You are on page 1of 39

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/341092798

The determinants of public acceptance of autonomous vehicles: An


innovation diffusion perspective

Article  in  Journal of Cleaner Production · October 2020


DOI: 10.1016/j.jclepro.2020.121904

CITATIONS READS

47 943

4 authors:

Kum Fai Yuen Yiik Diew Wong


Nanyang Technological University Nanyang Technological University
140 PUBLICATIONS   2,411 CITATIONS    205 PUBLICATIONS   3,487 CITATIONS   

SEE PROFILE SEE PROFILE

Fei Ma Xueqin Wang


Chang'an University Chung-Ang University
48 PUBLICATIONS   731 CITATIONS    60 PUBLICATIONS   1,126 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

RESEARCH STUDY INTO REVIEWING CROSS SECTION OF THE ROADS View project

All content following this page was uploaded by Kum Fai Yuen on 02 November 2020.

The user has requested enhancement of the downloaded file.


The determinants of public acceptance of autonomous
vehicles: An innovation diffusion perspective

The determinants of public acceptance of autonomous vehicles: An


Title
innovation diffusion perspective

Author(s) Yuen, Kum Fai; Wong, Yiik Diew; Ma, Fei & Wang, Xueqin

Yuen, K. F., Wong, Y. D., Ma, F., & Wang, X. (2020). The determinants
Citation of public acceptance of autonomous vehicles: An innovation diffusion
perspective. Journal of Cleaner Production, 121904.

Date 2020

URL https://www.sciencedirect.com/science/article/pii/S095965262031951X

Version Preprint

1
The determinants of public acceptance of autonomous
vehicles: An innovation diffusion perspective
Abstract

The adoption of autonomous vehicles is reported to confer numerous


benefits such as improved safety, comfort, resource (i.e. land and
energy) use, and environmental protection to a society. Public
acceptance is necessary for the widespread adoption of autonomous
vehicles. Drawing on innovation diffusion, perceived value and trust
theories, we developed a theoretical model to identify the latent factors
influencing public acceptance of autonomous vehicles and examined
their interrelationships. A survey questionnaire was designed and
administered to 526 residents in Seoul, Republic of Korea. Thereafter,
structural equation modelling was employed to analyse the survey data.
The results reveal that the influence of the innovation diffusion
attributes (i.e. relative advantage, compatibility, complexity, trialability
and observability) on public acceptance is fully mediated by the public's
perceived value of autonomous vehicles. Further, the effect of perceived
value on public acceptance is partially mediated by the public's trust in
autonomous vehicles. This study differentiates from the literature by
employing a theory-driven approach to explain autonomous vehicles
acceptance. It applies three diverse research paradigms anchoring on
innovation diffusion, customer utility and social psychology. In
addition, it synthesises the theories and provides a coherent explanation
of individuals’ cognition process that leads to acceptance. In sum, this
study enriches existing theoretical research on the acceptance of
autonomous vehicles, offers insights into understanding and managing
technology–human interactions and implicates transport policies and
practices.

Keywords: Autonomous vehicles; Public acceptance; Innovation


diffusion theory; Perceived value theory; Trust theory;
Structural equation modelling

1. Introduction

Autonomous vehicles (AVs) are driverless vehicles that can sense their environment
without human conduction. They are proposed to be an integral part of an intelligent
transportation system that coordinates environmental awareness, planning decisions and
multi-level assisted driving. With the rapid progress in the research and development of
AVs as well as its supporting infrastructures such as the fifth-generation cellular network
technology, it is expected that AVs will occupy 25% of the global private vehicle market
by 2040.

2
Despite several potential drawbacks such as high AV prices, displacement of jobs in
the transportation and logistics sector and social inequalities (Milakis et al., 2017; Wey
and Huang, 2018), the introduction of AVs for use in the private vehicle market can
provide numerous benefits to the society. To name a few, these benefits include improved
traffic safety, urban liveability and users' experience (Becker and Axhausen, 2017;
Fagnant and Kockelman, 2015; Kim et al., 2019; Narayanan et al., 2020). Firstly, AVs
can improve safety by reducing the likelihood of vehicle collision. Previous research has
reported that most traffic accidents are caused by human errors and misjudgement. AVs
can overcome the limitations of human senses and reactions by employing reliable
technologies such as radar, GPS, infrared sensor and computer vision. Secondly, urban
liveability is expected to improve because of the (1) reduced traffic congestions due to
reduced vehicle ownership, (2) reduced transport infrastructure, such as carparks, which
frees up land for commercial or recreational activities, and (3) reduced air pollution (Liu
et al., 2019a), thereby improving public health because of optimised routing and smoother
braking and speed adjustments. Finally, users' experience is expected to improve as they
can partake in more productive activities while commuting to their destination. In
addition, the introduction of AVs can reduce travel time, allow users to enjoy more fuel
savings and create greater travelling opportunities for people experiencing travel
restrictions such as the elderly.

However, nationwide buy-in or acceptance of AVs is crucial for an economy to


successfully benefit from the potential economic, environmental and social benefits of the
deployment of AVs (Becker and Axhausen, 2017). There is presently considerable public
resistance towards the adoption of AVs (Haboucha et al., 2017). AVs represent a huge
perceived technological leap from conventional vehicles which can affect public
acceptance because of issues concerning safety, lack of control, steep learning curve, trust,
cybersecurity and underdeveloped legal and liability rules (König and Neumayr, 2017;
Kyriakidis et al., 2015; Schellekens, 2015).

In recent years, numerous studies have been conducted to analyse the factors
influencing public acceptance of AVs. Most studies have chiefly focused on respondent’s
social-demographic, travel preference and price sensitivity (Bansal et al., 2016; Dixit et
al., 2016; Hohenberger et al., 2016; Liljamo et al., 2018; Menon et al., 2020; Souris et

3
al., 2019). However, consistent with the review of Gkartzonikas and Gkritza (2019), the
following research gaps are noted.

Firstly, very few studies have focused on understanding public acceptance of AVs
using theories. The application of theories can provide a coherent, logical explanation to
public acceptance of AVs. For instance, Panagiotopoulos and Dimitrakopoulos (2018)
employed an extended technology acceptance model to study consumers' intention to use
AVs. Their research led to the discovery of latent, attitudinal variables such as perceived
usefulness and perceived ease of use of AVs that can increase public acceptance of AVs.
Secondly, most studies have only considered the direct relationships between the
explanatory variables and public acceptance of AVs (Chen and Yan, 2019; Nordhoff et
al., 2018). There is presently very limited knowledge about the interrelationships between
the explanatory variables. The present study argues that examining their interrelationships
is equally important as their effects on public acceptance could result in the identification
of mediators which consider not only the direct effects but also the indirect effects of the
explanatory variables on public acceptance of AVs. This can provide a better nomological
understanding of the linkages between the explanatory variables.

To address these gaps in the literature, this study aims to identify and synthesise the
following three theoretical lenses in explaining public acceptance of AVs: innovation
diffusion theory, perceived value theory, and trust theory. These theories are argued to be
appropriate as they address contemporary public issues and concerns and opinion about
AVs (Liljamo et al., 2018). For instance, innovation diffusion theory proposes that an
individual's acceptance of a new product or service is influenced by five factors, namely,
relative advantage, compatibility, complexity, trialability and observability (Wang et al.,
2018). The theory can be applied to this context because AVs can be considered a form
of innovation that revolutionises the transport of passengers or goods (Nieuwenhuijsen et
al., 2018). Next, perceived value theory postulates that the acceptance of AVs can be
improved if AVs offer the best utility (i.e. economical, functional, hedonic and social
benefits) to their users amongst existing alternatives (Cunningham et al., 2019; Yuen et
al., 2018b). Finally, another key issue regarding the acceptance of AVs concerns public
trust in their ability to perform their tasks safely and reliably (Lee and Kolodge, 2020).
Trust theory can be applied to reflect and address this concern. In this context, the theory
posits that trust is ‘a psychological state of the public, comprising the intention to accept

4
vulnerability in a situation involving risk, based on positive expectations of AVs' actions’
(Castelfranchi and Falcone, 2010). The formation of trust can be increased if AVs are
perceived by the public to exhibit competencies such as driving expertise, integrity and
benevolence. Consequently, the formation of trust can lead to public acceptance.

This study differentiates from the literature on AV acceptance by approaching the topic
from a theory-driven approach. It applies three theoretical concepts rooted on diverse
paradigms to explain AV acceptance. The paradigms include innovation diffusion,
customer utility and social psychology. In particular, the applied theories relating to
innovation diffusion and customer utility paradigms have not received much attention
from current researchers. In addition, another originality of this study is that it synthesises
the theories and provide a coherent explanation of individuals’ cognition that leads to AV
acceptance.

The remainder of this study is organised as follows. Firstly, this study introduces a
hybridised theoretical model which identifies the factors influencing public acceptance
and specifies their interrelationships. Thereafter, it describes the applied methodology
which includes the design and administration of a survey questionnaire to citizens in
Seoul, Republic of Korea. Next, using the obtained survey data, this study estimates and
depicts the parameters of the theoretical model using structural equation modelling.
Finally, implications to theories and policies to improve public acceptance of AVs are
drawn.

2. Literature Review

2.1 Theory and model

This study reviews and introduces three behavioural theories to identify the factors
influencing public acceptance of AVs and specifies their interrelationships. This study
proposes that a society would accept AVs if (1) they possess positive characteristics such
as relative advantage, compatibility, reduced complexity, trialability and observability,
which translate to (2) utility to the society, and consequently, (3) the formation of trust.
The theories introduced in this study are elaborated in Table 1

Table 1. A review of behavioural theories and factors influencing public


acceptance of AVs

5
Theory's Innovation diffusion Perceived value Trust theory
characteristics theory theory

Paradigm Innovation acceptance Consumer utility Social psychology

Basic assumption The speed of diffusion A rational user will The acceptance of an
and adoption of an choose or accept a innovation can be
innovation is product that offers the increased by
underpinned by its best utility amongst developing trust
characteristics market alternatives
Representative Relative advantage, Perceived value Trust
constructs compatibility,
complexity, trialability
and observability
Specific The theory can explain The theory can explain The theory can explain
contributions to how improving the how the five how the consistent
model five innovation innovation adoption portrayal of positive
adoption characteristics lead to values from AVs
characteristics of AVs the creation of value, fosters trust, which can
can serve as a basis for which can improve improve public
public acceptance of public acceptance of acceptance of AVs
AVs AVs

Based on the interpretation of the theories, a theoretical model is presented in Fig.


1, which identifies the determinants of public acceptance of AVs and specifies their
interrelationships.

6
Fig. 1. The theoretical model

This study presents three arguments to explain the network of hypotheses presented
in Fig. 1. The first argument is rooted in perceived value theory (Kotler and Armstrong,
2010). In this context, the theory proposes that value is a direct predictor of public
acceptance of AVs. The value comprises four dimensions: economic, functional, hedonic
and social utility (Yuen et al., 2019). Perceived value theory states that products or
services that possess superior value among market alternatives will motivate acceptance
or continual usage. This explains the positive link between perceived value and public
acceptance of AVs in Fig. 1; that is, if AVs can offer superior value to the public, it could
encourage public acceptance (i.e. H6).

The second argument relates to the determinants of perceived value. According to


existing research, the perceived value of a product or service is derived from the
evaluation of its characteristics. AVs are considered a form of transport innovation; thus,
the tenets of innovation diffusion theory could be used to identify the characteristics of
AVs that lead to the creation of value. Innovation diffusion theory proposes that the rate
of diffusion or adoption of an innovation is determined by five salient characteristics of
the innovation (i.e. relative advantage, compatibility, complexity, trialability and
observability) (Gkartzonikas and Gkritza, 2019; Rogers, 2003). In this study, these
characteristics are rationalised to generate economic, functional, hedonic and social
utility to the public, which increases the public's perceived value of AVs (i.e. H1–H5).

The third argument concerns the indirect effect of perceived value on public
acceptance of AVs via trust. Recent research has suggested that trust plays an important
role in directly influencing technology or innovation acceptance (Zhang et al., 2020).
Trust can be measured based on the perceived competency of the trustee which includes
expertise, integrity and benevolence. In this regard, AVs that can exhibit these
competencies will improve the public's trust and consequently, public acceptance of AVs
(i.e. H8). Further, according to trust theory, the degree to which an individual will trust an
innovation is influenced by the individual's expectation that the innovation will produce
positive results in exchange for being subjected to vulnerability (Castelfranchi and
Falcone, 2010). Therefore, the perceived value, which reflects the accumulation of utility
from the use of AVs (i.e. positive results), can improve the public's trust in AVs (i.e. H7).

7
In the subsequent subsections, the hypotheses presented in Fig. 1 are elaborated.

2.2 The effects of innovation diffusion characteristics on the public's perceived


value of AVs

2.2.1 The effect of relative advantage on the public's perceived value of AVs
Relative advantage refers to the degree to which AVs are perceived as being better
than existing vehicles that require human conduction (Gkartzonikas and Gkritza, 2019).
It is conditional on whether the public views AVs as being advantageous. AVs present
several advantages which can positively influence the public's perceived value (i.e.
economic, functional, hedonic and social utility) of AVs.

For instance, AVs are noted to be safer alternatives to conventional vehicles (Deb et
al., 2017). A majority of traffic accidents are caused by human error or misjudgement,
which can be circumvented by AVs employing technologies to sense the environment
reliably and navigate the vehicles (Montoro et al., 2019). This has a positive effect on
perceived value, in particular, economic utility, because the public can avoid the hefty
cost of deaths, injuries as well as insurance, repair costs and third-party liability
associated with an accident (Penmetsa et al., 2019). Improved safety can also enhance
functional utility because safety is an important consideration in transport. According to
Xu et al. (2018), AVs are perceived to be four to five times safer than conventional
vehicles.

Next, another advantage of AVs is that they reduce fuel consumption because of
their optimised routing and smoother braking and acceleration (Fagnant and Kockelman,
2015). This can translate into financial savings for the public. Further, optimised routing
and smoother braking and acceleration can provide greater comfort and shorten the
travelling time. They can also provide critical mobility for the elderly or disabled, and
present safer options for the public who are concerned with impaired driving (i.e. under
the influence of fatigue, alcohol or drug) (Payre et al., 2014). Finally, AVs can also
present greater opportunities for their owners; for instance, they can share their AVs and
engage them in crowdsourcing such as ride-hailing services while they are not using the
vehicles. This provides additional income for the owners.

8
The aforementioned advantages of using or owning AVs can confer economic or
functional benefits to the public.

H1: The relative advantages of AVs over non-autonomous vehicles have a


positive effect on the public's perceived value of AVs

2.2.2 The effect of compatibility on the public's perceived value of AVs


In the context of this study, compatibility refers to the degree to which AVs are
perceived as consistent with existing values, lifestyle, past experiences and transport
needs of the public (Yuen et al., 2018a).

With the growing trend towards a more mobile lifestyle, AVs are better aligned with
such lifestyle as their ability to self-park offer convenience for individuals who are
frequently engaged in trip-chaining (Lavasani et al., 2016). Time can be saved when the
AV owners can immediately alight at their destinations rather than searching or waiting
to park their vehicles (i.e. functional utility). In addition, such arrangement can save cost
for the owners, offering economic utility because AVs has the ability to self-navigate to
cheaper parking areas (Fagnant and Kockelman, 2015).

Further, with more members of a household joining the workforce or engaging in


other activities such as schooling, the transport need of a household is shifting towards
multiple pick-up and drop-off locations. In this regard, AVs which can drive and park
autonomously are better positioned to meet this changing need. Furthermore, this growing
transport need provides greater opportunities for carpooling or shared ownerships, which
can be more efficiently and effectively performed by AVs rather than non-autonomous
vehicles (Fagnant and Kockelman, 2018). It was reported that waiting time, transit time,
and travel cost can be reduced from the introduction of shared AVs.

Lastly, AVs are noted to be a greener alternative to non-autonomous vehicles


(Vasebi et al., 2018). A key motivation for societies to introduce AVs in the private
vehicle market is to reduce greenhouse gas emissions and alleviate traffic congestion.
With the rising public's awareness and concern over the sustainability of societies such
as improving environmental protection and public health, AVs are better aligned with the
changing social values of the public in providing social utility.

9
H2: The compatibility of AVs with the changing lifestyle, transport needs
and values of the public has a positive effect on the public's perceived
value of AVs

2.2.3 The effect of complexity on the public's perceived value of AVs


Complexity is the degree to which the public views AVs to be difficult to understand
or use. It is the antonym of ‘ease of use’ which is a key construct in the technology
acceptance model that predicts the adoption of new technologies (May et al., 2017; Wang
et al., 2018).

According to Miyazaki and Kijima (2000), the complexity associated with


technology such as AVs can be measured using two dimensions (i.e. object and people).
For the object dimension, the complexity of AVs increases if the number of components
in AVs and the number of relations between these components increased. The people
dimension reflects the users' interests, capabilities and perceptions of AVs.

Essentially, if AVs were perceived to be complex by the public, this would have a
negative effect on their perceived value of AVs because additional time and effort are
required for the public to learn and understand the processes, thereby reducing the
economic utility of AVs. It can create inconvenience and frustration, which would affect
the functional and hedonic utility of AVs, respectively.

H3: The complexity of AVs has a negative effect on the public's perceived
value of AVs

2.2.4 The effect of trialability on the public's perceived value of AVs


Trialability refers to how easily the public can try or test AVs before their actual
adoption (Yuen et al., 2018a). Trialability of new transport innovations such as AVs is
important as it provides a platform for users or the public to experiment with the
innovations on a limited basis and in a supervised and safe environment (Strömberg et
al., 2016). For instance, the trialability of AVs can be enhanced by engaging the public
through test drives, demonstrations, training or presentations. Especially with the growth
in the use of information, web and mobile technologies to provide public transport, the
public who are now generally more tech-savvy would be more receptive to AVs.

10
Consequently, they would be more willing to try and experiment with innovations such
as AVs.

The freedom to explore the mechanisms of AVs promotes learning. Prior experience
with AVs during trials allows the public to exert less effort in learning and interacting
with AVs when they are launched, providing economic (i.e. time and cost savings) and
functional (i.e. ease of use) utility (Al-Rahmi et al., 2019). Furthermore, trials create
awareness and inform the public about the benefits of AVs. Consequently, the public
would have a better judgement of the values that AVs can provide over non-autonomous
vehicles. In addition, the notion of experimenting with innovations such as AVs can
create unexpected, positive experiences and emotions such as excitement which
contribute to the hedonic value of AVs (Wang et al., 2018).

H4: The trialability of AVs has a positive effect on the public's perceived
value of AVs

2.2.5 The effect of observability on the public's perceived value of AVs


Observability refers to the degree to which the benefits and process of using AVs
can be easily observed by and communicated to others (Wang et al., 2018). As the public
is connected by a social network, being able to communicate with each other about the
usage of AVs and their benefits could accelerate the diffusion of such information within
a society (Talebian and Mishra, 2018; Zhang et al., 2020). This could enhance the
perceived value of AVs, and subsequently, their acceptance.

When the steps involved with interacting with AVs could be simply learned by
observing others and simply explained by others, or when the advantages of using AVs
are apparent, the perceived value of AVs is expected to increase because of similar
reasons with those for explaining the effect of trialability on the public's perceived value
of AVs. For instance, greater observability can offer greater economic (i.e. time and cost
savings) and functional (i.e. ease of use) utility to the public (Al-Rahmi et al., 2019). The
public would also have a better judgement of the social and environmental benefits of
using AVs, which contributes to social utility.

H5: The observability of AVs has a positive effect on the public's


perceived value of AVs

11
2.3 The direct effect of perceived value on acceptance of AVs

In this context, acceptance can be referred to as an intention or commitment to adopt


or use AVs in the future. The public's decision to accept the usage of AVs could be
explained by perceived value theory.

The theory posits that products or services that possess superior value among market
alternatives will motivate acceptance or continual usage. In this regard, if AVs were
perceived to possess greater economic (e.g. price and fuel savings), functional (e.g.
convenience and ease of use), hedonic (e.g. excitement and pleasure) and social benefits
(e.g. greenhouse gas emission and congestion reduction) than conventional vehicles, the
public is expected to adopt AVs. For instance, concerning hedonic value, Bernhard et al.
(2020) showed that valence (i.e. pleasantness) of the trip has a positive effect on user
acceptance of AVs. Similarly, the authors also found that environmental friendliness (i.e.
social value) and spaciousness (i.e. functional value) also have positive influence on
acceptance of AVs.

In addition, Talebian and Mishra (2018) stated that the full acceptance or adoption
of AVs is only possible if everyone is satisfied with them. Satisfaction connotes similar
meaning with perceived value.

H6: The perceived value of AVs has a positive effect on public acceptance
of AVs

2.4 The indirect effect of perceived value on acceptance of AVs

This study argues that perceived value has an indirect effect on public acceptance of
AVs via trust. In this context, trust is defined as the expectation that AVs will engender
actions designed to produce positive results for the public, in situations of consistent
perceived risk (Castelfranchi and Falcone, 2010). The definition denotes several
meanings. Firstly, it requires the public to believe in or trust the decision process and
actions of AVs. Secondly, some benefits can be expected from relying on AVs. Thirdly,
the decision to trust AVs places the public in a position of being vulnerable.

The definition suggests that the public's trust in AVs can increase if (1) the expected
benefits from trusting AVs increase, and (2) the vulnerability accepted by the public can
be reduced. These provide more incentives for the public to trust AVs.

12
The perceived value of AVs which represents the net utility from evaluating the five
innovation characteristics can positively influence the public's trust in AVs because it
increases the public's expected benefits from trusting AVs. For instance, increasing the
perceived reliability of AVs, which is an aspect of functional utility, raises the benefits
for users as they can allocate less buffer time to commute. These benefits or utility
generate more incentives for the public to trust AVs.

Moreover, in certain cases, raising the public's perceived value of AVs may also
reduce the vulnerability that the public has to accept for trusting AVs. For instance,
improving the perceived reliability of AVs can reduce the perceived likelihood of
accidents which, consequently, reduces the perceived vulnerability. This also generates
greater incentives for the public to trust AVs. Therefore, the following hypothesis is
proposed.

H7: The perceived value of AVs has a positive effect on the public's trust in
AVs

Finally, this study suggests that trust has a positive effect on public acceptance of
AVs. Trust plays an important role to facilitate human and technology adoption (Adnan
et al., 2018; Lee and Kolodge, 2020; Liu et al., 2019b). AVs are intelligent vehicles that
are fully automated without human intervention. Hence, users have to give up some
degree of control and have to place their trust in AVs to execute the driving task reliably
and safely.

Possessing a high level of trust in AVs would indicate that the public has developed
beliefs that the AV system is (1) predictable and understandable (i.e. system
transparency), (2) performs tasks accurately and correctly (i.e. technical competence), and
(3) provides avenues for users to regain control of the vehicles whenever desired (i.e.
situation management) (Choi and Ji, 2015; Molnar et al., 2018). The availability of these
elements would improve the public's confidence in AVs, thereby resulting in their
acceptance.

H8: The public's trust in AVs has a positive effect on public acceptance of
AVs

13
3. Methodology
To reiterate, the theoretical model (Fig. 1) incorporates three behavioural theories,
namely, innovation diffusion theory, perceived value theory and trust theory.
Accordingly, (1) innovation diffusion theory proposes five explanatory constructs which
are relative advantage, compatibility, complexity, trialability and observability, (2)
perceived value theory proposes an explanatory construct (i.e. perceived value of AVs),
and (3) trust theory proposes an explanatory construct (i.e. trust in AVs) to account for
public acceptance of AVs. In Section 2, the interrelationships between the constructs are
specified by hypotheses.

Structural equation modelling is adopted as the data analysis method to test the
theoretical model. It is selected for the following reasons. Firstly, it allows the analysis
of a model that contains multiple dependent constructs (i.e. perceived value, trust and
public acceptance). Secondly, the constructs are latent (i.e. unobservable and multi-
dimensional); hence, they must be operationalised by several measurement items (i.e.
observable variables). In this regard, structural equation modelling accounts for the
measurement errors and hence improves the accuracy of the model estimation. Lastly and
most importantly, structural equation modelling estimates the model's parameters
simultaneously. Consequently, the estimated effect of an explanatory construct on a
dependent construct shall also account for the effects of other explanatory constructs.

3.1 Measurement items

The following measurement items are developed to operationalise each construct in


this study: relative advantage, compatibility, complexity, trialability, observability,
perceived value, trust and acceptance of AVs. Table 2 shows the constructs, measurement
items and the relevant articles that were used to develop the measurement items.

Table 2. Constructs and measurement items


Construct ID Measurement item Adapted source
Relative advantage From 1 = strongly Jansson (2011)
(RLA) disagree to
7 = strongly agree
Petschnig et al.
RLA1 AVs would solve
(2014)
problems that I have
encountered with
conventional cars.

14
RLA2 AVs would reduce the
time that I need to get
to places.
RLA3 AVs would allow
better access to my
intended destinations.
RLA4 AVs would be an
environmentally
friendly option.
RLA5 AVs would be more
advantageous
compared to using
conventional cars.
Compatibility From 1 = strongly Moons and De
disagree to Pelsmacker (2015)
(COM)
7 = strongly agree
COM1 AVs would be in line
Petschnig et al.
with my beliefs.
(2014)
COM2 AVs would fit well
with my driving
habits.
COM3 AVs would be
compatible with my
mobility needs.
COM4 AVs would suit me
well.
COM5 AVs would be in line
with my everyday
life.
Complexity From 1 = strongly Yuen et al. (2018a)
disagree to
(CPL)
7 = strongly agree
Petschnig et al.
CPL1 AVs would be
(2014)
difficult to use.
CPL2 AVs would be
difficult to learn how
to use.
CPL3. AVs would be
frustrating to use.
CPL4 AVs would be
cumbersome to use.
CPL5 AVs would require a
lot of effort to use.
Trialability From 1 = strongly Petschnig et al.
(TRI) disagree to (2014)
7 = strongly agree

15
TRI1 Before I decide to buy
an AV, I would like to
test-drive it.
TRI2 Before I decide to buy
an AV, I would like to
borrow it for a day or
two.
TRI3 Before I decide to buy
an AV, I would like to
try a friend's AV.
TRI4 Before I decide to buy
an AV, I would like to
view a demonstration
of using an AV.
TRI5 Before I decide to buy
an AV, I would like to
receive training or
attend a course on
using an AV.
Observability From 1 = strongly Yuen et al. (2018a)
(OBS) disagree to
7 = strongly agree
Petschnig et al.
OBS1 I believe I can learn
(2014)
how to use AVs.
OBS2 I believe I can explain
to others how to use
AVs.
OBS3 I believe I can benefit
from using AVs.
OBS4 I would have no
difficulty telling
others about the
process of using AVs.
OBS5 I believe I could
communicate to
others the benefits of
using AVs.
Perceived value From 1 = strongly
(VAL) disagree to
Yuen et al. (2019)
7 = strongly agree
VAL1 I feel that using AVs
can better cater to my
mobility requirements
(e.g. safety,
reliability, security or
convenience needs).
VAL2 I feel that using AVs
can confer cost-

16
savings (e.g. savings
on fuel or optimised
trips).
VAL3 I feel that using AVs
is pleasant.
VAL4 I feel that using AVs
would have positive
effects on the
environment and
society.
Trust (TRU) From 1 = strongly Nordhoff et al. (2018)
disagree to
7 = strongly agree
Kaur and Rampersad
TRU1 I trust that AVs can
(2018)
drive without
assistance from me.
TRU2 I trust AVs to be safe
and reliable in severe
weather conditions.
TRU3 I would trust the
driving skills of AVs
more than my own
driving skills.
TRU4 AVs can be trusted to
carry out journeys
effectively.
TRU5 My trust in AVs will
be based on the car
manufacturer's
reputation for safety
and reliability.
TRU6 My trust in AVs will
be based on the
reliability of the
underlying
technologies.
Acceptance (ACP) From 1 = strongly Choi and Ji (2015)
disagree to
7 = strongly agree
ACP1 I would consider
using AVs when they
are available in the
market.
ACP2 I would recommend
AVs to my family and
peers.

17
ACP3 I would encourage
others to use AVs.
ACP4 I have positive things
to say about AVs.

3.2 Survey design and administration

The questionnaire comprises three segments. The first section introduces the
participants to the motivation and objective of the study. The level of automation is also
defined (i.e. SAE 4 and above) to ensure that the respondents have a common
understanding of the level of automation when completing the survey questionnaire. In
particular, AVs are introduced to the respondents based on the understanding that AVs
do not require the users to execute steering and acceleration or deceleration (i.e. ‘hands
off’), monitor the driving environment (i.e. ‘eyes off’), and perform ad-hoc intervention
when required by AVs (i.e. ‘minds off’). The questionnaire also assures the nondisclosure
of the respondents' identities under any circumstances. Moreover, their honest opinions
are appreciated. The second section collects the respondent's demographic information,
such as age, gender, annual income, education, number of vehicles owned and driving
experience. The third section includes all measurement items found in Table 2. Identical
but reversed items were also incorporated in the questionnaire for validation purpose.

The questionnaire was administered to residents in Seoul, Republic of Korea. The


English-version questionnaire was first translated to Korean and then back-translated to
English by an editor to ensure readability and equivalent meaning. Both English versions
were examined to identify any discrepancies or misunderstanding. Thereafter, corrections
were made to the Korean questionnaire according to the editor's suggestion. The finalised
survey questionnaire was created online on Google Form. Thereafter, a QR code that
directs users to the Google Form was generated.

The survey was conducted simultaneously at the exits of five subway stations with
high passenger flow in Seoul. These locations were randomly selected. They include
Sinchon, Apgujeong, Seoul, Suwon and Express Bus Terminal stations. Four student
helpers were assigned to each station. They were tasked to conduct street-intercept
surveys on the passengers. The approached passengers were asked if they were residing
in Seoul and willing to participate in the survey. If both questions were answered

18
positively, they can complete the questionnaire either online by scanning a QR code or
the printed version. A cash voucher was given to all who had participated. The survey
was conducted for seven consecutive days from 8 am to 8 pm in April 2019.

Finally, 723 completed survey questionnaires were collected. After discarding those
that were invalid, 526 questionnaires were used for further analysis. Of these
questionnaires, 242 were completed online, whereas 284 were completed using the paper-
and-pencil method.

3.3 Demographics of respondents

Table 3 shows the demographic profile of the 526 respondents. The proportion of
male (48%) and female (52%) respondents was evenly distributed. The sample's
proportion is quite close to the population's gender distribution of 50.1% male and 49.9%
female (https://www.statista.com/statistics/642480/south-korea-population-male-
female-ratio/).

Next, approximately half of the sample earned an annual salary of 10–40 million
Korean Won, KRW (42%) and owned a vehicle (49%). The sample's average vehicle
ownership by household is 0.73, which is slightly higher than the population's average
value of 0.45 (https://pulsenews.co.kr/view.php?year=2019&no=523078). This
difference is expected because the study was conducted in Seoul, which is the capital city
of South Korea. Hence, the sample's average vehicle ownership is expected to be
reasonably higher than the population's average.

In addition, most of the respondents (71%) hold a bachelor's degree and above. The
finding that most of the respondents are highly educated is consistent with the country's
emphasis on education. According to a report (https://wenr.wes.org/2018/10/education-
in-south-korea), 70% of 24- to 35-year-olds in the nation of 51.5 million people have
completed university education in 2018. Further, as the survey was conducted in Seoul,
a higher proportion of highly educated citizens are expected to reside there as compared
with other cities in Republic of Korea. Hence, it is not surprising that 71% of the
respondents hold a bachelor's degree. The above comparisons provide some validation
regarding the representativeness of the sample.

Finally, as the survey was conducted near subway stations, sampling issues may
occur because the respondents might tend to favour using public transport rather than

19
private vehicles. Further, previous research has shown that car availability has a positive
effect on the acceptance of AVs (Krueger et al., 2016). A t-test was conducted on public
acceptance of AVs by comparing the means of those who own and those who do not own
vehicles. The test is not significant (p > 0.05), indicating no significant difference between
the sample sub-groups.

Table 3. Respondents' profile


Characteristics Items Frequency Percentage
(n = 526) (%)
Gender Male 254 48
Female 272 52
Age 18* - 30 186 35
(years) 30 - 40 124 24
40 - 50 86 16
50 - 60 88 17
> 60 42 8
Education level Elementary or 68 13
lower
High School 86 16
Bachelor 256 49
Post Graduate 116 22
Annual income < 10 116 22
(million KRW) 10 – 40 223 42
(1 KRW = 40 – 80 147 28
0.0009 USD)
> 80m 40 8
Number of 0 210 40
vehicles owned
1 256 49
per household
2 50 10
>2 10 1
Driving No licence 101 19
experience
<1 56 11
(years)
1-5 133 25
5 - 10 156 30
> 10 80 15
Note: * The legal age for driving a vehicle in Republic of Korea is 18 years.

20
4. Results and Discussion
Structural equation modelling consists of conducting the measurement model
analysis (i.e. confirmatory factor analysis) and structural model analysis (Hair et al.,
2010). The purpose of the measurement model analysis is to examine the relationship
between the constructs and their measurement items (i.e. factor loadings) whereas the
structural model analysis is to examine the relationship between the constructs specified
by the hypotheses.

The measurement model analysis involves assessing the overall model fit, reliability,
convergent and discriminant validity. The structural model analysis involves the
simultaneous estimation of the hypothesised relationships between the constructs. To
yield greater theoretical and policy implications, the direct, indirect and total effect
analysis is also conducted.

4.1 Measurement model analysis

A measurement model analysis was conducted, and the results are presented in Table
4. The fit indices of the measurement model meet the threshold proposed by Hu and
Bentler (1999). For instance, Tucker-Lewis index (TLI) and comparative fit index (CFI)
are greater than the recommended value of 0.95 whereas standardised root mean square
residual (SRMR) and root mean square error of approximation (RMSEA) are lower than
0.10 and 0.08, respectively. This indicates that the measurement model possesses
sufficient model fit.

Table 4. Confirmatory factor analysis results

Construct Item λ AVE CR


Relative RLA1 0.72 0.65 0.90
advantage
RLA2 0.76
(RLA)
RLA3 0.84
RLA4 0.88
RLA5 0.82

21
Compatibilit COM1 0.73 0.73 0.93
y
COM2 0.86
(COM)
COM3 0.95
COM4 0.91
COM5 0.80
Complexity CPL1 0.86 0.66 0.91
(CPL)
CPL2 0.84
CPL3 0.76
CPL4 0.72
CPL5 0.88
Trialability TRI1 0.86 0.67 0.91
(TRI)
TRI2 0.90
TRI3 0.78
TRI4 0.80
TRI5 0.74
Observabilit OBS1 0.71 0.66 0.90
y
OBS2 0.74
(OBS)
OBS3 0.86
OBS4 0.90
OBS5 0.82
Value VAL1 0.79 0.70 0.90
(VAL)
VAL2 0.81
VAL3 0.90
VAL4 0.83
Trust TRU1 0.81 0.67 0.93
(TRU) TRU2 0.72
TRU3 0.79
TRU4 0.83
TRU5 0.92
TRU6 0.84
Acceptance ACP1 0.92 0.82 0.95
(ACP) ACP2 0.88
ACP3 0.93
ACP4 0.90
Note: Model fit indices: χ2/df = 2.02, (p < 0.05, df = 237); CFI = 0.97; TLI = 0.97; RMSEA =
0.07; SRMR = 0.06.

22
The results in Table 4 also indicate that the measurement items are reliable.
Accordingly, the factor loadings of the measurement items as well as the composite
reliabilities are greater than the recommended values of 0.70 and 0.80 (Hair et al., 2010).
Further, the results presented in Table 5 suggest that the measurement items are valid.
First, convergent validity is established as the average variance extracted of each
construct is above the recommended value of 0.50 (Hair et al., 2010). In addition,
discriminant validity is also ascertained because the average variance extracted of any
pair of constructs are greater than their squared correlations.

Table 5. AVE, correlations, and squared correlations of the constructs

RLA COM CPL TRI OBS VAL TRU ACP

RLA 0.65a 0.19c 0.03 0.03 0.17 0.38 0.05 0.02


COM 0.44b 0.73 0.02 0.01 0.35 0.21 0.01 0.03
CPL -0.16 -0.15 0.36 0.01 0.07 0.07 0.05 0.03
TRI 0.16 0.11 0.04 0.67 0.24 0.10 0.04 0.03
OBS 0.41 0.59 -0.26 0.49 0.66 0.38 0.04 0.06
VAL 0.62 0.46 -0.26 0.31 0.62 0.7 0.52 0.26
TRU 0.23 0.11 -0.24 0.21 0.19 0.72 0.67 0.42
ACP 0.13 0.16 -0.22 0.16 0.24 0.51 0.65 0.82
a bCorrelations
Note: AVE values are along the main diagonal; between constructs are below the main diagonal;
cSquared correlations between constructs are above the main diagonal.

As the data were collected from surveys, common method bias might be prevalent
in the data. A marker variable test proposed by Podsakoff et al. (2003) is employed to
examine the common method bias. The measurement items were regressed on the marker
variable, which is a dichotomous variable that reflects whether the survey was completed
by respondents on a hardcopy or online. This variable is chosen because it is expected to
be uncorrelated with the latent constructs presented in Fig. 1. Consequently, the
regression weights were estimated. It is observed that most regression weights are
significant, indicating the presence of common method bias. To remedy the bias, all
measures are corrected for the common method variance using data imputation in the
subsequent analysis.

23
4.2 Structural model analysis

Figure 2 presents the estimated structural model. Control variables which include
age, education, and income were also added into the model to account for their effects on
public acceptance of AVs. These variables were noted to influence the adoption of AVs
significantly (Haboucha et al., 2017; Menon et al., 2020).

Note: *indicates that the path estimate is significant (p <0.05); Model fit indices: χ2/df = 2.12 (p < 0.05, df = 275); TLI
= 0.97; CFI = 0.97; RMSEA = 0.07; SRMR = 0.06.

Fig. 2. Parameter estimation of the proposed model

Figure 2 suggests that the structural model possesses good fit (χ2/df = 2.12 (p < 0.05,
df = 275); TLI = 0.97; CFI = 0.97; SRMR = 0.06; RMSEA = 0.07). The squared multiple
correlation (R2) values of the endogenous variables, that is, perceived value, trust, and
public acceptance of AVs are greater than 0.5, which highlight the explanatory power of
the exogenous variables.

As presented in Fig. 2, public acceptance of AVs is regressed on the control variables


which include ‘age’, ‘education’ and ‘income’. Accordingly, their standardised
regression estimates are 0.03, 0.02 and 0.10. However, only the effect of ‘income’ on
public acceptance is significant (p < 0.05). The effect of ‘age’ on public acceptance is not
significant. This finding is unexpected because the older generations would be less tech-

24
savvy. The same applies to ‘education’. Individuals who are more educated should have
a better appreciation of the various benefits of using AVs and are more receptive to AVs.
The result for ‘income’ is expected because richer individuals are more likely working
professionals who would be commuting with their vehicles more frequently. Hence, using
AVs is more aligned with their working lifestyle. Regardless, the relatively weak effects
shown in the demographic variables as compared with the theoretical predictors presented
in Fig. 2 are aligned with the findings of Kaye et al. (2020). They found that domain-
specific attributes are stronger determinants of public acceptance of AVs as compared
with sociodemographic variables such as age and gender.

All five innovation diffusion variables, namely, relative advantage, compatibility,


complexity, trialability and observability, have significant effects on the public's
perceived value of AVs. Their standardised effects are 0.58, 0.32, -0.12, 0.18 and 0.43,
respectively. Therefore, H1, H2, H3, H4 and H5 are accepted. Together with the control
variables, these five variables explain for 72% of the variance in perceived value (R2 =
0.72). In general, the results are in agreement with the current study's argument that the
innovation diffusion variables lead to the creation of value for the public, which can be
segmented into economic, functional, hedonic or social utility.

For example, AVs confer relative advantages over conventional vehicles such as
safety, comfort, shorter transit-time, opportunities for mobility while impaired, and lower
fuel consumption (Chen, 2019). These advantages provide functional and economic
benefits to the public. Similarly, AVs are argued to be more compatible with the existing
values, lifestyle, past experiences and transport needs of the public, thereby providing
greater convenience for individuals with a mobile lifestyle which involves trip-chaining
and multiple pick-up and drop-off locations. In addition, AVs are also better positioned
to meet the growing demands of the public for greener transport. These provide
functional, economic and social utility for the public. Reducing the complexity of AVs
can also lead to value creation. If AVs were viewed to be complex by the public,
additional time and effort will be required to learn and understand the process of using
AVs. This will reduce the economic utility (i.e. more time spent on learning) as well as
hedonic utility (i.e. frustration) of AVs. Greater trialability also has positive effects on
the perceived value of AVs. Trialability of AVs can be enhanced by engaging the public
through test drives, demonstrations, training or presentations. This engages the public and
promotes learning. Consequently, the public requires less effort to learn and interact with

25
AVs, which confers economic (i.e. time and cost savings) and functional (i.e. ease of use)
utility. Further, trying innovations such as AVs can result in unexpected emotions such
as excitement which confers hedonic value. Finally, greater observability can generate
positive values and allows the public to communicate with each other about the benefits
and process of using AVs with ease, contributing to the rate of learning which confers
greater economic (i.e. time and cost savings) and functional (i.e. ease of use) utility for
the public.

Fig. 2 also depicts that perceived value has a significant, positive effect on public
acceptance of AVs (β = 0.45, p < 0.05). Hence, H6 is accepted. This finding aligns with
perceived value theory which suggests that the likelihood of AVs to be accepted by the
public shall increase if AVs offered superior value amongst existing alternatives (i.e.
conventional vehicles). In this regard, a rational user will tend to favour using AVs which
offer more utility than conventional vehicles.

In addition to the direct effect that perceived value has on public acceptance of AVs,
perceived value also has an indirect effect on public acceptance of AVs which is
channelled via the public's trust in AVs. Perceived value has a significant, positive effect
on the public's trust in AVs (β = 0.73, p < 0.05), which in turn has a significant, positive
effect on public acceptance of AVs (β = 0.47, p < 0.05). Therefore, both H7 and H8 are
accepted. The result corroborates trust theory which suggests that the public's trust in AVs
is influenced by the individual's expectation that the innovation will produce positive
results for the individual in exchange for being subjected to vulnerability (Castelfranchi
and Falcone, 2010). In other words, the public's trust in AV can develop if the expected
benefits from trusting AVs increased or the vulnerability accepted from trusting AVs
reduced. Perceived value can increase the expected benefits because it creates more
incentives for the public to trust the decision process and actions of AVs. Further, certain
aspects of perceived value, for example, relating to safety and reliability, can help the
public to understand that their risk or vulnerability can be reduced from using AVs. The
above explanations support a positive link between perceived value and trust. Lastly, trust
can contribute to the formation of public acceptance of AVs. Trust towards AVs is
developed when AVs are predictable and understandable, perform tasks accurately and
correctly and provide ways for users to regain control of the vehicles whenever desired
(Molnar et al., 2018). This can bolster the public's confidence in using AVs and hence
result in acceptance.

26
Collectively, the theoretical model accounts for 69% of the variance in public
acceptance of AVs (R2 = 0.69). This indicates that the model has slightly improved
compared with the existing models which have R2 values between 0.43 and 0.67 (May et
al., 2017; Panagiotopoulos and Dimitrakopoulos, 2018; Xu et al., 2018; Zhang et al.,
2019). This improvement is expected because the theoretical model has expanded, and
most of the variables that are included in the existing models are considered.

4.3 Direct, indirect and total effect analysis

A bootstrapping analysis was conducted, and the significance tests support mediation
effects whereby the innovation diffusion variables have no significant, direct effects but
only significant, indirect effects on public acceptance of AVs. This indicates that
perceived value fully mediates the effects of the five innovation diffusion variables on
public acceptance. Further, trust partially mediate the influence of perceived value on
public acceptance. In addition, all the exogenous variables presented in Fig. 2 have
significant, total effects on public acceptance.

Table 6 presents the effects of the exogenous variables on endogenous variables. For
the direct effects, the key predictors of perceived value are relative advantage (a11 = 0.58),
followed by observability (a51 = 0.43), compatibility (a21 = 0.32), trialability (a41 = 0.18)
and complexity (a31 = -0.12). The only direct predictor of trust is perceived value (a62 =
0.73). Lastly, for public acceptance, its direct predictors are perceived value (a63 = 0.45)
and trust (a73 = 0.47).

Table 6. Direct, indirect, and total effects

Endogenous Perceived value Trust Public acceptance


(j)
(1) (2) (3)
Exogenous
(i)
Direct effects (aij) of …
relative advantage (1) 0.58 — —
compatibility (2) 0.32 — —
complexity (3) −0.12 — —
trialability (4) 0.18 — —
observability (5) 0.43 — —
perceived value (6) — 0.73 0.45

27
trust (7) — — 0.47

Indirect effects (bij) of …


relative advantage (1) — 0.42 0.46
compatibility (2) — 0.23 0.26
complexity (3) — −0.09 −0.10
trialability (4) — 0.13 0.14
observability (5) — 0.31 0.34
perceived value (6) — — 0.34
trust (7) — — —

Total effects (cij) of …


relative advantage (1) 0.58 0.42 0.46
compatibility (2) 0.32 0.23 0.26
complexity (3) −0.12 −0.09 −0.10
trialability (4) 0.18 0.13 0.14
observability (5) 0.43 0.31 0.34
perceived value (6) — 0.73 0.79
trust (7) — — 0.45

As for the indirect effects, the relative advantage has the strongest influence on public
acceptance (b13 = 0.46). This is followed by observability (b53 = 0.34) and perceived value
(b63 = 0.34), compatibility (b23 = 0.26), trialability (b43 = 0.14) and complexity (b33 = -
0.10). As shown in Fig. 2, the effects of the innovation diffusion variables on public
acceptance are fully channelled through perceived value and partially through trust.

For the total effects, perceived value has the largest total effects (c63 = 0.79) on public
acceptance. This is followed by relative advantage (c13 = 0.46) and trust (c73 = 0.45).
Subsequently, observability has the fourth largest total effect (c53 = 0.34). Accordingly,
compatibility (c23 = 0.26), trialability (c43 = 0.14) and complexity (c33 = -0.10) are ranked
fifth, sixth and seventh.

28
5. Conclusion and Policy Implications

5.1 Summary

The objective of this study is to identify the factors influencing public acceptance of
AVs and examine their interrelationships.

Through the theoretical lenses of innovation diffusion theory, perceived value theory
and trust theory, this study presents a model that explains public acceptance of AVs. The
central arguments are that public acceptance is directly influenced by both the public's
perceived value of and trust towards AVs. Further, perceived value has a positive
influence on trust because it raises the public's perceived benefits and reduces the
perceived vulnerability from using AVs. Moreover, perceived value can be explained
using the tenets of innovation diffusion theory which suggests five innovation diffusion
characteristics that lead to the creation of value for the public.

A street-intercept survey was conducted at five major subway stations in Seoul,


Republic of Korea. The survey administration yielded 526 valid responses. The results
suggest a mediated relationship whereby the effects of the innovation diffusion variables
(i.e. relative advantage, compatibility, complexity, trialability and observability) on
public acceptance of AVs are fully mediated by perceived value. In addition, the effect
of perceived value on public acceptance of AVs is partially mediated by trust. The total
effects analysis revealed that perceived value has the largest influence on public
acceptance of AVs. This is followed by relative advantage, trust, observability,
compatibility, trialability and complexity.

5.2 Theoretical contributions

This study has made a significant contribution to academic research. Firstly, it fills
an important gap in the literature on public acceptance of AVs by introducing and
synthesising three theories, namely, the innovation diffusion theory, perceived value
theory, and trust theory, to understand the factors influencing public acceptance of AVs.
There is presently very limited theoretical research on the public acceptance of AVs.
Within the limited theoretical research, most have employed the technology acceptance
model, its variants such as automation acceptance model, or its extension by incorporating
concepts or theories such as trust or task-technology fit model (May et al., 2017). The

29
current study builds on the work of Acheampong and Cugurullo (2019) and provides an
alternative perspective or explanation to public acceptance of AV. For instance,
Acheampong and Cugurullo (2019) integrated socio-demographic variables and latent
behavioural factors anchored on the technology acceptance model, theory of planned
behaviour and innovation diffusion model. While there is some overlap, this study has
considered and applied other theories such as perceived value theory and trust theory.
Accordingly, these theories (i.e. innovation diffusion theory, perceived value theory and
trust theory) are rooted in diverse paradigms such as technology innovation acceptance,
customer utility and social psychology. Consequently, the theories offer a comprehensive
examination of the factors influencing public acceptance of AVs.

Next, this study contributes to the literature by identifying, proposing and


operationalising several factors influencing public acceptance of AVs, namely, relative
advantage, compatibility, complexity, trialability, observability, perceived value and
trust. These factors account for a huge proportion of the variance in public acceptance of
AVs (69%). This suggests that using a combination of theories is more effective than
using a single theory, which is common in the existing literature in explaining public
acceptance of AVs. More importantly, the results imply the complementary nature of the
theories which provide unique perspectives to explaining public acceptance of AVs.
Nonetheless, it should be noted that the explanatory power of the theories differs. Based
on the total effect analysis, innovation diffusion theory has the strongest explanatory
power, followed by perceived value theory and trust theory. In addition, the findings of
this study are consistent with those of Nordhoff et al. (2018) who found that domain-
specific attributes are stronger determinants of public acceptance of AVs as compared
with sociodemographic variables.

Further, this study provides a better nomological understanding of the relationships


between the factors influencing public acceptance of AVs. The results align with the
study's central arguments that relative advantage, compatibility, complexity, trialability
and observability are used by the public for the evaluation of value. Further, perceived
value and trust directly contribute to public acceptance of AVs. This finding is in line
with the core principles of consumer behaviour and motivation research whereby the
decision process of individuals begins with the evaluation of AVs' associated attributes,
which results in the creation of value. Consequently, value leads to the acceptance of AVs
directly or indirectly via trust.

30
In addition, this study has contributed to the measurement or operationalisation of
the constructs. While the measurement items for perceived value and acceptance of AVs
were directly adopted from standardised questionnaires (refer to Table 2), the
measurement items for the remaining constructs which include the five dimensions of
innovation diffusion theory and trust were developed by synthesising relevant research
articles on new technologies and modifying them to fit the context of AVs.

Finally, to some extent, this study has clarified the role of trust in explaining public
acceptance of AVs. Previous research has unequivocal views regarding the influence of
trust on public acceptance of AVs. For instance, Panagiotopoulos and Dimitrakopoulos
(2018) found that trust only has a direct effect on acceptance of AVs. Some others noted
that trust only has an indirect effect on acceptance of AVs via components of the
technology acceptance model which include perceived ease of use and perceived
usefulness (May et al., 2017; Xu et al., 2018). Finally, Choi and Ji (2015) found that trust
has direct and indirect effect on acceptance of AVs via components of the technology
acceptance model. The results of this study corroborate some previous studies which
detected significant, direct relationship between trust and acceptance of AVs. However,
contrary to previous studies, the current study argues that trust is influenced by perceived
value, which is in turn influenced by the components of innovation diffusion theory which
encapsulate those components found in the technology acceptance model. The notion that
trust is influenced by evaluating the characteristics of AVs rather than vice versa is
consistent with existing research on other technological fields such as self-service
technologies (Johnson et al., 2008) and e-commerce (Eid, 2011). This study proposes that
there must be some basis for an individual to trust AVs. In this regard, trust can be
developed if individuals consistently perceive that the benefits outweigh the risk and cost
of using AVs (i.e. perceived value). Perceived value is in turn influenced by individuals'
assessment of an AV's innovative attributes such as relative advantage, compatibility,
complexity, trialability and observability.

5.3 Policy implications

From the policy perspective, this study can provide some recommendations for
transport operators and policymakers in allocating scarce resources to improve public
acceptance of AVs. Compared with conventional vehicles, using AVs possesses many

31
advantages, which can offer benefits to various stakeholders including car manufacturers,
transport operators, users, societies and the environment. Relevant stakeholders should
recognise the importance of creating value to users when incentivising or designing AVs
to garner public acceptance of AVs. Furthermore, to a large extent, value fosters the
public's trust towards AVs and is influenced by the AV's aforementioned attributes.

The analysis of the total effects on public acceptance of AVs indicates that resources
could be first allocated to strengthening the perceived value. Therefore, transport
operators can consider subsidising the prices of AVs, improving the functional attributes
of AVs, enhancing the visual appearance and including attractive attributes in the AVs'
design to elicit positive effects, and highlighting the social and environmental benefits
from using AVs. AVs are often compared with other alternatives by users; therefore,
enhancing their perceived value based on these dimensions could promote AVs to be the
best choice. The evaluation of value is subjective to individuals; hence, transport
operators or policymakers can invest in marketing campaigns to educate and promote the
benefits or utility that AVs can offer to the public.

Thereafter, transport operators or policymakers can focus on improving or


advertising the innovation diffusion characteristics of AVs which, in order of descending
importance, are relative advantage, observability, compatibility, trialability and
complexity. For the relative advantage, relevant parties can position AVs as a better
alternative to conventional vehicles by highlighting AVs' advantages such as improved
safety, lower fuel consumption, greater comfort, shorter travel time and opportunities for
impaired driving and making an additional income by engaging AVs in crowdsourcing
or ride-hailing services. As for observability, marketing campaigns should be targeted at
helping the public to relate to the benefits of AVs. Some strategies that can be adopted
include performing a side-by-side comparison between AVs and conventional vehicles
and providing percentage improvements in various performance such as fuel
consumption, accident avoidance and pollution reduction. In addition, providing
testimonials and reporting ‘before and after’ cases of improvement in the life of potential
adopters might improve the observability of AVs. Further, AVs should be marketed to be
aligned with existing values, lifestyle, past experiences and transport needs of the public
to enhance their compatibility with the public. In particular, AVs can be targeted at
working professionals, active households and pro-environmental individuals who would
view AVs to be more aligned with their transport needs. As for trialability, it can be

32
enhanced by engaging the public through test drives, demonstrations, training or
presentations. This allows the public to have a better understanding of the process and
benefits of AVs in a safe and supervised environment before the full adoption of AVs.
Finally, for complexity, the process of using or interacting with AVs can be simplified by
reducing the number of AVs' components that require human interaction and the number
of relations between these components. The complexity of AVs can also be reduced by
stimulating interests of the public, increasing the public's capabilities of using AVs and
eliciting positive experiences from using AVs.

Lastly, transport operators and policymakers should also dedicate resources to foster
the public's trust towards AVs. Trust towards AVS can be developed if AV systems are
predictable and understandable (i.e. system transparency), perform tasks accurately and
correctly (i.e. technical competence) and provide avenues for users to regain control of
the vehicles whenever desired (i.e. situation management). These dimensions of trust
should be incorporated into the design and marketing of AVs.

5.4 Limitations and recommendations

This study has a few limitations. Firstly, this study is conducted in Seoul, Republic
of Korea. It is a densely populated city with a high proportion of working professionals
who favour the use of AVs. Therefore, the interpretation of the results should be exercised
with caution for other contexts, for instance, in suburban or rural areas or other countries.
AVs might have an advantage in urban regions such as Seoul due to high population
density. This enables sharing of AVs and carpooling that increase the benefits of using
AVs. Therefore, future studies can consider examining the generalisability of the research
model by cross-validating it with other contexts.

Another limitation of this study is that it has only introduced three theoretical lenses
to understand and explain the factors influencing public acceptance of AVs. Future
research can focus on introducing other theories or analyse the nuances (i.e. moderating
effects) in public acceptance of AVs (Hein et al., 2018).

The next limitation concerns the inclusion of trialability and observability in the
theoretical model because the development of AVs is still at their early stage.
Consequently, the limited availability of AVs would mean that observability is low and
trialability may be challenging to assess. However, a small portion of the population may

33
already be exposed to AVs from reading and searching relevant information about AVs
from the media or by participating in trials, experiments and educational seminars. In
addition, the survey questions for the two constructs are asked on the basis of the
population's perception. Grounded on the universal truth that perception is more
important than reality, this study argues that an individual's perception about the
trialability and observability is more important than the reality in influencing the
perceived value or acceptance of AVs. Therefore, including both constructs in the model
would be useful for the formulation of policies to increase the acceptance of AVs despite
them being at the early stage. Similar to many existing studies, the current study has
considered trust as a unidimensional construct. Future research can consider decomposing
trust into various facets and analyse their impact on public acceptance of AVs. This would
yield further implications on policy formulation to improve the acceptance of AVs.

Lastly, this study acknowledges that the obtained data are hypothetical due to the use
of stated preference survey. In addition, the data are cross-sectional. Future research can
conduct the surveys over an extended period to analyse the diffusion of AV acceptance
in society.

34
REFERENCES
Acheampong, R.A. and Cugurullo, F., 2019. Capturing the behavioural determinants
behind the adoption of autonomous vehicles: Conceptual frameworks and measurement
models to predict public transport, sharing and ownership trends of self-driving cars.
Transportation Research Part F: Traffic Psychology and Behaviour, 62, 349-375.
Adnan, N., Nordin, S.M., bin Bahruddin, M.A. and Ali, M., 2018. How trust can drive
forward the user acceptance to the technology? In-vehicle technology for autonomous
vehicle. Transportation Research Part A: Policy and Practice, 118, 819-836.
Al-Rahmi, W.M., Yahaya, N., Aldraiweesh, A.A., Alamri, M.M., Aljarboa, N.A.,
Alturki, U. and Aljeraiwi, A.A., 2019. Integrating Technology Acceptance Model With
Innovation Diffusion Theory: An Empirical Investigation on Students’ Intention to Use
E-Learning Systems. IEEE Access, 7, 26797-26809.
Bansal, P., Kockelman, K.M. and Singh, A., 2016. Assessing public opinions of and
interest in new vehicle technologies: An Austin perspective. Transportation Research
Part C: Emerging Technologies, 67, 1-14.
Becker, F. and Axhausen, K.W., 2017. Literature review on surveys investigating the
acceptance of automated vehicles. Transportation, 44(6), 1293-1306.
Bernhard, C., Oberfeld, D., Hoffmann, C., Weismüller, D. and Hecht, H., 2020. User
acceptance of automated public transport: Valence of an autonomous minibus experience.
Transportation Research Part F: Traffic Psychology and Behaviour, 70, 109-123.
Castelfranchi, C. and Falcone, R., 2010. Trust Theory: A Socio-Cognitive and
Computational Model. John Wiley & Sons.
Chen, C.F., 2019. Factors affecting the decision to use autonomous shuttle services:
Evidence from a scooter-dominant urban context. Transportation Research Part F:
Traffic Psychology and Behaviour, 67, 195-204.
Chen, H.-K. and Yan, D.-W., 2019. Interrelationships between influential factors and
behavioral intention with regard to autonomous vehicles. International Journal of
Sustainable Transportation, 13(7), 511-527.
Choi, J.K. and Ji, Y.G., 2015. Investigating the importance of trust on adopting an
autonomous vehicle. International Journal of Human-Computer Interaction, 31(10), 692-
702.
Cunningham, M.L., Regan, M.A., Horberry, T., Weeratunga, K. and Dixit, V., 2019.
Public opinion about automated vehicles in Australia: Results from a large-scale national
survey. Transportation Research Part A: Policy and Practice, 129, 1-18.
Deb, S., Strawderman, L., Carruth, D.W., DuBien, J., Smith, B. and Garrison, T.M., 2017.
Development and validation of a questionnaire to assess pedestrian receptivity toward
fully autonomous vehicles. Transportation Research Part C: Emerging Technologies, 84,
178-195.
Dixit, V.V., Chand, S. and Nair, D.J., 2016. Autonomous Vehicles: Disengagements,
Accidents and Reaction Times. PLOS ONE, 11(12), e0168054.
Eid, M.I., 2011. Determinants of e-commerce customer satisfaction, trust, and loyalty in
Saudi Arabia. Journal of Electronic Commerce Research, 12(1), 78.
Fagnant, D.J. and Kockelman, K., 2015. Preparing a nation for autonomous vehicles:
opportunities, barriers and policy recommendations. Transportation Research Part A:
Policy and Practice, 77, 167-181.
Fagnant, D.J. and Kockelman, K.M.J.T., 2018. Dynamic ride-sharing and fleet sizing for
a system of shared autonomous vehicles in Austin, Texas. 45(1), 143-158.

35
Gkartzonikas, C. and Gkritza, K., 2019. What have we learned? A review of stated
preference and choice studies on autonomous vehicles. Transportation Research Part C:
Emerging Technologies, 98, 323-337.
Haboucha, C.J., Ishaq, R. and Shiftan, Y., 2017. User preferences regarding autonomous
vehicles. Transportation Research Part C: Emerging Technologies, 78, 37-49.
Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E. and Tatham, R.L., 2010. Multivariate
Data Analysis. Prentice Hall Upper Saddle River, NJ.
Hein, D., Rauschnabel, P., He, J., Richter, L. and Ivens, B., 2018. What drives the
adoption of autonomous cars?, ICIS 2018 Proceedings.
Hohenberger, C., Spörrle, M. and Welpe, I.M., 2016. How and why do men and women
differ in their willingness to use automated cars? The influence of emotions across
different age groups. Transportation Research Part A: Policy and Practice, 94, 374-385.
Hu, L.T. and Bentler, P.M., 1999. Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A
Multidisciplinary Journal, 6(1), 1-55.
Jansson, J., 2011. Consumer eco‐innovation adoption: assessing attitudinal factors and
perceived product characteristics. Business Strategy and the Environment, 20(3), 192-
210.
Johnson, D.S., Bardhi, F. and Dunn, D.T., 2008. Understanding how technology
paradoxes affect customer satisfaction with self‐service technology: The role of
performance ambiguity and trust in technology. Psychology & Marketing, 25(5), 416-
443.
Kaur, K. and Rampersad, G., 2018. Trust in driverless cars: Investigating key factors
influencing the adoption of driverless cars. Journal of Engineering and Technology
Management, 48, 87-96.
Kaye, S.A., Lewis, I., Forward, S. and Delhomme, P., 2020. A priori acceptance of highly
automated cars in Australia, France, and Sweden: A theoretically-informed investigation
guided by the TPB and UTAUT. Accident Analysis and Prevention, 137.
Kim, J.H., Lee, G., Park, J.Y., Hong, J. and Park, J., 2019. Consumer intentions to
purchase battery electric vehicles in Korea. Energy Policy, 132, 736-743.
König, M. and Neumayr, L., 2017. Users’ resistance towards radical innovations: The
case of the self-driving car. Transportation Research Part F: Traffic Psychology and
Behaviour, 44, 42-52.
Kotler, P.J. and Armstrong, G.M., 2010. Principles of Marketing. Pearson Education.
Krueger, R., Rashidi, T.H. and Rose, J.M., 2016. Preferences for shared autonomous
vehicles. Transportation Research Part C: Emerging Technologies, 69, 343-355.
Kyriakidis, M., Happee, R. and de Winter, J.C.F., 2015. Public opinion on automated
driving: Results of an international questionnaire among 5000 respondents.
Transportation Research Part F: Traffic Psychology and Behaviour, 32, 127-140.
Lavasani, M., Jin, X. and Du, Y., 2016. Market Penetration Model for Autonomous
Vehicles on the Basis of Earlier Technology Adoption Experience. Transportation
Research Record, 2597(1), 67-74.
Lee, J.D. and Kolodge, K., 2020. Exploring Trust in Self-Driving Vehicles Through Text
Analysis. Human Factors, 62(2), 260-277.
Liljamo, T., Liimatainen, H. and Pöllänen, M., 2018. Attitudes and concerns on
automated vehicles. Transportation Research Part F: Traffic Psychology and Behaviour,
59, 24-44.
Liu, F., Zhao, F., Liu, Z. and Hao, H., 2019a. Can autonomous vehicle reduce greenhouse
gas emissions? A country-level evaluation. Energy Policy, 132, 462-473.

36
Liu, P., Yang, R. and Xu, Z., 2019b. Public acceptance of fully automated driving: effects
of social trust and risk/benefit perceptions. Risk Analysis, 39(2), 326-341.
May, K.R., Noah, B.E. and Walker, B.N., 2017. Driving acceptance: Applying structural
equation modeling to in-vehicle automation acceptance, Proceedings of the 9th
International Conference on Automotive User Interfaces and Interactive Vehicular
Applications Adjunct. ACM, pp. 190-194.
Menon, N., Zhang, Y., Rawoof Pinjari, A. and Mannering, F., 2020. A statistical analysis
of consumer perceptions towards automated vehicles and their intended adoption.
Transportation Planning and Technology, 1-26.
Milakis, D., Van Arem, B. and Van Wee, B., 2017. Policy and society related implications
of automated driving: A review of literature and directions for future research. Journal of
Intelligent Transportation Systems, 21(4), 324-348.
Miyazaki, K. and Kijima, K., 2000. Complexity in Technology Management: Theoretical
Analysis and Case Study of Automobile Sector in Japan. Technological Forecasting and
Social Change, 64(1), 39-54.
Molnar, L.J., Ryan, L.H., Pradhan, A.K., Eby, D.W., St. Louis, R.M. and Zakrajsek, J.S.,
2018. Understanding trust and acceptance of automated vehicles: An exploratory
simulator study of transfer of control between automated and manual driving.
Transportation Research Part F: Traffic Psychology and Behaviour, 58, 319-328.
Montoro, L., Useche, S.A., Alonso, F., Lijarcio, I., Bosó-Seguí, P. and Martí-Belda, A.,
2019. Perceived safety and attributed value as predictors of the intention to use
autonomous vehicles: A national study with Spanish drivers. Safety Science, 120, 865-
876.
Moons, I. and De Pelsmacker, P., 2015. An extended decomposed theory of planned
behaviour to predict the usage intention of the electric car: A multi-group comparison.
Sustainability, 7(5), 6212-6245.
Narayanan, S., Chaniotakis, E. and Antoniou, C., 2020. Shared autonomous vehicle
services: A comprehensive review. Transportation Research Part C: Emerging
Technologies, 111, 255-293.
Nieuwenhuijsen, J., Correia, G.H.d.A., Milakis, D., van Arem, B. and van Daalen, E.,
2018. Towards a quantitative method to analyze the long-term innovation diffusion of
automated vehicles technology using system dynamics. Transportation Research Part C:
Emerging Technologies, 86, 300-327.
Nordhoff, S., De Winter, J., Kyriakidis, M., Van Arem, B. and Happee, R., 2018.
Acceptance of driverless vehicles: results from a large cross-national questionnaire study.
Journal of Advanced Transportation, 2018.
Panagiotopoulos, I. and Dimitrakopoulos, G., 2018. An empirical investigation on
consumers’ intentions towards autonomous driving. Transportation Research Part C:
Emerging Technologies, 95, 773-784.
Payre, W., Cestac, J. and Delhomme, P., 2014. Intention to use a fully automated car:
Attitudes and a priori acceptability. Transportation Research Part F: Traffic Psychology
and Behaviour, 27, 252-263.
Penmetsa, P., Adanu, E.K., Wood, D., Wang, T. and Jones, S.L., 2019. Perceptions and
expectations of autonomous vehicles–A snapshot of vulnerable road user opinion.
Technological Forecasting and Social Change, 143, 9-13.
Petschnig, M., Heidenreich, S. and Spieth, P., 2014. Innovative alternatives take action–
Investigating determinants of alternative fuel vehicle adoption. Transportation Research
Part A: Policy and Practice, 61, 68-83.

37
Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y. and Podsakoff, N.P., 2003. Common
method biases in behavioral research: a critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), 879.
Rogers, E.M., 2003. Diffusion of Innovations, 5th edition ed. Simon and Schuster.
Schellekens, M., 2015. Self-driving cars and the chilling effect of liability law. Computer
Law & Security Review, 31(4), 506-517.
Souris, C., Theofilatos, A. and Yannis, G., 2019. Attitudes of Greek drivers towards
autonomous vehicles-a preliminary analysis using stated preference approach. Advances
in Transportation Studies, 48.
Strömberg, H., Rexfelt, O., Karlsson, I.C.M. and Sochor, J., 2016. Trying on change–
Trialability as a change moderator for sustainable travel behaviour. Travel Behaviour and
Society, 4, 60-68.
Talebian, A. and Mishra, S., 2018. Predicting the adoption of connected autonomous
vehicles: A new approach based on the theory of diffusion of innovations. Transportation
Research Part C: Emerging Technologies, 95, 363-380.
Vasebi, S., Hayeri, Y.M., Samaras, C. and Hendrickson, C., 2018. Low-level automated
light-duty vehicle technologies provide opportunities to reduce fuel consumption.
Transportation Research Record, 2672(24), 60-74.
Wang, X., Yuen, K.F., Wong, Y.D. and Teo, C.C., 2018. An innovation diffusion
perspective of e-consumers’ initial adoption of self-collection service via automated
parcel station. The International Journal of Logistics Management, 29(1), 237-260.
Wey, W.-M. and Huang, J.-Y., 2018. Urban sustainable transportation planning strategies
for livable City's quality of life. Habitat International, 82, 9-27.
Xu, Z., Zhang, K., Min, H., Wang, Z., Zhao, X. and Liu, P., 2018. What drives people to
accept automated vehicles? Findings from a field experiment. Transportation Research
Part C: Emerging Technologies, 95, 320-334.
Yuen, K.F., Wang, X., Ma, F. and Wong, Y.D., 2019. The determinants of customers’
intention to use smart lockers for last-mile deliveries. Journal of Retailing and Consumer
Services, 49, 316-326.
Yuen, K.F., Wang, X., Ng, L.T.W. and Wong, Y.D., 2018a. An investigation of
customers’ intention to use self-collection services for last-mile delivery. Transport
Policy, 66, 1-8.
Yuen, K.F., Wang, X., Wong, Y.D. and Zhou, Q., 2018b. The effect of sustainable
shipping practices on shippers’ loyalty: The mediating role of perceived value, trust and
transaction cost. Transportation Research Part E: Logistics and Transportation Review,
116, 123-135.
Zhang, T., Tao, D., Qu, X., Zhang, X., Lin, R. and Zhang, W., 2019. The roles of initial
trust and perceived risk in public’s acceptance of automated vehicles. Transportation
Research Part C: Emerging Technologies, 98, 207-220.
Zhang, T., Tao, D., Qu, X., Zhang, X., Zeng, J., Zhu, H. and Zhu, H., 2020. Automated
vehicle acceptance in China: Social influence and initial trust are key determinants.
Transportation Research Part C: Emerging Technologies, 112, 220-233.

38

View publication stats

You might also like