Determinants of students’ behavior in using light rail transit in Hanoi

. On November 6th, 2021, the first light rail transit (LRT) began operating in Vietnam after a decade of construction. This project aimed to alleviate traffic congestion in the road network of the Hanoi capital. However, despite the deployment of the transit service, passenger volume remains under-capacity, raising questions about how to encourage people, particularly young individuals, to use the new public transit. Hence, this study investigated students' behavior in choosing the LRT as a means of transport. Statistical analysis of the questionnaire survey showed that the satisfaction level was the dominant factor affecting the mode choices of students. Besides, distance from home to school and LRT station accessibility significantly influenced students' traveling behavior. These findings provide fundamental insight into young Vietnamese individuals' behavior regarding a new transit service. We believe these outcomes would help support policy-makers and service providers in improving the LRT projects' efficiency.


Introduction
Motorization has greatly affected urban socio-economic aspects. In addition to positive impacts, such as significantly increasing human mobility and inspiring the automobile industry, motorization is also a source of air pollution. As of 2019, the transport sectors accounted for about 27% of global emissions [1]. Most studies agree that heavy dependence on private motorized vehicles for travel is responsible for various negative impacts. These adverse influences include traffic congestion, wasted travel time, and high accident rates, threatening sustainable development goals achievement (SDGs). In response to this problem, promoting public transport (PT) usage by improving accessibility and capacity plays a critical role. For the former, the coverage of the PT 2 Data and methodology 2

.1 Data collection
To address the research questions, we conducted a questionnaire survey at Hanoi Architectural University from the first of February to the fourth of March, 2023. After one month of collecting, the valid responses were 423. Table 1 summarises the final data set.
Relating to the research objective, we define the dependent variable as the status of having used LRT in Hanoi or not (question C14 in the questionnaire). This variable has two values, 0 indicates never using LRT, and 1 presents the experiencing LRT at least once. Surprisingly, about 50% of respondents have never used LRT since the project operated (refer to Table 1). In terms of independent variables, we collected individual and traveling habit information. For the first aspect, the primary information includes major studying, gender, and studying grade (from C1 to C3 in table 1). We further collected the interviewee the status of vehicle ownership, and driving license status, represented by questions C5 and C6, respectively. Other factors that might affect individual travel behavior were also obtained, including the accommodation type, distance from home to school, and part-time job status (C7 to C9). Regarding traveling characteristics, we first observed the respondence traveling frequency, the average number of trips made in a day (C10). C11 indicates the usual travel mode of students. Likewise, C12 and C13 present the average travel time and cost, respectively. Besides, C17 denotes the distance from the student's home to the nearest LRT station. Finally, we reviewed the student's opinions about LRT service using question C18. In this question, interviewees express their viewpoints on five LRT aspects, including ticket price, safety and comfort, easy to use, accessibility, and waiting time. We then converted these factors into three levels, dissatisfied, somewhat satisfied, and satisfied. Notably, there was only about 7% of respondents dissatisfied with the LRT service.

Statistical analyses
Regarding the data structure and dependent variable type, we employed a logistic regression model to reveal the travel behavior of Hanoi students. The model was conducted using the R program version 4.2.2 and related packages [14]. Besides, we have tested the variance inflation factor value for each explanatory variable. The threshold was set at 3.3 to control the multicollinearity problem as recommended by [15]. Table 2 expresses the results of logistic regression. As shown in the table, the results indicate four significant variables were associated with the student's behavior in using LRT. These factors are the studying grade, the distance from home to school, the distance from home to the nearest LRT station, and the satisfaction level. We will use the odds ratio to interpret the results expressed below. Regarding studying grades, senior students are more likely to use LRT than first-grade students. Specifically, second and fourth-year students have associated with a 107.6% and 119.3% increase in odds ratio, respectively. However, there was no distinction between third-and fifth-year, and first-year students in choosing LRT as a transport mode.

Results
In terms of distance, we found that students who live far from school have a higher intention of using the LRT than those who live near the school. An increase in the 1 km distance from home to school would increase by a 9.1% odds ratio. By contrast, the distance from home to the nearest LRT station reduces the possibility of using the LRT. As results indicate, the odds ratio would decrease by 7.1% when the distance between home and station increases by 1 km.
Regarding satisfaction level, the results demonstrate that high satisfaction would significantly associate with using LRT. A student who is somewhat satisfied with the service would prefer to use LRT compared to a dissatisfied student. The odds ratio increased by 394.3% for these students. Likewise, the satisfied group saw a 639.5% increase in the odds ratio.
Remarkably, the common factors such as private vehicle ownership, driving license, travel time, and travel cost did not correlate with the student's behavior in choosing LRT as the mean of transport.

Discussions
Consistent with previous studies, our results demonstrate that satisfaction is the most significant factor motivating students to use LRT. For instance, a study in Kuala Lumpur claims that, besides service quality and perceived value, satisfaction positively impacts the public transport usage of Malaysians [16]. For Hanoi, the LRT project introduces a new public service with luxury features that differ from the existing public modes, including BRT and traditional buses. The new LRT takes advantage of a separate way, spacious trunk space, convenient ticket service, etc., to ensure reliability and comfortability, which satisfies not only young people but other commuters. Satisfaction, however, has a strong relationship with passengers' perception, which has an intrinsic relationship with service quality. Though the new LRT project is operating with a high grade of service, the condition may decrease gradually. Therefore, the LRT operator should maintain the system regularly. The measures may include improving infrastructure such as information provision, facilities, signage, safety, and security. On the other hand, enhancing services, for instance, ticketing systems, frequency, cleanliness, and comfort, would be worthwhile to preserve customer satisfaction [17,18].
Regarding student's grade, our study result differs from the finding of [19] and [12]. In these studies, senior students were less likely to use public transport than first-year students. The authors claimed that new students would rely on buses because of their quality, and older students may switch from public transport to private modes due to changes in their home location and job status. We argue that this phenomenon is not necessarily true. New students, who mainly migrate to Hanoi from suburban or rural areas, prefer choosing accommodation close to the university. Short home-school trips are more suitable for active modes, e.g., walking and cycling, than motorized modes, including buses or LRT. As they move to higher grades, they become familiar with the city network and confident with the new style of travel mode that does not exist in their hometown. Also, from the second year, students consider finding part-time jobs and new accommodation that adapts to their lifestyle. As a result, they are more likely to use public transport than before when the home is far away from school. However, some students prefer using motorcycles to public transport when they want to control their time flexibly.
The effect of travel distance in the current study is consistent with the findings of [20] in Italy. On the other hand, this result contradicts the studies of [21] in America and [12] in Vietnam. Note that the first two cases analyzed student mode choice in developed countries that might not be suitable to compare with the Vietnamese context. We suggest that this factor represents travel cost, which includes time value, comfortability, and health matters. Firstly, alternative travel modes may outperform LRT for a short trip because of less access and egress time. Likewise, commuters might face traffic congestion, air pollution exposure, and uncomfortable weather during a long journey. These problems are common in large Vietnamese cities when traveling with private modes or even buses. Thus, a long homeschool trip school could encourage students to choose LRT as a means of transport. Nonetheless, the distance seems to have a threshold. Some studies have declared that public transport modes outperform others when the trip length is longer than 7 km [22].
As expected, the easy accessibility positively affects students' LRT usage intention. This finding is also consistent with other studies in China and Thailand [23,24]. A short distance from home to the LRT station enables commuters to access the LRT by walking. Consequently, they can be independent of vehicles to move from home to the station. Note that Park and Ride (P&R) infrastructures are currently unavailable for the Hanoi LRT project. Thus, proximity to the station is one of the convenient factors for passengers using public transport. Furthermore, this finding suggests that local administrators should invest in P&R facilities and develop transport-oriented development plans to enhance the effectiveness of the LRT project.
Remarkably, besides statistically significant factors, the current study did not find a relationship between some common variables and students' behavior in choosing LRT as a travel mode. Of these insignificant variables, gender and private vehicle ownership are interesting factors. For gender, although this finding is consistent with previous studies from Indonesia and America [21,25], other researchers have reported a distinction between male and female students in choosing public transport as a means of transport in Lebanon and China [22,26]. Additionally, other studies have suggested that physical health or concerns about sexual and gender harassment may cause differences between girls and boys in using public transport [12]. We argue this issue is still ambiguous and needs further investigation and discussion.
Although vehicle ownership and driving license status are typically associated with low levels of PT usage [19,25], this does not seem to be the case among students in Hanoi. There may be a possibility that this phenomenon is due to a change in young people's perceptions. As the data illustrated, approximately 80% of interviewees owned a motorized vehicle and driving license (Table 1). Additionally, traffic conditions in Hanoi are worsening due to insufficient PT and a low-capacity network. Therefore, commuting by private mode often results in congestion, excess time, and air pollution exposure, which can be extremely distressing for commuters. The operation of the LRT highly remedies these problems and may influence young people to switch from private mode to PT. However, the extent of this effect is still unclear, suggesting a need for further investigation.
Living with family, having a part-time job, and high travel costs are commonly associated with driving to school instead of using PT [25,26]. Ordinarily, students who live in the city have high confidence in commuting and are more likely to self-driving to school or share a ride with family members. Similarly, having a part-time job may encourage students to control their time and be independent of PT, leading to self-driving. However, these factors were insignificant in the current study, which is inconsistent with the findings of [12]. We propose that the difference in objectives between LRT and bus might account for this inconsistency.

Limitations and conclusions
The current study has examined the students' behavior using a new LRT project in Hanoi city. The logistic regression model results on the 423 respondents indicate that the satisfaction level was the dominant determinant of choosing LRT as a travel mode for Hanoi architectural university students. Besides, the distance from home to the nearest LRT station and school is also the main factor affecting the student mode choice preference. These results demonstrate that the long school trip and short access to a station would encourage students to use LRT. The current study's findings would contribute to the practice and academics at two points. First, they provide a good reference for policy-makers and LRT service operators in enhancing the service quality and thus improving the passenger volume. Second, to our knowledge, this is the first study that examines the students' travel behavior regarding the new transit mode in Hanoi. We propose that these outcomes would be worthy for underconstruction projects in Vietnamese cities and developing countries.
Some limitations exist in the study, including the observation size and homogeneity. Since the data collecting time was short, the respondent population was mainly observed at Hanoi architectural university. This fact may affect the quality of the data when the individual characteristics seem to be homogeneous. Thus, we propose to expand the targeted interviewee to other educational institutions in future studies. Besides, we also plan to investigate the effect of LRT on the travel mode choice of young people. This action might help address the issue of how to shift private vehicle dependence to public transport.