of cases, the number of trips in the past year was less than six, accounting for 62.6%. The purpose of trips was mainly business travel and returning home or going to school, accounting for 31.6% and 28.4%, respectively. The proportion of trips by direct high-speed rail accounted for 44%. The purchased seats were mostly second-class seats and economy-class seats, accounting for 86.2%. The distribution of travel distance and cost was relatively average, as shown in

Table 2. Passenger travel information statistics.

Classification and Proportion
Number of trips		Seats
≤6	62.6	Business class	1.7
7–14	22.1	First class	3.7
≥15	20.9	Second/economy class	86.2
Travel purpose		Hard seat/soft berth	5.5
Business travel	31.6	Others	2.9
Visiting relatives/friends	13.3	Travel distance
Returning home/Going to school	28.4	<500 km	23.7
Tourism and others	26.7	5000–1000 km	26.7
Travel mode		1000–1500 km	23.0
Direct flight	29.8	>1500 km	26.7
Direct high-speed rail	44.0	Travel expense
ARIS	12.1	<500 RMB	30.8
Air bus intermodal	4.9	500–800 RMB	25.0
Train	5.0	800–1000 RMB	18.4
Others	4.2	>1000 RMB	25.8

.

In terms of ARIS travel information, most passengers reserved 20–29 min for the shuttle bus, 10–19 min for check-in and baggage check, and 10–19 min for the security check, accounting for 35.4%, 33.3%, and 39.3%, respectively. The interval between high-speed rail arrival and flight departure was mainly 120 min, accounting for 32.9%, as shown in Figure 2.

4. Classification of ARIS Passengers’ Reservation Time

The perception of the reservation time for ARIS transfer varies from individual to individual. There may be many reasons, such as different predictions of the travel time by each person and different cognitions of the transfer process by people with different travel experiences [31]. Therefore, analyzing the differences in reservation time is helpful for understanding the travel preferences of different people and the transfer links that need to be optimized.

This study defines the transfer reservation time of ARIS passengers, including the shuttle bus reservation time, the check-in and baggage check-in reservation time, the security check reservation time, and the waiting reservation time. The interval between high-speed rail arrival and flight departure (regardless of delays) is the relative travel time. Through the difference between the transfer reservation time and the relative travel time, the perception of passengers’ reservation time is classified. Based on the preliminary analysis of the questionnaire data, combined with the total reservation time, we divided the difference in the perception time of ARIS passengers into three groups at an interval of 30 min. When the value is less than 30 min, the traveler is a class I traveler; when the value is between 30 and 60 min, the traveler is a class II traveler; and when the value is greater than 60 min, the traveler is a class III traveler. Specifically, the reservation time of class I travelers is close to the travel time, which means that this kind of traveler has a more accurate perception of the transfer process. The reservation time of class II travelers is slightly different from the travel time, with the risk of being late or arriving too early. The reservation time of class III travelers is quite different from the travel time. This kind of traveler has an inaccurate perception of the transfer time and needs relevant prompt services to avoid delaying the trip. The three types of reservation time perceptions are shown in Figure 3. The blue line represents class I, the purple line represents class II, and the green line represents class III.

By classifying the perception of the reservation time of passengers with different personal attributes and travel characteristics, this study finds that the average transfer time of air-rail intermodal passengers is from 102 to 128 min. More than half of passengers’ perceptions of the reservation time fall under class I; that is, for most passengers, the ARIS reservation time is reasonable. For personal attributes, the difference between reservation times by gender and income is relatively small. Passengers under the age of 50 have a slightly lower perception of the rationality of the reservation time than passengers over the age of 50. Those with high education have a more accurate prediction of the reservation time than those with low education. There is no obvious difference in the unreasonable perceptions of passengers holding different occupations, and the reasonable perception of low-income and higher-income people is slightly lower than that of lower-income and high-income people.

In terms of travel characteristics, passengers with different travel distances within one year have few differences in their perceptions of ARIS transfer time. Passengers with lower travel frequency have an unreasonable perception of the reservation time, and there is an imbalance between the reservation time and the travel time for passengers whose travel purposes are mainly to return home or go to school and tourism. Compared with other modes, the reservation time of passengers who mainly travel by air–bus intermodal transport and ordinary railways is unreasonable. Passengers who prefer first-class seats, hard seats, or soft berths have a weak perception of ARIS transfer links. The accuracy of the travel time for passengers with low travel costs is higher than that of passengers with high travel costs.

In general, the transfer schedule of most passengers is reasonable. It is worth noting that older passengers have a more reasonable perception of reservation time than younger passengers, which may be due to their rich travel experience. Similarly, this can also explain why passengers with lower travel frequency have a less accurate perception of transfer time. Low-educated and low-income people need more ARIS transfer guidance information. The proportion of passengers who spend less on a single trip is large, and they have a more accurate time perception, which indicates that passengers with more travel experience prefer cost-effective travel products, and the ARIS market can be expanded by introducing discount products.

5. Influencing Factor Analysis

5.1. Model Construction

This study analyzes the impact of personal attributes and travel characteristics on the reservation time of ARIS passengers. The dependent variable is the reservation time of each link in the transfer process, which is a multilevel ordered variable. The ordered logistic regression model is widely used to analyze dependent variables with ordinal relationships; therefore, it is reasonable to use an ordered logistic regression model to analyze the relevant influencing factors. Ordered logistic regression essentially divides multiple classifications of dependent variables into multiple binary logistic regression [32]. The model is expressed as follows:

$$ln\left[\frac{P\left(y\le j|x\right)}{1-P\left(y\le j|x\right)}\right]=t_j-{{\displaystyle \sum }}_{k=1}^K{\beta }_k{x}_k,j=1,2,\cdots J-1$$

(1)

where J is the number of categories of sequential dependent variables; $t_j$ is the threshold value, $t_{1 }$<$t_{2 }$<$\cdots t_{J-1}$; and ${\beta }_k$ is the regression coefficient of the kth explanatory variable $x_k$.

The cumulative probability can be predicted by Equation (2):

$$P\left(y\le j|x\right)=\frac{e^{t_j-{{\displaystyle \sum }}_{k=1}^K{\beta }_k{x}_k}}{1+e^{t_j-{{\displaystyle \sum }}_{k=1}^K{\beta }_k{x}_k}}$$

(2)

After the cumulative probability is calculated, the probability, P (y = j), belonging to a certain category, can be calculated. The maximum likelihood estimation method is used to solve the parameters of the ordered logistic regression model.

The reservation time of each transfer process is the dependent variable of this study, and the independent variables are the personal attributes and travel characteristics of ARIS passengers. The assignments of each variable are shown in

Table 3. Variables and assignments.

Category	Variable	Assignment
Dependent variable	Shuttle bus reservation time	10–19 min = 1; 20–29 min = 2; 30–39 min = 3; 40–49 min = 4; 50–59 min = 5; 60 min and above = 6
	Check-in and baggage check reservation time	0–9 min = 1; 10–19 min = 2; 20–29 min = 3; 30–39 min = 4; 40 min and above = 5
	Security check reservation time	0–9 min = 1; 10–19 min = 2; 20–29 min = 3; 30–39 min = 4; 40 min and above = 5
Personal attributes	Gender	Male = 1; Female = 2
	Age	≤29 = 1; 30–49 = 2; ≥50 = 3
	Education	Technical secondary school or below = 1; College or undergraduate = 2; Master degree or above = 3
	Occupation	Student = 1; Civil servant = 2; Employee or self-employed = 3; Worker and others = 4
	Income	<2000 RMB = 1; 2000–6000 RMB = 2; 6000–10,000 RMB = 3; >10,000 RMB = 4
Travel characteristics	Number of trips	≤6 = 1; 7–14 = 2; ≥15 = 3
	Travel purpose	Business travel = 1; Visiting relatives/friends = 2; Returning home/Going to school = 3; Tourism and others = 4
	Travel mode	Direct flight = 1; Direct high-speed rail =2; ARIS =3; Air bus intermodal = 4; Train = 5; Others = 6
	Seats	Business class = 1; First class = 2; Second/economy class = 3; Hard seat/soft berth = 4; Others = 5
	Travel distance	<500 km = 1; 5000–1000 km = 2; 1000–1500 km = 3; >1500 km = 4
	Travel expense	<500 RMB = 1; 500–800 RMB = 2; 800–1000 RMB = 3; >1000 RMB = 4

.

5.2. Analysis of the Influencing Factors of the Transfer Process Reservation Time

An ordered logistic regression model is established for the reservation time of each transfer process to analyze the influencing factors. After several iterations, the factors with statistical significance are retained, and the nonsignificant variables are removed. The calculation results are shown in

Table 4. Regression analysis of the influencing factors of shuttle bus reservation time.

Variable	Category	Coefficient	OR Value	Significance
Gender (Reference group: Female)	Male	0.387	1.473	0.005
Education (Reference group: Master’s degree or above)	Technical secondary school or below	0.004	1.004	0.988
	College or undergraduate	0.393	1.481	0.017
Occupation (Reference group: Worker and others)	Student	−0.574	0.563	0.013
	Civil servant	−0.505	0.603	0.047
	Employee or self-employed	−0.894	0.409	0.000
Travel expense (Reference group: >1000 RMB)	<500 RMB	−0.509	0.601	0.010
	500–800 RMB	−0.109	0.897	0.575
	800–1000 RMB	−0.400	0.670	0.048

,

Table 5. Regression analysis of the influencing factors of the check-in and baggage check reservation time.

Variable	Category	Coefficient	OR Value	Significance
Gender (Reference group: Female)	Male	−0.407	0.670	0.003
Travel mode (Reference group: Others)	Direct flight	0.357	1.429	0.336
	Direct high-speed rail	0.681	1.976	0.043
	ARIS	0.592	1.808	0.140
	Air–bus intermodal	0.920	2.509	0.046
	Train	0.797	2.219	0.081
Travel expense (Reference group: >1000 RMB)	<500 RMB	−0.518	0.596	0.012
	500–800 RMB	−0.616	0.540	0.002
	800–1000 RMB	−0.582	0.559	0.006

and

Table 6. Regression analysis of the influencing factors of the security check reservation time.

Variable	Category	Coefficient	OR Value	Significance
Gender (Reference group: Female)	Male	−0.457	0.633	0.001
Education (Reference group: Master’s degree or above)	Technical secondary school or below	−0.252	0.777	0.316
	College or undergraduate	−0.390	0.677	0.013
Income (Reference group: >10,000 RMB)	<2000 RMB	−0.544	0.580	0.014
	2000–6000 RMB	−0.279	0.756	0.180
	6000–10,000 RMB	−0.062	0.939	0.766
Travel expense (Reference group: >1000 RMB)	<500 RMB	−0.854	0.426	0.000
	500–800 RMB	−0.427	0.653	0.031
	800–1000 RMB	−0.796	0.451	0.000

. OR is the odds ratio; the relationship between the OR value and the regression coefficient, β, is OR = e^β. The OR value indicates the possibility for the experimental group to select a higher-level option (reservation time) compared with the reference group.

The regression results of the shuttle bus reservation time show that the significance of the parallel line test is p = 0.151 > 0.05, and the significance of model fitting is p < 0.001, indicating that the model is suitable for ordered logistic regression. The gender, education, occupation, and travel costs of ARIS passengers have a significant impact on the shuttle bus reservation time. When other factors remain unchanged, male passengers are 1.473 times more likely to choose a longer shuttle bus reservation time than female passengers, and female passengers tend to reserve a shorter shuttle bus reservation time. This shows that male passengers pay more attention to the connections between high-speed rail and flights, thus reserving more time for shuttle buses. Higher-educated people often have better job opportunities and, thus, higher income, which also affects travel costs. They also have different perceptions of the connection process. Compared with those with low education, higher-educated people tend to reserve a shorter shuttle bus reservation time. In terms of occupation, the shuttle bus reservation time, from short to long, is as follows: employees or self-employed individuals, students, civil servants, workers, and others. In the past year, passengers with travel costs of less than 500 RMB tend to have a shorter shuttle bus reservation time, while passengers with a travel cost of more than 800 RMB tend to have a longer shuttle bus reservation time.

The regression results of the check-in and baggage check reservation time show that the significance of the parallel line test is p = 0.375 > 0.05, and the significance of model fitting is p = 0.001 < 0.05, indicating that the model is suitable for ordered logistic regression. The gender, travel mode, and travel costs of ARIS passengers have a significant impact on the check-in and baggage check reservation time. When other factors remain unchanged, compared with female passengers, male passengers tend to have shorter check-in and baggage check reservation times. The reason may be that female passengers tend to carry more luggage, while male passengers tend to travel light. Passengers using direct high-speed rail and air–bus intermodal transportation as the main travel modes reserve a longer time than passengers using other modes. Table 2 shows that high-speed rail is still the main mode of travel in China, and most passengers believe that the check-in time of aircraft is longer than that of high-speed rail. Passengers with travel costs of less than 500 RMB have a shorter check-in and baggage check reservation time than passengers with travel costs more than 1000 RMB, followed by passengers with travel costs of 800–1000 RMB and passengers with travel costs of 500–800 RMB. The reason may be that low-cost passengers prefer low-cost airlines, which often do not include luggage.

The regression results of the security check reservation time show that the significance of the parallel line test is p = 0.099 > 0.05, and the significance of the model fitting is p < 0.001, indicating that the model is suitable for ordered logistic regression. The gender, education, income, and travel costs of ARIS passengers have a significant impact on the security check reservation time. When other factors remain unchanged, male passengers tend to reserve a shorter security check time, which may be because male passengers carry less luggage. Highly educated people tend to have longer security check reservation times. A monthly income of more than 2000 RMB has no significant impact on the security check reservation time, while low-income people tend to have a shorter security check reservation time. The travel cost also has a significant impact, and the reservation time of passengers who spend less than 1000 RMB is shorter.

Although gender has an impact on the reservation times of different ARIS transfer links, Table 3 shows that there is no obvious difference between the rationality of male and female passengers’ perceptions of the transfer reservation time. It may be that male passengers tend to reserve a longer shuttle bus reservation time, while female passengers tend to have a longer check-in and security check reservation time, thus shortening the time difference. Providing passengers with more connection guidance and baggage self-service can reduce the connection time. Compared with highly educated passengers, passengers with a low educational background have a less reasonable perception of the transfer reservation time, which may be due to an excessively long shuttle bus reservation time and an excessively short check-in and security check reservation time. Employees or self-employed passengers have a less reasonable perception of the transfer reservation time, which may be due to a shuttle bus reservation time that is too short. It is necessary to promote ARIS products to employees or the self-employed with low education, and more low-cost products will help promote ARIS products. The travel cost has a significant impact on the reservation time of all transfer links. Passengers who spend less than 800 RMB on a single trip have a more reasonable perception of the transfer reservation time. Passengers who pay attention to cost-effective travel products have more travel experience, while passengers with high travel consumption have less information about ARIS products. Adding self-service check-in equipment at the check-in counter of low-cost airlines can improve transfer efficiency. The unreasonable perception of passenger reservation times with ARIS as the main travel mode is slightly higher than that of other passengers. Additionally, it has no significant impact on the reservation time of each transfer link, which also shows that the popularity of ARISs in China needs to be improved.

6. Conclusions

In this study, an SP survey was used to obtain statistics on the distribution of ARIS passenger reservation times in each transfer link. Then, based on the difference between the reservation time and the total transfer time, the perceived rationality of ARIS passenger reservation times was classified. An ordered logistic regression model was used to analyze the influencing factors of different transfer links. We found that more than half of the passengers had a reasonable perception of the reservation time, but the perception rationality of ARIS passengers was slightly lower than that of passengers of other travel modes. Most passengers still estimate the reservation time based on previous traditional travel experiences, which indicates that ARIS products are not popular in China. Among personal attributes, gender, education, occupation, and income and, among travel characteristics, the travel mode and cost have a significant impact on the reservation time of different transfer links.

Transfer time is the quantification of the transfer process, and shortening the transfer time can effectively improve the ARIS experience. The classification of passenger reservation time is conducive to the formulation of differentiated policies and provides personalized and customized services for passengers, which reflects the people-centered perspective. For instance, male passengers need more shuttle guidance services, while female passengers need more baggage check-in services; passengers with low education, employees, or self-employed passengers need more comprehensive ARIS information and suggestions to reduce the risk of missing or arriving early for connecting travel. ARIS products should be promoted for high-income and high-consumption travel groups. We also found that low-income and low-travel consumer groups have a reasonable perception of ARIS reservation times and tend to have shorter check-in times, which means that low-income people are still the main customers of the ARIS market, and adding self-service check-in equipment at the check-in counter of low-cost airlines can improve transfer efficiency. This study analyzed the differences in passenger perceptions of ARIS transfer times, which is helpful in understanding the differences in passengers’ attitudes toward transfer services and their demands for information. This study holds a certain guiding significance for improving the service quality of air-rail intermodal transportation and improving transfer efficiency and provides a theoretical basis for personalized and customized services for air-rail intermodal passengers.