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

net/publication/265883429

Acceptability of Electric Vehicles: Findings from a driver survey

Article · January 2014

CITATIONS READS

15 3,277

5 authors, including:

Fakhra Jabeen Doina Olaru

Jazan University University of Western Australia
16 PUBLICATIONS   194 CITATIONS    110 PUBLICATIONS   1,307 CITATIONS   

SEE PROFILE SEE PROFILE

Brett W Smith Thomas Braunl

University of Western Australia University of Western Australia
55 PUBLICATIONS   540 CITATIONS    197 PUBLICATIONS   2,110 CITATIONS   

SEE PROFILE SEE PROFILE

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

Access and Accessibility View project

Robotics & Automation View project

All content following this page was uploaded by Doina Olaru on 15 January 2015.

The user has requested enhancement of the downloaded file.

ACCEPTABILITY OF ELECTRIC VEHICLES: FINDINGS FROM
A DRIVER SURVEY
Fakhra Jabeen1, Doina Olaru1, Brett Smith1, Thomas Braunl2, Stuart Speidel2
1
The University of Western Australia, Business School, M261, Perth, Australia
2
The University of Western Australia, School of Electrical and Computer Engineering, Perth, Australia
Email for correspondence: 20873104@student.uwa.edu.au

Abstract
Plug-in Electric Vehicles (EV) offer a clean and cost effective means in the long run of driving
short to medium distances within the city, even with the current high purchase cost. In Australia
EV may be attractive as a second car in the multicar household. The acceptance of EV requires
a change in behaviour – instead of re-fuelling, this vehicle requires battery charging each 140-
160km, either at home or at specialised charging stations.

A limited number of EVs are being driven in Perth as part of the Western Australia Electric
Vehicle trial (WA EV trial). The trial monitors the performance, benefits, infrastructure and
practical implications of EV fleet. This paper explores the opinions and experiences of 43 of the
participants. Factor analysis and multiple regression are applied to identify the main motivators
and barriers in purchasing and using an EV.

Ninety per cent of respondents are confident about driving the EV; more than 45% take trips of
more than 30km. While zero tailpipe emissions is the most desirable feature of EV, followed
closely by home charging, the limited range of the vehicle is regarded as the most serious
barrier to EV uptake. The overall satisfaction with the EV performance is high (an average score
of 3.96 out of 5), although 13 participants experienced at least one technical difficulty, when
driving the EVs in the trial.

Two latent constructs reflecting environmental concerns, and technology learning, along with
EV benefits and technical difficulties experienced while driving an EV explain 59.2% of the
variability of the willingness to purchase an EV as the next vehicle.

Key words: Electric vehicle, multivariate analysis, drivers’ attitudes.

 2012 ATRF Proceedings

1. Introduction
The increased demand for fossil fuels requires investigation of other energy sources in transport
planning. The plug-in Electric Vehicle (EV) is driven by electricity, using an electric motor
instead of a petrol or diesel engine. EV has distinct characteristics, for example limited driving
range, battery re-charging and zero tailpipe emissions. In addition, EV brings benefits in terms
of low running costs. People’s acceptance of new fuels and vehicles are determinants of the
EV’s place in the ensemble of vehicle technologies. The number of kilometres travelled on one
charge and the need for frequent charging are factors influencing the purchase and use of an
EV, along with the efficiency of the vehicle (weekly $ amount spent on travelling) and comfort.
Individuals are likely to trade-off these features and their decision is also affected by attitudes,
preferences, and habits.

Many Australian households use more than one car (ABS, 2008) so that the range limitation of
EV may not be considered an issue when there is a second car available for long distance trips.
With the low travel cost, EVs have the greatest potential for short trips within the city, but the
charging requires good trip planning.

A limited number of EVs are in use as part of an EV trial in Perth, Western Australia. The trial
monitors the performance, benefits, infrastructure, and practical implications of EV fleets. This
study aims to find the perceived barriers to the purchase and use of both converted and
commercially manufactured EV. A questionnaire was presented to the drivers in the WA EV
trial. Because the vehicles in the trial are all converted EVs, only four respondents use
manufactured EVs, with one having experience with both converted and commercially available
EV. In terms of sample size, number of manufactured EV drivers is small due to the limited
availability of EV in the Western Australian market.

In general, most of the drivers are confident in operating the EV, although 13 participants
experienced at least one technical difficulty when driving the converted EVs in the trial. The
overall satisfaction with the EV performance is still high with average score being 3.96 out of 5.

The two techniques used in this study include factor analysis and multiple linear regression. The
results of the survey are analysed by testing a set of hypotheses through the regression model.

1.1 Aims of the Study

This study explores the drivers’ behaviour through a survey with the following aims:
• Identifying drivers’ perceptions about EV, and their willingness to purchase an EV;
• Ascertaining participants’ attitudes towards the environment and adoption of new
technologies;
• Informing the research program and assisting in refining the design of the questionnaire
for the household survey that will be conducted separately. The EV driver survey serves
thus as a pilot, testing two sections of the household questionnaire: a stated choice
experiment and household attitudes towards EV. This study will assist in distinguishing
the most relevant characteristics for EV purchase, as well as testing the reliability of
several latent constructs necessary in capturing households’ preference heterogeneity.

The next section discusses the literature about EV uptake, followed by a conceptual model for
the adoption of EV (Section 3), and the data and methodology (Section 4). The findings of this
research are discussed next (Section 5) and the last section conveys the conclusions of the
study.
 Acceptability of Electric Vehicles: Findings from a Driver Survey

2. Previous Studies on the Uptake of Electric Vehicle

Considerable literature on the operating characteristics of EV (e.g. Voelker, 2009) and the work
at UWA (Mullan et al., 2010) has established that standard car models converted to EV can
give excellent performance.

The studies to explore the potential demand for EV have started in different regions of the
world. Most of the research work for EV uptake is in the USA. Kurani and Turrentine (1996)
compared petrol and CNG with the hybrid and “neighbourhood” EVs (for 454 households) and
found home-recharging will be successful. Half of the households mentioned that they would
buy EV as their next new vehicle in multi-vehicle households. Kurani and Turrentine (1996)
were also amongst the first researchers to incorporate attitudinal data in their modelling.
Golob and Gloud (1998), with 69 individuals, applied regression analysis comparing petrol and
EV, and found EV likely to be used if average vehicle mileage is less than 28 miles/day.
Another study in California (Hess et al., 2006) comparing internal combustion engine vehicles,
EV and hybrid vehicles, suggested that EV can only compete in the market if they have a range
greater than 353 miles – thus recommending increased driving range for EV acceptance.
Bolduc et al. (2008) conducted an experiment in Canada with 866 individuals, comparing petrol,
alternative fuel, hydrogen fuel cell vehicle and hybrid EV. They used hybrid choice models
including perceptions and attitudes and the structural and measurement equations for latent
variables were simulated together. The hybrid choice model demonstrated its capabilities to
capture: i) the environmental concerns; and ii) the appreciation of new car features. The
behaviour towards charging of electric vehicles was not discussed; however, the latent
constructs enriched the model’s explanatory power.

Recent study by Lieven et al. (2011) in Germany applied correspondence analysis to rank eight
types of cars (city, small, van, sports, luxury, etc.) for six types of uses (first vehicle for all uses,
second for leisure, etc.). Their findings tell that price is the top priority for both conventional and
EVs, with range ranked second. Performance, durability, environment, and convenience are
given less priority. Only 4.2% of first car buyers chose EV and they rated price and range as a
lower priority than non-EV potential buyers. Another recent research in vehicle type choice
modelling is by Kuwano et al. (2012) in Japan, they designed a two stage model. In the first
stage of decision making respondent was given a brief overview of EV features, and then asked
whether to keep EV as one of the available choices. If the respondent decided to keep EV in the
choice sets, a set of scenarios containing gasoline, hybrid-electric, and EV in the choice sets
was displayed to the respondent; otherwise scenarios with only gasoline and hybrid-electric
vehicle were given to the respondent. In this way social conformity was reflected in their model,
and heterogeneity in the preferences was explained by the use of latent class models. In
addition to the attributes that were considered by similar studies (such as purchase price,
range, charging time, and operation costs), Kuwano et al. (2012) had market share as an
attribute in their stated preference choice sets. With 384 respondents in Japan, Kuwano et al.
(2012) found that 10% of respondents prefer to own an EV, while 20.2% considered EV as an
alternative in the choice experiments. They obtained three latent classes: EV share rise, EV
purchase price reduction, and EV performance improvement (Kuwano et al., (2012); page 7).

In summary, studies of EV acceptance have been increasing since their start more than ten
years ago (Kurani and Turrentine, 1996; Brownstone et al., 2000; Ahn et al., 2008), with the
most recent research in this area being in the USA (Hidrue, 2010), Switzerland (Ziegler, 2010),
Germany (Lieven et al., 2011), and Japan (Kuwano et al., 2012). The technology at the core of
this study embodies significant advances and the study has the task of assessing how much
these advances will improve acceptability of EV.

3
 2012 ATRF Proceedings

2.1 Consumer Behaviour Models on the Adoption of New Technologies

EV is a significant new technology; this makes it pertinent to explore EV adoption as “new
technology” adoption. In the literature we find that technology adoption research includes
variations for market inventions, in the field of information technology (IT), or both.

2.1.1 Technology Acceptance Model

Davis (1989) theorizes in the Technology Acceptance Model (TAM), that behavioural intention
to use a system is determined by two factors: perceived usefulness and perceived ease of use.
The term system here was taken as any Information System, and perceived usefulness is the
extent to which an individual believes that the system will help to enhance his performance. The
ease of use similarly indicates the extent to which an individual believes that using the system
will not require extra effort to learn first. A theoretical extension of this model as TAM2 is
defined (Venkatesh and Davis, 2000); it contributes by adding social influence constructs and
also explores how perceived ease of use can be increased by helping the user to learn the
system. This model has been used in different studies, as Lee et al. (2003) summarises its use
in literature from 1986 till 2003.

2.1.2 Technology Readiness

The concept of technology readiness (Parasuraman, 2000) refers to the people’s propensity to
embrace and use new technologies for accomplishing goals in home life and at work.
Parasuraman (2000) in collaboration with a company in the United States developed a
Technology Readiness Index (TRI) as part of a technology readiness research program. Focus
groups and interviews were conducted with the customers of companies from a variety of
different technologies (e.g., financial services, e-commerce, online services, and
telecommunications). After a number of analyses, a technology readiness scale was designed
with four dimensions. The two positively supporting dimensions: Optimism and Innovativeness
were classified as drivers, whereas the other two Discomfort and Insecurity were classified as
inhibitors. The items in this TRI were further used by many researchers as a scale to measure
self-service (e.g., ATM, bank by phone, and online banking) technologies adoption (James et
al., 2005, Meuter et al., 2003), and also to explore the Internet home usage (Matthing et al.,
2006). Both TAM and TRI consider the positive drivers of technology, however, in addition to
TAM, TRI incorporates constructs with a negative effect in the adoption of new systems.

2.1.3 Technology Adoption Propensity

Ratchford and Barnhart (2011) reported on the assessment of consumer propensity to adopt
new technologies. This research primarily considers the adoption of new technology by
consumers in the market, while TRI focused mainly on specific technologies (for example,
computers, or Internet). When buying a new technology the decision is made based on the
benefits, and the time and effort consumers spend in learning and absorbing the new
technology (Ratchford and Barnhart, 2011). The precise forecasting of technology products
requires measurement of both positive and negative attitudes towards the technology.
Ratchford and Barnhart (2011) recently developed a Technology Adoption Propensity (TAP)
index containing 14 items, significantly shorter than TRI with 36 items.

2.1.4 Post Adoption Behaviour

Huh and Kim (2008) studied the role of post-adoption behaviour and experimented with young
people and early adopters. On the other side, Son and Han (2011) indicated that technology
readiness of the consumer (i.e. how well a consumer is prepared for the new technology) has
an impact on the post-adoption behaviour. Gatignon and Robertson's (1985) suggested that
diffusion of technological innovations will depend on consumers developing new knowledge and
new patterns of experience.
 Acceptability of Electric Vehicles: Findings from a Driver Survey

3 A Conceptual Model for the Adoption of Plug-in Electric Vehicles

Drawing on the above literature, a set of latent constructs were identified through which the
acceptability of Plug-in Electric Vehicles by the drivers’ in the WA EV trial can be assessed.
Thus, the objective of this study is to determine what contributes for the drivers’ attitudes and
perceptions of EV, and also to find which EV driving experiences can affect their propensity to
adopt EV.
The specific questions we explore in this research refer to: the direct impact of EV benefits,
technical difficulties experienced while driving EV, along with effects of the attitudes towards
environment and technology adoption (measured using latent constructs) on the willingness of
the drivers to recommend and purchase an EV. The survey instrument was designed according
to the conceptual model given in Figure 1. While the purpose of the overall research is to test a
mediating model (EV benefits and barriers, environmental concern, and technology learning
impact on the overall satisfaction while driving an EV, which in turn allows predicting the
willingness to recommend and purchase an EV) for this paper, we test a direct model with all
predictors affecting the willingness to recommend and purchase an EV.

EV Benefits Figure 1: Conceptual Model

EV Barriers

Willingness to
Environmental Satisfaction in
recommend and
Concerns driving an EV
purchase an EV

Technology
Savviest

The primary hypotheses of this study include:

H1: Drivers confident in the environmental performance and efficient use of energy of EV are
more likely to recommend and purchase an EV.
H2: Drivers showing concerns for environmental changes are more likely to recommend and
purchase an EV.
H3: Drivers ready to adopt and learn new technologies are more likely to recommend and
purchase an EV.
H4: Perceived EV benefits influence positively the willingness to recommend and purchase an
EV.
H5: Experienced technical difficulties while driving an EV influence negatively the willingness to
recommend and purchase an EV.
H6: Overall, drivers’ satisfaction with EV reflects the willingness to adopt EV as a future car. For
this paper the satisfaction with driving an EV is tested as one of the independent variables, as
this is not mediating model rather a direct model is tested with all predictors affecting the
willingness to adopt EV as a future car.

5
 2012 ATRF Proceedings

4 Data and Methodology

In order to design the survey questionnaire, a focus group was conducted in November 2011
with 11 EV drivers at The University of Western Australia. The drivers discussed their EV
driving experiences and perceptions towards EV as a new technology. Overall, they were
satisfied with the trial EV performance and showed confidence towards its acceptance. The
participants indicated the pros and cons of EV in the trial. The advantages of EV as discussed
in the focus group include: smooth and quiet operating drive, good torque, resource
management, sustainability, being a new technology (innovative) but appearing or driving like a
normal car, clean energy with no emissions, low running cost, minimal service cost or no need
to go for oil-checks, free reserved parking, efficiency. The drivers also discussed the drawbacks
and concerns that they had while driving EV: limited range, finding a charging station,
recharging time, trip planning, range indicator problems, and technical problems like
regenerative braking, acceleration etc. These barriers also affected the willingness of other
drivers to become part of the trial, when presented in the induction process for EV usage. The
participants also indicated the factors that might affect EV performance in the market, such as
range, performance, place and time required for recharging, substantial price, limited choice of
EV models, and their resale value.

In December 2011, an online survey was deployed and sent to all EV drivers in Perth, WA. The
experiences of the drivers in the focus group helped the design of the questionnaire. The
instrument included four sections: 1) EV characteristics; 2) drivers’ experiences; 3) attitudinal
questions; and 4) background questions. The socio-demographics in the survey included the
age, sex, education of the respondents, and number of cars at home. Since the drivers in the
trial did not purchase the EVs themselves, the income variable was deemed irrelevant. The
questionnaire also asked drivers about the technical problems encountered when driving the
EV, as well as what do they perceive the most and the least desirable features of EV. The
vehicles in the trial are all converted EVs, thus only a limited number of drivers outside the trial
had experiences with manufactured EVs. The overall satisfaction of driving EV was also
included in the questionnaire.

4.1 Survey Design and Data Collection

The drivers in the EV trial filled in an online survey, with 43 respondents completing all
questions. Although this is a small number of respondents, the response rate was high (and the
sample appropriate for representing the EV drivers in WA) considering that only few
organisations in the trial have started to use EV, with not all the respondents using it on a
regular basis. Among these 43 respondents, four respondents experienced driving
commercially manufactured EV, while rest of respondents are drivers of the converted EV..

The socio-demographics in survey (Table 1) show that the majority of respondents are male
drivers (67.4%), and a number of respondents (73%) own 2 or more cars. Twenty-two
respondents are over 40 years and 28 have tertiary education.

More than 80% of drivers showed satisfaction in driving EV, with 34.1% being extremely
satisfied. This is a positive indication towards EV acceptance in the WA EV trial, where 24% of
respondents drive more than 50km, 39% drive 21-50km, 27% drive 10 to 20km, and only 11%
drive less than 10km in a single trip.
 Acceptability of Electric Vehicles: Findings from a Driver Survey

Table 1: Information about Respondents

Variable % Count
Gender
Male 67.4% 29
Female 32.6% 14
Age
17-22 9.3% 4
23-29 20.9% 9
30-39 18.6% 8
40-49 18.6% 8
50-59 20.9% 9
60+ 11.6% 5
What is your highest level of education?
Year 12 9.3% 4
College/Professional qualification 25.6% 11
University Bachelor Degree 48.8% 21
Masters or PhD 16.3% 7
How many vehicles do you have at home?
1 27.9% 12
2 48.8% 21
3 or more 23.3% 10

“Zero-tail-pipe emissions” was considered the most desirable feature suggesting that the
drivers are concerned about the environment, followed by “low running cost”, then “reliability”,
“low-maintenance”, and “home-charging”. “Low level of noise” is also suggested as a desirable
feature of EV by the drivers in the trial. In terms of perceived barriers for EV uptake, the
respondents indicated the “limited range” and “purchase cost” as the most serious limitations,
followed by “recharging infrastructure” and “recharging time”, with “reliability” the least serious
barrier.

As informed by the focus group, the questionnaire presented a list of technical problems with
EV, from which the participants had to select the ones they encountered while driving EV. Forty-
two respondents answered this question, 52% respondent indicated “Power-steering failure”,
“no regenerative braking” and “range indicator errors”, while 10 respondents reported other
faults that are related to charging, braking faults, motor overloading, and gearbox problems.

Recognising the role of attitudes and preferences in explaining behaviours, the survey included
a set of latent constructs regarding EV benefits, environmental concerns, adoption of new
technologies, and willingness to recommend and purchase an EV. Since the objective of this
survey is to investigate and test the role of these latent constructs against the willingness to
purchase an EV, the analysis included two stages: i) exploratory factor analysis to test the
validity of the latent constructs (latent factor scores were derived for use in the subsequent
analysis); ii) multiple linear regression, for simultaneous assessment of the linear
interrelationships between predictors for willingness to purchase EV.

4.1 Exploratory Analysis of Attitudes towards Electric Vehicle

To test the drivers’ behaviours and attitudes towards EV, items reflecting several latent
constructs were included in the survey. These constructs refer to: EV benefits, environmental
concerns, adoption of new technologies, and willingness to recommend and purchase an EV.

7
 2012 ATRF Proceedings

The latent constructs’ items were designed as a set of five level Likert-Scale questions ranging
from strongly agree to strongly disagree. After an Exploratory Factor Analysis (EFA) stage, uni-
dimensional constructs were tested.
During the analysis of the constructs, it has been found that few construct items were weak and
they will be redefined for the household survey. Each construct is discussed in detail below.

4.1.1 Environmental Concern

This construct showed strong relationships among the variables. The basic assumptions of
factor analysis are satisfied, with a Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy
of 0.707 indicating a strong construct. The alpha factoring extraction method was used to
maximise the construct reliability; factor loadings of each element in this construct are above
0.5, as shown in Table 2.

The analysis of results showed that 90% respondents agreed that it is now the real time to
worry about our environment and this requires our immediate efforts. A large number (69.8%) of
respondents believed that climate change is not a myth; this shows that respondents are
concerned about climate change and air pollution effects. Approximately 63% of respondents
showed willingness to spend extra time or pay more for products and services, only to save the
environment.

Table 2: Environmental Concern Factor Loadings

Items Factor Loadings

Now is the real time to worry about the effects of air pollution. 0.795
I am concerned that future generations may not be able to
0.757
enjoy the world as we know it currently.
Saving the environment requires our immediate efforts. 0.718
I am willing to pay more for products or services only to save
0.714
the environment.
I am willing to spend extra time only to save the environment. 0.622
Vehicle emissions can destroy our flora and fauna. 0.534

For this construct, the reliability coefficient, Cronbach's Alpha has a value 0.832, suggesting
consistency of the entire scale (Hair et al., 2010).

4.1.2 Technology Adoption

This is a very important construct, already tested in literature investigating the adoption of EV as
new technology (Ewing and Sarigollu, 2000). Our analysis showed that multiple constructs may
emerge (the items were not correlated significantly for a unidimensional factor), and we
selected here to report the strongest one – “technology learning”.

Overall, the survey responses are convincing about the relevance of technology adoption in
further uptake of EV. For example, 90% respondents believed that using new technologies
makes our life easier, and 70% respondents felt that new technologies give more control over
our daily life. Nearly 77% of respondents showed an excitement for learning new technologies,
while 80% of the drivers agreed that keeping up with the new knowledge or technologies is
necessary.

When exploring the trendy or being fashionable tendency of the respondents, we found that
almost 30% of respondents are savvy-trendy adopters, based on their response that “taking up
new technologies makes one trendy”, and that “being fashionable means having up-to-date
knowledge of the techno-world”. Approximately 44% of respondents did not agree that new
technologies cause more problems than they solve.

As indicated, the EFA suggested more than one dimension, but only three items, with higher
commonalities and factor loadings were further retained. They are shown in Table 3.
 Acceptability of Electric Vehicles: Findings from a Driver Survey

Table 3: Technology Learning Factor Loadings

Items Factor Loadings
I am excited to learn to use new technologies. 0.758
Reverse (Things have become so complicated today that it is hard
0.703
to understand what is going on in this techno-world)
I love gadgets 0.601

The measure of sampling adequacy (KMO) value 0.669 and a Cronbach's Alpha of 0.703
indicated that this structure for the one-dimensional Technology Learning construct is
appropriate.

4.1.3 EV Benefits and Challenges

The most important EV benefits, identified by respondents, included: convenience of home
battery recharging and reduced average travel cost per trip. The respondents are also
comfortable with recharging their EV at public stations, although almost half of the respondents
need to do a lot of planning of activities when they drive EV.

In regard to EV technical difficulties, only 20% of the respondents believed that EVs have
problems with the acceleration; while 29% disagreed that EVs incur significant maintenance
costs.
None of these two constructs, EV benefits or Technical problems associated with EV had
adequate reliability in this sample, and consequently they were not used in this analysis.

4.1.4 Willingness to Recommend and Purchase an EV

This construct showed strong relationships among the variables (KMO=0.725). Factor loadings
of the elements in this construct (all above 0.8) are given in Table 4. The Cronbach's Alpha had
the highest value of all constructs, 0.910.

Table 4: Willingness to recommend and purchase an EV Factor Loadings

Items Factor Loadings
I prefer to use EV over any other type
0.911
of cars.
I would recommend EV to others. 0.828
I would buy an EV as my next car. 0.837
The results of the analysis show that approximately 65% of respondents would recommend EV
to others. Buying an EV as a next car is chosen by 27.9% of respondents, while 35% of
respondents would prefer to use EV over any other cars. This percentage of driver’s showing a
preference to use EV over any other type of cars indicates a positive attitude towards EV and
acceptability of the electric car.

5 Regression Model for EV Adoption

Once all the possible factors were identified, the next step was to quantify the effect of different
factors in the willingness to adopt EVs. As suggested in the hypotheses, the set of independent
variables identified for this model include: environmental concern, attitudes towards technology
learning, EV benefits, EV technical problems, being a savvy-trendy adopter, and having
confidence in driving EV. The socio-demographics considered in the analysis include age,
gender, and education.

The regression model initially tested all the independent variables, but the high correlations
among the explanatory variables resulted in multicollinearity issues (Hair et al., 2010). The
9
 2012 ATRF Proceedings

correlations between independent variables and the willingness to purchase and recommend an
EV are given in Table 5.

Table 5: Correlations between Independent Variables and Willingness to Recommend

and Purchase an EV
Willingness to
recommend Significant Cross
and purchase Correlation Coefficients
Independent Variables
an EV between potential
Correlation explanatory variables
Coefficients
AGE What is your age (years)? 0.143
What is your highest level of -0.152
HE
education?
TechL Technology learning construct 0.157
EnvC Environmental concern construct 0.250
How confident are you in the 0.561** EV_B1 (0.448*),
Conf environmental performance and EV_B2 (0.434*),
efficient use of energy of EV? OvSat (0.475**)
New technologies give more control -0.004
Tech_B
over our daily life.
Being fashionable means having 0.077
TFas up-to-date knowledge of the techno-
world.
0.476** AccP (0.454*),
Reverse (I spent a significant
EV_B1 (0.482**),
LessM amount of money to fix my EV in
EV_B2 (0.449*),
the last 3 months).
OvSat (0.441*)
I believe EV has no problems with 0.346* LessM (0.454*)
AccP
acceleration.
0.594** Conf (0.448*),
Battery recharging at home is LessM (0.482**),
EV_B1
convenient for my EV. EV_B2 (0.509**),
OvSat (0.552)
0.491** Conf (0.434*),
EV driving reduces my average LessM (0.449*),
EV_B2
travel cost/trip. EV_B1 (0.509**),
OvSat (0.560**)
0.634** Conf (0.475**) ,
Overall, how satisfied are you LessM (0.441*),
OvSat
driving an EV? EV_B1 (0.552**),
EV_B2 (0.560**)
* p<.05 ** p<.01

Table 5 shows that all independent variables (Conf, LessM, EV_B1, EV_B2, OvSat) have
moderate correlations with each other. Overall satisfaction in driving an EV (OvSat) is related to
EV Benefits (EV_B1, EV_B2), and to being confident in environmental performance and
efficient use of EV energy (Conf). Similarly, a lower amount of money spent to fix EV in last 3
months (LessM) has a positive impact on the overall satisfaction (OvSat), and perceived EV
benefits (EV_B1, EV_B2).

One of the remedies for multicollinearity is to omit one or more highly correlated variables, and
identify other independent variables to help the prediction (Hair et al., 2010). To address
multicollinearity and given the reduced sample size, a backwards elimination procedure was
applied. Two different models were tested, with overall satisfaction and EV benefits being the
response variables (Tables 6 and 7).
 Acceptability of Electric Vehicles: Findings from a Driver Survey

5.1 Multiple Linear Regressions’ Results

With a coefficient of determination R2 =0.643, the regression model presented in Table 6
confirms a subset of our hypotheses. The standardised coefficients indicate the relative
importance of predictors in the same units or standards, regardless of the measurement scale
used for the independent variables (Hair et al., 2010). When considering the socio-
demographics, age played a significant positive role in the model, with younger people less
likely to recommend and purchase an EV (beta for AGE is 0.185). This might be due to the
reason that more than 30% of respondents have an age of 50 years or above. The AGE
variable has even more significant value in Table 7 where beta is 0.260.
The first hypothesis of this study (drivers confident in the environmental performance and
efficient use of energy of EV are more likely to recommend and purchase an EV), is confirmed
with the standardised coefficient as 0.262. The third hypothesis shows mixed results with one
positive coefficient (technology learning 0.198) and a negative one (control given by
technologies -0.287). Hypothesis 5 is also confirmed with a significant negative coefficient and
the highest beta in absolute terms (0.367). The satisfaction variable (OvSat) comes next
(0.336), confirming hypothesis 6 that overall, drivers’ satisfaction with EV reflects the willingness
to adopt EV as a future car.

Table 6: Regression Model with Satisfaction Variable as Predictor

Dependent Variable: Unstandardized Standardized
Willingness to Coefficients Coefficient
recommend and B Beta Significance
Std.
purchase an EV Error

Independent Variables
(Constant) -0.716 0.815 0.385
AGE What is your age (years)? 0.127 0.072 0.185 0.086
How confident are you in the 0.370 0.177 0.262 0.044
environmental performance
Conf (H1) and efficient use of energy of
EV?
Tech_B New technologies give more -0.371 0.146 -0.287 0.016
(H3-A) control over our daily life
TechL 0.281 0.174 0.198 0.114
Technology learning construct
(H3-B)
I spent a significant amount of -0.387 0.125 -0.367 0.004
Tech_Diff
money to fix my EV in the last
(H5) 3 months
OvSat Overall, how satisfied are you 0.338 0.131 0.336 0.014
(H6) driving an EV?
Note: Parameters significant at 0.05 level in bold.

As discussed in more detail in the next section, satisfaction is a mediator between the EV
benefit, EV barriers, and technology learning constructs, and the willingness to recommend and
purchase an EV.

The regression model in Table 7 also tests hypotheses of this study, but this time after
excluding the overall satisfaction from the list of predictors; independent variables that were not
significant were removed from the model, one at a time, while exploring the impact of the rest of
the variables. The final model, containing only significant variables, is given below. It has the R2

11
 2012 ATRF Proceedings

value of 0.592, this indicates that variables in this model explain 59.2% of the variability in the
willingness to recommend and purchase an EV.

The second hypothesis in this study (drivers showing concerns for environmental changes are
more likely to recommend and purchase an EV) is not confirmed by the model, but this may be
due to the sample size and limited variability in the construct (the average factor score is 3.71,
with a standard deviation of 1.02). Ewing and Sarigollu (2000) found that the consumers
accepted the environmental impact of clean fuel vehicles, but the vehicle’s standards cannot be
compromised.

Again, hypothesis 3 does not have full support with the question on technology’s control over
lives displaying a negative relationship. This negative coefficient was unexpected, however it
might be due to the fact that most of the respondents in this study have an experience of driving
converted EVs, and not commercially manufactured EVs. Another possible reason might be the
word “control”. This item needs to be reconsidered for the household survey and perhaps
instead of “control over our daily life”, the question needs to be reformulated to include “enable
us” or another positive phrase (for example “Using new technologies in our daily lives makes life
easier.”)

Table 7: Final Regression Model

Dependent Variable: Unstandardized Standardized
Willingness to Coefficients Coefficient
B Std. Beta Significance
recommend and
Error Level
purchase an EV
Independent Variables
(Constant) 0.411 1.416 0.773
What is your age
AGE 0.180 0.082 0.260 0.036
(years)?
EnvC Environmental Concern
0.224 0.172 0.150 0.201
(H2) Construct
New technologies give
Tech_B
more control over our -0.382 0.172 -0.299 0.034
(H3-A)
daily life
TechL Technology learning
0.387 0.178 0.278 0.037
(H3-B) construct
Battery recharging at
EV_B1
home is convenient for 0.266 0.124 0.308 0.040
(H4-A)
my EV.
EV_B2 EV driving reduces my
0.284 0.147 0.268 0.062
(H4-B) average travel cost/trip.
I spent a significant
Tech_Diff amount of money to fix
-0.305 0.151 -0.289 0.051
(H5) my EV in the last 3
months
Note: Parameters significant at 0.05 level in bold.

The fourth hypothesis (H4) of the study (perceived EV benefits influence positively the
willingness to recommend and purchase an EV) is confirmed, with EV_B1 and EV_B2
presenting beta coefficients of 0.308 and 0.268, among the highest in the model. Thus, this
demonstrates that perceived EV benefits (low driving cost and home-charging) influence
positively the willingness to recommend and purchase an EV. This is consistent with the
previous literature: e.g., Kurani and Turrentine (1996) identified the “home-charging” as a key
benefit of EV.
 Acceptability of Electric Vehicles: Findings from a Driver Survey

The fifth hypothesis (H5), regarding the relationship between experienced technical difficulties
while driving an EV and the willingness to recommend and purchase an EV, is confirmed as
well, with a negative coefficient and a beta value of -0.289. Technical difficulties experienced
while driving an EV act as a deterrent for EV uptake. This is well supported by the literature.
Dagsvik et al. (2002) indicated that alternative fuel vehicles can compete with petrol cars if
maintenance and refuelling infrastructures for alternative fuel vehicles are well established.
Again these coefficient values could be different if there were more number of respondents
driving commercially manufactured EVs (with less technical difficulties) instead of converted
EVs.

5.2 Discussion and Future Research

The independent variables taken into account in this study were derived from literature and
were further refined after the focus group. This study primarily explored the behaviours and
experiences of the drivers already using the EV, in the WA EV trial. With a limited number of
respondents (N=43) a number of hypotheses were tested and confirmed. One of the limitations
of this study is that among small set of respondents (N=43) the majority of drivers used
converted EVs, only 4 drivers had experience of driving manufactured EVs. Thus, the results
would intuitively be different if the number of commercially manufactured EV drivers was larger.
At the same time, this limitation does not impact the main objective of the study that is to
discover the drivers’ perceptions and attitudes towards EVs, and to determine how their
experiences might affect acceptability of Electric Vehicles. The weakness of the few constructs
was also noted as another limitation and these constructs will be revised for the upcoming
household survey.

Since the satisfaction variable seems to be a mediator between perceived EV benefits, EV

technical difficulties, attitudes towards technologies constructs and willingness to recommend
and purchase an EV, the next step will be to assess these relationships using structural
equation modelling (SEM) approach (Meyers et al., 2006). On account of small sample size,
this was not currently possible, but with a higher number of respondents from the household
survey it might be possible in the future.

6 Conclusion
This research explores the EV drivers’ behaviour and their perceptions and attitudes towards
new technologies. Experiences of drivers in the trial are useful for exploring the impact of EV
benefits and of their technical difficulties on the acceptance of EV. The drivers showed
confidence in the EV’s environmental performance and efficient use of energy. The range is a
serious barrier to EV uptake, with almost half of drivers indicating that they require significant
trip planning especially for trips longer than 30km.

The analysis of the drivers’ survey also aimed to refine the latent constructs such as technology
adoption and environmental concern. With the data from the drivers’ survey the reliability of the
constructs was assessed and items with low value of loadings are being revised. Although the
environmental concern appeared non-significant in the regression models, the literature
identified it as a key construct, and we will consider it in the household survey. Another
supporting argument for environmental concern construct is that the “Zero-tail pipe emissions”
is ranked as the most desirable feature of EV by the drivers in the trial.
The results of this analysis will inform the household survey, and these constructs will be
presented with further improvements, in the pilot household survey.

13
 2012 ATRF Proceedings

7 References

ABS. (2008). 4102.0 - Australian Social Trends Retrieved 27 May 2011, from
http://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/4102.0Chapter10102008.
Ahn, J., Jeong, G., & Kim, Y. (2008). A forecast of household ownership and use of alternative fuel
vehicles: A multiple discrete-continuous choice approach. Energy Economics, 30(5), 2091-2104.
Bolduc, D., Boucher, N. and Alvarez-Daziano, R. (2008) “Hybrid Choice Modeling of New Technologies
for Car Choice in Canada”, Transportation Research Record, 2082, 63-71.
Brownstone, D., Bunch, D.S., and Train, K. (2000) “Joint mixed logit models of stated and revealed
preferences for alternative-fuel vehicles”, Transportation Research B, 315-338.
Dagsvik, J.K., Wennemo, T., Wetterwald, D.G. and Aaberge, R. (2002) “Potential demand for alternative
fuel vehicles”, Transportation Research B 36, 361-384.
Davis, F. D. 1989. Perceived usefulness, perceived ease of use, and user acceptance of information
technology. MIS Qutart. 13 319-339.
Ewing, G. and Sarigollu, E. (2000) “Assessing consumer preferences for clean-fuel vehicles: a discrete
choice experiment”, Journal of Public Policy and Marketing 19, 106–118.
Gatignon H, Robertson TS. A propositional inventory for new diffusion research. J Consum Res 1985;
11(4):849–67.
Golob T.F. and Gould J. (1998) “Projecting use of Electric Vehicles from household vehicle trials”,
Transportation Research A 32, 441–454.
Hair, J.F., William Jr., Black, C., Barry J., Anderson, R.E., (2010), Multivariate Data Analysis, 7/e., 2010,
Pearson Prentice Hall.
Hess, S., Train, K.E., Polak, J.W. (2006) “On the use of a Modified Latin Hypercube Sampling (MLHS)
method in the estimation of a Mixed Logit Model for vehicle choice”, Transportation Research B
40, 147-163.
Hidrue, M.K. (2010) “The Demand for Conventional and Vehicle-to-Grid Electric Vehicles: A Latent Class
Random Utility Model”, PhD Dissertation UMI Number: 3440476.
Huh, Y.E., and Kim, S.H. (2008) “Do early adopters upgrade early? Role of post-adoption behaviour in
the purchase of next-generation products”, Journal of Business Research (61), 40–46.
James M. Curran, Matthew L. Meuter, (2005) "Self-service technology adoption: comparing three
technologies", Journal of Services Marketing, Vol. 19 Iss: 2, pp.103 – 113.
Kurani, K.S. and Turrentine, D.S. (1996) "Testing electric vehicle demand in hybrid households' using a
reflexive survey", Transportation Research D 1(2), 131-150.
Lee Y, Kozar KA, Larsen KRT. The technology acceptance model: past, present, and future.
Communications AIS 2003,(12):752–80.
Lieven, T., Muhlmeier, S., Henkel, S., and Waller, J.F. (2011) “Who will buy electric cars? An empirical
study in Germany”, Transportation Research D 16, 236–243.
Masashi Kuwano, Makoto Tsukai, and Tsukasa Matsubara (2012) “Analysis on promoting factors of
th
electric vehicles with social conformity”, Presented at the 13 International Conference on Travel
Behaviour Research, Toronto, Canada, 15-20.
Matthing J, Per Kristensson, Anders Gustafsson, A. Parasuraman, (2006) "Developing successful
technology-based services: the issue of identifying and involving innovative users", Journal of
Services Marketing, Vol. 20 Iss: 5, pp.288 – 297.
Meuter, M. L., Ostrom, A. L., Bitner, M. J., & Roundtree, R. (2003). The influence of technology anxiety
on consumer use and experiences with self-service technologies. Journal of Business Research,
56(11), 899-906. doi: 10.1016/s0148-2963(01)00276-4.
Meyers, L.S., Gamst, G., Guarino, A.J., (2006), Applied Multivariate Research: Design and Interpretation,
Sage Publications, USA.
Mullan, J., Whitely, S., Harries, D. and Bräunl, T. (2010) "Electrification of Personal Transport in Western
Australia, An Impact Study on the Rapid Introduction of Plug-in Battery Electric Vehicles and
 Acceptability of Electric Vehicles: Findings from a Driver Survey

Plug-in Hybrid-Electric Vehicles in Western Australia", EV-Report, CREST, 2010, available at:
http://www.crest-au.com/docs/EV-Report-20-04-12.pdf (accessed on: May 2011).
Parasuraman A. Technology readiness index (TRI): a multiple item scale to measure readiness to
embrace new technologies. J Serv Res 2000; 2(4): 307– 20.
Ratchford, M. and Barnhart, M. (2011) “Development and validation of the technology adoption
propensity (TAP) index”, Journal of Business Research (In press).
Son, M., and Han, K (2011) “Beyond the technology adoption: Technology readiness effects on post-
adoption behaviour”, Journal of Business Research (In press).
Venkatesh, V., & Davis, F. D. (2000). A Theoretical Extension of the Technology Acceptance Model: Four
Longitudinal Field Studies. Management Science, 46(2), 186-204.
Voelker, J. (2009) “How Green is My Plug-In?”, IEEE Spectrum 48-49 (6), 42- 45.
Ziegler, A. (2010) “Individual Characteristics and Stated Preferences for Alternative Energy Sources and
Propulsion Technologies in Vehicles: A Discrete Choice Analysis”, Economics Working Paper
10/125, available at SSRN: http://ssrn.com/abstract=1585142, CER-ETH, Switzerland.

15

View publication stats