Investigation into the Antecedents of Autonomous Car Acceptance using an Enhanced UTAUT Model

Full text

Investigation into the Antecedents of Autonomous
Car Acceptance using an Enhanced UTAUT Model
Nathan Kettles and Jean-Paul Van Belle
CITANDA, Department of Information Systems
University of Cape Town
Cape Town, South Africa
nathan@kettles.co.za, Jean-Paul.VanBelle@uct.ac.za
Abstract— This paper investigates the intention of South
African drivers to use Autonomous Vehicles (AVs), vehicles
where the full-time performance can be handled by an
automated driving system without or with minimal human
input. The behaviour intention of people to use such vehicles is
of interest due to questions of trust, safety, performance, and
enjoyment. Little to no research of this kind had been done in
South Africa before, and the field of study is still relatively new
globally. An appropriately extended version of the Unified
Theory of Acceptance and Use of Technology (UTAUT) was
used as the basis for this research. A survey was completed by
121 respondents. Hedonic Motivation (β = -.19, p < .05) and
Performance Expectancy (β = -.053, p < .05) were the only
statistically significant predictors of Behavioural Intentions to
use Autonomous Vehicles. However, Trust in Safety and Social
Influence also exhibit strong correlations. The results validate
the findings of similar studies.
Keywords—Autonomous Vehicles (AVs), Self-driving Cars,
Hedonic Motivation, Unified Theory of Acceptance and Use of
Technology (UTAUT), South Africa.
I. INTRODUCTION
Autonomous vehicles, in some form or another, have been
around for many years, at least since the early 1990’s [1].
From automated cruise control, to parking assistance and lane
guidance, the prevalence of autonomy in vehicles at basic
levels is in many cases universally available wherever cars are
available [1]. However, in more recent years, the industry’s
momentum towards higher levels of automation has increased
– with a publicly announced focus from large traditional
vehicle manufacturers as well as big players in the consumer
electronics industry – the likelihood of fully or highly
automated vehicles becoming the new norm appears to be
growing [2][3][4].
Fully autonomous vehicles as defined by the Society of
Automotive Engineers (SAE) level 5 is “the full-time
performance by an automated driving system of all aspects of
the dynamic driving task under all roadway and environmental
conditions that can be managed by a human driver” [5] which
is to say that the vehicle’s autonomous system would be
responsible in all modes and for all tasks of driving, including
monitoring the environment, without the need for any human
intervention. SAE level 4 is similar to level 5 except that it
does not support all driving modes: the autonomy may be
limited to specific areas (such as within certain geographical
areas) or only available in certain circumstances such as
bumper-to-bumper traffic. Level 4 cars are already in limited
testing phases across the United States, European countries
(including the United Kingdom).
These vehicles are expected to bring many benefits across
environmental, economic, and social spheres. Some of these
benefits include fewer accidents, reduced congestion, shorter
travel times, lower emissions, reduced fuel consumption,
greater mobility and independence for mobility-challenged
seniors and those with physical, cognitive, or mental
disabilities [3][6]. There are some negatives though, such as
the high cost of autonomous vehicles, loss of feeling of
control, lower driving enjoyment, less pleasure from the
overall experience, and the likely change to reduced car
ownership or a car-free future [7].
Autonomous vehicles would need to gain mainstream user
acceptance to get high levels of market penetration [3][6]. It is
this, the user acceptance of autonomous vehicles, that this
study will focus on. While there is much discourse and
attention in the public media towards the challenges to
develop a market-ready product of level 4 autonomous car (as
well as a level 5 fully autonomous vehicle), the fact that there
is still little empirical knowledge of the level of acceptance of
autonomous vehicles is the reason we need to separate out the
two broad issues of product-readiness (and everything it
entails), and the user acceptance of autonomous vehicles
[3][8][9].
It is known that level 4 cars are already in limited-testing
phases, however, there is still a long way to go before these
vehicles will be ready for full use on any given road or weather
condition. At the same time, the rate of technological
innovation is extremely high, and we should not
underestimate the pace at which these companies, particularly
those like Waymo and Uber, can make significant progress
[4][6][10]. It is with this in mind that research around
acceptance of autonomous vehicles is both relevant and
important at this time.
The research question is: What are the drivers and
perceived issues in user acceptance of autonomous cars? In
answering the research question, the following objectives will
be achieved:
 Determine the influence of the determinants (of
performance expectancy, effort expectancy, social
influence, and enjoyment) on behavioural intention to use
autonomous cars
 Determine the degree to which gender, age, and
experience play a role.

II. LITERATURE REVIEW
Autonomous vehicles, as an emerging and developing
technology, has the potential to revolutionise the transport
industry. Although the date may still be uncertain, the likely
and eventual replacement of traditional vehicles with
autonomous ones will have far reaching consequences [11].
Many researchers as well as automotive companies and
organisations are studying the various aspects of autonomous
vehicles, including the legal, technological, regulatory, social,
and financial aspects. And the reality is that autonomous
vehicles are no longer a fantasy, they are will soon become
reality for many as the technology matures [2].
In a large survey across 109 different countries, 69% of
respondents believe that autonomous vehicles will reach 50%
market share between now and 2050 [12]. The same study also
found that 51% of respondents believe that automated driving
will become so advanced and widespread that they would not
be allowed to drive manually in 30 years [12]. Understanding
the acceptance of these vehicles is thus a very important
research topic [13].
Autonomous vehicles fall into a broad category of new and
emerging transportation technologies. These include
autonomous vehicles, electric vehicles, and pod platoons (a
combination of autonomous and electric vehicles specific for
public transit) [4]. SAE International, an automotive
standardisation body, originally released a classification
system of six levels in 2014. It has since been slightly updated
in 2016. The levels range from Level 0 (completely manual
driving with full human interaction) to Level 5 (fully
automated driving without any human interaction) [5]. This is
the most broadly-used classification standard, with the likes of
the United States’ National Highway Traffic Safety
Administration (NHTSA) and others choosing to use the
autonomy level definitions released by SAE International
instead of their own or other classification systems.
Understanding the levels as defined by SAE International
is important when reviewing the literature and studies of
autonomous vehicles. SAE International [5] defines the
following levels: Level 0, named “no automation”, where the
human driver is responsible for all driving tasks. Level 1,
named “driver assistance”, where the human driver receives
some driver assistance in the form of steering or
acceleration/deceleration. Level 2, named “partial
automation” is where the automation assists both with steering
and acceleration/deceleration using information gathered
from the driving environment. The human driver is still in full
control. Level 3, named “conditional automation” is the level
where the automation system starts to monitor the driving
environment. In this level, the vehicle can drive itself in
certain conditions, and where unable to will request the human
driver to intervene (if the human driver does not, the car will
gradually come to a stop). Level 4, named “high automation”,
is where the automation system can execute all driving tasks
in specific conditions (perhaps certain geo-locations, etc.),
regardless of whether the human drivers responds to a request
to intervene. Level 5, named “full automation”, is the full-time
handling of all driving modes by the automation system,
essentially replacing the human driver completely. A vehicle
at Level 5 would never require a human to engage in driving
activities. [5]
A. Benefits of autonomous vehicles
Autonomous vehicles are expected to bring many benefits
across environmental, economic, and social spheres. With
regards to safety, existing advanced driving assistance
systems are already able to capture, analyse, and react to data;
reacting faster and in a far more reliable manner than humans
[3]. Since over 40% of fatal crashes involve a combination of
drunk driving, drugs, driver distraction, tiredness, and fatigue,
fully autonomous vehicles should ultimately lead to a very
significant reduction in the number of vehicle accidents [6].
This is especially important, given that traffic accidents are the
leading cause of death among young adults aged 15 to 29 years
old [11]. In addition, general driver error is not reflected in the
causes above [6]. Driver error — which includes crashes due
to aggressive driving, speeding, slow reaction times, lack of
driving experience, poor attention to road/driving conditions
— is believed to be the main reason behind over 90% of all
crashes in the United States [6]. Thus the potential safety
benefits of autonomous vehicles are substantial, given that
over 30 000 people die each year in traffic accidents, 2.2
million crashes result in injury, and the annual economic cost
of crashes is a staggering 277 billion US dollars.
This large cost has a ripple effect throughout society,
including medical and emergency service costs, workplace
losses and productivity, lost quality of life, as well as legal,
insurance, and administration costs [14]. While there are
currently still technical challenges to get autonomous vehicles
to perform safely in nearly all situations, including complex
ones, many analysts predict that autonomous vehicles will
overcome these obstacles, with fatality rates reaching those
seen in aviation or rail, and eventually an end goal of crash-
less cars [2][6].
Autonomous vehicles are also expected to relieve
congestion and reduce fuel consumption. As an example, they
could sense and predict the braking and acceleration actions
of vehicles in front of them, which could result in smoother
braking and speed adjustment, thus reducing fuel consumption
[6]. In a similar way, autonomous vehicles would use existing
roads and lanes more efficiently, driving closer together in an
almost platoon-like way, and choosing more efficient routes,
resulting in relieved congestion [6]. While a discussion on the
effect of autonomous vehicles on congestion is speculative or
hypothetical, it is argued that, on the whole, autonomous
vehicle technology could lead to significantly less
congestion[11].
While autonomous vehicles will likely increase vehicle
miles travelled (VMT) per capita, they would also drastically
increase vehicle throughput [11]. Because autonomous
vehicles are able to constantly monitor their surroundings,
they will be able to travel at much higher speeds with reduced
space between other vehicles. Autonomous vehicles would
help avoid the inefficient start-and-stop conditions of
congestion, caused by delayed or slow responses of human
drivers in high speed and high throughput conditions. In
addition, as argued by other researchers, the increase in safety
(mentioned previously) would also have a positive impact on
congestion, as traffic incidents account for about 25% of all
congestion delays [11][14].
Due to less causing less congestion, autonomous vehicles
are also expected to have positive environmental benefits [11].
The optimised driving and technology gained from
autonomous vehicles results in “eco-driving”, which is
proving to increase fuel economy by 4-10% [14]. However,

the greatest benefit for environment will come when
combining autonomous vehicles with electric vehicles. In that
case, the environmental benefits are expected to be very
significant [14].
An eventual outcome benefit of autonomous vehicles is
greater mobility for those who are physically, mentally, or
otherwise impaired [11][14]. The many positive benefits of
AVs for those who currently cannot drive for themselves are
also evident [8].
B. Factors accounting for autonomous vehicle acceptance
There is a growing but still small amount of research
around autonomous vehicle adoption [15][16][17][18]. Using
an adapted Technology Acceptance Model (TAM), [3]
postulated that personal driving enjoyment and perceived
traffic safety were the two opposing factors that might play a
major role in acceptance. Specifically, they highlighted that
many respondents enjoy the physical act of driving, and that
autonomous vehicle manufacturers should look to emphasise
the hedonic aspects afforded by AVs.
The UTAUT model was combined with some other
models/frameworks to develop a full acceptance model for
autonomous vehicles, though at the time of publishing it had
not been empirically validated yet [4]. Additional constructs
were added to UTAUT to arrive at what was proposed as the
Car Technology Acceptance Model (CTAM) [19]. A study in
Austin, Texas (a city where autonomous vehicles have been
tested), concluded that for 50% of the population, the adoption
rates of autonomous vehicles would depend on the adoptions
rates of their friends and neighbours [2]. This survey of 347
people found that the average willingness to pay for fully
automated vehicles is much higher than that of the automation
level below that. In their study, more than 80% of respondents
indicated that they were interested in owning fully automated
vehicles. Interestingly, they found that the biggest benefit of
autonomous vehicles would be lower rates of vehicle crashes
(safety), and that less congestion would be the least likely
benefit. As with other studies, they found that people who
drive more are more likely to buy and adopt autonomous
vehicles [2].
Another study looked at the acceptance and adoption of
Automated Road Transport Systems (ARTS) in the context of
the EU-funded project CityMobil2 [13]. ARTS are form of
transport complimentary to public transport, which add extra
supply options to areas of low or dispersed demand (Madigan
et al., 2016). The vehicle looks like a small pod of a bus, holds
12 people, and travels at speeds between 12-45km/h. Using
the Unified Theory of Acceptance and Use of Technology
(UTAUT), [13] collected 349 valid responses across France
and Switzerland. They found that all three constructs of
UTAUT have an impact on the intention to use ARTS.
Further, they found that the most important factor for
predicting usage was how well people believed ARTS
performed in comparison to other public transport systems.
They suggested that UTAUT doesn’t currently capture all the
factors that could influence an individual’s behaviour to use
autonomous vehicles, stating that the hedonic motivation is
perhaps an important determinant when looking at
behavioural intention to use in consumer settings.
III. METHODOLOGY
This research aims to extend and add to some of the
research already done in other countries with regards to
autonomous vehicles, specifically in our case with regards to
autonomous cars.
Given the extended UTAUT research model, the
following hypotheses can be formulated.
• H1: Performance expectancy, effort expectancy,
social influence, trust in safety, hedonic motivation, resources
& knowledge have an impact of behavioural intention to use
autonomous cars.
• H2: Gender, age and driving experience have
moderating influences on the relationships between the above
antecedents and intention to use autonomous cars.
The research philosophy used in this research is
positivism: the researcher is external, objective, and
independent of any social actors [20]. A quantitative approach
will be taken by means of a predetermined questionnaire and
structured data collection technique; this fits with the
positivist philosophy. We will be making use of validated
theory to work and conduct our research, a theory which lays
out a set of concepts and relationships that explain our
phenomenon of interest; this implies a deductive approach
The research timeframe is cross-sectional as it studies
particular or phenomena at a particular time [20].
The research model that was used is the UTAUT Model
by Venkatesh and others [21][22]. It was slightly expanded to
be more applicable to the study. It was decided to take the
modified UTAUT model in a 2016 study in a similar field [13]
and add the determinant of ‘Hedonic Motivation’ (also known
as ‘enjoyment’) to it, thus given a tested model with an added
determinant “Resources and Knowledge” as supported by the
literature review [3].
Fig. 1. Research Model based on UTAUT [21]
The research instrument was a questionnaire be
administered online via the internet. The questionnaire items
were adapted from the referenced studies [3][13][22]. The test
items were measured on a 7-point Likert scale ranged from 1
= “strongly disagree” to 7 = “strongly agree”.
This study targets people who drive in South Africa. A
non-probability convenience sampling approach was used:
people in the financial and technology industries were
approached via email, social media, and similar methods. This
sampling method is not be fully representative of the South
African population, however this method allows us to use
minimal resources and is also of a practical nature given the
time and budget constraints. A further research limitation is
that autonomous cars (SAE levels 4 or 5) are not currently

being tested in South Africa, so the respondents answered with
very limited experiential knowledge of what autonomous cars
are. With this in mind, the questionnaire assumed no prior
knowledge of autonomous cars and explained clearly what
AVs are.
The research and the instrument was approved by the UCT
ethics committee. All respondents were over the age of 18, and
privacy and confidentiality were safeguarded. Participation in
the research was voluntary, and the respondent could
withdraw at any time.
IV. 4. DATA ANALYSIS AND FINDINGS
A. Descriptive analysis
Out of the 121 respondents, there were 65 females (54%)
and 55 males (45%); 1 respondent preferred not to answer.
Fig. 2. Respondents’ age
The respondents were aged 25-34 years (36.3%), 35-44
years (25.6%), 45-54 years (19%) (Fig. 2). Very few were
aged 55 years and few aged 18-24 years. A fifth (20.8%) of
respondents had less than 10 years of driving experience, a
third (34.8%, the modal category) had between 11 to 20 years
of such experience, while a quarter (24.4%) had between 21
to 30 years’ experience. A fifth (20.0%) had more than 30
years of driving experience. Eight out of ten (81.8%)
respondent’s indicated that they had motorised vehicle that
they drive themselves. Very few of the respondents used a
train, walk, bus, and taxi.
Six out of every ten (60.6%) respondents felt that they did
not plan to use AVs within the first six months of local
availability; a fifth (20%) felt so and another fifth (20%) were
uncertain. The majority felt that if it were affordable, they
would use AVs in the future (82%), while almost three
quarters (75%) felt they intended to use AVs in the future. Few
(37%) felt they were prepared to pay a premium price to use
AVs when compared to similar non-autonomous vehicles,
43% were prepared and a fifth uncertain. Respondents
generally agreed that if AVs were affordable, they would use
AVs in the future (M=5.5) and that they intend to use AVs in
the future (M=2.2). They were generally unsure if they were
prepared to pay a premium price to use AVs when compared
to similar non-autonomous vehicles.
Fig. 3. Box plot for observed construct values
The mean for Hedonic Motivation is the highest
(Mean=5.3), followed by Performance Expectation (M=5.0),
Effort Expectancy (4.8) and Trust in Safety (M=4.7) (Fig.3).
The three lowest ranked were Behaviour Intention (M=4.4),
Social Influence (M=4.2) and least Resources and Knowledge
(M=3.9). These results suggest that the respondents were more
in agreement for HM, and PE; weak agreement for EE and TS,
doubtful towards slight agreement for BI, SI and RK.
B. Reliability tests
TABLE I. CRONBACH'S ALPHA, MEAN SCORES
Variable
Obs
alpha
Effort expectancy (EE)
120
0.789
Hedonic Motivation (HM)
119
0.753
Trust in Safety (TI)
114
0.752
Performance Expectancy (PE)
109
0.717
Social Influence (SI)
109
0.753
Resources and Knowledge
(RK)
109
0.809
Behavioural Intention (BI)
109
0.714
Mean (unstandardized items)
0.784
The alpha coefficient of all the variables is above the
recommended 0.7 which is considered acceptable. This
confirms that the internal consistency of the questions was
acceptable to use as an instrument, which gives us reliability.
This was not surprising given that the test items associated
with the instrument have been widely used and tested.
C. Antecedents for Behavioural Intention (BI)
The first analysis looked at the correlation between the
hypothesized antecedents and intention to adopt AVs. The
Pearson correlation coefficient measures the strength and
direction of the relationship between the two variables.
TABLE II. PEARSON'S CORRELATION WITH INTENTION
Antecedent
(sorted by correlation)
r with BI
Strength
Performance Expectancy (PE)
0.7115*
Very strong
Hedonic Motivation (HM)
0.5717*
Strong
Trust in Safety (TI)
0.4894*
Strong/moderate
Social Influence (SI)
0.4785*
Strong/moderate
Resources and Knowledge (RK)
0.2846*
Weak
02468
Effort expectancy Hedonic Motivation
Trust in Safety Performance Expectancy
Social Influence Resources and Knowledge
Behavioral Intention

Antecedent
(sorted by correlation)
r with BI
Strength
Effort expectancy (EE)
0.2692*
Weak
Experience
-0.2025*
Weak
Age Group
-0.1996*
Weak
Gender
-0.1125
Not significant
The analysis (Table II) shows that all hypothesized BI
antecedents are positively and statistically significantly
correlated with Behavioural Intention (BI), all with a
correlation exceeding 25%. PE (r = .71, p < .05) had the
highest and positive association with Behaviour Intention
(BI), results significant at 5% level. A unit increase in PE is
associated with a 0.71 times unit increase in BI. This is
followed by HM (r = 0.57, p < .05), also significant and
moderate association, which could have as much as 0.57 times
influence on BI. SI (β = .48, p < .05) and TS (β = .49, p < .05)
could have as much as 0.48 times positive influence on BI. EE
(β = .27, p < .05) and RK (β = .28, p < .05) has weak positive
associations with BI, a unit increase in these could lead to a
potential about .27 times increase/influence in BI. Age group,
experience, and gender seem to have negative weak
associations with BI, but the gender correlation is not
statistically significant.
The Pearson correlation analysis analyses each variable
individually. However, a number of antecedents are also
correlated with each other. To ignore the effect of the
collinearities, a SEM analysis was also conducted. Figure 4
shows a pathway of factors that drivers and perceived issues
in user acceptance of autonomous cars, taking into account the
correlations between the hypothesized antecedents.
Fig. 4. SEM Pathway analysis
Hedonic Motivation (β = -.19, p < .05) and Performance
Expectancy (β = -.053, p < .05) were the only statistically
significant predictors of Behavioural intentions to use
Autonomous Vehicles. These results suggest that PE and HM
have 53% and 19% positive direct effect respectively on the
respondents’ behavioural intentions to use Autonomous
Vehicles.
Effort Expectancy (β =-.18, p > .05) and Social Influence
(β =-.12, p < .05) were not statistically significant predictors
of Behavioural Intention to use Autonomous Vehicles. These
results suggest that EE and SI could potentially have 18% and
12% positive direct effects respectively on behavioural
intentions to use Autonomous Vehicles.
Trust in Safety (β = -.0.10, p > .05) and Resources and
Knowledge (β = -.03, p > .05) had the lowest effects as
predictors of Behavioural intentions to use Autonomous
Vehicles. However, these results provide an insight that PE
and HM could potentially have 11% and 3% positive effects
respectively on the respondent’s behavioural intentions to use
Autonomous Vehicles.
V. CONCLUSION
This paper evaluated the antecedents relating to
autonomous car acceptance by using an adapted UTUAT
model. Using the constructs of Effort expectancy, Hedonic
Motivation, Trust in Safety, Performance Expectancy, Social
Influence, and Resources and Knowledge, we probed the
extent to which they influenced or impact the Behaviour
Intention to use Autonomous Vehicles.
The research confirms that Performance Expectancy and
Hedonic Motivation (also known as Enjoyment) are the most
significant predictors of Behaviour Intention; this also
confirms the findings of prior research elsewhere. Thus, South
African drivers are not unique in wanting to get productivity
and personal enjoyment (or pleasure) out of autonomous cars.
This illustrates the importance that motor manufacturers
should place on these things as the autonomous car industry
evolves.
Trust in Safety and Social Influence also show a strong
correlation with intentions to buy AVs. Resources &
Knowledge, and Effort Expectancy are also antecedents of
Behaviour Intention, but to a much smaller level of similar
research studies in other countries. Thus H1 was accepted.
Driving experience – and the highly correlated variable
age - did play in role in behaviour intention, but had a negative
correlation: older, more experienced drivers have slightly
lower intentions to adopt AVs. Gender had no significant
impact. Thus H2 was rejected.
A. Limitations of this Research
With a low sample size (just over 100 observations), it is
difficult to generalise from this research. There are millions of
drivers in South Africa, and as such it cannot be expected that
our research is truly representative.
Another limitation of this research was that autonomous
vehicles are not widely or well known in South Africa, and we
don’t have any publicly known trials happening either. This
means that most of the knowledge and information that
respondents were aware of is not first-hand and could be
biased by the media or the channels they got the information
from. Conducting this research with people in cities where
there are autonomous car trials could be more beneficial to the
body of research.
B. Recommendations for Further Research
Studies similar to this one using TAM or UTUAT are far
more common than qualitative ones at this stage. It is believed
that the qualitative research could supplement this field of
study in a good way. A hybrid research study consisting of
both qualitative and quantitative research methods would
yield deeper insights. Also, it would also be advised to do a

wider scale study and investigate the age and gender effects
on the behavioural intention to use autonomous cars.
REFERENCES
[1] JCF de Winter, R Happee, MH Martens, & NA Stanton. 2014. Effects
of adaptive cruise control and highly automated driving on workload
and situation awareness: A review of the empirical evidence.
Transportation Research Part F: Traffic Psychology and Behaviour, 27,
196–217. http://doi.org/10.1016/j.trf.2014.06.016
[2] P Bansal, KM Kockelman, & A Singh. 2016. Assessing public opinions
of and interest in new vehicle technologies: An Austin perspective.
Transportation Research Part C, 67(C), 1–14.
http://doi.org/10.1016/j.trc.2016.01.019
[3] CPH Ernst, & P Reinelt. 2017. Autonomous Car Acceptance: Safety
vs. Personal Driving Enjoyment (pp. 1–8). Presented at the Twenty-
third Americas Conference on Information Systems, Boston.
[4] S Nordhoff. 2015. A Conceptual Model to Explain, Predict, and
Improve User Acceptance of Driverless Vehicles (pp. 1–19). Presented
at the 95th Annual Meeting of the Transportation Research Board,
January 2016 Washington D.C.
[5] SAE International, J3016. 2016. Taxonomy and Definitions for Terms
Related to Driving Automation Systems for On-Road Motor Vehicles.
SAE International. Retrieved from
https://www.sae.org/standards/content/j3016_201609
[6] DJ Fagnant, & K Kockelman. 2015. Preparing a nation for autonomous
vehicles: opportunities, barriers and policy recommendations.
Transportation Research Part A: Policy and Practice, 77, 167–181.
http://doi.org/10.1016/j.tra.2015.04.003
[7] S Nordhoff, J de Winter, M Kyriakidis, B van Arem, & R Happee.
2018. Acceptance of Driverless Vehicles: Results from a Large Cross-
National Questionnaire Study. Journal of Advanced Transportation,
2018, 1–22. http://doi.org/10.1155/2018/5382192
[8] P Bazilinskyy, M Kyriakidis, & J de Winter. 2015. An International
Crowdsourcing Study into People's Statements on Fully Automated
Driving. Procedia Manufacturing, 3(C), 2534–2542.
http://doi.org/10.1016/j.promfg.2015.07.540
[9] Y Sun, D Olaru, B Smith, S Greaves, & A Collins. 2017. Road to
autonomous vehicles in Australia: an exploratory literature review.
Road Transport Research, 26(1), 34.
[10] F Favarò, S Eurich, & N Nader. 2018. Autonomous vehicles’
disengagements: Trends, triggers, and regulatory limitations. Accident
Analysis and Prevention, 110, 136–148.
http://doi.org/10.1016/j.aap.2017.11.001
[11] JM Anderson, K Nidhi, KD Stanley, P Sorensen, C Samaras, & OA
Oluwatola. 2014. Autonomous Vehicle Technology: A Guide for
Policymakers. Santa Monica, California: Rand Corporation.
[12] M Kyriakidis, R Happee, & JCF de Winter. 2015. Public opinion on
automated driving: Results of an international questionnaire among
5000 respondents, 32(C), 127–140.
http://doi.org/10.1016/j.trf.2015.04.014
[13] R Madigan, T Louw, M Dziennus, T Graindorge, E Ortega, M
Graindorge, & N Merat. 2016. Acceptance of Automated Road
Transport Systems (ARTS): An Adaptation of the UTAUT Model.
Transportation Research Procedia, 14, 2217–2226.
http://doi.org/10.1016/j.trpro.2016.05.237
[14] SA Bagloee, M Tavana, M Asadi, & T Oliver. 2016. Autonomous
vehicles: challenges, opportunities, and future implications for
transportation policies. Journal of Modern Transportation, 24(4), 284–
303. http://doi.org/10.1007/s40534-016-0117-3
[15] S Osswald, D Wurhofer, S Trösterer, E Beck, & M Tscheligi. 2012.
Predicting information technology usage in the car: towards a car
technology acceptance model (pp. 51–9). Presented at the Proceedings
of the 4th International Conference on Automotive User Interfaces and
Interactive Vehicular Applications (AutomotiveUI '12), Portsmouth,
NH: ACM Press. http://doi.org/10.1145/2390256.2390264
[16] K Ruggeri, O. Kácha, IG Menezes, M Kos, M Franklin, L Parma, P
Langdon, B Matthews, B. an& J Miles. 2018. In with the new?
Generational differences shape population technology adoption
patterns in the age of self-driving vehicles. Journal of Engineering and
Technology Management, 50, 39-44.
[17] R Shabanpour, A Shamshiripour, & A Mohammadian. 2018. Modeling
adoption timing of autonomous vehicles: innovation diffusion
approach. Transportation, 45(6), 1607-1621.
[18] LM Hulse, H Xie, & ER Galea. 2017. Perceptions of autonomous
vehicles: Relationships with road users, risk, gender and age. Safety
Science, 102, 1–13. http://doi.org/10.1016/j.ssci.2017.10.001
[19] R Shabanpour, N Golshani, A Shamshiripour, & AK Mohammadian.
2018. Eliciting preferences for adoption of fully automated vehicles
using best-worst analysis. Transportation Research Part C: Emerging
Technologies, 93, 463-478.
[20] MNK Saunders, P Lewis, & A Thornhill. 2009. Research Methods for
Business Students (5 ed.). London: Pearson Education.
[21] V Venkatesh, MG Morris, FD Davis, & GB Davis. 2003. User
Acceptance of Information Technology: Toward a Unified View. MIS
Quarterly, 27(3), 425–478. http://doi.org/10.2307/30036540
[22] V Venkatesh, JYL Thong, & X Xu. 2012. Consumer Acceptance and
Use of Information Technology: Extending the Unified Theory of
Acceptance and Use of Technology. MIS Quarterly, 36(1), 157–178.