Understanding the role of risk perception and health measures in ridesourcing usage in the post-COVID-19 era

Abstract

The COVID-19 pandemic has had a significant impact on how people perceive health risks when using ridesharing services, as there is an increased risk of transmission. To mitigate this risk, individuals have implemented self-imposed preventive measures, while ridesharing service providers have introduced health measures to reduce the likelihood of transmission and encourage ridership. To understand the impact of COVID-19 on ridesharing usage and willingness to use it, we propose a framework to study the relationship between influencing factors (such as risk perception, vaccination status, and self-imposed or service-imposed non-pharmaceutical health measures like masks and/or sanitizing) and ridesharing-related decisions and intention. A web-based survey in China collected over 2300 responses, and Multiple Indicators and Multiple Causes models were estimated to investigate the influencing factors’ impacts on ridesourcing service usage at the later stage of the pandemic when China was still under Zero-COVID policy and their willingness to use it after the pandemic. The estimation results indicate that risk perception, self-imposed and service-imposed health measures, post-pandemic expectations of these measures are key influencing factors, along with certain sociodemographic and travel behavior factors. These findings can help decision-makers develop strategies to support the recovery of ridesharing services in the post-COVID-19 era.

Full text

Understanding the role of risk perception and health measures in ridesourcing usage in the post-COVID-
1
19 era
2
3
Xinghua Li, Ph.D.
4
Professor
5
Key Laboratory of Transport Industry of Comprehensive Transportation Theory, College of Transportation
6
Engineering, Tongji University
7
Shanghai, 201804, China
8
Email: xinghuali@tongji.edu.cn
9
10
Yueyi Yang, Ph.D. Student
11
Key Laboratory of Transport Industry of Comprehensive Transportation Theory, College of Transportation
12
Engineering, Tongji University
13
Shanghai, 201804, China
14
Email: 2210179@tongji.edu.cn
15
16
Yuntao Guo, Ph.D., Corresponding Author
17
Assistant Professor
18
Key Laboratory of Transport Industry of Comprehensive Transportation Theory, College of Transportation
19
Engineering, Tongji University
20
Shanghai, 201804, China
21
Email: yuntaoguo@tongji.edu.cn
22
23
Dustin Souders, Ph.D.
24
Assistant Professor
25
Department of Psychology
26
Clemson University, SC, USA, 29634
27
28
Jian Li, Ph.D.
29
Associate Professor
30
Key Laboratory of Transport Industry of Comprehensive Transportation Theory, College of Transportation
31
Engineering, Tongji University
32
Shanghai, 201804, China
33
Email: jianli@tongji.edu.cn
34
35

2
Abstract
1
The COVID-19 pandemic has had a significant impact on how people perceive health risks when using
2
ridesharing services, as there is an increased risk of transmission. To mitigate this risk, individuals have
3
implemented self-imposed preventive measures, while ridesharing service providers have introduced health
4
measures to reduce the likelihood of transmission and encourage ridership. To understand the impact of COVID-
5
19 on ridesharing usage and willingness to use it, we propose a framework to study the relationship between
6
influencing factors (such as risk perception, vaccination status, and self-imposed or service-imposed non-
7
pharmaceutical health measures like masks and/or sanitizing) and ridesharing-related decisions and intention. A
8
web-based survey in China collected over 2300 responses, and Multiple Indicators and Multiple Causes models
9
were estimated to investigate the influencing factors’ impacts on ridesourcing service usage at the later stage of
10
the pandemic when China was still under Zero-COVID policy and their willingness to use it after the pandemic.
11
The estimation results indicate that risk perception, self-imposed and service-imposed health measures, post-
12
pandemic expectations of these measures are key influencing factors, along with certain sociodemographic and
13
travel behavior factors. These findings can help decision-makers develop strategies to support the recovery of
14
ridesharing services in the post-COVID-19 era.
15
Keywords: Ridesourcing services; COVID-19; Travel behavior, Risk perception, Health measures
16
17
18

3
1. Introduction
1
Prior to the emergence of the COVID-19 pandemic in December 2019, popular ridesourcing platforms such
2
as Uber and Lyft were experiencing rapid growth rates in many major economies (Tirachini, 2020). These
3
platforms utilize online applications to connect passengers with drivers, creating employment opportunities for
4
drivers and providing flexible and convenient transportation options for travelers (Wang & Noland, 2021).
5
Ridesourcing services are often viewed by travelers as a more cost-effective and convenient mode of
6
transportation than public transit, as well as a more economical alternative to owning a personal vehicle in many
7
urban centers (Jin et al., 2018; Ghaffar et al., 2020). These services embody a manifestation of the global trend
8
towards shared economy models in the transportation industry (Jin et al., 2018).
9
The outbreak of the COVID-19 had a significant impact on the transportation sector, causing shockwaves and
10
uncertainties due to its rapid transmission (Ivanov, 2020; Sobieralski, 2020; De Vos, 2020). This has led to the
11
implementation of various measures by governments to restrict individual mobility, such as regional lockdowns
12
and service hour restrictions, as well as increased fear of contracting COVID-19 while traveling with others
13
(Shokouhyar et al., 2021). As a result, ridesourcing services, along with public transportation and airline services,
14
experienced a sharp decline in ridership, while private vehicle usage and active travel saw an increase throughout
15
the pandemic (Guo et al., 2021). A recent study conducted in the Greater Toronto Area revealed that over 70%
16
of respondents had never used ridesourcing services during the pandemic, in contrast to less than 50% before
17
the pandemic (Loa et al., 2022).
18
Governments worldwide are gradually lifting COVID-19 restrictions and transitioning to a "new normal" by
19
coexisting with the virus. However, the effects of COVID-19 on people's travel behavior, including their use of
20
ridesourcing services, may still persist (Loa et al., 2022; Nguyen-Phuoc et al., 2021; Nguyen-Phuoc et al., 2022;
21
Yu et al., 2022). Although not many studies have focused on ridesourcing service usage, they have found
22
consistent results that many travelers have reduced their use of ridesourcing services due to their fear of exposure
23
to the virus when sharing a ride with a driver or unknown passengers. Restoring people's trust in the safety of
24
ridesourcing services to provide a low-risk environment for passengers against COVID-19 and future
25
transmissible diseases may be challenging. Additionally, other factors that may influence such a decision include
26
people's sociodemographic characteristics, attitudinal factors, built environment factors, and infection rate of the
27
virus (Rahimi et al., 2021).
28
Several ridesourcing services have imposed health measures to reduce the risk of COVID-19 transmission and
29
increase their ridership (Liu, 2021; Shin et al., 2021; Liu et al., 2022; Wu et al., 2022; Zhao & Gao, 2022). These
30
measures include passenger screening, enhancing vehicle and driver hygiene, and requiring drivers to undergo
31
daily testing (Kungwola et al., 2022). In addition, many travelers have developed a hygiene routine to minimize
32
their risk of contracting COVID-19 while traveling, such as using hand sanitizer and alcohol wipes to clean seats.
33
The availability of vaccines may aid the recovery of ridesourcing services, as a recent study shows that
34
vaccinated travelers are more likely to use public transit as they perceive themselves to be at a lower risk of
35
infection (Fall, 2022; Fan et al., 2022). Despite these efforts, people's willingness to use ridesourcing services
36
might still depend on their perceived risk, especially for non-essential travel.
37
To sum up, earlier studies have offered valuable insights into various factors that influence the use of
38
ridesourcing during the COVID-19 pandemic and people's inclination to utilize it in the post-COVID-19 era.
39
However, none of the current studies have extensively examined the collective effects of travelers' risk perception,
40
vaccination status, self-imposed and service-imposed health measures on individuals' use of ridesourcing
41

4
services at the conclusion of the pandemic (short-term impacts) and their readiness to use it after the pandemic
1
(long-term impacts).
2
This study aims to investigate the factors that influence people's use of ridesourcing services during the later
3
stages of the COVID-19 pandemic and their willingness to continue using these services after the pandemic. The
4
study considers various factors that affect people's decision-making process, including their perception of risk,
5
vaccination status, current and expected post-pandemic non-pharmaceutical health measures like masks and/or
6
sanitizing (both self-imposed and provider-imposed), attitudes toward COVID-19 vaccines, and
7
sociodemographic and behavioral characteristics. To analyze the relationship between these factors and
8
ridesourcing service usage-related decisions and intentions, the study proposes a research framework and
9
Multiple Indicators and Multiple Causes (MIMIC) models were estimated to evaluate the framework and capture
10
subgroup heterogeneities. By estimating the proposed model framework, the study aims to evaluate the potential
11
short-term and long-term impacts of COVID-19 on ridesourcing services. The model estimation results and
12
insights gained from the study can assist decision-makers in designing strategies to facilitate the recovery of
13
ridesourcing services in the late stages of the pandemic and beyond.
14
The structure of this study is as follows: in Section 2, we provide a review of the literature related to this topic.
15
Section 3 outlines the methodologies used in this study, including the model construct, hypotheses, and survey
16
design. In Section 4, we present the descriptive statistics, model estimation results, and interpretation of the
17
model parameters. The empirical in-depth discussions are presented in Section 5. Finally, in Section 6, we
18
provide a summary of our concluding comments, policy implications, and study limitations.
19
20
2. Literature review
21
The COVID-19 pandemic has had a profound impact on people's travel behavior and urban mobility across
22
the world. A growing body of literature has examined the short-term impacts of the pandemic worldwide. A
23
significant amount of literature has investigated the pandemic's short-term effects on travel behavior, such as
24
travel distance and mode choice, and various transportation sectors, such as day-to-day operations and ridership
25
(Abdullah et al., 2020; Bucsky, 2020; De Haas et al., 2020; De Vos, 2020; Hotle et al., 2020; Li et al., 2021;
26
Mogaji, 2020; Yen et al., 2020; Shakibaei et al., 2021; Li et al., 2023a). These studies have generally concluded
27
that the potential risk of exposure to COVID-19 felt by both the drivers and passengers is one of the primary
28
drivers behind the decline in ridesourcing service usage and service offered, along with public transit and other
29
modes of travel that involve drivers and/or other passengers (Loa et al., 2022; Nguyen-Phuoc et al., 2021;
30
Nguyen-Phuoc et al., 2022; Yu et al., 2022). Several studies have also focused on the factors that influence
31
people's frequency of self-reported ridesourcing service usage and their willingness to use it, including their
32
attitudes or perceptions regarding COVID-19, preventive measures, mobility services, vaccination status, and
33
sociodemographic and travel-related characteristics (Aguilera-García et al., 2022; Zhang & Liu, 2022).
34
Risk perception has long been recognized as a critical attitudinal factor influencing the use of ride-sourcing
35
services, even prior to the COVID-19 pandemic (Neuburger & Egger, 2021; Faruque et al., 2022; Alonsa-
36
Almeida, 2022; Guo et al., 2022a; Feng et al., 2023; Hasselwander et al., 2023; He et al., 2023). Sharing a
37
confined space with strangers often acts as a deterrent for shared mobility, primarily due to concerns about
38
personal safety and privacy. The advent of COVID-19 has further compounded these issues, as fears of potential
39
infection and car cleanness now form an additional layer of risk. Moreover, in countries with contact tracing
40
measures like China, people may also worry about being identified as close contacts of infected individuals.
41

5
(Feng et al., 2023). Previous studies have found that people who perceive COVID-19 as a significant risk are
1
less likely to choose modes of travel with higher transmission risks or a greater likelihood of contracting the
2
virus, preferring instead to take self-imposed preventive measures to reduce their risk, such as wearing facemasks
3
and using hand sanitizers (Ding et al., 2020; Mansilla et al., 2020; He et al., 2021; Park et al., 2022). Loa et al.
4
(2022) and Zhang and Liu (2022) focused on shared mobility service usage during the pandemic and found that
5
people who have a high-risk perception of COVID-19 are less likely to use shared mobility services. However,
6
the effects of exposure risks on people's travel behavior and ridesourcing service usage may persist over a more
7
extended period. Further research is needed to investigate the potential long-term impacts of the risks brought
8
about by the pandemic.
9
In addition to risk perception, other influencing factors encompass attitudes towards mobility services and
10
health measures (Semple et al., 2021; Downey et al., 2022; Mashrur et al., 2022; Zhang et al., 2022; Feng et al.,
11
2023; Lopez et al., 2023). These studies suggest that most people agree that ride-sourcing providers should
12
implement relevant non-pharmaceutical health measures to mitigate the risk and impact of the COVID-19
13
pandemic, and they are willing to undertake similar actions themselves. Moreover, people are likely to use ride-
14
sourcing services as frequently as they did prior to the pandemic if they trust the effectiveness of these health
15
measures, exhibit a strong intention to use ride-sourcing services, or shift from using public transit to more
16
frequent ride-sourcing usage (Alemi et al., 2018; Koh et al., 2021; Mallinas et al., 2021; Nguyen et al., 2020;
17
Loa et al., 2022; Nguyen-Phuoc et al., 2021; Nguyen-Phuoc et al., 2022; Su et al., 2021; Yu et al., 2022).
18
Furthermore, individuals who adopted non-pharmaceutical health measures during the pandemic are highly
19
likely to maintain these practices even after the pandemic concludes (Yamamura & Tsutsui, 2022).
20
As the COVID-19 vaccine becomes more widely available, people’s attitudes towards the vaccine and their
21
vaccination status are increasingly important in influencing ridesourcing service usage. Ideally, the vaccine
22
should increase the vaccination rate and improve confidence in using ridesourcing services due to reduced risk
23
perception (Gursoy et al., 2021; Gursoy & Chi, 2021). However, due to widespread misinformation, people have
24
polarized opinions regarding the COVID-19 vaccine and these attitudes have been shown to impact their
25
approach to travel during and after the pandemic (He et al., 2021; Hsieh & Hsia, 2022; Currie et al., 2021; Gursoy
26
et al., 2021). Some individuals who do not believe in the COVID-19 vaccine (and may not believe that COVID-
27
19 is real) are more likely to forgo vaccination and are also more likely to resume their travel routine as they
28
perceive the risk of COVID-19 to be low, making them less cautious (Gursoy et al., 2022). Conversely,
29
individuals who believe in the COVID-19 vaccine’s efficacy and get vaccinated may be less likely to resume
30
their travel routine due to their fear of contracting and/or spreading COVID-19, despite their low-risk status
31
against the disease (Ma et al., 2021). In summary, people's attitudes towards the vaccine and vaccination status
32
may have complex impacts on their risk perception and usage of ridesourcing services, which may also vary
33
across regions (Yang et al., 2021; Masters et al., 2022; Yang et al., 2023). Therefore, it is crucial to understand
34
both the direct and indirect impacts of people's attitudes towards the vaccine and vaccination status on
35
ridesourcing service usage-related decisions.
36
Regarding sociodemographic and travel-related characteristics, factors such as gender, age, education, income,
37
and residence significantly affect residents' risk perception and willingness to use ridesourcing during the
38
pandemic (Nguyen-Phuoc et al., 2022; Loa et al., 2022). Most studies suggest that women, the elderly, wealthy
39
individuals, and those with underlying health conditions are more likely to perceive a higher risk associated with
40
traveling during the COVID-19 pandemic (Bruine de Bruin, 2021; He et al., 2021; Park et al., 2022; Neubauer
41

6
et al., 2019). However, a few recent studies have highlighted variations among these groups (Abdullah et al.,
1
2021; Khaddar & Fatmi, 2021), particularly towards the end of the pandemic. Additional studies are needed to
2
better understand their impacts.
3
To sum up, existing studies offer valuable insights to relevant stakeholders in designing policies for post-
4
pandemic recovery in the travel industry at large, and Table 1 summarizes some of the key findings in the
5
literature. To build on these understandings, a comprehensive framework is still needed to capture the
6
interrelationships among risk perception, attitudes towards COVID-19 vaccines, vaccination status, attitudes and
7
behaviors towards COVID-19 health measures by both individuals and service providers, and their direct and
8
indirect impacts on short- and long-term decisions related to ridesourcing usage in a post-COVID-19 era.
9
This study aims to develop such a framework to understand these interconnected relationships. Based on this
10
literature review, the proposed framework includes 23 hypotheses. To validate the framework, a stated and
11
revealed preference survey has been designed and distributed among ridesourcing service users aged 18 or above
12
in China. Multiple Indicators Multiple Causes (MIMIC) models were estimated to quantify the direct and indirect
13
impacts of various factors on people's ridesourcing service usage towards the end of the COVID-19 pandemic
14
and their willingness to use it after the pandemic. The model estimation results and insights may help decision-
15
makers design effective strategies to influence ridesourcing service usage in a post-COVID-19 era.
16
17
3. Methodology
18
3.1. Methodological review
19
Structural Equation Modelling (SEM) is a versatile multivariate statistical method that is linear in parameters
20
and includes both latent and observed variables. Observed variables are collected directly, such as responses to
21
survey questions, while latent variables (e.g., attitudes, perceptions, and emotions) are used to measure
22
unobserved constructs that can be inferred from the observed variables (Kaplan et al., 2022). This approach
23
allows for modeling multiple dependent variables with both observed and latent independent variables
24
simultaneously (Bowen & Guo, 2011; Batomen et al., 2022). Therefore, SEM is an ideal method for this study
25
as it involves several observed and latent variables with dependent variables such as people's ridesharing service
26
usage frequency and willingness to use it after the pandemic.
27
SEM models have two primary components: a measurement model and a structural model (see Fig. 1). The
28
measurement model specifies how well various measured exogenous (observed) variables identify latent
29
(unobserved) variables (Skrondal & Rabe‐Hesketh, 2005). Each latent variable has its own latent construct,
30
which contains exogenous variables and their factor loadings. The structural model is used to capture the
31
interrelationships among latent variables and variables that do not belong to any latent constructs. These
32
relationships are developed based on existing literature or hypotheses and collectively form a conceptual model,
33
which is then validated through model estimation results (Washington et al., 2020; Zheng et al., 2020).
34
This study uses a special form of SEM model, the MIMIC model (Jöreskog & Goldberger, 1975), to examine
35
potential heterogeneity among different subgroups with sociodemographic characteristics such as gender and
36
age. This is accomplished by incorporating regressors or covariates in the model framework (see Fig. 1).
37
Compared to traditional methods that partition samples into several groups, this method enables analysis of
38
subpopulation differences with one sample (Kline, 2015). Gender, age, education, income, employment type,
39
and location of residence are evaluated as possible regressors to explore possible heterogeneous behavior among
40

7
respondents.
1

8
Table 1 Literature review summary related to various factors affecting people’s mode choice and
ridesourcing usage during the COVID-19.
Reference
Factors
Region
Sample
Sociodemographic
Risk Perception
Attitude
Health Measures
Vaccination
Size
Mode Choice
Abdullah et al. 2020
Gender +
✓+
✓+
Global
1,203
Age ×
Education ×
Occupation ×
Income ×
Bhaduri et al. 2020
Gender -
India
498
Age -
Income +
Education +
Abdullah et al. 2021
Gender +
✓+
Pakistan
1,516
Age -
Education -
Occupation -
Income +
Neuburger & Egger
Age ×
✓+
DACH region
(Germany,
Austria,
Switzerland)
1,158
2021
Gender +
Education ×
Travel frequency +
Scorrano & Danielis
Age -
✓-
✓+
Italy
11,922
2021
Gender +
Employment +
Shakibaei et al. 2021
Gender ×
Istanbul
144
Age +
Education ×
Occupation +
Income ×
Abdullah et al. 2022
Gender -
✓+
Pakistan
671
Age ×
Income -
Education ×
Zhao & Gao 2022
Gender +
✓-
✓+
China
1,360
Age -
Education -
Occupation +
Income +
Downey et al.2022
Age -
✓+
✓+
Scotland
994
Occupation +
Education +
Mashrur et al. 2022
Age -
✓+
✓+
Greater
Toronto Area
(GTA)
905
Gender +
Occupation -

9
Reference
Factors
Region
Sample Size
Sociodemographic
Risk Perception
Attitude
Health Measures
Vaccination
Trip Frequency & Willingness to Travel
Hotle et al. 2020
Gender -
✓-
✓+
✓+
U.S.
3,604
Age ×
Income -
Region +
Employment ×
Aaditya & Rahul 2021
Gender +
✓+
✓+
/
410
Age -
Region ×
Education +
Occupation -
Fan et al. 2022
Gender -
✓-
✓+
✓+
China
1,423
Age -
Education +
Income +
Park et al. 2022
Age -
U.S.
393
Gender +
Employment status +
Education -
Income +
Region -
Ridesourcing Usage & Intention
Alemi et al. 2018
Gender +
✓+
California
1,975
Age +
Education +
Income +
Employment status ×
Region ×
Nguyen-Phuoc et
Gender ×
✓-
✓+
Vietnam
545
al. 2021
Age +
Education ×
Income ×
Loa et al. 2022
Age +
✓+
✓×
✓-
Greater
Toronto
Area
920
Employment status ×
Education -
Nguyen-Phuoc et al.
Gender ×
✓+
✓+
✓+
Vietnam
562
2022
Age ×
Education ×
Zhang & Liu 2022
Gender ×
✓-
China
964
Age -
Education -
Income -
(×indicates the factor was investigated in the study but was not found to be significant; + indicates the factor in
the study was significantly positive; - indicates opposite).

10
1
Fig. 1. MIMIC model framework.
2
3
A literature-based five-step method (Allen et al., 2018; Jöreskog & Goldberger, 1975; Washington et al., 2020;
4
Zheng et al., 2020) is utilized to establish the proposed framework using the MIMIC modeling method. The
5
steps are as follows: (i) estimate Herman's single factor score and determine whether the information of a latent
6
variable comes from only one factor; (ii) apply Exploratory Factor Analysis (EFA) to evaluate the relationships
7
between latent and observed variables, followed by Confirmatory Factor Analysis (CFA) to assess the model's
8
validity and internal consistency; (iii) measure the construct's internal reliability; (iv) evaluate the degree of
9
conceptual overlap among formative indicators to avoid multicollinearity; and (v) assess the model's goodness-
10
of-fit.
11
12
3.2. Model construct and hypothesis development
13
The proposed model framework consists of 23 hypotheses aimed at capturing the factors that influence
14
people's post-pandemic expectations regarding self-imposed and service-imposed health measures, their usage
15
frequency of ridesourcing services during and after the pandemic, and the interrelationships among these factors
16
(see Fig. 2). These hypotheses were developed based on a literature review and the interrelationship among the
17
nine factors that the authors seek to validate. These factors can be classified into four types: (i) perception/attitude
18
related to COVID-19 and its vaccines (risk perception toward COVID-19, confidence in vaccines, and
19
respondents' vaccination status); (ii) self-imposed and service-imposed health measures to combat COVID-19
20
at the time of the survey; (iii) expected post-pandemic self-imposed and service-imposed health measures, and
21
post-pandemic willingness to use ridesourcing services; and (iv) ridesourcing service usage frequency both
22
before the COVID-19 outbreak and currently (e.g., frequency of using ridesourcing services before December
23
2019 and in the past three months).
24
It is expected that people who are concerned about the COVID-19 infection and its health-related
25

11
consequences were less likely to report traveling during the pandemic but more likely to consider getting
1
vaccinated once it is/was available (Hotle et al., 2020). Similarly, people who believe in the effectiveness of
2
COVID-19 vaccines are more likely to get vaccinated and maintain their pre-pandemic travel behavior (Fan et
3
al., 2022). Therefore, this study defines risk perception as the degree of people's concern and fear of COVID-19
4
infection and its physical health-related consequences. Attitudes toward vaccines are designed to capture
5
passengers' confidence in the development, effectiveness, and safety of COVID-19 vaccines.
6
As rates of vaccination were high in China (nearly 80%) at the time of the survey’s distribution (China Release,
7
2021), ‘boosted’ is introduced as a binary indicator to observe whether the respondent has received a booster
8
shot of vaccines. It is assumed that individuals possessing a more positive attitude towards vaccines are likely
9
to be more inclined towards vaccination. This hypothesis finds support in the outcomes of studies such as Adane
10
et al. (2022) and Paul et al. (2021). However, the situation in China, as opposed to other countries, strongly
11
encourages students and full-time workers to get vaccinated in certain regions. One recent study shows that six
12
months after the national vaccination campaign (launch on December 2020) in China, over 80% of the survey
13
participants who had a prior COVID-19 vaccination intention before the campaign actually received the vaccine,
14
while, more impressively, around 60% of those without a prior intention got vaccinated (Wang et al., 2022). Such
15
results suggest that the effectiveness of this vaccination campaign and may indicates an overall high vaccination
16
rate at the time of our study. As a result, the "boosted" variable (referring to people who received the COVID-
17
19 vaccine booster) was used instead of the "vaccinated" variable to better encapsulate more voluntary behavior
18
related to vaccination. Regarding its correlation with people's willingness to use ride-sourcing services, recent
19
studies such as Almokdad et al., 2023; Morar et al., 2022, suggest that the receipt of a booster shot have
20
statistically significant impact on travel behavior and mode choices.
21
To better capture the protective behavior of ridesourcing service users towards COVID-19 and potential
22
health-related concerns beyond the pandemic, we introduce self-imposed and service-imposed health measures
23
at the time of survey (now) and post-pandemic (expected). These measures reflect individuals' likelihood of
24
utilizing self-imposed health measures (e.g., wearing a mask, social distancing) during and after the pandemic,
25
as well as their expectations for ridesourcing service providers to implement necessary health measures to
26
prevent the spread of COVID-19. Additionally, we introduce three potential exogenous variables which represent
27
individuals' usage of ridesourcing services three months before December 2019 (pre-pandemic frequency), three
28
months before December 2021 (current frequency), and after the pandemic (post-pandemic willingness).
29
Eight hypotheses are proposed to describe the interrelationship among the aforementioned factors in this study
30
as follows:
31
H1. The greater the risk perception,
32
(a) the more likely individuals are to use self-imposed health measures when using ridesourcing services at
33
the time of the survey (current self-imposed health measures).
34
(b) the more likely individuals are to expect provider-imposed health measures when using ridesourcing
35
services at the time of the survey (current provider-imposed health measure expectations).
36
(c) the more likely individuals are to use ridesourcing services at a lower frequency at the time of the survey
37
(current frequency).
38
(d) the more likely individuals are to keep service-imposed health measures when using ridesourcing
39
services after the pandemic (post-pandemic service-imposed health measure expectations (provider)).
40
(e) the more likely individuals are to keep health measures that were once service-imposed when using
41

12
ridesourcing services after the pandemic (post-pandemic expectations (self)).
1
(f) the lower willingness to use ridesourcing services after the pandemic (post-pandemic willingness).
2
(g) the higher likelihood of having received a COVID-19 booster (boosted).
3
H2. The more positive attitudes toward vaccines, the higher likelihood of having received a COVID-19 booster.
4
H3. People with a COVID-19 booster,
5
(a) are more likely to use self-imposed health measures currently when using ridesourcing services (current
6
self-imposed health measures).
7
(b) are more likely to expect service-imposed health measures at the time of the survey (current expectation
8
(service)).
9
(c) are more likely to have a lower ridesourcing service usage frequency at the time of the survey (current
10
frequency).
11
(d) are more willing to use ridesourcing services after the pandemic (post-pandemic willingness).
12
H4. The greater the likelihood of using self-imposed health measures when using ridesourcing services at the
13
time of the survey,
14
(a) the more likely individuals are to expect provider-imposed health measures when using ridesourcing
15
services at the time of the survey (current provider-imposed health measure expectations).
16
(b) the more likely individuals are to use ridesourcing services at a lower frequency at the time of the survey
17
(current frequency).
18
(c) the more likely individuals are to keep service-imposed health measures when using ridesourcing
19
services after the pandemic (post-pandemic provider-imposed health measure expectations (provider)).
20
(d) the more likely individuals are to keep service-imposed health measures when using ridesourcing
21
services after the pandemic (post-pandemic expectations (self)).
22
H5. The higher likelihood of expecting service-imposed health measures when using ridesourcing services at
23
the time of the survey,
24
(a) the lower the ridesourcing service usage frequency at the time of the survey (current frequency).
25
(b) the higher likelihood of expecting service-imposed health measures after the pandemic (post-pandemic
26
expectation (service)).
27
H6.
28
(a) The higher likelihood of keeping self-imposed health measures after the pandemic, or
29
(b) the higher likelihood of keeping service-imposed health measures after the pandemic,
30
the greater willingness to use ridesourcing services after the pandemic (post-pandemic willingness).
31
H7. The higher the pre-pandemic ridesourcing service usage frequency,
32
(a) the higher the ridesourcing service usage frequency at the time of the survey (current frequency).
33
(b) the greater willingness to use ridesourcing services after the pandemic (post-pandemic willingness).
34
H8. The higher the ridesourcing service usage frequency at the time of the survey, the greater willingness to
35
use ridesourcing services after the pandemic (post-pandemic willingness).
36
In addition to examining people's perceptions of risk, attitudes, and health measures, this study employs six
37
sociodemographic variables (all binary) to investigate heterogeneity within subpopulation groups using the
38
MIMIC model framework. These variables include gender (with male coded as 1 and female as 0), age (with
39
millennials or younger coded as 1, otherwise 0), level of education (with a college degree or higher coded as 1,
40
otherwise 0), pre-tax annual self-imposed income (with ¥75,000 or higher coded as 1, otherwise 0), employment
41

13
status (with employed coded as 1, otherwise 0), and location of residence (with urban areas coded as 1, otherwise
1
0). Socio-demographic variables such as age, education, and income are encoded as binary variables for two
2
primary reasons. Socio-demographic variables such as age, education, and income are encoded as binary
3
variables for two primary reasons. First, in comparing models that employ binary socio-demographic variables
4
with those that use ordered socio-demographic variables, we found that the former outperform the latter across
5
multiple criteria: model goodness-of-fit, AIC, and BIC metrics (Appendix). Secondly, our approach strikes a
6
balance between optimizing the model's goodness-of-fit and aligning with insights from existing literature. For
7
instance, when it comes to the "age" variable, previous travel mode choice-related studies (Delbosc and
8
Nakanishi, 2017; Guo et al., 2021; Nayum & Nordfjærn, 2021; Zhang et al., 2022; Jia et al., 2023) have suggested
9
that the travel preferences of millennials and younger generations differ substantially from older cohorts.
10
Younger demographics are often seen as a 'generation in transition' (Garikapati et al., 2016), which influences
11
their transportation choices. Figure 3 provides an illustration of the relationship between these sociodemographic
12
factors and the measurement model, using "current self-imposed health measures (CSHM)" as an example.
13
When it comes to age, millennials and younger generations tend to be more acquainted with ridesourcing and
14
other mobility services that are smartphone-related than their older counterparts. In this study, individuals born
15
after 1981 (i.e., under 41 years of age at the time of the survey), are categorized as such. Regarding pre-tax
16
annual income, ¥75,000 is roughly the average amount earned by Chinese adults in 2021 (National Bureau of
17
Statistics of the People's Republic of China, 2022).
18
19

14
Fig. 2. The framework of the proposed model for exploring the impacts of COVID-19 on ridesourcing usage, post-pandemic expectation to health
measures, and willingness to use ridesourcing services post-pandemic.
(+ suggests a positive correlation, while - suggests a negative correlation).

15
1
2
Fig.3. The example of measurement model with a latent variable (yellow CSHM), its observed indicators
3
(yellow CSHM1, CSHM2, CSHM3, CSHM4, and CSHM5), and six travel and sociodemographic variables
4
(blue).
5
6
3.3. Survey design and distribution
7
A survey was conducted to validate the proposed model framework, utilizing both stated preference and
8
revealed preference methods. The survey was divided into six sections. The first section collected
9
sociodemographic information, including gender, age, residence type (urban, suburban, or rural), income,
10
education, income, and COVID-19 vaccination status. The second section focused on risk perception and
11
attitudes towards the COVID-19 pandemic, while the third section assessed respondents' knowledge of the
12
pandemic. The fourth section examined respondents' attitudes towards precautions and policies in their province
13
and their confidence in vaccines. The fifth section of the survey comprises both stated preference and revealed
14
preference survey questions. Questions related to participants' attitudes and behaviors towards ridesourcing
15
service usage before and during the pandemic can be categorized as self-reported revealed preference survey
16
questions. As for their post-COVID-19 behavior, a stated preference method was employed to capture their
17
willingness to utilize ridesourcing services after the pandemic has subsided.
18
Table 3 lists all the questions related to variables and hypotheses outlined in section 3.2. Questions related to
19
perception and attitude were measured using a 5-point Likert scale ranging from 1=“strongly disagree” to
20
5=“strongly agree”.
21
22

16
Table 3 Measurement items for each variable of the MIMIC construct.
Variables
Items
Attitudes toward vaccines (VA)
What is your level of confidence in the development and approval processes of COVID-19 vaccines? (VA1)
What is your level of confidence about the COVID-19 vaccine’s effectiveness? (VA2)
What is your level of confidence about the COVID-19 vaccine’s safety? (VA3)
Current self-imposed health measures
(CSHM)
Now, I would sanitize my seat/space in a ridesourcing service vehicle by myself before sitting down. (CSHM1)
Now, I would wear a mask when I am using ridesourcing services. (CSHM2)
Now, I engage in preventive behavior (e.g., hand hygiene and cleaning your seats) more often when traveling due to COVID-19.
(CSHM3)
Current service-imposed expectation
(CSE)
I expect ridesourcing services to implement cleaning/disinfecting measurements to combat COVID-19. (CSE1)
I expect ridesourcing services to display or communicate when the last cleaning/disinfecting occurred. (CSE2)
I expect ridesourcing services to display or communicate what type of cleaning/disinfecting occurred. (CSE3)
I expect ridesourcing services to provide physical barriers among passengers/drivers. (CSE4)
I expect ridesourcing service drivers to wear a mask. (CSE5)
Risk perception (RP)
The COVID-19 is widespread and highly contagious. (RP1)
Getting COVID-19 is fatal. (RP2)
I am afraid of being infected. (RP3)
I am afraid the people I care about may be affected. (RP4)
The pandemic is terrible. (RP5)
Current frequency (CF)
I used ridesourcing services like Didi or Caocao _____ times per week in the past 3-months. (CF)
Pre-pandemic frequency (PF)
I used ridesourcing services _____ times per week before the pandemic. (PF)
Post-pandemic expectation (self) (PE-Self)
Even after the COVID-19 ends, I expect ridesourcing services to maintain the current cleaning/disinfecting measurements. (PE-Self)
Post-pandemic expectation (service)
(PE-Service)
Even after the COVID-19 ends, I expect to maintain my preventive behavior (e.g., hand hygiene and cleaning your seats) when
traveling. (PE-Service)
Boosted (Boosted)
Have you been boosted? (Boosted)
Post-pandemic willingness (PW)
Once the COVID-19 ends, I would use ridesourcing services as often as I did before COVID-19. (PW)

17
The anonymous survey was distributed via Questionnaire Star (https://www.wjx.cn/) between November 22,
1
2021, and January 6, 2022. This platform is one of the largest survey platforms in China, with millions of
2
potential respondents. To participate in the survey, respondents had to be at least 18 years old and residing in
3
China at the time of the survey. Any completed surveys that were too short (less than five seconds per question)
4
or that failed any of the three attention check questions were removed. A total of 2,369 valid responses were
5
collected. Participants who completed a valid survey response were paid 8-yuan through the survey platform
6
(around 1.2 USD).
7
8
4. Results
9
4.1. Descriptive statistics of the respondents’ sociodemographic and travel behavior characteristics
10
Table 4 presents some of the key sociodemographic and travel behavior characteristics of the survey
11
respondents. From the data, a few observations can be made. Firstly, our participants skew towards female or
12
millennials or belong to younger generations (those born after 1980 or under 41 years old at the time of the
13
survey). They are generally well-educated, with a college degree or higher, or are employed full-time. They also
14
belong to the high-income group, with a national average disposable income (total personal income minus
15
current income taxes) of around 35,000 yuan or $5,057 per person. Additionally, they are mainly living in urban
16
areas. Although the actual demographics of the ridesourcing service users in China, this study sample’s
17
demographics is similar to other ridesharing-related studies, particularly pertain to those in China (Feng et al.,
18
2023; Zhang et al., 2022; Ren et al., 2023). Given access to the actual demographics of ridesourcing service
19
users, future studies would certainly compare them with the results of this study. Secondly, over 90% of the
20
respondents reported being fully vaccinated, and approximately 30% of them had received booster shots. This is
21
relatively high compared to the national average and other parts of the world. This is likely because the
22
respondents were predominantly younger, and the vaccination rate among the elderly remained low throughout
23
the pandemic compared to other age groups. As of July/August 2022, approximately 70% of individuals in China
24
who are over 80 years old have completed the primary series of vaccinations (Wang et al., 2023). Thirdly,
25
concerning travel behavior, public transportation (subway and bus) was the most used mode for commute before
26
COVID-19 and at the time of the survey. However, its overall usage decreased sharply, especially for buses, with
27
more people choosing to drive themselves, bike, or walk. This trend was observed globally during the pandemic.
28
Lastly, carpooling or ridesharing services were used by fewer people, but the difference is minor.
29
30
Table 4 Descriptive statistics of the respondents’ sociodemographic and travel behavior characteristics (n = 2369).
31
Socio-demographic Variables
Items
Percentage
Sample Size
Gender
Male
44.6
1056
Female
55.4
1313
Age
15-20
5.9
140
21-30
52.3
1240
31-40
33.5
793
41-50
5.8
138
Above 60
2.4
58
Residence type
Urban
81.9
1941
Suburban
14.0
332
Rural
4.1
96
Education
Less than primary school
0.3
8

18
Primary or secondary school
5.7
135
College / university
85.4
2022
Graduate school
8.6
204
Self-imposed Annual Pretax Income
Less than ￥35, 000
15.7
371
￥35, 000 to ￥55, 000
10.1
239
￥55, 000 to ￥75, 000
13.7
325
￥75, 000 to ￥95, 000
12.9
305
￥95, 000 to ￥150, 000
26.8
636
￥150, 000 to￥300, 000
17.5
415
￥300, 000 or more
3.3
78
Employment Status
Student without a part-time job
5.8
137
Student with a part-time job
5.6
133
Employed full time
84.8
2009
Employed part time
1.6
39
Unemployed looking for work
0.8
19
Unemployed not looking for work
0.6
14
Retired
0.8
18
Boosted
Fully vaccinated with booster
31.1
737
Fully vaccinated without booster
62.9
1490
Not fully vaccinated
6
142
Most commonly used mode for
commute before COVID-19
Subway
28.7
681
Ridesourcing or carpooling
17.0
402
Driving
21.4
506
Bus
20.6
488
Bike or walk
11.2
267
Others
0.7
16
Don’t need to go to work or school
0.4
9
Most common commute mode choice at
the time of the survey
Subway
23.2
549
Ridesourcing or carpooling
15.3
363
Driving
29.4
696
Bus
14.7
348
Bike or walk
15.7
372
Others
1.1
25
Don’t need to go to work or school
0.7
16
Pre-pandemic Frequency
Less than once a month
19.4
459
Once a month
10.6
251
2-4 times a month
36.9
875
2-3 times a week
22.1
524
4-6 times a week
8.6
203
At least once a day
2.4
57
Current Frequency
Never
6.0
142
Less than once a month
21.9
518
Once a month
13.3
315
2-4 times a month
31.4
743
2-3 times a week
18.5
439
4-6 times a week
6.7
159
At least once a day
2.2
53
1
4.2. Model construct validation
2
For model validation and MIMIC model estimation, SPSS 24.0 and Smart-PLS 3.0 were used. Initially,
3

19
Herman's single factor score was used to assess the information source, which revealed that it did not come from
1
a single factor. An exploratory factor analysis (EFA) was then conducted, which showed that four factors should
2
be extracted for further analysis. These results are displayed in Table 5, with the Kaiser-Mayer-Olkin criterion
3
being 0.86 and the Bartlett's test p-value being less than 0.001. After that, confirmatory factor analysis (CFA)
4
was conducted to determine if the collected data matched the model, which is also presented in Table 5. The
5
empirical results indicated that the model is a good fit for the survey data, as evidenced by the following values:
6
root mean square error of approximation (RMSEA) = 0.059 (recommended RMSEA < 0.080), standardized root
7
mean square residual (SRMR) = 0.044 (recommended SRMR < 0.080), comparative fit index (CFI) = 0.913
8
(recommended CFI ≥ 0.9), and Tucker-Lewis Index (TLI) = 0.901 (recommended TLI ≥ 0.900) (Hu & Bentler,
9
1999).Subsequently, reliability and validity tests were conducted on variables with multiple measurement items,
10
utilizing Cronbach's Alpha values and average variance extracted (AVE). All constructs nearly met the criterion,
11
as shown in Table 5, with the values of Cronbach's Alpha being greater than 0.7 and AVE being greater than 0.5
12
(Chiu & Wang, 2008). Finally, variance inflation factors (VIFs) were used to assess the potential multicollinearity,
13
with all values meeting the criterion requirement (recommended VIF < 3.3) (Petter et al., 2007). Based on these
14
results, the model was deemed to be well-fitted for further estimation.
15
16
Table 5 The reliability and validity results of measurement model for each latent variable.
17
Variables
Items
Factor
loadings
Cronbach's
Alpha
Composite
Reliability (C.R.)
Average Variance
Extracted (AVE)
Attitudes toward Vaccines (VA)
VA1
0.852
0.817
0.890
0.729
VA2
0.892
VA3
0.816
Risk Perception (RP)
RP1
0.705
0.727
0.851
0.534
RP2
0.715
RP3
0.723
RP4
0.735
RP5
0.773
Current service-imposed expectation
(CSE)
CSE1
0.707
0.740
0.859
0.549
CSE2
0.743
CSE3
0.720
CSE4
0.730
CSE5
0.800
Current Self-imposed Health Measures
(CSHM)
CSHM1
0.804
0.762
0.826
0.613
CSHM2
0.739
CSHM3
0.804
18
4.3. Model estimation results
19
Based on the reliability and validity results presented above, we utilized the regression method of partial least
20
squares (PLS) and bootstrapping to test the hypotheses and further examine the impact of health and risk
21
perception, expectation for precautions of ridesourcing services, and other latent variables. Given the complexity
22
of the MIMIC model structure, we chose to use the PLS-SEM method for further data analysis, with
23

20
bootstrapping employed to verify the significance of the path in the model structure using SmartPLS with 5000
1
subsamples (Hair et al., 2019).
2
Table 6 presents the model estimation results, including standardized path coefficients (std. estimate), standard
3
errors, and p-values. We considered the hypothesis supported if p < 0.05, indicating that a latent variable exerts
4
a significant effect on a connected variable. As shown in section 4.2 above, the results suggest that the proposed
5
model has a good model fit. The model results, presented in Fig. 6, show that all proposed hypotheses were
6
tested and supported except for hypotheses H1(c), (f)-(g), H3(b)-(d), H4(b), and H6(b). Regarding perception
7
towards COVID-19 and vaccines, Table 6 shows that risk perception has a significant and positive effect on
8
respondents' Current Self-imposed Health Measures (CSHM) (β=0.468, p<0.001), Current Service-imposed
9
Expectation (CSE) (β=0.444, p<0.001), Post-pandemic Expectation (Service) (PE-Service) (β=0.204, p < 0.001),
10
and a negative effect on Post-pandemic Expectation (Self) (PE-Self) (β=-0.055, p<0.001). Thus, hypotheses H1(a)
11
and H1(c)-(e) are supported. Moreover, the empirical results show that Boosted (Boosted) significantly and
12
positively affects CSHM (β=0.035, p<0.001), which supports H3(a). The results also verify that Attitudes
13
towards Vaccines (VA ) have a significant and positive effect on Boosted (β=0.094, p<0.001), thereby supporting
14
H2.
15
Although the hypotheses of H1(c) and H1(f)-(g) have been verified, there are no statistically significant direct
16
impacts of Risk Perception (RP) on Post-pandemic Willingness (PW), Current Frequency (CF), and Boosted.
17
However, RP still has significantly indirect impacts on CF and PW through CSHM, CSE, PE-Service, and PE-
18
Self. Furthermore, Boosted and VA significantly and indirectly influence people's Post-Pandemic Expectation
19
(both service- and self-imposed) and PW to use ridesourcing services.
20

21
Fig. 6. Results of MIMIC model estimation for evaluating the COVID-19 impacts on the Post-pandemic Willingness as often as pre-pandemic when after
the pandemic.
Notes. *** Significant at the p < .01 level, while ** significant at the p < .05 level. + suggests a positive impact while - suggests a negative impact.

22
Table 6 verifies that both during and after the pandemic, individuals' intentions to take health measures and
1
use ridesourcing services are significantly impacted. Specifically, CSHM has a significant and positive effect on
2
PE-Service (β=0.006, p< 0.001) and PE-Self (β=0.105, p<0.001), which supports H4(c)-(d). Additionally, H4(a)
3
is supported, meaning that CSHM has a significant and positive effect on CSE (β= 0.253, p<0.001). The model
4
results also indicate that CSE has a significant effect on CF (β= 0.083, p< 0.05) and PE-Service (β=-0.032,
5
p<0.001), supporting H5. Furthermore, PE-Service has a significant and positive effect on PW (β=0.011, p<0.05),
6
while PE-Self does not have a statistically significant impact on users’ willingness (H6(b)). The model results
7
suggest that the remaining variables do not have statistically significant impacts on the decision variables.
8
Regarding other variables, the results show that PF significantly and positively impacts CF (β= 0.025, p<0.001)
9
and PW (β= 0.012, p<0.05), and CF also has a significant and positive effect on consumers' PW (β= 0.033,
10
p<0.05), thereby supporting H7.
11
12
Table 6 Results of hypothesis testing (Direct effects)
13
Hypothesis
Path Coefficients
Std. Error
T Statistics
p-Value
Conclusion
H1(a): RP -> CSHM
0.468
0.024
11.069
0.000
Supported
H1(b): RP -> CSE
0.444
0.021
12.25
0.000
Supported
H1(c): RP -> CF
0.008
0.016
0.346
0.729
Not Supported
H1(d): RP -> PE-Service
0.204
0.021
3.933
0.000
Supported
H1(e): RP -> PE-Self
-0.055
0.02
5.222
0.000
Supported
H1(f): RP -> PW
-0.147
0.021
0.542
0.588
Not Supported
H1(g): RP -> Boosted
0.272
0.022
1.475
0.140
Not Supported
H2: VA -> Boosted
0.094
0.019
7.737
0.000
Supported
H3(a): Boosted -> CSHM
0.035
0.02
4.555
0.000
Supported
H3(b): Boosted -> CSE
0.676
0.016
1.547
0.122
Not Supported
H3(c): Boosted -> CF
0.051
0.015
0.772
0.440
Not Supported
H3(d): Boosted -> PW
0.069
0.02
1.642
0.101
Not Supported
H4(a): CSHM -> CSE
0.253
0.02
23.71
0.000
Supported
H4(b): CSHM -> CF
0.269
0.018
0.423
0.672
Not Supported
H4(c): CSHM -> PE-Service
0.006
0.026
7.964
0.000
Supported
H4(d): CSHM -> PE-Self
0.105
0.019
23.36
0.000
Supported
H5(a): CSE -> CF
0.083
0.019
2.965
0.003
Supported
H5(b): CSE -> PE-Service
-0.032
0.027
10.038
0.000
Supported
H6(a): PE-Service -> PW
0.011
0.023
2.281
0.023
Supported
H6(b): PE-Self -> PW
0.089
0.023
1.525
0.127
Not Supported
H7(a): PF -> CF
0.025
0.014
48.024
0.000
Supported
H7(b): PF -> PW
0.012
0.03
3.166
0.002
Supported
H7(c): CF -> PW
0.033
0.03
2.308
0.021
Supported
Notes:
14
1. Statistically significant positive impact: Red; Statistically significant negative impact: Blue; Not statistically
15
significant: Green.
16
2. RP: Risk Perception; CSHM: Current Self-imposed Health Measures; CSE: Current Service-imposed
17
Expectation; PE-Service: Post-pandemic Expectation (service); PE-Self: Post-pandemic Expectation (self);
18
PW: Post-pandemic Willingness; CF: Current Frequency; PF: Pre-pandemic Frequency; Boosted: Boosted;
19

23
VA: Attitudes toward Vaccines.
1
2
Table 7 displays the estimation results of the MIMIC model, which show the impact of sociodemographic
3
characteristics on latent variables. The results indicate that gender has significant effects on several variables,
4
including CSHM, CSE, PE-Service, Boosted, and PW. In comparison to their male counterparts, female
5
respondents are more likely to implement health measures and receive booster shots of COVID-19 vaccines to
6
reduce virus exposure. They also expect more health-related and preventative measures from ridesourcing
7
operators or drivers. However, they have a lower willingness to use ridesourcing services as often as they did
8
before the pandemic. Additionally, the "Age" factor significantly affects most of the variables. Millennials and
9
younger respondents have a more positive view of COVID-19 vaccines and a higher risk perception of the
10
pandemic. They also use ridesourcing services more frequently than older people, both before the pandemic and
11
in the past three months. However, there were no statistically significant differences in post-pandemic
12
expectations for health measures. The model estimation results show that the "Edu" factor only significantly
13
impacts consumers' attitudes toward vaccines but does not significantly affect the rest of the variables. Most
14
people with a college degree or above rate COVID-19 vaccines more positively. Moreover, those with relatively
15
higher annual self-imposed pretax income (over 75,000 RMB) travel by ridesourcing more frequently before
16
December 2019 and have a higher willingness to use ridesourcing as often as they did before once the pandemic
17
is no longer considered a threat. The "Employment Status" factor has significant impacts on variables related to
18
health and safety measures and the frequency of using ridesourcing before the pandemic. Those with fixed
19
income (employed full time and with pensions) are more likely to use ridesourcing services more frequently
20
before December 2019. They are also more concerned about preventative measures to reduce the risk of close
21
contact with others or infection when using ridesourcing and are more likely to take precautions when using
22
ridesourcing services. Regarding the "Residence" factor, those living in urban areas are more likely to show a
23
higher concern and expectation for health-related measures provided by ridesourcing services. They also use
24
ridesourcing services more frequently both before and after the pandemic. However, they have a lower
25
expectation to implement preventive measures to reduce the risks when the pandemic is over compared to those
26
living in suburban or rural regions. The rest of the relationships between latent variables and sociodemographic
27
characteristics do not have statistical significance.
28
29

24
Table 7 Sociodemographic characteristics’ relationship with latent variables (bold results suggest a statistically significant relationship).
Std.
Estimate
Std.
Error
p -
value
Gender: most male respondents are more likely to … compared to female respondents
… does not have a statistically significant impact on the attitude toward COVID-19 vaccines...
0.026
0.021
0.221
… use self-imposed health measures when using ridesourcing services at the time of the survey...
-0.086
0.02
0.000
… expect service-imposed health measures when using ridesourcing services at the time of the survey...
-0.075
0.021
0.000
… does not have a statistically significant impact on the ridesourcing service usage frequency at the time of the survey...
0.015
0.021
0.481
… keep self-imposed health measures after the pandemic...
-0.054
0.02
0.008
… expect service-imposed health measures after the pandemic ...
-0.077
0.021
0.000
… does not have a statistically significant impact on the pre-pandemic ridesourcing service usage frequency...
0.004
0.02
0.826
… does not have a statistically significant impact on the risk perception...
0.004
0.021
0.841
… receive a COVID-19 booster...
-0.046
0.021
0.027
… have greater willingness to use ridesourcing services after the pandemic...
0.057
0.02
0.005
Age: most millennials or younger are more likely to …. compared to old generations…
… have more positive attitudes toward COVID-19 vaccines…
0.062
0.025
0.012
… does not have a statistically significant impact on whether use self-imposed health measures when using ridesourcing services at the time of
the survey…
0.013
0.02
0.510
… does not have a statistically significant impact on whether expect service-imposed health measures at the time of the survey…
0.042
0.025
0.100
… have higher ridesourcing service usage frequency at the time of the survey…
0.111
0.021
0.000
… does not have a statistically significant impact on whether use self-imposed health measures after the pandemic…
0.002
0.02
0.907
… does not have a statistically significant impact on whether expect service-imposed health measures after the pandemic…
0.005
0.021
0.794
… have higher pre-pandemic ridesourcing service usage frequency…
0.096
0.02
0.000
… have lower risk perception.…
-0.056
0.025
0.027
… does not have a statistically significant impact on whether receive a COVID-19 booster.…
0.005
0.02
0.790
… does not have a statistically significant impact on the willingness to use ridesourcing after the pandemic…
0.021
0.018
0.262
Education: most respondents with a college degree or above are more likely to … compared to those who don’t have…
… have more positive attitudes toward COVID-19 vaccines…
0.067
0.029
0.022
… does not have a statistically significant impact on whether use self-imposed health measures when using ridesourcing services at the time of
the survey…
0.019
0.027
0.490
… does not have a statistically significant impact on whether expect service-imposed health measures at the time of the survey…
0.042
0.03
0.165
… does not have a statistically significant impact on the ridesourcing service usage frequency at the time of the survey…
-0.013
0.023
0.559

25
… does not have a statistically significant impact on whether use self-imposed health measures after the pandemic…
0.024
0.023
0.301
… does not have a statistically significant impact on whether expect service-imposed health measures after the pandemic…
0.031
0.026
0.233
… does not have a statistically significant impact on the pre-pandemic ridesourcing service usage frequency…
-0.008
0.024
0.740
… does not have a statistically significant impact on the risk perception.…
0.042
0.029
0.143
… does not have a statistically significant impact on whether receive a COVID-19 booster.…
0.034
0.021
0.099
… does not have a statistically significant impact on the willingness to use ridesourcing after the pandemic…
-0.001
0.022
0.950
Annual self-imposed pretax income: most respondents with relatively high annual self-imposed pretax income (more than 75,000 RMB) are more
likely to … compared to those who have relatively low annual self-imposed pretax income.…
… does not have a statistically significant impact on the attitude toward COVID-19 vaccines…
0.001
0.024
0.963
… does not have a statistically significant impact on whether use self-imposed health measures when using ridesourcing services at the time of
the survey…
0.018
0.023
0.450
… does not have a statistically significant impact on whether expect service-imposed health measures at the time of the survey…
0.000
0.024
0.999
… have higher ridesourcing service usage frequency at the time of the survey…
0.076
0.023
0.001
… does not have a statistically significant impact on whether keep self-imposed health measures after the pandemic…
-0.013
0.023
0.564
… does not have a statistically significant impact on whether expect service-imposed health measures after the pandemic…
-0.012
0.023
0.601
… have higher pre-pandemic ridesourcing service usage frequency…
0.109
0.024
0.000
… does not have a statistically significant impact on the risk perception.…
-0.006
0.024
0.810
… does not have a statistically significant impact on whether receive a COVID-19 booster.…
0.039
0.024
0.100
… have higher willingness to use ridesourcing…
0.11
0.024
0.000
Employment Status: most respondents who have relatively fixed income are more likely to … compared to those who do not have fixed income…
… does not have a statistically significant impact on the attitude toward COVID-19 vaccines…
0.016
0.025
0.532
… use self-imposed health measures when using ridesourcing services…
0.085
0.025
0.001
… does not have a statistically significant impact on whether expect service-imposed health measures at the time of the survey…
0.034
0.026
0.198
… have higher ridesourcing service usage frequency at the time of the survey…
0.085
0.023
0.000
… keep self-imposed health measures after the pandemic…
0.07
0.024
0.004
… expect service-imposed health measures after the pandemic…
0.049
0.025
0.049
… have higher pre-pandemic ridesourcing service usage frequency…
0.094
0.022
0.000
… does not have a statistically significant impact on the risk perception.…
0.027
0.025
0.282
… does not have a statistically significant impact on whether receive a COVID-19 booster.…
-0.018
0.023
0.432
… does not have a statistically significant impact on the willingness to use ridesourcing after the pandemic…
0.024
0.024
0.320
Residence: most respondents who live in urban region are more likely to … compared to people living in suburban or rural regions.…
… does not have a statistically significant impact on the attitude toward COVID-19 vaccines…
0.014
0.022
0.520

26
… does not have a statistically significant impact on whether use self-imposed health measures when using ridesourcing services at the time of
the survey…
-0.017
0.021
0.399
… expect service-imposed health measures when using ridesourcing services at the time of the survey…
0.039
0.023
0.044
… have higher ridesourcing service usage frequency at the time of the survey…
0.08
0.021
0.000
… keep self-imposed health measures after the pandemic…
-0.043
0.021
0.044
… does not have a statistically significant impact on whether expect service-imposed health measures after the pandemic…
-0.003
0.022
0.884
… have higher pre-pandemic ridesourcing service usage frequency…
0.064
0.021
0.003
… does not have a statistically significant impact on the risk perception.…
0.039
0.023
0.089
… does not have a statistically significant impact on whether receive a COVID-19 booster.…
-0.037
0.021
0.079
… does not have a statistically significant impact on the willingness to use ridesourcing after the pandemic
0.025
0.021
0.242

27
5. Discussion
1
Drawing upon the model estimation results detailed in Section 4, this section presents relevant discussions
2
and policy insights from four critical perspectives: (i) impacts of COVID-19 pandemic on ridesourcing usage
3
and willingness to use it; (ii) the effects of risk perception and attitudes of COVID-19 pandemic; (iii) the effects
4
of self-imposed health-related measures and expectations for ridesourcing services; and (iv) the effects of
5
sociodemographic and travel behavior characteristics.
6
7
5.1. Impacts of COVID-19 pandemic on ridesourcing usage and willingness to use it (short- and long-term effects)
8
5.1.1 The short-term effects of COVID-19 pandemic on ridesourcing usage
9
Based on the results presented in Figure 7, the usage of ridesourcing services (such as the frequency of use or
10
willingness to use these services) has been impacted by the COVID-19 pandemic. Of the respondents, 35.6%
11
reduced their frequency of using ridesourcing services (either significantly or slightly) compared to before
12
December 2019, while 51% reported no change in their frequency of use. However, these findings suggest that
13
a post-pandemic rebound in ridesourcing service usage will be likely as 63.7% of respondents expressed a higher
14
willingness to use ridesourcing services as often as before the pandemic, once COVID-19 is no longer viewed
15
as a health threat. Those who used ridesourcing services more frequently before the pandemic (i.e., 2-3 times a
16
week or 4-6 times a week) reported reducing their usage, while some respondents who used ridesourcing services
17
less frequently (i.e., below once a month or once a month) reported increased their usage. In terms of willingness
18
to use ridesourcing services, respondents with both high and low usage frequencies expressed a positive or
19
neutral attitude (87%) towards recovering their usage of these services before the pandemic. Furthermore, of the
20
respondents who reduced their ridesourcing trips after the outbreak of the pandemic, 23% expressed a greater
21
willingness to use ridesourcing services as often as before December 2019. However, less than 5% of respondents
22
who showed a tendency to reduce their usage of ridesourcing services before the pandemic may refuse to recover
23
their usage.
24
These findings suggest that over half of the respondents may be willing to recover their usage or preference
25
for ridesourcing services as before the pandemic, which could help the recovery and development of the
26
ridesourcing industry. Respondents who had more frequent ridesourcing service usage before the pandemic and
27
reduced their usage are more likely to return to using these services as often as before. Additionally, the pandemic
28
may have longer-term impacts on people’s willingness to use public transport mode and ridesourcing services.
29
Therefore, it is important to focus on providing and improving the level of travel services to increase customers’
30
loyalty and dependence on ridesharing services. For consumers who are less willing to use ridesourcing services,
31
it is important to investigate the factors that are contributing to their resistance and how they can be addressed.
32
33
5.1.2 The long-term effects of COVID-19 pandemic on willingness to use ridesourcing services
34
As the results indicate, the COVID-19 pandemic has had a direct and indirect impact on ridesourcing users’
35
frequency and post-pandemic willingness to use these services. Those that thought that COVID-19 was very
36
risky expressed a greater willingness to resume ridesourcing service use, but only with performing their self-
37
imposed health measures (masking, sanitizing) AND/OR if the service intended on implementing some sort of
38
COVID-19 countermeasures as well. If they perceive shared vehicles as clean and safe, they are more likely to
39
resume pre-pandemic ridesourcing usage. This is likely because consumers with high-risk perception are more
40

28
sensitive to the uncertainty and exposure risks of traveling with strangers. To improve the post-pandemic
1
willingness of passengers with high-risk perception, ridesourcing services should provide a clear, convenient,
2
and timely communication platform for riders to reassure them that hygiene measures are being implemented
3
and drivers have been tested for COVID-19 recently and are not sick themselves. Additionally, ridesourcing
4
service providers should enhance driver assessments and training, including mask covering, cleaning between
5
passengers, maintaining a clean and tidy vehicle environment (people may simply gave a lower rating in the past
6
but now they are more likely to refuse using the service), and providing satisfactory and safe services to meet
7
passengers’ expectations for service-imposed health measures and rebuild consumer confidence in ridesourcing.
8
On the other hand, those who had high pre-pandemic and current frequency of ridesourcing usage are more likely
9
to use these services as often as before the pandemic. The results demonstrate the significant impact of past and
10
current trip experiences on people’s willingness to use the same services. Considering the influence of pre-
11
pandemic and current frequency on post-pandemic willingness to use ridesourcing, service providers should pay
12
attention to developing distinct measures for different types of users to attract low-frequency users and strengthen
13
relationships with regular customers. This plays a significant role in re-establishing a preference and/or habit for
14
using ridesourcing services. Additional studies are needed to better design these measures.
15
16
5.2. Risk perception and attitudes of COVID-19 pandemic
17
Based on the estimation results, risk perception plays a crucial role in ridesourcing usage, even after the end
18
of the pandemic, through factors related to self-imposed prevention. People's attitudes towards the spread of the
19
pandemic and the consequences of being diagnosed can go hand-in-hand. The majority of respondents (73.7%)
20
are concerned about the risk and consequences of exposure and infection and are taking precautions to avoid the
21
risk. Interestingly, the results reveal that risk perception does not have a significant and direct negative impact
22
on the willingness to use ridesourcing services after the end of the pandemic. However, it does significantly
23
impact self-imposed sanitary habits related to ridesourcing, such as wearing masks, frequent hand washing, or
24
refusing to ride with others not wearing masks. These habits, in turn, affect the intention to use ridesourcing
25
services in the future. It can be concluded that respondents are more likely to use self-imposed preventative
26
measures to protect themselves if there is an increased perceived risk of contracting the virus in transit systems,
27
especially since the threat of COVID-19 has not completely disappeared.
28
The study results show that people high in risk perception that are taking precautions now do not necessarily
29
think they will need to take precautions themselves in the future, but they will expect ridesourcing service
30
providers to continue COVID-19 countermeasures. This may be due to passengers' preference to return to the
31
liberty and comfort of pre-pandemic normal as soon as possible but still want to enjoy the improved cleanness
32
of the ridesourcing services throughout the pandemic.
33
34

29
Fig. 7. Cross-tabulation analysis of (a) post-pandemic willingness to use ridesourcing services and pre-pandemic frequency; (b) post-pandemic
willingness to use ridesourcing services and current frequency; (c) change of ridesourcing usage and post-pandemic willingness to use ridesourcing; (d)
change of ridesourcing usage and pre-pandemic frequency.

30
Table 8 The indirect effects of route analyzing
1
Path Coefficients
Std. Error
T Statistics
p-Value
VA -> CSHM
0.013
0.004
3.467
0.001
VA -> CSE
0.010
0.003
2.931
0.003
VA -> CF
0.001
0.002
0.571
0.568
VA -> PE-Self
0.006
0.002
3.406
0.001
VA -> PE-Service
0.005
0.002
3.372
0.001
VA -> Boosted
-
-
-
-
VA -> PW
0.005
0.003
1.732
0.083
CSHM -> CSE
-
-
-
-
CSHM -> CF
-0.026
0.009
2.953
0.003
CSHM -> PE-Self
-
-
-
-
CSHM -> PE-Service
0.128
0.014
9.088
0.000
CSHM -> PW
0.031
0.011
2.857
0.004
CSE -> CF
-
-
-
-
CSE -> PE-Service
-
-
-
-
CSE -> PW
0.010
0.007
1.545
0.122
CF -> PW
-
-
-
-
PE-Self -> PW
-
-
-
-
PE-Service -> PW
-
-
-
-
PF -> CF
-
-
-
-
PF -> PW
0.047
0.020
2.301
0.021
RP -> CSHM
-0.003
0.002
1.361
0.173
RP -> CSE
0.124
0.013
9.844
0.000
RP -> CF
-0.019
0.006
2.973
0.003
RP -> PE-Self
0.118
0.012
9.704
0.000
RP -> PE-Service
0.157
0.014
11.285
0.000
RP -> Boosted
-
-
-
-
RP -> PW
-0.018
0.007
2.750
0.006
Boosted -> CSHM
-
-
-
-
Boosted -> CSE
0.042
0.009
4.472
0.000
Boosted -> CF
-0.003
0.002
1.662
0.097
Boosted -> PE-Self
0.040
0.009
4.435
0.000
Boosted -> PE-Service
0.036
0.008
4.502
0.000
Boosted -> PW
0.004
0.002
2.285
0.022
2
The empirical results indicate that people with a more positive attitude towards the development and
3
effectiveness of COVID-19 vaccines are more likely to receive a booster shot. Interestingly, those who are
4
concerned about the risks of exposure and infection from the COVID-19 virus may not necessarily be receptive
5
to receiving booster shots, even if they have a high level of recognition regarding the effectiveness and safety of
6
the vaccine. It cannot be inferred that those who perceive more risk from COVID-19 are more likely to receive
7
a booster shot. The reason for this may be that (i) the effectiveness of the regular COVID-19 vaccine has already
8
been significantly verified in reducing the risk of contracting the virus by the time the booster shot becomes
9
available, and (ii) they may feel that they may have limit exposure to other people and feel the booster is not
10
warranted.
11
However, those who are fully vaccinated, including those who have received boosters, prefer to have access
12

31
to more cautious and thoughtful preventative measures during the COVID-19 pandemic. They also have a greater
1
intention to continue taking these measures to reduce potential risks when using ridesourcing services after the
2
end of the pandemic. These findings may be because fully vaccinated passengers are more concerned and
3
cautious of exposure to public spaces and risks when riding with others, while those who are not vaccinated may
4
not concern about the threats caused by the pandemic. As a result, the behavioral characteristics and intentions
5
of fully vaccinated passengers change as follows: (i) they have higher expectations for traveling in a safe and
6
clean environment; (ii) they are more likely to implement hygiene measures to protect themselves; (iii)
7
individuals demonstrate greater concern and caution when deciding whether to utilize ride-sourcing services as
8
the perceived risk of contracting COVID-19 in a confined space with drivers and other passengers is perceived
9
to be heightened; and (iv) they have a more lasting impact on their health awareness and habits related to daily
10
transit. Additionally, the results also confirm that vaccination significantly and indirectly affects passengers' post-
11
pandemic expectations for health measures and their willingness to use ridesourcing services through the current
12
health measures provided by themselves and ridesourcing operators. Those that are vaccinated expect to take
13
precautions and have ridesourcing service providers do so as well. Doing so can potentially eliminate public
14
concern and encouraging people to travel by ride-sharing services more after the pandemic.
15
16
5.3. Self-imposed health-related measures and expectations for ridesourcing services
17
This study provides several implications for promoting the further development of the ridesourcing services
18
market. Nearly all the respondents (89.6%) have expressed a desire to engage in preventive behavior, such as
19
hand washing and cleaning seats, more often when traveling due to the risks of exposure to public space and
20
interaction with others during the COVID-19 pandemic. The level of sanitization service provided by
21
ridesourcing companies has a significant impact on passengers' willingness to use these services after the
22
pandemic. According to the model results, people who have higher expectations for cleanliness related to
23
vehicles are less likely to reduce their ridesourcing usage and would like to use these services as often as before
24
the pandemic when the threat of COVID-19 has passed. The tendency of passengers with greater risk perception
25
to consider using preventative measures implies that providing more hygiene-related services related to
26
ridesourcing may help build public confidence in the safety of these services, which is an important way to
27
realize the recovery of ridesourcing ridership both during and after the pandemic.
28
As the above results suggest, people still tend to take precautions and retain current preventive measures when
29
using ridesourcing, even after the COVID-19 pandemic. This may be due to the fact that, both during and after
30
the pandemic, passengers view ridesourcing as a means of satisfying their mobility needs as an alternative to
31
public transit. Therefore, it is essential for passengers to reduce potential risks and ensure safety while traveling
32
with ridesourcing by implementing preventative measures such as providing details of the last sanitization,
33
physical barriers, and drivers' health-related information. The urgency of reducing potential risks or other health
34
threats should be considered as soon as possible to increase users' willingness to use ridesourcing services and
35
rebuild their confidence in them.
36
These findings indicate that risk sensitivity may lead passengers to choose modes with lower risk of infection
37
and implement preventative measures by themselves. Furthermore, people's post-pandemic expectation for a
38
comfortable and clean environment when using ridesourcing services highlights the role of providing a low-risk
39
environment in the shared mobility industry’s recovery. These factors have the potential to impact health habits
40
and travel behavior both during and after the pandemic.
41

32
1
5.4. Sociodemographic and travel behavior characteristics
2
Female respondents in our sample reported a greater likelihood of perceiving health and safety threats and
3
tended to implement preventive measures to minimize their exposure to risks compared to their male
4
counterparts. Similarly, female participants have higher post-pandemic expectations of implementing preventive
5
measures by themselves and service providers. However, a subset of female respondents reported a reduced
6
willingness to continue their ridesourcing usage as often as before COVID-19, even after the pandemic has ended.
7
It can be concluded that female respondents' concerns about contracting the virus appear to be deterring them
8
from using ridesourcing services, and they are more concerned about the cleanliness of the vehicles and travel
9
safety. These findings also provide insights into the main factors that individuals are more concerned about, and
10
they would be more willing to use ridesourcing services both during the COVID-19 recovery period and after
11
the COVID-19 if additional health and safety measures were implemented.
12
Additionally, age is found to be significant in most factors, including attitudes toward vaccines, risk perception,
13
and ridesourcing usage both before and after the pandemic. Millennials and younger people have more positive
14
attitudes toward vaccines and less perceived risks compared to older respondents in this Chinese sample. This
15
may be because young people may receive information related to the COVID-19 through different sources and
16
have a more comprehensive understanding of the severity of the outbreak and the effectiveness of vaccines. On
17
the other hand, their overall perceptions of new things are more open to living with potential risks, as supported
18
by the results showing higher frequency of ridesourcing usage both before and after the pandemic. In contrast,
19
members of older generations are more afraid of the consequences of the COVID-19, as they are more likely to
20
be vulnerable to the virus due to underlying health conditions and less resistance. However, older people with
21
limited access to travel reduced their outdoor trips and are less exposed to potential contracting risks, as
22
evidenced by the trip reduction after the pandemic. Therefore, the importance of retaining precautions provided
23
by self-imposed and ridesourcing services after the outbreak may be influenced by the accessibility and
24
frequency of travel for older people.
25
Regarding education and income, people with a college degree or above are more likely to endorse the
26
development and effectiveness of COVID-19 vaccines. Those with relatively high annual self-imposed pretax
27
income (more than 75,000 RMB) and a relatively fixed income per month are more likely to use ridesourcing
28
services frequently in commuting before the pandemic and have a more frequent ridesourcing usage currently.
29
Moreover, people with higher income also tend to avoid the risks of exposure and infection by implementing
30
preventive measures both during the COVID-19 period and after the pandemic and are likely to expect travelling
31
in a relatively clean and safe environment in a vehicle.
32
Finally, this study shows that current ridesourcing usage is significantly associated with willingness to use
33
after the COVID-19 pandemic. Passengers who had more frequent ridesourcing trips in their daily lives both
34
before and after the pandemic tend to maintain ridesourcing usage as often as before once the outbreak has ended.
35
The results may be consistent with the fact that past travel behavior has more lasting impacts on people’s
36
willingness to maintain such behavior in the future.
37
38
6. Conclusions and policy implications
39
In this study, we investigated how the COVID-19 pandemic has affected the usage of ridesourcing services
40

33
and people's willingness to use them in the future. We conducted a web-based survey in China that focused on
1
factors such as risk perception, attitudes towards preventative measures, travel behavior, and sociodemographic
2
characteristics. Through our analysis, we were able to assess and quantify the pandemic's impact on various
3
dimensions, and we used a MIMIC model structure to explore both the direct and indirect effects of COVID-19
4
on residents' sociodemographic information. These results are crucial for understanding the key factors in the
5
development of ridesourcing in the later days of the pandemic and providing policy recommendations.
6
The results of this study have the potential to contribute to academia, policymakers, and the ridesourcing
7
industry, as well as the shared mobility sector, in various ways. Firstly, the study highlights the importance of
8
potential riders’ risk perception of health and safety, both during and after the COVID-19 pandemic. Given the
9
altered context of the pandemic, the risk of disease transmission through ridesharing has led to public concern
10
about the safety of ridesourcing services. This concern is associated with passengers' willingness to use shared
11
transportation with strangers. To restore public confidence in shared vehicles, ridesourcing service providers
12
may consider providing more details about service preventative measures and implementing stricter cleaning
13
management to reduce the potential risk of disease transmission in small vehicles. Policymakers may also
14
consider policies such as (i) updating information on the consequences of COVID-19 for the public to increase
15
awareness and reduce concerns, (ii) developing more stringent sanitization policies related to ridesourcing transit
16
to improve services in this industry, and (iii) implementing other measures to reduce the risk of disease
17
transmission in public spaces.
18
Secondly, the study provides insights into the impact of preventive measures on ridesourcing usage and
19
willingness to use shared vehicles during and after the pandemic. Most existing studies have not considered the
20
effects of preventive measures and vaccines, especially the intention to maintain current precautions once the
21
outbreak has ended. The usage of shared vehicles has decreased during the pandemic due to the higher perceived
22
risks, and more passengers have started implementing sanitized measures such as wearing masks and frequent
23
sterilization. In addition, people expect ridesourcing service providers to provide a relatively safe and clean
24
environment, whether the outbreak is over or not, and vaccination also has a significant and positive impact. For
25
ridesourcing service providers, it is important to develop or improve prevention-related services to encourage
26
passengers to alleviate public concerns and choose ridesourcing for daily mobility. This includes stricter checks
27
of drivers’ health information, timely disclosure of cleaning details, and providing a fresh and clean environment
28
in shared vehicles. Governments could also consider funding research to improve the effectiveness of vaccines,
29
promoting their benefits to increase vaccination rates, encouraging the public to get fully vaccinated and keep
30
current with boosters that address new variants, and formulating effective disinfection standards for the
31
ridesourcing industry.
32
Lastly, this study sheds light on the impacts of sociodemographic and travel behavior characteristics on
33
attitudes towards COVID-19, intentions to use prevention measures, and willingness to use ridesourcing services
34
after the pandemic. Policymakers can use these insights to develop regulations and support policies for
35
disadvantaged groups, and to create convenient and extensive infrastructure for daily trips to increase social
36
participation among vulnerable or older groups and reduce resistance to emerging modes of transportation.
37
People who use shared vehicles more frequently may be incentivized by high-quality transport services,
38
customized offerings, and occasional discounts. Meanwhile, ridesourcing providers should focus on identifying
39
underlying causes of low market share among people who use shared vehicles less frequently or not at all, such
40
as concerns about accident liability, ease of use, and perceived risks.
41

34
There are several limitations that future studies could address. Firstly, data collection was performed via a
1
web-based survey. This methodology inherently carries the limitation of under-representing groups with limited
2
access or familiarity to technology, such as the elderly and individuals without tertiary education, a constraint
3
well-documented in previous studies (Tang et al., 2023; Guo et al., 2022b). This could potentially affect our
4
understanding of these groups' responses to ride-sourcing services during and post COVID-19, leading to their
5
under-representation in policy recommendations. Future research can target these populations to better
6
understand their behavior and responses in the post-COVID-19 era. Future research could incorporate additional
7
data types, such as travel mobility and mobile data, to further investigate the topic. Secondly, this study utilized
8
a MIMIC model structure to examine the effects of COVID-19 on ridesourcing usage and willingness to use it.
9
In future studies, a mode choice framework or analysis of changes in ridesourcing usage could be used to explore
10
how risk perception and other attitudinal factors (e.g., attitudes towards regulation effectiveness) impact different
11
characteristics of the transportation network. Thirdly, several studies (Mohamed, 2022; Cipolletta et al., 2022;
12
Lo et al., 2022) have shown that personality characteristics have significant effects on self-imposed risk
13
perception, and these could be introduced to differentiate between various groups observed or potential
14
heterogeneities. Lastly, the COVID-19 situation is constantly evolving, and related policies change quickly. The
15
different pandemic periods have distinct characteristics, and future research could investigate how various
16
factors impact the transportation system in cases of varying severity by dividing the post-outbreak period into
17
different segments.
18
19
CRediT authorship contribution statement
20
Xinghua Li: Conceptualization, Funding acquisition, Resources, Supervision, Project administration, Writing
21
– review & editing. Yueyi Yang: Conceptualization, Data curation, Investigation, Methodology, Formal analysis,
22
Software, Validation, Visualization, Writing – original draft. Yuntao Guo: Conceptualization, Data curation,
23
Supervision, Methodology, Writing – review & editing. Dustin Souders: Writing – review & editing. Jian Li:
24
Conceptualization, Supervision, Writing – review & editing.
25
26
Declaration of Competing Interest
27
The authors declare that they have no known competing financial interests or self-imposed relationships that
28
could have appeared to influence the work reported in this paper.
29
30
Acknowledgments
31
This study was supported by the National Natural Science Foundation of China (52272322) and the
32
Fundamental Research Funds for the Central Universities (22120220124, 2022-5-YB-02, and 2023-4-YB-04).
33
34
35

35
Appendix
1
Table 1 Comparison of models with binary sociodemographic variables and models with ordered
2
sociodemographic variables (AIC and BIC values)
3
Variable
BIC
AIC
Binary
variable
Ordered
variable
Binary
variable
Ordered
variable
Attitudes toward Vaccines
52.793
61.646
12.401
21.254
Current Self-imposed Health Measures
-171.962
-162.882
-223.894
-194.814
Current service-imposed expectation
-280.202
-269.930
-337.904
-327.632
Current Frequency
-284.403
-243.980
-353.646
-313.223
Post-pandemic Expectation (service)
-184.076
-179.397
-236.008
-231.329
Post-pandemic Expectation (self)
-192.674
-168.302
-250.376
-226.004
Pre-pandemic Frequency
-52.631
-40.101
-93.023
-80.493
Risk Perception
78.053
90.022
37.661
49.630
Boosted
56.284
66.579
4.352
14.647
Post-pandemic Willingness
-25.257
-12.336
-100.270
-87.349
4
5

36
References
1
Abdelrahman, M. (2022). Personality Traits, Risk Perception, and Protective Behaviors of Arab Residents of
2
Qatar During the COVID-19 Pandemic. Int J Ment Health Addiction, 20, 237–248.
3
https://doi.org/10.1007/s11469-020-00352-7
4
Abdullah, M., Ali, N., Hussain, S. A., Aslam, A. B., & Javid, M. A. (2021). Measuring changes in travel behavior
5
pattern due to COVID-19 in a developing country: A case study of Pakistan. Transport Policy, 108, 21-33.
6
https://doi.org/10.1016/j.tranpol.2021.04.023
7
Abdullah, M., Dias, C., Muley, D., & Shahin, M. (2020). Exploring the impacts of COVID-19 on travel behavior
8
and mode preferences. Transportation Research Interdisciplinary Perspectives, 8, 100255.
9
https://doi.org/10.1016/j.trip.2020.100255
10
Adane, M., Ademas, A., & Kloos, H. (2022). Knowledge, attitudes, and perceptions of COVID-19 vaccine and
11
refusal to receive COVID-19 vaccine among healthcare workers in northeastern Ethiopia. BMC Public
12
Health, 22(1), 128. https://doi.org/10.1186/s12889-021-12362-8
13
Aguilera-García, Á., Gomez, J., Velázquez, G., & Vassallo, J. M. (2022). Ridesourcing vs. traditional taxi
14
services: Understanding users’ choices and preferences in Spain. Transportation Research Part A: Policy
15
and Practice, 155, 161-178. https://doi.org/10.1016/j.tra.2021.11.002
16
Alemi, F., Circella, G., Handy, S., & Mokhtarian, P. (2018). What influences travelers to use Uber? Exploring
17
the factors affecting the adoption of on-demand ride services in California. Travel Behaviour and Society,
18
13, 88-104. https://doi.org/10.1016/j.tbs.2018.06.002
19
Allen, J., Muñoz, J. C., & Ortúzar, J. d. D. (2018). Modelling service-specific and global transit satisfaction
20
under travel and user heterogeneity. Transportation Research Part A: Policy and Practice, 113, 509-528.
21
https://doi.org/10.1016/j.tra.2018.05.009
22
Almokdad, E., Kiatkawsin, K., & Lee, C. H. (2023). Antecedents of Booster Vaccine Intention for Domestic and
23
International Travel. Sustainability, 15(8), 6399. https://doi.org/10.3390/su15086399
24
Alonso-Almeida, M. D. M. (2022). To use or not use car sharing mobility in the ongoing COVID-19 pandemic?
25
Identifying sharing mobility behaviour in times of crisis. International Journal of Environmental Research
26
and Public Health, 19(5), 3127. https://doi.org/10.3390/ijerph19053127
27
Batomen, B., Cloutier, M. S., Palm, M., Widener, M., Farber, S., Bondy, S. J., & Di Ruggiero, E. (2023). Frequent
28
public transit users views and attitudes toward cycling in Canada in the context of the COVID-19 pandemic.
29
Multimodal Transportation, 2(2), 100067. https://doi.org/10.1016/j.multra.2022.100067
30
Bowen, N. K., & Guo, S., 2011. Structural equation modeling: Oxford University Press. New York, NY.
31
Bruine de Bruin, W. (2021). Age differences in COVID-19 risk perceptions and mental health: Evidence from a
32
national US survey conducted in March 2020. The Journals of Gerontology: Series B, 76(2), e24-e29.
33
https://doi.org/10.1093/geronb/gbaa074
34
Bucsky, Péter. (2020). Modal share changes due to COVID-19: The case of Budapest. Transportation Research
35
Interdisciplinary Perspectives, 8, 100141. https://doi.org/10.1016/j.trip.2020.100141
36
China Release. (2021). National Health Commission of the People’s Republic of China: Complete the entire
37

37
vaccination of the new coronavirus vaccine more than 1.1 billion people. Retrieved from
1
http://news.china.com.cn/2021-11/30/content_77902414.html. Accessed October 10, 2022
2
Chiu, C. M., & Wang, E. T. (2008). Understanding Web-based learning continuance intention: The role of
3
subjective task value. Information and management, 45(3), 194-201.
4
https://doi.org/10.1016/j.im.2008.02.003
5
Cipolletta S., Andreghetti G. R., Mioni G. (2022). Risk Perception towards COVID-19: A Systematic Review
6
and Qualitative Synthesis. International Journal of Environmental Research and Public Health, 19(8):4649.
7
https://doi.org/10.3390/ijerph19084649
8
Currie, G., Jain, T., & Aston, L. (2021). Evidence of a post-COVID change in travel behaviour – Self-reported
9
expectations of commuting in Melbourne. Transportation Research Part A: Policy and Practice, 153, 218-
10
234. https://doi.org/10.1016/j.tra.2021.09.009
11
De Haas, M., Faber, R., & Hamersma, M. (2020). How COVID-19 and the Dutch ‘intelligent lockdown’ change
12
activities, work and travel behaviour: Evidence from longitudinal data in the Netherlands. Transportation
13
Research Interdisciplinary Perspectives, 6, 100150. https://doi.org/10.1016/j.trip.2020.100150
14
De Vos, J. (2020). The effect of COVID-19 and subsequent social distancing on travel behavior. Transportation
15
Research Interdisciplinary Perspectives, 5, 100121. https://doi.org/10.1016/j.trip.2020.100121
16
Delbosc, A., & Nakanishi, H. (2017). A life course perspective on the travel of Australian millennials.
17
Transportation Research Part A: Policy and Practice, 104, 319-336.
18
https://doi.org/10.1016/j.tra.2017.03.014
19
Ding, Y., Du, X., Li, Q., Zhang, M., Zhang, Q., Tan, X., & Liu, Q. (2020). Risk perception of coronavirus disease
20
2019 (COVID-19) and its related factors among college students in China during quarantine. PloS one,
21
15(8), e0237626. https://doi.org/10.1371/journal.pone.0237626
22
Downey, L., Fonzone, A., Fountas, G., & Semple, T. (2022). The impact of COVID-19 on future public transport
23
use in Scotland. Transportation Research Part A: Policy and Practice, 163, 338-352.
24
https://doi.org/10.1016/j.tra.2022.06.005
25
Fall, A. N. (2022). Analysis of social acceptability in the implementation of a congestion pricing area in Senegal.
26
Multimodal Transportation, 1(4), 100036. https://doi.org/10.1016/j.multra.2022.100036
27
Fan, X., Lu, J., Qiu, M., & Xiao, X. (2022). Changes in travel behaviors and intentions during the COVID-19
28
pandemic and recovery period: A case study of China. Journal of Outdoor Recreation and Tourism, 100522.
29
https://doi.org/10.1016/j.jort.2022.100522
30
Faruque, S., Fonzone, A., & Fountas, G. (2022). Explaining expected non-shared and shared use of driverless
31
cars in Edinburgh. Transportation research procedia, 62, 286-293.
32
https://doi.org/10.1016/j.trpro.2022.02.036
33
Feng, F., Li, X., Guo, Y., & Cheng, C. (2023). Understanding factors that impact ridesourcing service usage
34
frequency: a case study in Shanghai. Transportation Planning and Technology, 46(4), 462-481.
35
https://doi.org/10.1080/03081060.2023.2194875
36
Garikapati, V. M., Pendyala, R. M., Morris, E. A., Mokhtarian, P. L., & McDonald, N. (2016). Activity patterns,
37

38
time use, and travel of millennials: a generation in transition?. Transport Reviews, 36(5), 558-584.
1
https://doi.org/10.1080/01441647.2016.1197337
2
Ghaffar, A., Mitra, S., & Hyland, M. (2020). Modeling determinants of ridesourcing usage: A census tract-level
3
analysis of Chicago. Transportation Research Part C: Emerging Technologies, 119, 102769.
4
https://doi.org/10.1016/j.trc.2020.102769
5
Guo, Y., Qian, X., Lei, T., Guo, S., & Gong, L. (2022a). Modeling the preference of electric shared mobility
6
drivers in choosing charging stations. Transportation Research Part D: Transport and Environment, 110,
7
103399. https://doi.org/10.1016/j.trd.2022.103399
8
Guo, Y., Peeta, S., Agrawal, S., & Benedyk, I. (2022b). Impacts of Pokémon GO on route and mode choice
9
decisions: exploring the potential for integrating augmented reality, gamification, and social components in
10
mobile apps to influence travel decisions. Transportation, 1-50. https://doi.org/10.1007/s11116-021-10181-
11
9
12
Guo, Y., Souders, D., Labi, S., Peeta, S., Benedyk, I., & Li, Y. (2021). Paving the way for autonomous Vehicles:
13
Understanding autonomous vehicle adoption and vehicle fuel choice under user
14
heterogeneity. Transportation Research Part A: Policy and Practice, 154, 364-398.
15
https://doi.org/10.1016/j.tra.2021.10.018
16
Gursoy, D., & Chi, C. G. (2021). Celebrating 30 years of excellence amid the COVID-19 pandemic — An update
17
on the effects of COVID-19 pandemic and COVID-19 vaccines on hospitality industry: overview of the
18
current situation and a research agenda. Journal of Hospitality Marketing & Management, 30(3), 277-281.
19
https://doi.org/10.1080/19368623.2021.1902052
20
Gursoy, D., Can, A. S., Williams, N., & Ekinci, Y. (2021). Evolving impacts of COVID-19 vaccination intentions
21
on travel intentions. The Service Industries Journal, 41(11-12), 719-733.
22
https://doi.org/10.1080/02642069.2021.1938555
23
Gursoy, D., Ekinci, Y., Can, A. S., & Murray, J. C. (2022). Effectiveness of message framing in changing
24
COVID-19 vaccination intentions: Moderating role of travel desire. Tourism Management, 90, 104468.
25
https://doi.org/10.1016/j.tourman.2021.104468
26
Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to use and how to report the results of PLS-
27
SEM. European Business Review, 31(1), 2-24. https://doi.org/10.1108/EBR-11-2018-0203
28
Hasselwander, M., Nieland, S., Dematera-Contreras, K., & Goletz, M. (2023). MaaS for the masses: Potential
29
transit accessibility gains and required policies under Mobility-as-a-Service. Multimodal Transportation,
30
2(3), 100086. https://doi.org/10.1016/j.multra.2023.100086
31
He, L., Li, J., & Sun, J. (2021). How to promote sustainable travel behavior in the post COVID-19 period: A
32
perspective from customized bus services. International Journal of Transportation Science and Technology.
33
https://doi.org/10.1016/j.ijtst.2021.11.001
34
He, L., Li, J., Guo, Y., & Sun, J. (2023). Commuters’ intention to choose customized bus during COVID-19
35
pandemic: Insights from a two-phase comparative analysis. Travel Behaviour and Society, 33, 100627.
36
https://doi.org/10.1016/j.tbs.2023.100627
37

39
He, S., Chen, S., Kong, L., & Liu, W. (2021). Analysis of risk perceptions and related factors concerning COVID-
1
19 epidemic in Chongqing, China. Journal of Community Health, 46(2), 278-285.
2
https://doi.org/10.1007/s10900-020-00870-4
3
Hotle, S., Murray-Tuite, P., & Singh, K. (2020). Influenza risk perception and travel-related health protection
4
behavior in the US: Insights for the aftermath of the COVID-19 outbreak. Transportation Research
5
Interdisciplinary Perspectives, 5, 100127. https://doi.org/10.1016/j.trip.2020.100127
6
Hsieh, H. S., & Hsia, H. C. (2022). Can continued anti-epidemic measures help post-COVID-19 public transport
7
recovery? Evidence from Taiwan. Journal of Transport & Health, 26, 101392.
8
https://doi.org/10.1016/j.jth.2022.101392
9
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional
10
criteria versus new alternatives. Structural equation modeling: a multidisciplinary journal, 6(1), 1-55.
11
https://doi.org/10.1080/10705519909540118
12
Ivanov, D. (2020). Predicting the impacts of epidemic outbreaks on global supply chains: A simulation-based
13
analysis on the coronavirus outbreak (COVID-19/SARS-CoV-2) case. Transportation Research Part E:
14
Logistics and Transportation Review, 136, 101922. https://doi.org/10.1016/j.tre.2020.101922
15
Jia, W., & Chen, T. D. (2023). Investigating heterogeneous preferences for plug-in electric vehicles: Policy
16
implications from different choice models. Transportation Research Part A: Policy and Practice, 173,
17
103693. https://doi.org/10.1016/j.tra.2023.103693
18
Jin, S. T., Kong, H., Wu, R., & Sui, D. Z. (2018). Ridesourcing, the sharing economy, and the future of cities.
19
Cities, 76, 96-104. https://doi.org/10.1016/j.cities.2018.01.012
20
Jöreskog, K. G., & Goldberger, A. S. (1975). Estimation of a Model with Multiple Indicators and Multiple Causes
21
of a Single Latent Variable. Journal of the American Statistical Association, 70(351a), 631-639.
22
10.1080/01621459.1975.10482485
23
Kaplan, S., Tchetchik, A., Greenberg, D., & Sapir, I. (2022). Transit use reduction following COVID-19: The
24
effect of threat appraisal, proactive coping and institutional trust. Transportation Research Part A: Policy
25
and Practice, 159, 338-356. https://doi.org/10.1016/j.tra.2022.03.008
26
Khaddar, S., & Fatmi, M. R. (2021). COVID-19: Are you satisfied with traveling during the pandemic?
27
Transportation Research Interdisciplinary Perspectives, 9, 100292.
28
https://doi.org/10.1016/j.trip.2020.100292
29
Kline, R. B. (2015). Principles and practice of structural equation modeling. Guilford publications.
30
Koh, T. J. W., Ling, A. H. Z., Chiang, C. L. L., Lee, G. S. J., Tay, H. S. E., & Yi, H. (2021). Attitudes towards
31
COVID-19 precautionary measures and willingness to work during an outbreak among medical students in
32
Singapore: a mixed-methods study. BMC Medical Education, 21(1), 317. https://doi.org/10.1186/s12909-
33
021-02762-0
34
Kungwola, K., Trerattanaset, P., & Liudmila Guzikova. (2022). Airline Safety measures to prevent the COVID-
35
19 pandemic that affect the confidence of passenger’s decision making to travel with domestic low-cost
36
airlines during the pandemic. Transportation Research Procedia, 63, 2485-2495.
37

40
https://doi.org/10.1016/j.trpro.2022.06.285.
1
Li, X., Yang, Z., Qian, X., Guo, Y., & Yang, C. (2023). Investigating the Impacts of Property Walking
2
Accessibility on Housing Affordability and Equity: Evidence from Shanghai, China. Journal of Urban
3
Planning and Development, 149(4), 05023035. https://doi.org/10.1061/JUPDDM.UPENG-4197
4
Li, H., Zhang, Y., Zhu, M., & Ren, G. (2021). Impacts of COVID-19 on the usage of public bicycle share in
5
London. Transportation Research Part A: Policy and Practice, 150, 140-155.
6
https://doi.org/10.1016%2Fj.tra.2021.06.010
7
Lingmou INSIGHT. (2022). China Internet Taxi Market Research Insights Report 2022. Retrieved from
8
https://www.sgpjbg.com/baogao/98713.html. Accessed July 11 2023.
9
Liu, P. L. (2021). COVID-19 information on social media and preventive behaviors: Managing the pandemic
10
through self-imposed responsibility. Social Science & Medicine, 277, 113928.
11
https://doi.org/10.1016/j.socscimed.2021.113928
12
Liu, X., Kortoçi, P., Motlagh, N. H., Nurmi, P., & Tarkoma, S. (2022). A survey of COVID-19 in public
13
transportation: Transmission risk, mitigation and prevention. Multimodal Transportation, 1(3), 100030.
14
https://doi.org/10.1016/j.multra.2022.100030
15
Lo Presti S., Mattavelli G., Canessa N., Gianelli C. (2022). Risk perception and behaviour during the COVID-
16
19 pandemic: Predicting variables of compliance with lockdown measures. PLoS ONE, 17(1): e0262319.
17
https://doi.org/10.1371/journal.pone.0262319
18
Loa, P., Hossain, S., Liu, Y., & Habib, K. N. (2022). How has the COVID-19 pandemic affected the use of ride-
19
sourcing services? An empirical evidence-based investigation for the Greater Toronto Area. Transportation
20
Research Part A: Policy Practice, 155, 46-62. https://doi.org/10.1016/j.tra.2021.11.013
21
Lopez, J., Baringer, K., Souders, D. J., (2023). Returning to Ridesharing: Passengers’ Preferred Covid-19
22
Countermeasures. Submitted to Journal of Urban Mobility.
23
Ma, T., Heywood, A., & MacIntyre, C. R. (2021). Travel health seeking behaviours, masks, vaccines and
24
outbreak awareness of Australian Chinese travelers visiting friends and relatives – Implications for control
25
of COVID-19. Infection, Disease & Health, 26(1), 38-47. https://doi.org/10.1016/j.idh.2020.08.007
26
Mallinas, S. R., Maner, J. K., & Ashby Plant, E. (2021). What factors underlie attitudes regarding protective
27
mask use during the COVID-19 pandemic? Personality and Individual Differences, 181, 111038.
28
https://doi.org/10.1016/j.paid.2021.111038
29
Mansilla Dominguez, J. M., Font Jimenez, I., Belzunegui Eraso, A., Pena Otero, D., Diaz Perez, D., & Recio
30
Vivas, A. M. (2020). Risk perception of COVID− 19 community transmission among the Spanish
31
population. International Journal of Environmental Research and Public Health, 17(23), 8967.
32
https://doi.org/10.3390/ijerph17238967
33
Mashrur, S. M., Wang, K., & Habib, K. N. (2022). Will COVID-19 be the end for the public transit? Investigating
34
the impacts of public health crisis on transit mode choice. Transportation Research Part A: Policy and
35
Practice, 164, 352-378. https://doi.org/10.1016/j.tra.2022.08.020
36
Masters, N. B., Zhou, T., Meng, L., Lu, P.-J., Kriss, J. L., Black, C., … & Singleton, J. A. (2022). Geographic
37

41
Heterogeneity in Behavioral and Social Drivers of COVID-19 Vaccination. American Journal of Preventive
1
Medicine. https://doi.org/10.1016/j.amepre.2022.06.016
2
Mogaji, E. (2020). Impact of COVID-19 on transportation in Lagos, Nigeria. Transportation Research
3
Interdisciplinary Perspectives, 6, 100154. https://doi.org/10.1016/j.trip.2020.100154
4
Morar, C., Tiba, A., Jovanovic, T., Valjarević, A., Ripp, M., Vujičić, M. D., ... & Lukić, T. (2022). Supporting
5
tourism by assessing the predictors of COVID-19 vaccination for travel reasons. International Journal of
6
Environmental Research and Public Health, 19(2), 918. https://doi.org/10.3390/ijerph19020918
7
National Bureau of Statistics of the People's Republic of China. (2022). Resident income and consumer spending
8
in 2021 in China. Retrieved from http://www.stats.gov.cn/tjsj/zxfb/202201/t20220117_1826403.html.
9
Accessed June 9 2022.
10
Nayum, A., & Nordfjærn, T. (2021). Predictors of public transport use among university students during the
11
winter: A MIMIC modelling approach. Travel behaviour and society, 22, 236-243.
12
https://doi.org/10.1016/j.tbs.2020.10.005
13
Neuburger, L., & Egger, R. (2021). Travel risk perception and travel behaviour during the COVID-19 pandemic
14
2020: A case study of the DACH region. Current Issues in Tourism, 24(7), 1003-1016.
15
https://doi.org/10.1080/13683500.2020.1803807
16
Nguyen, T. T. P., Nguyen, L. H., Le, H. T., Vu, G. T., Hoang, M. T., Nguyen, D. N., … & Ho, C. S. (2020).
17
Perceptions and attitudes toward COVID-19-related national response measures of Vietnamese:
18
Implications for pandemic prevention and control. Frontiers in Public Health, 8, 589053.
19
https://doi.org/10.3389/fpubh.2020.589053
20
Nguyen-Phuoc, D. Q., Oviedo-Trespalacios, O., Nguyen, M. H., Dinh, M. T. T., & Su, D. N. (2022). Intentions
21
to use ride-sourcing services in Vietnam: What happens after three months without COVID-19 infections?
22
Cities, 126, 103691. https://doi.org/10.1016/j.cities.2022.103691
23
Nguyen-Phuoc, D. Q., Oviedo-Trespalacios, O., Vo, N. S., Thi Le, P., & Van Nguyen, T. (2021). How does
24
perceived risk affect passenger satisfaction and loyalty towards ride-sourcing services? Transportation
25
Research Part D: Transport and Environment, 97, 102921. https://doi.org/10.1016/j.trd.2021.102921
26
Park, K., Chamberlain, B., Song, Z., Esfahani, H. N., Sheen, J., Larsen, T., … & Christensen, K. (2022). A double
27
jeopardy: COVID-19 impacts on the travel behavior and community living of people with disabilities.
28
Transportation Research Part A: Policy and Practice, 156, 24-35. https://doi.org/10.1016/j.tra.2021.12.008
29
Paul, E., Steptoe, A., & Fancourt, D. (2021). Attitudes towards vaccines and intention to vaccinate against
30
COVID-19: Implications for public health communications. The Lancet Regional Health–Europe, 1.
31
https://doi.org/10.1016/j.lanepe.2020.100012
32
Petter, S., Straub, D., & Rai, A., 2007. Specifying Formative Constructs in Information Systems Research. Mis
33
Quarterly, 31(4), 623-656. https://doi.org/10.2307/25148814
34
Rahimi, E., Shabanpour, R., Shamshiripour, A., & Mohammadian, A. (2021). Perceived risk of using shared
35
mobility services during the COVID-19 pandemic. Transportation Research Part F: Traffic Psychology
36
and Behaviour, 81, 271-281. https://doi.org/10.1016/j.trf.2021.06.012
37

42
Ren, X., Chen, Z., Liu, C., Dan, T., Wu, J., & Wang, F. (2023). Are vehicle on-demand and shared services a
1
favorable solution for the first and last-mile mobility: Evidence from China. Travel Behaviour and
2
Society, 31, 386-398. https://doi.org/10.1016/j.tbs.2023.01.008
3
Semple, T., Fountas, G., & Fonzone, A. (2021). Trips for outdoor exercise at different stages of the COVID-19
4
pandemic in Scotland. Journal of Transport & Health, 23, 101280.
5
https://doi.org/10.1016/j.jth.2021.101280
6
Shakibaei, S., De Jong, G. C., Alpkökin, P., & Rashidi, T. H. (2021). Impact of the COVID-19 pandemic on
7
travel behavior in Istanbul: A panel data analysis. Sustainable Cities and Society, 65, 102619.
8
https://doi.org/10.1016/j.scs.2020.102619
9
Shin, S. H., Ji, H., & Lim, H. (2021). Heterogeneity in preventive behaviors during COVID-19: Health risk,
10
economic insecurity, and slanted information. Social Science & Medicine, 278, 113944.
11
https://doi.org/10.1016/j.socscimed.2021.113944
12
Shokouhyar, S., Shokoohyar, S., Sobhani, A., & Gorizi, A. J. (2021). Shared mobility in post-COVID era: New
13
challenges and opportunities. Sustainable Cities and Society, 67, 102714.
14
https://doi.org/10.1016/j.scs.2021.102714
15
Skrondal, A., & Rabe‐Hesketh, S. (2005). Structural equation modeling: categorical variables. Encyclopedia of
16
statistics in behavioral science. https://doi.org/10.1002/0470013192.bsa596
17
Sobieralski, J. B. (2020). COVID-19 and airline employment: Insights from historical uncertainty shocks to the
18
industry. Transportation Research Interdisciplinary Perspectives, 5, 100123.
19
https://doi.org/10.1016/j.trip.2020.100123
20
Su, D. N., Nguyen-Phuoc, D. Q., & Johnson, L. W. (2021). Effects of perceived safety, involvement and
21
perceived service quality on loyalty intention among ride-sourcing passengers. Transportation, 48(1), 369-
22
393. https://doi.org/10.1007/s11116-019-10058-y
23
Tang, T., Guo, Y., Wang, H., Li, X., & Agrawal, S. (2023). Determinants of Helmet Use Intention Among E-
24
Bikers in China: An Application of the Theory of Planned Behavior, the Health Belief Model, and the Locus
25
of Control. Transportation Research Record, 03611981231176290.
26
https://doi.org/10.1177/03611981231176290
27
Tirachini, A. (2020). Ride-hailing, travel behaviour and sustainable mobility: an international
28
review. Transportation, 47, 2011–2047. https://doi.org/10.1007/s11116-019-10070-2
29
Tsinghua University. (2021). Research Report on Travel Platforms in China's First-Tier Cities, 2021. Retrieved
30
from https://www.docin.com/p-2666477115.html. Accessed July 11 2023.
31
Wang, G., Yao, Y., Wang, Y., Gong, J., Meng, Q., Wang, H., Wang, W., Chen, X., & Zhao, Y. (2023).
32
Determinants of COVID-19 vaccination status and hesitancy among older adults in China. Nature Medicine,
33
1-1. https://doi.org/10.1038/s41591-023-02241-7
34
Wang, J., Zhu, H., Lai, X., Zhang, H., Huang, Y., Feng, H., ... & Fang, H. (2022). From COVID-19 vaccination
35
intention to actual vaccine uptake: A longitudinal study among Chinese adults after six months of a national
36
vaccination campaign. Expert Review of Vaccines, 21(3), 385-395.
37

43
https://doi.org/10.1080/14760584.2022.2021076
1
Washington, S., Karlaftis, M., Mannering, F., & Anastasopoulos, P. (2020). Statistical and econometric methods
2
for transportation data analysis. CRC press.
3
Wu, G., Deng, X., & Liu, B. (2022). Managing urban citizens' panic levels and preventive behaviours during
4
COVID-19 with pandemic information released by social media. Cities, 120, 103490.
5
https://doi.org/10.1016/j.cities.2021.103490
6
Yamamura, E., & Tsutsui, Y. (2022). How does the impact of the COVID-19 state of emergency change? An
7
analysis of preventive behaviors and mental health using panel data in Japan. Journal of the Japanese and
8
International Economies, 64, 101194. https://doi.org/10.1016/j.jjie.2022.101194
9
Yang, Y., Cao, M., Cheng, L., Zhai, K., Zhao, X., & De Vos, J. (2021). Exploring the relationship between the
10
COVID-19 pandemic and changes in travel behaviour: A qualitative study. Transportation Research
11
Interdisciplinary Perspectives, 11, 100450. https://doi.org/10.1016/j.trip.2021.100450
12
Yang, Z., Li, X., Guo, Y., & Qian, X. (2023). Understanding active transportation accessibility's impacts on
13
polycentric and monocentric cities' housing price. Research in Transportation Economics, 98, 101282.
14
https://doi.org/10.1016/j.retrec.2023.101282
15
Yen, M. Y., Schwartz, J., Chen, S. Y., King, C. C., Yang, G. Y., & Hsueh, P. R. (2020). Interrupting COVID-19
16
transmission by implementing enhanced traffic control bundling: implications for global prevention and
17
control efforts. Journal of Microbiology, Immunology and Infection, 53(3), 377-380.
18
https://doi.org/10.1016/j.jmii.2020.03.011
19
Yu, J., Xie, N., Zhu, J., Qian, Y., Zheng, S., & Chen, X. (2022). Exploring impacts of COVID-19 on city-wide
20
taxi and ride-sourcing markets: Evidence from Ningbo, China. Transport Policy, 115, 220-238.
21
https://doi.org/10.1016/j.tranpol.2021.11.017
22
Zhang, W., & Liu, L. (2022). Exploring non-users' intention to adopt ride-sharing services: Taking into account
23
increased risks due to the COVID-19 pandemic among other factors. Transportation Research Part A:
24
Policy and Practice, 158, 180-195. https://doi.org/10.1016/j.tra.2022.03.004
25
Zhang, X., Shao, C., Wang, B., & Huang, S. (2022). The impact of COVID-19 on travel mode choice behavior
26
in terms of shared mobility: a case study in Beijing, China. International Journal of Environmental
27
Research and Public Health, 19(12), 7130. https://doi.org/10.3390/ijerph19127130
28
Zhao, P., & Gao, Y. (2022). Public transit travel choice in the post COVID-19 pandemic era: An application of
29
the extended Theory of Planned behavior. Travel Behaviour and Society, 28, 181-195.
30
https://doi.org/10.1016/j.tbs.2022.04.002
31
Zheng, L., Guo, Y., Peeta, S., & Wu, B. (2020). Impacts of information from various sources on the evacuation
32
decision-making process during no-notice evacuations in campus environment. Journal of Transportation
33
Safety & Security, 12(7), 892-923. https://doi.org/10.1080/19439962.2018.1549643
34
35