Information & Management 53 (2016) 668–681

Contents lists available at ScienceDirect

Information & Management

journal homepage: www.elsevier.com/locate/im

A study on the impact of design attributes on E-payment service utility

Eric W.K. See-To a, Kevin K.W. Ho b,
*
a
Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Kowloon, Hong Kong
b
School of Business and Public Administration, University of Guam, Mangilao, GU 96923, United States

A R T I C L E I N F O A B S T R A C T

Article history: Prior research has shown how consumer perception, when risk is an important factor, may affect the
Received 10 June 2014 adoption and the usage of E-payment service. Given this general knowledge, this study investigates the
Received in revised form 6 February 2016 more specific issues of the mix of design attributes to drive consumer choice in using E-payment
Accepted 18 February 2016
services. Using six design attributes defined by a group of practitioners and E-payment service users
Available online 3 March 2016
using the Delphi method, an online conjoint experiment is conducted. We find out how these design
attributes may affect consumer choice under risk and the rank order of the magnitudes of their effects.
Keywords:
ß 2016 Elsevier B.V. All rights reserved.
E-payment service
Website design
B2C E-commerce
Perceived risk
Conjoint analysis

1. Introduction E-payment services, reduces perceived usefulness and the intent to

The development of E-business technologies in the past two
decades has significantly changed the way of conducting retail
businesses. Using the latest E-business technologies, business-to-
consumer (B2C) E-commerce operators can develop their retail
businesses in the virtual world with minimal cost. Some
companies, such as Amazon and those E-vendors operating on
eBay, do not have any physical store. This virtual business model
provides B2C E-commerce operators with the opportunities to be
connected with their online customers from all over the world on a
24  7 basis. However, some online customers only surf these
websites but do not complete their transactions online because
they perceive making payment online using E-payment service is
risky [1,2]. Therefore, many electronic marketplaces, such as eBay,
have developed specific design attributes, such as a peer-
evaluation system, to reduce the risk of making online payments
[3,4]. Most of these design attributes increase the probability of
closing online transactions by reducing the perceived risk of
participating in online transactions.
Prior research has shown that both hedonic and functional
qualities of E-service have positive impacts on its perceived values
[5]. However, the perceived risk of using E-services, including
* Corresponding author. Tel.: +1 671 735 2501; fax: +1 671 734 5362.
E-mail addresses: eric.see-to@polyu.edu.hk (Eric W.K. See-To),
kevinkho@triton.uog.edu (Kevin K.W. Ho).
use E-services by users [6]. In addition, Im et al. [7] also reported
that the perceived ease of use of a technology has a more
significant impact on a user’s adoption of a risky technology
compared with a less risky technology; and perceived usefulness of
a technology has a more significant impact on a user’s adoption of a
less risky technology than a risky technology. In a review
conducted by Darley et al. [8], it is reported that perceived risk
is an important factor affecting online consumer behavior. Pavlou
[9] and Pavlou and Gefen [10] showed that perceived risk has a
negative impact on consumers’ intention to undertake a transac-
tion with an E-retailer. Gupta and Kim [11] also reported that
perceived risk has a negative impact to purchase intention in an
online environment through the perceived value of the E-service
concerned. On the contrary, marketing researchers have suggested
that the attributes of a service have significant impacts on its
perceived risk levels [12]. Thus, by manipulating the attributes of a
service, it is possible to vary the level of impact of perceived risk on
the intent to use the service. Hence, if an E-payment service
engineer develops some design attributes that can reduce the level
of perceived risk for using such services, these attributes will
reduce the overall negative impact of perceived risk on the intent
to use E-service by users. This will convert more web surfers into
online shoppers.
The core research issue in this study is to investigate issues
related to the service design of an information system. Prior
research has shown the importance of service attributes toward
user satisfaction on E-services (see [13]). Venkatesh [14] suggested

http://dx.doi.org/10.1016/j.im.2016.02.004
0378-7206/ß 2016 Elsevier B.V. All rights reserved.
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
IS researchers should conduct research studies on ‘‘the right mix of
attributes to drive consumer choice’’ in service design research
using statistical techniques such as conjoint analysis. In this study,
we investigate the impacts of six design attributes of an E-payment
service selected by a group of practitioners and E-payment service
users through the Delphi Method [15]. Those attributes are
believed to have impacts on the risk propensity of using E-payment
services. The incorporation of the ideas of the group of practi-
tioners and users into our research design echoes the suggestion by
Lee [16] who suggests that ‘‘the starting point of IS research need
not be existing theory’’ but can be the wisdom or craftsmanship of
IS professionals and practitioners, which eventually will enhance
the practical implications of this research. Hence, our main task is
to find the right mix of design attributes to drive consumer choice
in using E-payment services [14], which can help researchers and
practitioners to gain a better understanding on how technology
interacts with social and behavioral aspects of consumers in the
E-payment service context. To achieve our research goal, conjoint
analysis is selected as the statistical technique used in this study
[17], which allows us to rank the impacts of different design
attributes based on their relative importance, and to understand
their relative impacts on the consumers’ utility of using E-payment
services.
The results of this study have several theoretical contributions
and practical implications and one methodological contribution.
First, they extend knowledge on how design attributes of an
E-payment system may affect the associated consumers’ utility
through their influence on the perceived risk level. Second, this
study shows the value of conjoint analysis, as a methodology, in
analyzing the non-linear impact of constructs on information
system adoption. For practical perspectives, our results help
E-commerce operators to improve and better plan for the designs
of their E-payment services by enhancing consumers’ utility of
using these E-payment service websites. In addition, they provide
insights to E-commerce operators on the effectiveness of different
design attributes on the consumers’ utility of using these
E-payment service websites, which can be indirectly measured
as a monetary value [18].
2. Research model and hypothesis development
2.1. Research model
We investigate the impacts of different design attributes of
E-payment services on the consumers’ utility of using E-payment
services. The design attributes concerned are developed to reduce
the negative effect (i.e., negative impact on utility) of different
dimensions of perceived risk on the consumers’ utility of using
E-payment services, which directly affect and increase the intention
to adopt the service [19] due to the increase of risk propensity of
consumers using the service. Prior research studies on perceived risk
on consumer behavior can be dated back to the seminal article of
Bauer [20], which defines that risk ‘‘will produce consequences
which (consumer) cannot anticipate with anything approximating
certainty, and some of which at least are likely to be unpleasant.’’
Other researchers, such as Peter and Ryan [21], enriched the
definitions of risk by further elaborating Bauer’s definition. While
some early researchers analyzed perceived risk as a single construct,
other researchers, such as Stone and Grønhaug [22], suggested that
perceived risk can be analyzed as a multidimensional construct [6]
or measured by different aspects [23].
IS researchers have already studied the impact of perceived risk
on IS for over two decades. In particular, they are interested in the
impact of perceived risk of the intention to adopt E-services, such
as E-banking service [24] and E-payment service [6]. Some of these
IS studies treat perceived risk as a single construct in studying
669
consumers’ intention to transact with E-retailers and observe the
negative impact of perceived risk of transaction intention
[9,10]. Nicolaou and McKnight [25] also researched into the
impact of perceived risk on performing online transactions in an
exchange and showed that perceived information quality reduces
the level of perceived risk of users. Kim et al. [26] also used
perceived risk as a single construct to study the impact of risk and
trust on the intention of online consumers to purchase. More
recently, Chang and Wu [27] revisited the impact of perceived risk
on purchase intention in an online environment and reported that
the impact of perceived risk to purchase intention is moderated by
the decision-making style.
Other IS studies investigated perceived risk based on its
different risk dimensions. Ho and Ng [28] investigated the level of
perceived risk of different types of E-payment services using five
risk dimensions. They showed that the level of perceived risk of
different risk dimensions depends on the types of E-payment
service concerned, and the amount of money involved in the
business transactions. Hann et al. [18] focused their investigation
on the impact of privacy risk on consumers’ utility of E-commerce
website design using a conjoint analysis. They suggested that it is
possible to offset the negative effect of privacy risk by monetary
reward and future convenience provided by the websites. Lou
et al. [29] studied the impact of perceived risk on E-banking
services and showed that perceived risk has a negative impact on
the intention to adopt E-banking services and the performance
expectancy of E-banking services. Featherman and his colleagues
also conducted several studies on E-payment services by
operationalizing perceived risk by several risk dimensions
[6,19,30]. In this study, we define perceived risk using the
perceived risk dimensions proposed by Featherman and Wells [6],
as reported in Table 1.
In this study, we focused our investigation on six design
attributes of E-payment service, that is, (i) monetary reward per
transaction; (ii) consumers’ liability for losses per transaction; (iii)
physical control feature through a physical token; (iv) anonymous
payment record; (v) online information transfer; and (vi)
acceptability of E-payment services, on the consumer’s utility of
using E-payment services. These six attributes were selected based
on the recommendations from the group of practitioners and
E-payment service users through the Delphi Method. This group
consisted of 15 members, including 10 practitioners and five users.
Four of the practitioners belonged to the senior management of
financial industries who had >10 years of experience and
involvement in formulating banking policies in information
security, compliance, credit and debit card services, and online
enquiry services in relation to the E-banking services. The other
three practitioners were senior management of the E-business
sectors who had over 10 years of experience and involvement in
the design of their corporate website for retailing industries, which
were able to handle online orders. The other three practitioners
belonged to the senior management of E-service providers, whose
companies were providing either E-payment service or web-portal
services. These 10 practitioners were involved to join the study
based on their job relevance, as all of them were involved in the
provision or usage of E-payment service. Besides, they have
experience across different industries, which allow us to gain their
insights of the E-payment service requirement across industries.
The five users involved in the Delphi Study included two business
school doctoral students, two MBA students who had >5 years of
full-time working experience in major retailing companies, and
one undergraduate student. The doctoral and MBA students were
having more than ten years of experience in making purchases
online and using E-banking services, and the undergraduate
student had around 3 years of similar experiences. We anticipate
the doctoral students can provide us with opinions through the
670 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681

Table 1
Definition of six dimensions of perceived risk of using the E-payment service.

Dimension of perceived risk Definition

Financial risk Consumer assessment of potential financial losses because of potential Internet fraud
Performance risk Consumer assessment of potential performance problems, malfunctioning, transaction processing errors, reliability, and/or security
problems, that cause the E-payment service to not perform as expected.
Privacy risk Consumer assessment of potential losses to the privacy and confidentiality of their personal identifying information and that
E-payment usage exposes them to potential identity theft.
Psychological risk Consumer assessment of potential losses to their self-esteem, peace of mind or self-perception (ego) because of worrying, feeling
frustrated, or foolish a result of using E-payment service.
Social risk Consumer assessment of potential losses to their perceived status in their social group as a result of using an E-payment service. The
assessment of the probability that consumers believe that they will look foolish to important others.
Time risk Consumer assessment of potential losses to convenience, time, and effort caused by wasting time researching, purchasing, setting up,
switching to, and learning how to use the E-payment service.

Notes: (1) The information on this table is extracted from Table 1 of Featherman and Wells [6]. (2) Social risk and psychological risk are not included in our research model.

theoretical lens, the MBA students can provide us with opinions
through the practical perspectives, and the undergraduate student
can provide us with opinions of young people. Similar to other IS
study [15,31,32], we employed the Delphi method to identify the
important factors in the development of E-payment services from
the practitioners’ side at the beginning of this research. This
arrangement can help us gain some insights in our research design
from the experts and professional contexts [33]. Through three
rounds of interview, we were able to boil down their suggestions
and insights into the six design attributes that we mentioned.
Afterward, we started our in-depth literature review to find the
theoretical support for including these six attributes into our
research model, which will be reported in the following subsec-
tions. Therefore, the selection was based on both the theoretical
implications and practical values of these attributes. In the
development of our research model, we were also concerned
about the practical implications of our study of our theory
development [34–36], which echoes Lee’s comment [16] men-
tioned previously.
According to the theory of reasoned action (TRA), the attitude
toward a behavior (including the attitude toward the use of an IS) is
developed based on an expectancy–value model, which is a
measurement of attitude based on its utility, as shown in Eq. (1)
[37]:
X
n
A¼ bi ei
i¼1
these two risk dimensions by varying design attributes. We also
included three control variables in our research model, that is,
gender, age, and educational background, as well as a system
feature as moderator. Fig. 1 shows our research model.
2.2. Financial risk: monetary reward and consumers’ liability for
losses per transaction
The first two design attributes in our research model are
designed to reduce the negative impact of the financial risk on the
consumers’ utility of using E-payment services. Financial risk, in
the context of the usage of an E-payment service, is the assessment
conducted by the consumers on ‘‘the potential financial losses
because of potential Internet fraud’’ [6]. Ho and Ng [28] suggested
that the impact of financial risk can be reduced by mechanisms
such as a money-back guarantee. In this study, we investigated
into how the amount of monetary reward and consumers’ liability
for losses affect the negative impact of financial risk on the
consumers’ utility of using E-payment services. These two design
attributes have financial impacts to their users [38], which can be
explained using the prospect theory [39,40].
The provision of monetary reward has two impacts on
consumers’ utility of using E-payment services. First, it offsets
part of the financial loss of consumers if they face an Internet fraud.
Besides, it helps lock in customers for using a particular E-payment
(1) service, as the monetary reward is only usable on the particular
E-payment service concerned. On the contrary, the provision of

Therefore, when applying TRA to the adoption of E-services, the

intention to adopt an E-payment service can be estimated by the
Design Features
sum of the product of the evaluation of its design attributes (i.e., ei)
and its subjective probability related to the corresponding design
Monetary Reward
attributes toward the use of the E-payment service (i.e., bi), and the (H1.1)
intention to adopt the E-payment service is measured based on its Moderator

consumers’ utility. Here, we define the consumers’ utility of using Consumers’

RFID Card (H5)
Liability for Losses
an E-payment service as the benefits gained from using the (H1.2)
E-payment service by the consumer subtract the corresponding
perceived usage risks [19]. Physical Control
We manipulated the six design attributes of the E-payment Feature (H2)
Consumers’ Utility on E-
service in this study to facilitate investigation into whether they Payment Service
reduce the negative effect of different dimensions of perceived risk Payment Method Control Variables
(H3.1)
on the attitude of consumers on the intention to adopt an Age
E-payment service. In other words, we would like to observe
Information Transfer
whether these design attributes can increase the risk propensity of Method (H3.2) Gender
the attitude of consumers on the intention to adopt these
E-payment services. The six design attributes of E-payment Educational
services proposed by our group of practitioner and E-payment Acceptability (H4) Background

service users do not have any impact on social risk and

psychological risk, as it is difficult to manipulate the impacts of Fig. 1. Research model.
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
low consumers’ liability for losses attracts customers, as it provides
financial protection as a money-back guarantee to and limited
liability for the buyers, in case they fall victim to an Internet fraud.
In the E-payment service perspective, the provision of a
monetary reward of a consumer is actually a discount on the
price of goods and service sold. Thus, it is an operating expense.
Further, the consumers’ liability for losses imposed by E-payment
service providers has a financial impact on the companies as it
affects the money to be written-off for the losses related to Internet
frauds. As both types are operating expenses for E-payment service
providers, it is important for E-payment service providers to
understand which of these two mechanisms is more attractive to
consumers, that is, having a higher impact on the consumers’
utility of using E-payment services. Another possible alternative is
to optimize the net effect on the consumers’ utility of using
E-payment services by developing a suitable mix of these two key
attributes, which encourages consumers to participate in online
transactions by reducing the negative attitude on the intention to
use E-payment services.
Monetary rewards per transaction, such as cash rebates or
coupons, are frequently provided for E-payment services to their
consumers to encourage them to complete their transactions
online. Apart from that, this arrangement also encourages
customers to continue using their services in the future. Prior
literature [38] shows that the provision of monetary reward has a
significant but diminishing impact on consumers’ decision making.
As suggested by the prospect theory [39,40], there is a diminishing
sensitivity to gain and loss. As shown in the first quadrant of Fig. 2,
people usually underestimate the expected value for high gain and
overestimate the expected value for low gain (see the comparison
between x1 and x2, and f(x1) and f(x2)). As a result, the slope or f0 (xi),
which is the per-unit effect of gain, is smaller for high gain when
compared with the low gain, which is a diminishing effect (i.e.,
f0 (x2) > f0 (x1)). Based on this idea, we conjecture that monetary
reward reduces the negative impact of the financial risk on the
consumers’ utility of using E-payment services in a diminishing
rate, that is, the per-dollar effect of monetary reward on reducing
the negative impact of financial risk. Thus, there is a positive but a
diminishing impact of monetary reward on consumers’ utility of
using E-payment services.
H1.1. The monetary reward provided by an E-payment service has
a positive, but a diminishing impact on the consumers’ utility of
using the E-payment service.
Value
f(X2 )
f(X1)
X4 X3
Losses
X1 X 2
f(X3 )
f(X4 )
Fig. 2. A hypothetical example of value function in prospect theory.
671
While monetary reward stimulates consumers to use
E-payment services through enhancing their utility on using the
service, to cap the consumers’ liability for losses is an alternative
approach for attracting consumers to use the E-payment services,
which reduce the uncertainty (i.e., the potential level of financial
lost) in using E-payment services [39–41]. Due to the risk of facing
Internet fraud [2], consumers need additional incentives to
compensate for their potential losses caused by using E-payment
services to purchase goods and services online. Thus, many
E-payment service websites and third-party payment systems
have set an upper limit of financial liability for their consumers.
Based on the premises of the prospect theory [39,40], we suggest
that the negative utility of the per-unit liability for losses also has a
diminishing return. As shown in the third quadrant of Fig. 2, people
usually underestimate the expected value for high loss and
overestimate the expected value for low loss (see the comparison
between x3 and x4, and f(x3) and f(x4)). As a result, the slope or f0 (xi),
which is the per-unit effect of loss, is smaller for high loss when
compared with the low loss, which is a diminishing effect (i.e.,
f0 (x4) > f0 (x3)). Based on this reason, we make the conjecture that
there is a negative, but a diminishing impact on the amount of
consumers’ liability for losses on the consumers’ utility of using
E-payment services.
H1.2. The consumers’ liability for losses set by an E-payment
service has a negative, but a diminishing impact on the consumers’
utility of using the E-payment service.
2.3. Performance risk: physical control feature
The second group of design attributes in our research model is
designed to reduce the negative impact of performance risk on the
consumers’ utility of using E-payment services. Performance risk,
in the context of the usage of an E-payment service, is the
assessment conducted by the consumer on ‘‘potential performance
problems, malfunctioning, transaction processing errors, reliabili-
ty and/or security problems’’ that prevent an E-payment service
being able to perform as expected [6]. As suggested by our group of
practitioners and E-payment service users, we concentrated on the
security problem of the performance risk, as this is the most
important facet of performance risk, which most online customers
care about.
One of the common methods to reduce the security problem in
general is to implement an authentication process into the system.
In this study, we investigated into the impact of using a physical
security feature in the authentication process on consumers’ utility
of using E-payment services. Prior literature shows that the
provision of security features has a significant impact on
consumers’ utility in using telephone shopping [42], which has
a high similarity to today’s E-payment services. In addition, the use
of physical security devices for authentication is a common
practice in information security [43]. Therefore, we anticipate that
the use of a physical security feature for authentication reduces the
negative impact of performance risk on the consumers’ utility of
using the E-payment service. Hence, we have our H2 as follows:
Gains
H2. Incorporating a physical control feature into an E-payment
service has a positive effect on consumer utility, as compared to
E-payment services without a physical control feature.
2.4. Privacy risk: anonymity payment method and offline transfer of
information
The third group of design attributes in our research model is
designed to reduce the negative impact of privacy risk on
consumers’ utility of using E-payment services, which is receiving
672 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681

an increased concern by Internet users [44]. Privacy risk, in the
context of the usage of E-payment services, is the assessment
conducted by the consumer on ‘‘potential losses to the privacy and
confidentiality of personally identification information’’ which can
lead to potential identity theft [6]. Privacy risk, and in particular
identity theft, is one of the main causes that makes consumers
hesitate to complete their transactions online. When personal data,
such as name, social security number, and credit or debit card
information, are transferred online, it is always possible that these
data are unlawfully intercepted by criminals. This issue has
become a predominant problem as more and more users are using
wireless fidelity (Wi-Fi) networks to get online. Tsai et al. [45]
showed that some online consumers are more willing to purchase
items from a website with privacy protection even if they need to
pay a price premium. Hence, privacy protection provides
consumers’ utility of the online purchase processes. Therefore,
we conjecture that E-payment users prefer to have their payments
handled anonymously, as it reduces the risk of identity theft and
provides privacy protection. In this case, even if criminals
successfully intercept the information transmitted online or
through the Wi-Fi, they cannot have full access to the information
and cannot steal the identity of the users. For example, the E-
payment service users can use their pre-assigned token for the
processing of their payment and thus, they do not need to provide
their personal data in the transaction processes. Therefore, the
payment can be processed with anonymity. We also conjecture
that E-payment service users prefer to have their information
transferred offline compared to transferring their information
online as it provides even further protection to their privacy. This
arrangement can be done by allowing consumers to send payment
requests from their computers to the E-payment service provider
(which has the personal information and credit card information in
the database when the consumer registered the service through a
paper-based application) and the E-payment service provider will
transfer such information to the E-vendor directly. Such arrange-
ment can help E-payment service users to enjoy the benefits of
online payment (i.e., easier to search for products online and to
complete the transactions in a fast and paperless way) as well as
protecting their personal identity (i.e., their personal information
is transferred from the E-payment service provider to the E-vendor
directly through an offline channel). Thus, we have the following
two hypotheses:-
H3.1. Consumers’ utility of using an E-payment service with the
anonymity payment method is higher than that with the named
payment method.
transaction as they do not need to waste their time in searching,
setting up, and switching to E-payment service when they make
their purchase with different online merchants, and thus, reduces
the time risk.
We anticipate that if the level of the acceptability of E-payment
services on the Internet is high, the consumers’ utility of using the
E-payment service is also high. It is because the users of the
E-payment service enjoy more convenience of online shopping, if
the E-payment services concerned are highly accepted by
E-vendors [42,46] and the time risk is reduced. This can help to
log in customers for future usage [47,48]. In this study, we tested
the impact of the acceptability of E-payment services by E-vendors
on consumers’ utility at two levels, that is, the E-payment service is
either accepted by all websites or accepted only by some pre-
selected websites. Thus, H4 is as follows:
H4. The level of acceptability of E-payment service by E-vendors
has a positive impact on the consumer’s utility of using the
E-payment service.
Table 2 summarizes the six attributes of E-payment service
studied in this research.
2.6. The Impact of radio-frequency identification -based stored-value
smart-card type
In this study, we used a radio-frequency identification (RFID)-
based stored-value smart card as the E-payment service platform
for our investigation. This RFID-based stored-value smart card is
used commonly in the city in which we conducted our investiga-
tion. This smart card is initially developed as a stored-value card
for the subway system in the city, which has now evolved as the
most predominant off-line payment smart-card system in the
metropolis which carries most of the features of a debit card. While
it can be used for paying transportation fares for nearly all
transportation systems in the city, it can also be used for settling
transactions in supermarkets and convenience stores. The smart
card users can top up the stored value of the card by cash at train
stations, supermarkets, and convenience stores, or by direct debit
authorization via their bank or credit card accounts. There are two
types of smart cards: an anonymous card (with a unique card ID)
Table 2
Design attributes of E-payment service.
Design attribute Level

Monetary reward per  US$0.64

H3.2. Consumers’ utility of using an E-payment service with the transaction (H1.1)  US$1.28
offline transfer of information is higher than that with the online  US$6.41
transfer of information. Consumers’ liability for  US$64.1
losses per transaction  US$128.2
(H1.2)  US$256.4
2.5. Time risk: acceptability
Physical control  A user is required to present his/her smart
feature (H2) card to complete the payment process.
The last attribute that we studied is the acceptability of  A user is not required to present his/her
E-payment services by E-vendors, which we conjecture reduces smart card to complete the payment
the negative impact of time risk on consumers’ utility of using process.
E-payment services. Featherman and Wells [6] defined time risk as Payment method (H3.1)  The payment reveals user’s identity.
the assessment by consumers on ‘‘the potential loss of conve-  The payment is anonymous.
nience, time and effort caused by wasting time searching,
Information transfer  A user inputs and sends information
purchasing, setting up, switching to, and learning how to use method (H3.2) over the Internet.
the E-payment service.’’ Ho and Ng [28] also suggested that time  A user does not need to input and to
risk is the risk of ‘‘taking up more time to complete a transaction’’ send information over the Internet.
online. In this study, we define acceptability as the level of Acceptability of E-payment  The payment method is accepted
acceptance of the E-payment service by the online merchants. If services (H4) everywhere online.
more online merchants are willing to use a particular E-payment  The payment method is accepted by a few
web sites only.
service, the users will take less time to complete an online
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
without direct debit authorization function and users needing to
use cash to top up its value and a personalized card with the name
of the user imprinted on the card and embedded with the direct
debit authorization function. When a user is using these cards to
conduct a transaction, the system will only capture the ID of the
card and transmit it through the payment system. Therefore, the
personal information of the user will not be transmitted
throughthe system. It is obvious that the direct debit authoriza-
tion function of the personalized card will inherit a higher level
of risk to its users when the card is being stolen, as there is a
chance that the personal information of the users will be stolen
and the culprits will be able to steal more money through the
direct debit authorization function. Hence, we conjecture users
of the personalized card would be the group of users who are
more readily to take risks, whereas the users of the anonymous
card would be more likely to be risk averse users. As prior
research reports perceived risk level has an impact on a user’s
intention to purchase online [26], we anticipate that these two
groups of users would have different attitudes toward the
E-payment service we investigate in this study. Hence, we have
our H5 as follows:
H5. The users of anonymous card and the users of personalized
card will have different attitudes toward the different design
features investigated in this study.
3. Research methodology
3.1. Design of the online experiment
In our study, the experimental subjects were first presented
with a proposed version of the E-payment service to be provided by
the RFID-based stored-value smart card that included a combina-
tion of particular levels within the six design attributes that we
wanted to investigate. As suggested by our group of practitioners
and E-payment service users, we studied the impact of monetary
reward on consumers’ utility using E-payment services with three
different monetary levels, that is, US$0.64 (i.e., the price of a can of
soft drink), US$1.28 (approximately the price of a local magazine),
and US$6.41 (approximately the price of a set meal in a café),
respectively. We also tested the impact of consumers’ liability to
losses per transaction on consumers’ utility using E-payment
services with three different monetary levels as proposed by the
group, that is, US$64.1, US$128.2, and US$256.4. We studied the
impact of physical control on consumers’ utility by incorporating
the said smart card as a physical security feature for authentication
during a transaction (Code = 1), compared with conducting a
transaction without such authentication procedure (Code = 0).
Concerning our investigation on the privacy risk, we tested its
impact by two methods, that is, the anonymous payment method
(Code = 0) versus a named payment method (Code = 1), and offline
transfer of information (Code = 0) versus online transfer of
information (Code = 1). Finally, we tested the level of acceptability
of E-payment services by proposing that the smart card system can
be used for settling online payment for all websites (Code = 1) or
for a limited number of websites (Code = 0).
We used a conjoint analysis to analyze the impact of the six
design attributes of E-payment services on the consumers’ utility
of E-payment service. Conjoint analysis is one of the methods that
researchers in economics [49], psychology [50], e-commerce
product development [51], and IS use [18] for analyzing the
relationship between the utility and the importance of different
attributes of a product or a service in various marketing research
studies [52]. There are three major types of conjoint analysis,
namely full profile conjoint, adaptive and hybrid conjoint, and
673
choice-based conjoint [53]. In a full profile conjoint study, the
experimenter presents the product or service as complete profiles
consisting of all the attributes to the subjects and the subjects are
asked to indicate their rank and utility score for the product or
service [54]. In an adaptive and hybrid conjoint study, which is
designed to handle scenarios having an overwhelming number of
attributes in the product/service, the subjects will be first asked to
state their preferences and attribute importance before they
review the profiles so that the researchers can reduce the number
of profiles which the subjects need to review [55]. However, in a
choice-based conjoint study, which applies the discrete choice
model in the conjoint analysis, the subjects will be going through a
sequence of experiments asking them to select their responses
from a choice set made of different profiles [56].
In this study, we used the full profile conjoint to analyze our
data as many attributes included in our model. IS studies using
conjoint analysis analyze problems such as IT application services
outsourcing [57], adoption of new computing architectures [58]
and decision support systems [12], and adoption of computer
systems [59]. The data collected from a full conjoint analysis can be
analyzed by multiple regression analysis.
Based on the premises of conjoint analysis, each of these
proposed versions of attributes is called a conjoint stimulus. In
total, there are a maximum of 3  3  2  2  2  2 = 144 conjoint
stimuli available in this study. To make our study manageable, we
used the optimal orthogonal design proposed by Addelman [60] to
select 18 optimal designs for collecting our experimental data via
an online experiment. The selection of these optimal designs can
also be done using an appropriate statistical software (such as
SPSS) or performed manually based on the methods reported in
the literature [17,60,61]. The 18 optimal orthogonal designs used
in this study are reported in Appendix A.
3.2. Control variables: gender, age, and educational background
Apart from studying the impacts of the six design attributes of
E-payment services on consumers’ utility, we also included three
control variables in our data analysis, that is, gender, age, and
educational background. Prior IS research revealed that gender
factors have an impact on the intention to adopt IS [46,62–
66]. Other behavioral researchers reported that gender factors
have significant impacts on decisions under risk [67], in handling
tangible and intangible claims [68], and in considering objective
and subjective claims when facing the risk [69]. Gender was coded
as a dichotomous variable, that is, male = 0, female = 1, in our
analysis.
Age has been shown as a mediator on the adoption of
technology [66]. Thus, it has been included in our model as a
control variable and was coded as a continuous variable
[70]. Similarly, educational background was coded as a dummy
variable, that is, high school graduate or below = 1, college
graduate (undergraduate and associate degree) = 2; and graduate
degree holders = 3.
4. Data collection
To study the impacts of the six design attributes on the
consumers’ utility of E-payment services, we conducted an online
experiment using the Government-to-Citizen (G2C) E-government
portal of the municipal government concerned. We invited all the
160,000 subscribers of the G2C E-government portal to participate
in this online experiment, which is one of the three different online
experiments that we conducted simultaneously. To be eligible for
participating in our online experiment, the participants were
required to use their smart card to register on the online
experiment website. Some small-value cash prizes as incentives
674 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681

were provided to our participants through a lucky draw. We confirmed their selections, we asked them to rate their selected
collected over 2500 responses from this online experiment during stimuli with a score ranging from 0 to 100. Based on the theory of
the 4-week data collection period. After cleaning our data for conjoint analysis, the scores reported were the proxy of their
missing or contradictory information, we got 1795 subjects, consumers’ utility. Thus, it was used as the dependent variable in
providing a total of 10,879 pieces of usable responses. The this study.
demographic data are summarized in Table 3. When we compared
the demographic profiles of our subjects with the local census
about the Internet users, we noted that there is no significant 5. Data analysis
difference between the age, gender, and educational background.
The only difference is that our subjects have more experience in The data collected from our online experiment were analyzed
conducting online purchase. This difference may be because our using a conjoint analysis with the following regression model
subjects are E-government portal users, who usually use the portal (i.e., Eq. (2)):
to obtain information and pay for government bills. Therefore, they X
are more likely to have prior experience in conduct online Utility ¼ a þ OutcomeMonetary Rewardj PW Monetary Rewardj
payment. As this research is investigating the design attributes j 2 f1:28;6:41g
of E-payment service, we are of the view that our subjects have X
þ OutcomeLiabilityj PW Lost Protection
more online purchase experience is good for this study as they can j 2 f128:2;256:4g
j

better understand the processes related to E-payment service.

þ OutcomeEverywhere PW Everywhere þ OutcomeAuthentication
When we conducted our experiment, the operator of the RFID-
based smart card was considering launching a new service for PW Authentication þOutcomeIdentity PW Identity þOutcomeonline
their customers by extending its service to E-payment services. PW online þ Gender þ Age þ Educatonal Background
Therefore, the operator wanted to have a detailed market analysis
on the impacts of different payment protection attributes on The coefficients of the regression equation are the marginal
consumers’ utility using the card through conducting an online utility of the respective attributes in respect to the consumers’
experiment. As this was a proposed E-payment service, we utility of using E-payment services. The results of the regression
anticipated that our subjects might be unfamiliar with the models developed for the conjoint analyses for the overall sample
proposed format of operation. Therefore, to ensure our subjects and for both types of smart card are presented in Table 4. As shown
had an equivalent understanding of this new business initiative, in Table 4, the F-statistics for the three regression models are
an online video was provided to demonstrate the operation of between 21.639 and 51.783, with the range of p-value of 1044 to
each of these attributes before they participated in the online 10112, which indicates that the F-statistics are significant.
experiment. The content of the online video had been viewed by However, the R2 values and the R2adj values are relatively low.
our group of practitioners and E-payment service users and other The significant F-statistics indicate that these three conjoint
independent IS researchers. It was agreed that the content models fit the respective dataset well. As the significance of the
provided gave the plain facts of the functionality of the design coefficients of the regression models obtained is possibly due to
attributes and would not create any bias toward the design the huge sample size (R2 is small), we follow the procedures
attributes. suggested by Murphy et al. [72] and conducted a test for the
Our online experiment consisted of two parts. First, our minimum-effect hypotheses on these regression models using a
subjects were asked to provide their demographic data and their 0.01 level of effect size, a 0.05 level of alpha, and 0.80 for power. As
smart card information. Afterward, we introduced the six design the R2 values of our regression models are all greater than the
attributes of the E-payment services concerned in this study threshold value of 0.015, we can conclude that the R2 values of
through the video previously mentioned. Then our subjects were these three conjoint analyses are significant, and the significance of
asked to review all 18 sets of optimal orthogonal conjoint stimuli the coefficients is owing to the effect of the factors studied, rather
selected using the method mentioned above [60], and select those than the huge sample size. Indeed, other conjoint analyses in
conjoint stimuli, which they considered using for making online E-commerce also report relatively small R2 values, close to our
payments, that is, the acceptable designs [71]. After they findings [73]. As suggested by Harrison et al. [74], the low adjusted
R2 values in conjoint analysis are usually caused by the differences
in each subject’s subjective rating for the same service, wherein
additional variation is introduced to the regression models due to
Table 3 the aggregating effect. As the dependent variable of our study, that
Demographic data.
is, utility, is a psychometric measurement and is usually spread out
Overall Male Female in a relatively board range among different individuals, we assume
Number of subjects 1795 920 875 that the low R2 and adjusted R2 values are not due to the weak
(51.3%) (48.7%) relationship between our dependent variable with our indepen-
Type of RFID-based stored-value smart card dent variables, but due to the variance within the dependent
 Anonymous card 958 459 499 variable as suggested by prior literature [75].
(47.9%) (52.1%)
As there are two types of smart card, we first analyzed our
 Personalized card 837 461 376
(55.1%) (44.9%) results with the overall sample, and then we split the dataset into
Average age 29.0 30.2 27.6 two groups based on the card type to check if there was any impact
Educational background related to the card type. For the overall sample, all coefficients of
 High school or below 1246 590 656
the regression model, except gender types, have significant results,
(64.1%) (75.0%)
 Undergraduate degree 447 257 190
that is, p < 0.05. The coefficient of gender types is marginally
(27.9%) (21.7%) significant with p < 0.1. The ranking of the relative importance of
 Graduate degree 102 73 29 these attributes is monetary reward per transaction (32.9%),
(8.0%) (3.3%) information transfer method (18.8%), acceptability of E-payment
With online purchase experience 1096 636 460
services (17.1%), consumers’ liability for losses per transaction
(61.1%) (69.1%) (52.6%)
(13.4%), payment method (11.2%), and physical control feature
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681 675

Table 4
Result of conjoint analyses: overall sample and different types of smart card users.

Attributes Levels Overall Anonymous card Personalized card

Coeff. R.I. (%) Coeff. R.I. (%) Coeff. R.I. (%)

Monetary reward per transaction US$0.64 n.a. 32.9% n.a. 30.7% n.a. 34.8%
US$1.28 1.761*** 1.332** 2.277***
US$6.41 5.528*** 4.392*** 6.856***

Consumers’ liability for losses per transaction US$64.1 n.a. 13.4% n.a. 9.7% n.a. 16.6%
US$128.2 1.368*** 1.011* 1.751***
US$256.4 2.251*** 1.379** 3.274***

Physical control feature No control n.a. 6.6% n.a. 7.4% n.a. 6.1%
Payment with authentication 1.115*** 1.061** 1.197**

Payment method Anonymous payment n.a. 11.2% n.a. 12.2% n.a. 10.1%
Payment with identity 1.885*** 1.749*** 1.986***

Information transfer method Offline data transfer n.a. 18.8% n.a. 18.2% n.a. 19.4%
Online data transfer 3.159*** 2.603*** 3.828***

Acceptability of E-payment services Selected websites n.a. 17.1% n.a. 21.8% n.a. 13.0%
Everywhere online 2.865*** 3.113*** 2.563***

Control variables Gender 0.559* 0.312 0.767

Age 0.166*** 0.160*** 0.171***
Educational background 1.818*** 1.837*** 1.779***

R2 0.050 0.040 0.064

Adj. R2 0.049 0.038 0.062
F-Value 51.783*** 21.639*** 31.795***

Notes: ‘‘n.a.’’ stands for not applicable; ‘‘Coeff.’’ stands for coefficient; ‘‘R.I.’’ stands for relative importance.
***
p < 0.01.
**
p < 0.05.
*
p < 0.1.

(6.6%) respectively. In summary, our basic hypotheses, that is, H2,
H3.1, H3.2, and H4 are supported by our empirical data.
When we split our dataset based on the types of smart card, we
noticed that the ranking of relative importance of these attributes
for the two types of smart card were different. First, for users of the
anonymous card, their ranking of the relative importance
is monetary reward per transaction (30.7%), acceptability of
E-payment services (21.8%), information transfer method
(18.2%), payment method (12.2%), consumers’ liability for losses
per transaction (9.7%), and physical control feature (7.4%). In
addition, consumers’ liability for losses per transaction at the first
level (i.e., with US$128.2 excess) is marginally significant, that is,
p < 0.1, in the regression model. For users of a personalized card,
the ranking of the relative importance of these attributes is
monetary reward per transaction (34.8%), information transfer
method (19.4%), consumers’ liability for losses per transaction
(16.6%), acceptability of E-payment services (13.0%), payment
method (10.1%), and physical control feature (6.1%). Hence, the
card type interacts with the ranking of the relative importance of
these attributes, and thus, supports our H5.
To analyze the diminishing returns of monetary reward and the
consumers’ liability to losses proposed in our H1.1 and H1.2,
respectively, we further analyzed the results obtained. The result is
presented in Table 5, where the ratios of per-dollar impact of
monetary reward per transaction on consumers’ utility for US$1.28
and US$6.41 of our overall sample and for both types of smart card
range between 1.52 and 1.66. The results also show that the per-
dollar impact of monetary reward for US$6.41 for all three cases is

Table 5
Diminishing return of monetary reward and consumers’ liability for losses.

Category Overall Anonymous card Personalized card

Monetary reward Per US$ impact on consumers’ utility of US$1.28 reward per transaction (a) 1.38 1.04 1.78
Per US$ impact on consumers’ utility of US$6.41 reward (b) 0.862 0.685 1.07
Ratio of per US$ impacts on consumers’ utility 1.60 1.52 1.66
(c) = (a)/(b)

Consumers’ liability for losses Per US$ impact on consumers’ utility of US$128.2 liability (d) 0.0107 0.00789 0.0137
Per US$ impact on consumers’ utility of US$256.4 liability (e) 0.00878 0.00538 0.0128
Ratio of per US$ impacts on consumers’ utility 1.22 1.47 1.07
(f) = (d)/(e)

Notes:
(1) The values of the first row are calculated by dividing coefficients of the respectively conjoint analysis by US$1.28. For example, the per-US$ impact on consumers’ utility of
US$1.28 reward per transaction is calculated by dividing the coefficient shown in Table 4 (i.e., 1.761) by US$1.28, that is, 1.761/1.28 = 1.38.
(2) The values of the second row calculated by dividing coefficients of the respectively conjoint analysis by US$6.41. For example, the per-US$ impact on consumers’ utility of
US$6.41 reward per transaction is calculated by dividing the coefficient shown in Table 4 (i.e., 5.528) by US$6.41, that is, 5.528/6.41 = 0.862.
(3) The values of the third row are calculated by dividing coefficients of the respectively conjoint analysis by US$128.2. For example, the per-US$ impact on consumers’ utility
of US$128.2 liability is calculated by dividing the coefficient shown in Table 4 (i.e., 1.368) by US$128.2, that is, 1.368/128.2 = 0.0107.
(4) The values of the fourth row are calculated by dividing coefficients of the respectively conjoint analysis by US$256.4. For example, the per-US$ impact on consumers’
utility of US$256.4 liability is calculated by dividing the coefficient shown in Table 4 (i.e., 2.251) by US$256.4, that is, 2.251/256.4 = 0.00878.
676 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
less than the same per-dollar impact of monetary reward for
US$1.28. This reflects that the per-dollar impact of monetary
reward is positive, but it decreases as the amount of reward
increases, which supports our H1.1. Similar impacts are observed
for consumers’ liability for losses per transaction, ranging from
1.07 to 1.47. The result also shows that the per-dollar effects of the
negative impact of consumers’ liability for losses per transaction
for US$256.4 for all three cases are less than the same per-dollar
impact for US$128.2. This shows that the per-dollar effect of the
negative impact of consumers’ liability for losses per transaction is
negative, but it decreases as the liability increases for the overall
sample and for anonymous users, which supports H1.2. We also
notice that the impact of diminishing return is stronger for the
monetary reward than the consumers’ liability for losses.
6. Discussion
6.1. The analysis of relative importance: overall sample
Table 6 and Fig. 3 summarize the ranking of the relative
importance of all design attributes of the regression models that
we developed in this study, that is, the results for the overall
sample, anonymous card, and personalized card, respectively. We
observe that monetary reward per transaction is the first rank in
the relative importance of the respective conjoint analyses. This
result indicates that monetary reward is the most important
attribute of E-payment services and it attracts more consumers to
use E-payment services. It is probably because cash rebates and
coupons provide immediate rewards to online consumers. Thus, it
brings the most significant positive impact to consumers’ utility in
using E-payment services. Besides, we note that its impact is
diminishing, that is, the positive per-US$ impact of monetary
reward per transaction is reduced when the amount of reward is
increased.
Our result for the overall sample shows that the information
transfer method is the second most important feature for
affecting their consumers’ utility of E-payment services.
Basically, our subjects prefer to have their transaction informa-
tion be transferred offline as predicted in our H3.2. Nowadays,
many people who used to participate in online transactions are
reluctant to do so due to the hazards of identity theft. Thus, even
though it is more efficient to transfer their data online, our
subjects prefer not to transfer their online transaction informa-
tion via the Internet while using the E-payment service, that is,
they do not want to type in their personal and financial
information in the browser and send it to the E-vendor. They
prefer the E-payment service provider to transfer such
information to the E-vendor directly.
Acceptability of E-payment services by E-vendors is the third
most important feature affecting consumers’ utility of E-payment
services in our overall result. This result may arise from the impact
of the convenience provided to the consumers as proposed in H4.
Table 6
Relative importance of the attributes.
Ranking of relative Overall sample
importance
1st Monetary reward per transaction (32.9%)
2nd Information transfer method (18.8%)
3rd Acceptability of E-payment services (17.1%)
4th Consumers’ liability for losses per
transaction (13.4%)
5th Payment method (11.2%)
6th Physical control feature (6.6%)
40.0%
35.0%
30.0%
Relative Importance
25.0%
20.0%
15.0%
10.0%
5.0% Anonymous
Card
0.0% Personalized
Monetary Consumers' Physical Payment Information Acceptability Card
Reward Liability to Control Method (3.1) Transfer (H4)
(H1.1) Losses Feature (H2) Method (3.2) Overall Result
(H1.2)
Fig. 3. Relative importance of different design attributes.
Consumers’ liability for losses per transaction is the fourth most
important feature, which also has a diminishing effect similar to
the impact of monetary reward per transaction, whereas payment
method is the fifth most important feature affecting the
consumers’ utility of using E-payment services. The latter result
is also proposed in H3.1.
It is noted that the physical control feature is the least
important, but still a significant feature among these six features of
E-payment services, which supports H2. As the RFID-based stored-
value smart card concerned is designed for micro-payments, its
users are used to using it only for micropayments. Thus, they
probably feel that there is no imminent need to use the physical
security feature in settling their micropayments. Hence, our
subjects rank it as the least important feature among the six design
attributes investigated in this study.
6.2. The impact of RFID-based stored-value smart card type
We note that the relative importance of the design attributes is
related to the RFID-based stored-value smart card type, which
supports our H5 as mentioned. To further analyze this effect, we
conducted two moderating regression analyses. The results of the
moderating regression analyses are shown in Table 7. While the
first moderating regression analysis (i.e., Moderating Regression
(A)) includes all the possible moderating effects from the RFID-
based stored-value smart card, which we used to identify the
factors that are being moderated, the second one (Moderating
Regression (B)) only includes those moderating relationships
having significant effects. As shown in our Moderating Regression
(A), only three interaction terms, that is, US$6.41 reward per
transaction (M2), US$256.4 consumers’ liability of losses per
transaction (L2), and online data transfer (ODT), possibly interact
with the smart card type. To test for whether the moderating
effects studied in Moderating Regression (B) are significant or not,
we test for the significance of the change of DR2 using the method
Anonymous card Personalized card
Monetary reward per transaction (30.7%) Monetary reward per transaction (34.8%)
Acceptability of E-payment services (21.8%) Information transfer method (19.4%)
Information transfer method (18.2%) Consumers’ liability for losses per
transaction (16.6%)
Payment method (12.2%) Acceptability of E-payment services (13.0%)
Consumers’ liability for losses Payment method (10.1%)
per transaction (9.7%)
Physical control feature (7.4%) Physical control feature (6.1%)
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681 677

Table 7
Result of moderating regression analysis.

Independent variable Main effect Moderating regression (A) Moderating regression (B)

Coeff. t-Value Coeff. t-Value Coeff. t-Value

US$1.28 reward per transaction (M1) 1.763*** 4.331 1.337** 2.392 1.770*** 4.351
US$6.41 reward per transaction (M2) 5.529*** 13.391 4.400*** 7.770 4.591*** 8.701
US$128.2 liability (L1) 1.362*** 3.348 1.013* 1.813 1.355*** 3.333
US$256.4 liability (L2) 2.242*** 5.321 1.383** 2.397 1.477*** 2.682
Payment with authentication (PA) 1.112*** 3.171 1.060** 2.198 1.120*** 3.195
Payment with identity (PI) 1.889*** 5.421 1.745*** 3.661 1.870*** 5.369
Online data transfer (ODT) 3.159*** 8.838 2.604*** 5.318 2.607*** 5.494
Accepted everywhere online (AE) 2.872*** 8.199 3.110*** 6.510 2.862*** 8.174
Personalized card (PER) 0.354 1.076 0.935 0.983 0.787 1.313
M1  PER – – 0.937 1.149 – –
M2  PER – – 2.454*** 2.966 2.037*** 2.898
L1  PER – – 0.732 0.899  
L2  PER – – 1.881** 2.230 1.679** 2.185
PA  PER – – 0.134 0.190  
PI  PER – – 0.249 0.357  
ODT  PER – – 1.220* 1.704 1.213* 1.819
AE  PER – – 0.545 0.776  
Constant 65.550*** 77.952 65.260*** 67.571 65.342*** 74.892

Gender 0.523 1.560 0.520 1.550 0.522 1.556

Age 0.165*** 8.162 0.166*** 8.177 0.166*** 8.174
Education 1.808*** 6.527 1.813*** 6.545 1.809*** 6.534

R2 0.050 0.051 0.051

Adj. R2 0.049 0.050 0.050
F-Value 47.566*** 29.453*** 39.078***

Notes: ‘‘Coeff.’’ stands for coefficient.

***
p < 0.01.
**
p < 0.05.
*
p < 0.1.

proposed by Carte and Russell [76]. We noted that the F-statistics
concerned is 4.919, with p < 0.01. Therefore, we confirmed that the
moderating effects are significant. Based on the results in Table 7,
we observed that the impact of US$6.41 monetary reward per
transaction, US$256.4 consumers’ liability of losses per transac-
tion, and online data transfer on consumers’ utility of E-payment
service is moderated by the smart card type.
We suggest that the impact of smart card types on the relative
importance of the design attributes is related to the risk aversion
level of the respective types of users, as prior IS research showed
that risk aversion has an impact on the adoption of E-payment
services [77]. Users of anonymous cards face a lower level of risk in
using that for an E-payment service as the anonymous version of
the card, which does not capture any personal information on the
card. In addition, as it does not have the direct debit authorization
feature, the financial risk of using it is smaller than for a
personalized card. However, users of personalized cards face more
privacy risk (as their cards have their personal information) and
financial risk (as they may lose more money due to the direct debt
authorization arrangement). This is also consistent with the result
of the moderating regression analysis, which shows that person-
alized card users are more sensitive to the attributes related to
financial risk (i.e., more positive and negative responses to
monetary reward and consumers’ liability for losses, respectively)
and privacy risk (i.e., more negative responses to online data
transfer of information).
When we further investigated the relative importance of users
of anonymous cards and users of personalized cards, we noticed
that their rankings of relative importance of these design attributes
are significantly different. While both put monetary reward per
transaction and physical control feature as their first rank and sixth
rank, respectively, users of anonymous cards rank acceptability of
E-payment services by E-vendors, information transfer method,
payment method, and consumers’ liability for losses per transac-
tion as their second to fifth ranks, whereas those personalized card
users rank information transfer method, consumers’ liability for
losses, acceptability of E-payment services by E-vendors, and
payment method as their second to fifth ranks, respectively. These
differences in the rankings of relative importance can also be
explained by the differences in the design of these two types of
smart card.
As the users of anonymous card have limited financial risk,
consumers’ liability for losses brings limited consumers’ utility of
using the E-payment service. As these users also face limited
privacy risk, due to the fact that no personal information is stored
on their cards, whether the online payment method is anonymous
or not and whether the transaction information is transferred
online or not, does not contribute much to the impact level of
consumers’ utility of using the E-payment services. As suggested
by Im et al. [7], perceived usefulness of a technology has a more
significant impact on a user’s adoption of a less risky technology,
we can expect that the users of anonymous card (which is a less
risky technology compared with the personalized card) will rank
those features which are more useful for them to complete the
online payment. Therefore, for these users, acceptability of
E-payment services by E-vendors, which is the feature related to
the usefulness of the anonymous card, becomes the second most
important factor affecting the consumers’ utility of using the
E-payment services to settle online payments.
However, the users of personalized cards face a higher level of
privacy risk and financial risk when they use it to conduct online
payments via the E-payment services. Therefore, consumers’
liability for losses per transaction and online transfer of informa-
tion become two important design attributes for protecting their
wealth and privacy, respectively. Hence, we observe that these two
factors have higher impacts on consumers’ utility of using
678 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
Table 8
Comparison of anonymous and personalized card users in the diminishing returns.
Category
Monetary reward Per-US$ impact on consumers’ utility of US$1.28 reward
Per-US$ impact on consumers’ utility of US$6.41 reward
Consumers’ liability for losses Per-US$ impact on consumers’ utility of US$128.2 excess
Per-US$ impact on consumers’ utility of US$256.4 excess
E-payment services. We also observe that the impact of payment
method on consumers’ utility of using the E-payment services is
relatively low for these users. It is because, if the data are already
transferred offline, the privacy of users would already have been
protected. Hence, the positive impact of the anonymous payment
method on consumers’ utility of using the E-payment service to
settle online payment is reduced as the chance for leaking their
personal information during the transactions is already minimized.
Finally, when we compared the per-dollar impact on con-
sumers’ utility of using the E-payment service for monetary reward
and consumers’ liability for losses, we notice that the impacts are
smaller for users of anonymous cards compared with users of
personalized cards (see Table 8). This result reinforces our
observation that users of personalized card are more sensitive
to the financial impacts compared with users of anonymous card.
7. Theoretical and methodological contributions and practical
implications
This paper has several theoretical and methodological contri-
butions. First, it responds to Venkatesh’s question [14] on the need
to conduct service design research, which probes into the impacts
of different aspects of service attributes toward consumer choice.
This study demonstrates how to study the IS service design
question using a conjoint analysis as a methodology. Our result
shows that IS researchers and practitioners can together investi-
gate service design issues, thus, the research outcomes along with
their theoretical and practical implications can be addressed
[16,34–36]. As suggested by prior research [78], conjoint analysis is
a suitable method to study the joint effect of attributes of a product
or services, and the full conjoint analysis, that is, the one we used in
this study, provides ‘‘exacting’’ result compared with the other two
types of conjoint analysis. Besides, even though other regression
methods, such as ordinal regression, could be used to analyze the
data collected, the results generated by other regression methods
would be difficult for users to interpret the relative importance of
the factors. Therefore, the result of this study shows the possibility
of using conjoint analysis, and in particular, the full conjoint
analysis, to study the design of an IS service.
We also show that conjoint analysis provides an additional
flexibility in analyzing the impact of the design attributes of IS.
Prior studies in TAM [65,66] usually used structural equation
modeling (SEM) to study the impacts of various factors on the
intention of adopting the IS; these factors could be the design
attributes of an IS service measured using a survey. For most of
these studies, the impacts of the factors are assumed to be linear. If
the researchers assume that a factor would have a non-linear
effect, they would need to explore such effect using the squared
term in their SEM developed. However, it is easier for the
researchers to use the conjoint analysis to study such effects. First,
the conjoint analysis needs a different research design, which is
more simple and direct as the respondents are providing their
utility of a particular set of design, instead of responding to a
multiple sets of multiple-item survey constructs. Second, the
design of the conjoint analysis is flexible and is easy for researchers
to probe into nonlinear relationships, such as the diminishing
Anonymous card (a) Personalized card (b) Ratio (b)/(a)
1.04 1.78 1.71
0.685 1.07 1.56
0.00789 0.0137 1.74
0.00538 0.0128 2.38
impact of the monetary reward on the consumers’ utility of using
E-payment services, in exploring this type of information system
design problem. This makes the conjoint analysis has an edge for
studies related to the IS service research with non-linear impacts.
The results of our conjoint analyses reveal consumers’
preferences in the choice of design features, which is the major
theoretical contribution of the study. E-payment service users
prefer to have a high level of monetary reward with minimal (if
possible no) consumers’ liability for losses, that is, reducing
financial risk. In addition, they prefer the E-payment services to be
able to help them to settle online transactions in all websites, that
is, reducing the time risk. Further, they enjoy a high degree of
protection of their personal information as they prefer to have their
transaction data transferred offline with anonymity, that is,
reducing the privacy risk. In brief, this research shows the
possibility of using the conjoint analysis to probe into the design
of a website. We also show that it is possible to use different design
attributes of a website to reduce the negative impact of risk on the
consumers’ utility on E-payment service. In particular, we show
that the provision of monetary reward, which targeted the
negative impact of financial risk, and the method of information
transfer (in anonymity) and the wide acceptance of the E-payment
service, which target to reduce the privacy risk and time risk
respectively, can significantly enhance consumers’ utility on E-
payment service. Therefore, our results provide a framework for
further research in developing theories in service design. Our
results, in particular, have impacts on the literature related to
studies relating to the business impact of IT service design.
Apart from the above theoretical and methodological implica-
tions, our results have two practical implications, which are useful
for practitioners. First, we observed different levels of diminishing
returns on the monetary reward and consumers’ liability for losses
with types of smart card users, which probably arise from the
different levels of risk propensity of the two types of smart card
users. These results provide E-payment service operators and
designers with a yardstick to refine their designs of their E-
payment services. In particular, these results provide them with an
insight into how to allocate their resources for providing monetary
reward and consumers’ liability for losses to their potential
consumers. For example, if customers anticipate that the chance of
facing Internet fraud is lower, the E-payment service provider
should consider providing a low consumers’ liability for losses. It is
because, from the consumers’ perspective, such liability brings a
negative impact on their utility on using E-payment services, and
thus, it will reduce their intention to adopt the E-payment service
concerned. Besides, consumers are very concerned about their
privacy (as reflected by the concern of the information transfer
method) and the acceptance of the E-payment service by other
online vendors. Therefore, practitioners should also incorporate
these features into the E-payment service and emphasize the
existence of these design attributes when they promote their
service to the potential consumers.
Second, the moderating effect of the type of smart card has a
significant implication to practitioners. Our results show that
consumers with different types of smart card who faced different
levels of risk have different preferences on ranking the relative
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681 679

importance of these design attributes of E-payment services, when they purchase different types of product; previous studies
which means that practitioners may need to consider developing showed that customers’ valuation for the customization strategies of
different types of E-payment services for different types of E-retailing has a correlation with the product types [80].
consumers. More importantly, we show that the choice of payment
media, that is, the choice of smart card, of this study, can be a proxy 9. Conclusion
of the risk level faced by consumers. These findings can definitely
help practitioners to enhance their design of E-payment services. This is a pioneering study on E-payment service design
research. We developed our research model based on recommen-
dations from a group with practitioners and users of E-payment
8. Limitations and future research
service through the Delphi method. These design attributes were
developed to counter the negative impacts of various facets of
Similar to other research, this study also has limitations. First,
perceived risk. From the result of our conjoint analyses, we showed
we only considered one type of E-payment service, that is, an RFID-
that we could enhance the consumers’ utility of using E-payment
based stored-value smart card, which is designed for micropay-
services by operationalizing the design attributes of E-payment
ments. Therefore, our results may not be directly applicable to
services. We also showed that conjoint analysis is a promising
other types of E-payment services, such as a credit card payment
method for conducting service design research. Our findings also
system or a third-party payment system. However, as the features
have several significant practical implications. Practitioners may
investigated in this study are commonly used by other types of E-
gain a better understanding of the consumers’ utility implications
payment services, we anticipate that our results are still useful for
of E-payment service design attributes and move forward to better
other types of E-payment service. In addition, prior research results
E-payment service design.
have shown that conjoint analysis can achieve a reasonable level of
predictive accuracy [79]; hence, we assume that our results can
Acknowledgements
provide useful managerial implications.
Second, nonresponse bias is possible for our type of study. As
The work described in this paper was partially supported by a
mentioned in Section 4, we compared the demographic profiles of
grant from the Department of Industrial and Systems Engineering,
our subjects with the city’s demographics reported in a local census
Hong Kong Polytechnic University, Hong Kong, China (Project No.
and noted that they were very similar, except our subjects had more
G-UB98). The authors would like to thank for the help of the two
online purchase experience. This reflected that the samples collected
reviewers and the Associate Editor for helping them to improve the
were representative. Coupled with the high Internet penetration
manuscript. The authors would like to indicate that they have
rate (100%) and computer literacy rate (94%) in the city, we suggest
equal effort in developing this paper.
that the impact of nonresponse bias should be low.
For future research, IS researchers can consider investigating the
impact of culture and other design features on other types of online Appendix A. The conjoint used in this study
payment services, such as a third-party payment system. Another
possible extension of this research is to study the impact of The details of these 18 optimal orthogonal designs are reported
customization on the E-payment service design features by the users in the following table:

Design Monetary Consumers’ Acceptability of Physical control Payment method Information transfer method
rewards per liability for E-payment service feature
transaction losses

1 US$0.64 US$64.1 All websites Use smart card Anonymous payment Online transfer of information
2 US$0.64 US$128.2 Selected websites No card Anonymous payment Offline transfer of information
3 US$0.64 US$256.4 All websites Use smart card Payment with identity Online transfer of information
4 US$1.28 US$64.1 All websites No card Payment with identity Offline transfer of information
5 US$1.28 US$128.2 All websites Use smart card Anonymous payment Online transfer of information
6 US$1.28 US$256.4 Selected websites Use smart card Anonymous payment Online transfer of information
7 US$6.41 US$64.1 All websites Use smart card Anonymous payment Offline transfer of information
8 US$6.41 US$128.2 Selected websites Use smart card Payment with identity Online transfer of information
9 US$6.41 US$256.4 All websites No card Anonymous payment Online transfer of information
10 US$0.64 US$64.1 Selected websites Use smart card Payment with identity Online transfer of information
11 US$0.64 US$128.2 All websites Use smart card Anonymous payment Offline transfer of information
12 US$0.64 US$256.4 All websites No card Anonymous payment Online transfer of information
13 US$1.28 US$64.1 All websites Use smart card Anonymous payment Online transfer of information
14 US$1.28 US$128.2 All websites No card Payment with identity Online transfer of information
15 US$1.28 US$256.4 Selected websites Use smart card Anonymous payment Offline transfer of information
16 US$6.41 US$64.1 Selected websites No card Anonymous payment Online transfer of information
17 US$6.41 US$128.2 All websites Use smart card Anonymous payment Online transfer of information
18 US$6.41 US$256.4 All websites Use smart card Payment with identity Offline transfer of information
680 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
References
[1] D. Hoffman, T. Novak, M. Peralta, Building consumer trust online, Commun. ACM
42 (4), 1999, pp. 80–85.
[2] X. Hu, Z. Lin, A.B. Whinston, H. Zhang, Hope or hype: on the viability of Escrow
services as trusted third parties on online auction environments, Inf. Syst. Res. 15
(3), 2004, pp. 236–249.
[3] S. Ba, P. Pavlou, Evidence of the effect of trust building technology in electronic
markets, MIS Q. 26 (3), 2002, pp. 243–268.
[4] C. Dellarocas, The digitization of word of mouth: promises and challenge of online
feedback mechanisms, Manag. Sci. 49 (10), 2003, pp. 1407–1424.
[5] M. Bernardo, F. Marimon, M. del M. Alonso-Almeida, Functional quality and
hedonic quality: a study of the dimensions of e-service quality in online travel
agencies, Inf. Manag. 49 (7–8), 2012, pp. 342–347.
[6] M.S. Featherman, J.D. Wells, The intangibility of E-services: effects on perceived
risk and acceptance, Data Base Adv. Inf. Syst. 41 (2), 2010, pp. 110–131.
[7] I. Im, Y. Kim, H.-J. Han, The effects of perceived risk and technology type on user’s
acceptance of technologies, Inf. Manag. 45 (1), 2008, pp. 1–9.
[8] W.K. Darley, C. Blankson, D.J. Luethge, Toward an integrated framework for online
consumer behavior and decision making process: a review, Psychol. Mark. 27 (2),
2010, pp. 94–116.
[9] P. Pavlou, Consumer acceptance of electronic commerce: integrating trust and
risk with the technology acceptance model, Int. J. Electron. Commerce 7 (3), 2003,
pp. 101–134.
[10] P.A. Pavlou, D. Gefen, Building effective online marketplaces with institutional-
based trust, Inf. Syst. Res. 15 (1), 2004, pp. 37–59.
[11] S. Gupta, H.-W. Kim, Value-driven internet shopping: the mental accounting
theory perspective, Psychol. Mark. 27 (1), 2010, pp. 13–35.
[12] A. Money, D. Tromp, T. Wegner, The quantification of decision support benefits
within the context of value analysis, MIS Q. 12 (2), 1988, pp. 222–237.
[13] E.W.K. See-To, S. Papagiannidis, V. Cho, User experience on mobile video appre-
ciation: how to engross users and to enhance their enjoyment in watching mobile
video clips, Technol. Forecast. Soc. Change 79 (8), 2012, pp. 1484–1494.
[14] V. Venkatesh, Where to go from here? Thoughts on future directions for research
on individual-level technology adoption with a focus on decision making
Decis. Sci. 37 (4), 2006, pp. 497–518.
[15] C. Okoli, S.D. Pawlowski, The Delphi method as a research tool: an example,
design considerations and applications, Inf. Manag. 42 (1), 2004, pp. 15–29.
[16] A.S. Lee, Retrospective and prospects: information systems research in the last
and next 25 years, J. Inf. Technol. 25 (4), 2010, pp. 336–348.
[17] P.E. Green, V. Srinivasan, Conjoint analysis in marketing: new developments with
implications for research and practice, J. Mark. 54 (4), 1990, pp. 3–19.
[18] I.-H. Hann, K.-L. Hui, S.-Y.T. Lee, I.P.I. Png, Overcoming online information privacy
concerns: an information-processing theory approach, J. Manag. Inf. Syst. 24 (2),
2008, pp. 13–42.
[19] M.S. Featherman, P.A. Pavlou, Predicting E-services adoption: a perceived risk
facets perspective, Int. J. Hum.-Comput. Stud. 59 (4), 2003, pp. 451–474.
[20] R.A. Bauer, Consumer behavior as risk-taking, in: R.S. Hancock (Ed.), Dynamic
Marketing for a Changing World, American Marketing Association, Chicago, IL,
1960, pp. 389–398.
[21] J. Peter, M. Ryan, An investigation of perceived risk at the brand level, J. Mark. Res.
13 (2), 1976, pp. 184–188.
[22] R.N. Stone, K. Grønhaug, Perceived risk: further considerations for the marketing
discipline, Eur. J. Mark. 27 (3), 1993, pp. 39–50.
[23] P. Verdegem, L. De Marez, Rethinking determinants of ICT acceptance: towards an
integrated and comprehensive overview, Technovation 31 (8), 2011, pp. 411–423.
[24] M.L. Meuter, M.J. Bitner, A.L. Ostrom, S.W. Brown, Choosing among alternative
service delivery modes: an investigation of customer trial of self-service technol-
ogies, J. Mark. 69 (2), 2005, pp. 61–83.
[25] A.I. Nicolaou, D.H. McKnight, Perceived information quality in data exchanges:
effects of risk, trust, and intention to use, Inf. Syst. Res. 17 (4), 2006, pp. 332–351.
[26] D.J. Kim, D.L. Ferrin, H.R. Rao, A trust-based consumer decision-making model in
electronic commerce: the role of trust, perceived risk, and their antecedents,
Decis. Support Syst. 44 (2), 2008, pp. 544–564.
[27] M.-L. Chang, W.-Y. Wu, Revisiting perceived risk in the context of online shop-
ping: an alternative perspective of decision-making style, Psychol. Mark. 29 (5),
2012, pp. 378–400.
[28] S.S.M. Ho, V.T.F. Ng, Customers’ risk perceptions of electronic payment systems,
Int. J. Bank Mark. 12 (8), 1994, pp. 26–38.
[29] X. Lou, H. Li, J. Zhang, J.P. Shim, Examining multi-dimensional trust and multi-
faceted risk in initial acceptance of emerging technologies: an empirical study of
mobile banking services, Decis. Support Syst. 49 (4), 2010, pp. 222–234.
[30] M. Featherman, R.T. Wright, J. Thatcher, J.C. Zimmer, R. Pak, The influence of
interactivity on E-service offerings: an empirical examination of benefits and
risks, AIS Trans. Hum.-Comput. Interact. 3 (1), 2011, pp. 1–25.
[31] K. Peffers, T. Tuunanen, Planning for IS applications: a practical, information
theoretical method and case study in mobile financial services, Inf. Manag. 42 (3),
2005, pp. 483–501.
[32] J.L. Worrell, P.M. Gangi, A.A. Bush, Exploring the use of the Delphi method in
accounting information systems research, Int. J. Acc. Inf. Syst. 14 (3), 2013, pp.
193–208.
[33] J. Landeta, J. Barrutia, A. Lertxundi, Hybrid Delphi: a methodology to facilitate
contribution from experts in professional contexts, Technol. Forecast. Soc. Change
78 (9), 2011, pp. 1629–1641.
[34] I. Benbasat, R.W. Zmud, Empirical research in information systems: the practice of
relevance, MIS Q. 23 (1), 1999, pp. 3–16.
[35] K.C. Desouza, O.A. El Sawy, R.D. Galliers, C. Lobbecke, R.T. Watson, Beyond rigor
and relevance towards responsibility and reverberation: information systems
research that really matters, Commun. AIS 17, 2006, Article 16.
[36] K. Lyytinen, Empirical research in information systems: on the relevance of
practice in thinking of IS research, MIS Q. 23 (1), 1999, pp. 25–28.
[37] M. Fishbein, I. Ajzen, Belief, Attitude, Intention, and Behavior: An Introduction to
Theory and Research, Addison-Wesley, Reading, MA, 1975.
[38] V.L. Smith, J.M. Walker, Monetary rewards and decision cost in experimental
economics, Econ. Inq. 31 (2), 1993, pp. 245–261.
[39] D. Kahneman, A. Tversky, Prospect theory: an analysis of decision under risk,
Econometrica 47 (2), 1979, pp. 263–291.
[40] A. Tversky, D. Kahneman, Advances in prospect theory: cumulative representa-
tion of uncertainty, J. Risk Uncertain. 5 (4), 1992, pp. 297–323.
[41] R. Thaler, Mental accounting and consumer choice, Mark. Sci. 4 (3), 1985, pp. 199–
214.
[42] D.F. Cox, S.U. Rich, Perceived risk and consumer decision-making: the case of
telephone shopping, J. Mark. Res. 1 (4), 1964, pp. 32–39.
[43] M.E. Whitman, H.J. Mattord, Principles of Information Security, 3rd edition,
Course Technology, Boston, MA, 2009.
[44] W. Hong, J.Y.L. Thong, Internet privacy concerns: an integrated conceptualization
and four empirical studies, MIS Q. 37 (1), 2003, pp. 275–298.
[45] J.Y. Tsai, S. Egelman, L. Cranor, A. Acquisti, The effect of online privacy information
on purchase behavior: an experimental study, Inf. Syst. Res. 22 (2), 2011, pp. 254–
268.
[46] A. Bhatnagar, S. Misra, H.R. Rao, On risk, convenience, and Internet shopping
behavior, Commun. ACM 43 (11), 2000, pp. 98–105.
[47] K.B. Murray, G. Häubl, Explaining cognitive lock-in: the role of skill-based habits
of use in consumer choice, J. Consumer Res. 34 (1), 2007, pp. 77–88.
[48] G. Zauberman, The intertemporal dynamics of consumer lock-in, J. Consumer Res.
30 (1), 2003, pp. 405–419.
[49] G. Debreu, Topological methods in cardinal utility theory, in: S. Kalin, K.J. Arrow, P
Suppes (Eds.), Mathematical Methods in the Social Sciences, Stanford University
Press, Palo Alto, CA, 1960, pp. 16–26.
[50] R.D. Luce, J.W. Tukey, Simultaneous conjoint measurement: a new type of
fundamental measurement, J. Math. Psychol. 1 (1), 1964, pp. 1–27.
[51] M.G. Helander, J. Jiao, Research on e-product development (ePD) for mass
customization, Technovation 22 (11), 2002, pp. 717–724.
[52] V.-M. Mitchell, Consumer perceived risk: conceptualisations and models, Eur. J.
Mark. 33 (1–2), 1999, pp. 163–195.
[53] D. Bakken, C.L. Frazier, Conjoint analysis: understanding consumer decision-
making, in: R. Grover, M. Vriens (Eds.), The Handbook of Marketing Research:
Uses, Misuses, and Future Advances, (2006), Sage Publications, Inc., Thousand
Oaks, CA, 2006, pp. 288–311.
[54] P.E. Green, V.R. Rao, Conjoint measurement for quantifying judgmental data, J.
Mark. Res. 8 (3), 1971, pp. 355–363.
[55] R.M. Johnson, Adaptive conjoint analysis, in: Proceedings of the Sawtooth Soft-
ware Conference on Perceptual Mapping, Conjoint Analysis, and Computer Inter-
viewing, Sawtooth Software Inc., Orem, UT, 1987, pp. 253–265.
[56] J.J. Louviere, G.G. Woodworth, Design and analysis of simulated consumer choice
or allocation experiments: an approach based on aggregate data, J. Mark. Res. 20
(4), 1983, pp. 350–367.
[57] A. Schwarz, B. Jayatilaka, R. Hirschheim, T. Goles, A conjoint approach to under-
standing IT application services outsourcing, J. Assoc. Inf. Syst. 10 (10), 2009, pp.
748–781.
[58] A. Bajaj, A study of senior information systems managers’ decision models in
adopting new computing architectures, J. Assoc. Inf. Syst. 1 (4), 2000, pp. 1–56.
[59] L. Baker-Eveleth, R.W. Stone, Expectancy theory and behavioral intentions to use
computer applications, Interdiscip. J. Inf. Knowl. Manag. 3, 2008, pp. 35–146.
[60] S. Addelman, Orthogonal main-effect plans for asymmetrical factorial experi-
ments, Technometrics 4 (1), 1962, pp. 21–46.
[61] J.H. Steckel, W.S. DeSarbo, V. Mahajan, On the creation of acceptable conjoint
analysis experimental designs, Decis. Sci. 22 (2), 1991, pp. 435–442.
[62] R. Agarwal, J. Prasad, Are individual differences germane to the acceptance of new
information technology? Decis. Sci. 30 (2), 1999, pp. 361–391.
[63] K.A. Franek, Women and computing, Commun. ACM 33 (11), 1990, pp. 34–45.
[64] D. Gefen, D.W. Straub, Gender differences in the perception and use of E-mail: an
extension to the technology acceptance model, MIS Q. 21 (4), 1997, pp. 389–400.
[65] V. Venkatesh, M.G. Morris, P.L. Ackerman, A longitudinal field investigation of
gender differences in individual technology adoption decision-making processes,
Organ. Behav. Hum. Decis. Process 83 (1), 2000, pp. 33–60.
[66] V. Venkatesh, M.G. Morris, G.B. Davis, F.B. Davis, User acceptance of information
technology: towards a unified view, MIS Q. 27 (3), 2003, pp. 425–478.
[67] H. Fehr-Duda, M. de Gennaro, R. Schubert, Gender, financial risk, and probability
weights, Theory Decis. 60 (2–3), 2006, pp. 283–313.
[68] M.B. Holbrook, Beyond attitude structure, J. Mark. Res. 15 (4), 1978, pp. 546–556.
[69] W.K. Darley, R.E. Smith, Gender differences in information processing strategies:
an empirical test of the selective model in advertising response, J. Advert. 25 (1),
1995, pp. 41–56.
[70] J.H. Dulebohn, An investigation of the determinants of investment risk behavior in
employer-sponsored retirement plan, J. Manag. 28 (1), 2002, pp. 3–26.
[71] V.R. Rao, Theory and design of conjoint studies (Ratings based methods), Applied
Conjoint Analysis, Springer, Berlin, Germany, 2014 pp. 37–78(Chapter 2).
[72] K.R. Murphy, B. Myors, A. Wolach, Statistical Power Analysis, 3rd edition,
Routledge, New York, NY, 2009.
[73] Y.-H. Chen, I.-C. Hsu, C.-C. Lin, Website attributes that increase consumer pur-
chase intention: a conjoint analysis, J. Bus. Res. 63 (9–10), 2010, pp. 1007–1014.
 E.W.K. See-To, K.K.W. Ho / Information & Management 53 (2016) 668–681
[74] R.W. Harrison, A. Özayan, S.P. Meyers, A conjoint analysis of new food products
processed from underutilized small crawfish, J. Agric. Appl. Econ. 30 (2), 1998, pp.
257–265.
[75] F.M. Bass, D.J. Tigert, R.T. Lonsdale, Market segmentation: group versus individual
behavior, J. Mark. Res. 5 (3), 1968, pp. 264–270.
[76] T.A. Carte, C.J. Russell, In pursuit of moderation: nine common errors and their
solutions, MIS Q. 27 (3), 2003, pp. 479–501.
[77] J.M. Kolodinsky, J.M. Hogarth, M.A. Hilgert, The adoption of electronic banking
technologies by US consumers, Int. J. Bank Mark. 22 (4), 2004, pp. 238–259.
[78] R. Furlan, R. Corradetti, An empirical comparison of conjoint analysis model on a
same sample, Stat. Appl. 17 (2), 2005, pp. 141–158.
[79] T.G. Vavra, P.E. Green, A.M. Krieger, Evaluating EZ pass: using conjoint analysis to
assess consumer response to a new tollway technology, Mark. Res. 11 (2), 1999,
pp. 5–16.
[80] S. Thirumalai, K.K. Sinha, Customization strategies in electronic retailing: impli-
cations of customer purchase behavior, Decis. Sci. 40 (1), 2009, pp. 5–36.
Eric W.K. See-To is assistant professor at the Department of Industrial and Systems
Engineering, Hong Kong Polytechnic University, and is currently associate editor of
Electronic Commerce Research. Prior to the current appointment, Eric served as
senior lecturer in the Business School, University of Newcastle-upon-Tyne, UK, and
as lecturer in the Department of Management Science, Lancaster University
Management School, UK. Eric’s research focuses on the interplay between
technology and people, using a (big) data perspective. He studied the strategic
681
implications of technology, and consumer behaviors in mobile and digital
environments, using state of the art text mining and data analytical techniques.
His current projects include the application of text mining techniques to the
financial valuation of innovation and intangible assets, and the analysis of online
opinions and sentiments about social networks and data mining, and other practical
applications of business intelligence and knowledge mining. His papers have been
published in international and top-tier journals, such as Journal of Management
Information Systems (JMIS), Research Policy, Technological Forecasting and Social
Change, Electronic Commerce Research, Electronic Markets, Information Systems
Frontiers, and Journal of Electronic Commerce Research.
Kevin K.W. Ho is an associate professor at the School of Business and Public
Administration, University of Guam, Mangilao, Guam. Kevin’s research interests
include electronic service, information systems strategy, social media, green
information systems, and electronic government. He is an associate editor/editorial
board member in International Journal of Systems and Service-Oriented Engineering,
International Journal of Social and Organizational Dynamics in Information Technology,
International Journal of Technology and Human Interaction, and Journal of
Communication and Education. His papers have been published in Computers in
Human Behavior, Decision Support Systems, Information & Management, Journal of
Electronic Commerce Research, and Journal of Global Information Management, among
others.