Description
This paper presents an empirical investigation into the factors that shape the propensity to use the Internet for shopping and banking through application of bivariate probit regression techniques to data sourced from a survey of 259 respondents in Athens, Greece.
1
Internet shopping and Internet banking in sequence
Athanasios G. Patsiotis
a
, Tim Hughes
b
and Don J. Webber
c
a
Department of Marketing, Deree College, American College of Greece, Athens, Greece
b
Department of Business and Management, University of the West of England, Bristol, UK
c
Department of Accounting, Economics and Finance, University of the West of England, Bristol, UK
Abstract
This paper presents an empirical investigation into the factors that shape the
propensity to use the Internet for shopping and banking through application of
bivariate probit regression techniques to data sourced from a survey of 259
respondents in Athens, Greece. Based on the observation that Internet banking
usage typically requires familiarity with Internet shopping, we estimate the
marginal effects of the determinants of Internet banking use conditioned on the
determinants of Internet shopping use.
Our results suggest that not controlling for this conditioning will bias
estimates and could lead to incorrect policy-recommendations. For instance,
personal capacity is found to be an important determinant of the propensity to use
Internet banking in a non-sequential approach but it is found to have no
significant effect after conditioning. In particular, our results suggest that policy-
makers should emphasise usefulness attributes of computer-based innovations
when attempting to increase the use of the Internet for banking by people who
already use the Internet for shopping.
Keywords: Internet banking; Internet shopping; adoption rate; bivariate probit regression;
conditional marginal effects
Address for correspondence: Dr Athanasios Patsiotis, Department of Marketing, Deree
College, American College of Greece, Athens, Greece. Email: [email protected]
2
1. Introduction
The Internet represents a huge source of information that can be organized and retrieved in
many different ways based on individual users needs (Mahajan et al., 2000). It facilitates
communication and shopping through computer-mediated environments and it is a market
where a large variety of new technologies and interdependent products are introduced
(Mahajan et al., 2000; Varadarajan and Yadav, 2002). It can also be a discontinuous
innovation process and lead to new product developments. Following the introduction and
acceptance of Internet shopping, new technological interfaces developed by banks, such as
Internet banking, are innovative delivery and communication channels where new products
and services are introduced. These innovations have facilitated interaction and the building of
relationships between banks and their customers (Tapp and Hughes, 2004).
New technologies and especially the developments of self-service technologies
present several challenges for banks in terms of their customer relationships. Banks that offer
Internet banking services can benefit from lower costs due to the utilization of less human
and physical resources and the potential of economies of scale in bank operations (Shi et al.,
2008). Consumers’ transferring their decision making processes from traditional offline to
online can engender cost and time savings benefits (Shi et al., 2008) at the expense of various
risks (Durkin, 2007). Consequently banks need to alter their operations and internal and
external communication media, and such major changes can encounter resistance.
1
A number of articles present investigations of the determinants of Internet use for a
variety of services. Such studies typically examine either one Internet service in isolation or
assume away structure or order between Internet services. This paper purports that there is a
sequence of Internet usage choices, with consumers first becoming familiar with the Internet
for their shopping experience and, once proficiency in this area has been achieved, consumers
will then consider using the Internet for banking services. Based on the idea of a conditional
and sequential link between Internet shopping and Internet banking we proceed to examine
empirically the factors that influence the rate of Internet banking adoption. Our results
strongly support the assumption of association between Internet banking and Internet
shopping, and once sequencing has been integrated into the modelling approach we identify
potential conflicting results and important policy levers.
The paper is structured as follows. The next section presents a review of extant
literature and our conceptual model. Section 3 details the data and the modelling approach.
Section 4 is a discussion of findings and implications, and Section 5 concludes.
2. Theoretical background
Innovations can be defined in terms of the amount of behavioural change necessary to use the
innovation effectively; they can be classified along a continuum from the least to the most
disruptive which is related to the extent to which the innovation is functionally new
(Robertson, 1971). Technological innovations can be discontinuous (Moore, 1991), are
typically viewed as being rooted in new information and computer-based technologies, and
can disrupt existing patterns of behaviour (Fitzsimmons and Fitzsimmons, 2011; Littler,
1
For example, customers do not always welcome technology or they may increasingly use a combination of
banking methods. The theoretical literature on diffusion research of such technological innovations is well-
developed but it lacks empirical evidence on non-adopter behaviour and focuses mainly on mental behaviour
(Hernandez et al., 2009). There is limited evidence on the possible differences between pre-adoption and
usage behaviour (Hernandez and Mazzon, 2007).
3
2001; Veryzer, 1998b). Their usage can involve a very high degree of technological
uncertainty, a sequence of innovations based on the core application, a longer development
process, and a greater distance from the end user in terms of customer familiarity with the
innovation and the time it takes to evolve (Veryzer, 1998a, 1998b). Thus, technological
innovations can require a change in the behaviour of potential adopters and the development
of new skills.
The introduction of the core application of the Internet (i.e. linking computers in
networks) has spawned a sequence of Internet based innovations that have facilitated
communication and shopping. Such innovations include Internet shopping and Internet
banking that required major behavioural changes by potential adopters in their business and
personal relationships. Internet shopping is now widespread and typically includes books,
cosmetics, consumer durables and service operations, such as retailing, entertainment and
travel. Internet banking is also available and requires a more sophisticated application of
Internet technologies to satisfy consumers’ banking needs, and it differs from other self-
service innovations in financial services since it requires major changes in behaviour (i.e. a
completely new way of consumer banking).
An order of succession
There may be a logical dependence of engagement with innovations, and usage of Internet
shopping and Internet banking may be a prime example. This paper purports a sequence of
events whereby individuals make a series of decisions:
1) Decide (consciously or otherwise) to use the Internet;
2) After some familiarity of use with the Internet has been achieved, a next step in using
the Internet is for shopping;
3) After some familiarity of use with the Internet for shopping has been achieved, a next
step in using the Internet is for banking.
Decision (1) is beyond the scope of this paper. This paper examines empirically the factors
that contribute to decisions (2) and (3) for a sample of Internet users.
The decisions above are each associated with different, albeit potentially sequential
innovations that require interaction with technological interfaces and necessitate a degree of
behavioural change by potential adopters. The conceptual framework for this study, shown in
Figure 1, illustrates that someone must adopt the Internet first, and then must adopt Internet
shopping as a prerequisite to deciding whether to adopt Internet banking. The ordering
purported in Figure 1 corresponds to both the sequence of introduction of the above
technological interfaces (shopping and banking) and the degree of the behavioural change
required by potential adopters to use these functional innovations. As a result, there is a
sequence of two dichotomous decisions addressing the respective conditional link:
{Insert Figure 1 about here}
Factors influencing rates of adoption
The rate of adoption refers to the frequency of new users of the innovation out of its market
potential (Rogers, 2003). Diffusion research has explicitly considered the communication
process for the diffusion of a technological innovation using mathematical models (e.g. Bass,
1969; Fourt and Woodlock, 1960; Lilien et al., 1992; Mansfield, 1961). Following these
4
classic works, a number of models have been developed to capture other dynamics of the
innovation diffusion process, such as the influence of the marketing mix on new product
diffusion (Mahajan et al., 1990, 2000).
The Bass model and its revised forms have been used in marketing for forecasting
innovation diffusion (i.e. the lifecycle dynamics of a new product) in retail service and
consumer durable goods, among others (Mahajan et al., 1990). However, some of the
assumptions underlying the Bass model have been questioned, such as that market potential
remains constant over time and that adoption is an individual decision (Lilien et al., 1992).
Moreover, to explain consumer acceptance diffusion research has focused on i) the perceived
attributes of an innovation and ii) the potential adopter’s characteristics. The empirical
literature emphasises the importance of gender, age and income differences (e.g. Gan et al.,
2006) but offers little consistency on the importance of other individual characteristics in
explaining adoption of an innovation (e.g. Wang et al., 2008), and this inconsistency may be
based on the variety of research contexts, the nature of the innovation, the sample’s
representation of the target population and the geographic context. The research also
indicates that the perceived attributes of an innovation are stronger predictors of adoption
than the personal characteristics of potential adopters (Gatignon and Robertson, 1989;
Lockett and Littler, 1997; Moore and Benbasat, 1991; Rogers, 2003). Although the perceived
attributes of an innovation can be used to explain adoption rates, researchers have devoted
little effort in examining how these factors affect adoption rates (Rogers, 2003).
Within the context of Internet-based innovations, the literature has examined
empirically the factors that influence consumer attitudes and their effects on intentions
towards adopting these services. Many of these studies are based on the works of Rogers
(2003) (innovation characteristics), Davis (1989) and Davis et al., (1992) (technology
acceptance model - TAM), Parasuraman (2000) (technology readiness index - TRI),
Dabholkar (1996) (service quality) and extensions and combinations of these theories. In
addition, the constructs of perceived risk (Bobbitt and Dabholkar, 2001; Cunningham et al.,
2005; Curran and Meuter, 2005), interactivity (to understand future buyer-seller activities in
the electronic marketplace) (Sawhney et al., 2005; Varadarajan and Yadav, 2002; Yadav and
Varadarajan, 2005a; Yadav and Varadarajan, 2005b), human interaction (Gilbert et al., 2004;
Makarem et al., 2009; Simon and Usunier, 2007), perceived ability/capacity (Bitner et al.,
2002; Ellen et al., 1991; Walker et al., 2002), time and energy savings (Walker and Johnson,
2006), aspects of service convenience (Berry et al., 2002) and consumer demographics and
personal characteristics (Dabholkar and Bagozzi, 2002; Elliott and Hall, 2005; Lee et al.,
2010) have all been found to be important factors in explaining adoption.
The above studies and related theories could be examined in order to identify what
influences the initial adoption of innovations (Mahajan et al., 2000) but equally there is space
to examine post-adoption or usage behaviour (Hernandez et al., 2009; Mahajan et al., 2000),
which is particularly important for service innovations that are used frequently after adoption.
The construct of an e-service quality is an important factor contributing to repeat purchases
from websites (Zeithaml et al., 2002) and the measurement of this construct must rest on
perception data which is based on the use (adopters) or non-use (non-adopters) of the Internet
for shopping. The most frequently found e-service quality perception factors are: reliability,
privacy/security, site design, ease of use, responsiveness, control, accessibility, speed of
delivery, enjoyment and accuracy (Bobbitt and Dalholkar, 2001; Jauda et al., 2002; Jun and
Cai, 2001; Shamdasani et al., 2008; Zeithaml et al., 2002).
There are some overlaps between aspects and constructs with reliability, privacy and
security concerns being aspects of trust (Yousafzai et al., 2009) and with reliability also being
an aspect of perceived risk (Curran and Meuter, 2005). Davis’s constructs of usefulness and
5
ease of use are similar to Rogers’ constructs of perceived relative advantage and complexity,
respectively, while aspects of relative advantage and convenience are related. Rogers’
compatibility construct is broadly defined and captures knowledge aspects, such as personal
ability/capacity and cultural values.
2
There is limited empirical evidence on non-adopter behaviour, which is particularly
important for situations where adopters of the core application, such as the Internet, have not
adopted or have adopted only some of its applications. The next section presents an empirical
development of the adoption model to test for and explain a sequential link between adoption
rates of Internet shopping and adoption rates of Internet banking, and hence to fill this gap in
the literature.
3. Data and method
This study draws on a data set collected via questionnaires distributed to organizations in
Athens, Greece, and has been described in detail in Patsiotis et al. (2012). That study
employed categorical data to capture respondent demographics and 7-point Likert scales to
measure respondent perceptions on characteristics of innovation. Descriptions of the
variables used in this study are presented in Table 1.
{Insert Table 1 about here}
Modelling approach
Our proposition is that there is a clear sequence of decisions relating to the adoption of
computer-based innovation technologies: first the Internet user must make the decision to use
the Internet for shopping (Yes = 1; No = 0) and then make a similar decision to use the
Internet for banking (Yes = 1; No = 0). These two dichotomous choices are traditionally
modelled separately as dependent variables in regression; we continue to do this but we
model them simultaneously initially and then sequentially.
An alternative way to think about these questions is to consider it as a (potential)
sample selection issue: if these are sequential decisions then individuals who are not Internet
shoppers are much less likely to consider the Internet for banking. Hence any efforts to
identify the contributory effect of explanatory variables on the decision to use the Internet for
banking services may produce biased results, as part of the sample are not considering using
the Internet for banking, as they have not first engaged enough in shopping on the Internet.
Therefore, estimates of the effect of explanatory variables on Internet banking should be
conditional on the factors that influence adoption of the Internet for shopping.
An appropriate method to employ in this instance is the bivariate probit regression
approach and conditional marginal effects can be obtained where P(Internet Banker = 1 |
Internet Shopper = 1). Given marginal effect estimates of this conditional probability, it
2
Some of these constructs may not be relevant to usage behaviour. For instance, technology readiness factors
focus on consumer traits and personal orientation, such as innovativeness, but do not necessarily predict
technology usage (Meuter et al., 2003). Human interaction may be an inhibitor to Internet shopping and
Internet banking and is related more to pre-adoption behaviour. As people use a combination of shopping
modes, Internet facilities may only serve as a substitute channel for users and hence it is less likely for
Rogers’ observability to be an important factor (Black et al., 2001; Lassar et al., 2005). Conversely, Rogers’
trialability construct (the lack of trial opportunities to start using Internet banking) may be an important
inhibitor to adoption.
6
would be possible to identify whether respondent demographics and perceived attributes
contribute either to the decision to participate in Internet banking or to the decision to
participate in Internet shopping, or to both decisions.
We adopt the formal model for estimating the probabilities according to Greene
(2003). Let
i
y
1
be a latent variable that denotes the probability that an individual is an
Internet banker, which is dependent on a range of contributory factors,
i
X
1
. Also let
i
X
2
be a
latent variable that denotes the probability that the individual is an Internet shopper, where
this is also dependent upon a range of factors,
i
X
2
. The model is represented as follows:
i i i
X y
1 1 1 1
c | + =
i i i
X y
2 2 2 2
c | + =
where the values for
i
y
1
are observable and related to the following binary dependent
variables, on the basis of the following conditions:
0 , 1
1
> =
i i
y if banker Internet 0 , 0
1
s =
i i
y if banker Internet
and
0 , 1
2
> =
i i
y if shopper Internet 0 , 0
2
s =
i i
y if shopper Internet
where 1 =
i
shopper Internet denotes that the individual is an Internet shopper, and
1 =
i
banker Internet denotes that the individual is an Internet banker. The errors ) , (
2 1 i i
c c are
assumed to have the standard bivariate normal distribution, with ) ( 1 ) (
2 1 i i
V V c c = = and
µ c c = ) , (
2 1 i i
Cov . Thus the individual’s probability of being an Internet banker can be written
as:
) ( banker Internet P ) 1 , 1 ( = = =
i i
shopper Internet banker Internet P
) , (
2 2 1 1 i i i i
x X x X P < < =
i i
x x
i i
dz dz z z
i i
2 1 2 1
2 1
) ; , ( 2
} }
· ÷ · ÷
= µ |
) ; , (
2 2 1 1
µ | |
i i
X X F =
where F denotes the bivariate standard normal distribution function with correlation
coefficient µ .
3
The bivariate probit model has full observability if
i
banker Internet and
i
shopper Internet are both observed in terms of all their four possible combinations [i)
‘Internet banker
i
= 1, Internet shopper
i
= 1’, ii) ‘Internet banker
i
= 1, Internet shopper
i
= 0’,
iii) ‘Internet banker
i
= 0, Internet shopper
i
= 1’, and iv) ‘Internet banker
i
= 0, Internet
shopper
i
= 0’]. Category (ii) is always equal to zero in our sample. If there is a clear sequence
3
Greene (2003) shows that the density function is given by:
2 / 1 2 ) 1 /( ) 2 )( 2 / 1 (
2
) 1 ( 2 /
2
2 1
2
2
2
1
µ t |
µ µ
÷ =
÷ ÷ + ÷
i i i i
x x x x
e .
7
of decisions in adopting Internet shopping and then Internet banking then it is appropriate to
deem this to be a naturally constrained complete set of observations and effectively provides
us with full observability in our data. It is known that full observability leads to the most
efficient estimates (Ashford and Sowden, 1970; Zellner and Lee, 1965).
The following section presents the results from application of the bivariate probit
method to the case explained above, and specifically will discuss the conditional marginal
effects obtained from P(Internet Banker
i
= 1 | Internet Shopper
i
= 1).
4. Results
A sequential structure
Figure 2 shows the structure of our purported sequential structure. It shows that
approximately one-third of Internet users do not use the Internet for shopping. More
importantly for our research, it illustrates that no one uses the Internet for banking if they do
not use the Internet for shopping. In our sample, less than 30 percent of respondent use the
Internet for banking, and this represents more than 40 percent of the respondents who use the
Internet for shopping. Figure 2 corroborates the perspective that the decisions to use the
Internet for shopping and for banking may be sequential as it is in line with the proposition
that using the Internet for shopping is a prerequisite for using the Internet for banking.
{Insert Figure 2 about here}
We draw our data from Patsiotis et al.’s (2012) study which captures the
multidimensional nature of each construct, but inclusion of all dimensions in our regression
approach would produce excessive multicollinearity and potentially incorrect coefficient
estimates due to included variable bias.
4
To circumvent these problems we apply principal
component analysis to each construct as a data reduction technique. We retain each
construct’s principal component and use these as regressors into our bivariate probit
analysis.
5
Table 2 presents descriptions of these principal components.
{Insert Table 2 about here}
Bivariate regression results
Our regression results are presented in Table 3. They indicate there is a very strong
association between the choices to use the Internet for shopping and to use the Internet for
banking; this is illustrated by the highly statistically significant result of the log-likelihood of
rho (chi
2
= 43.2469, p < 0.000).
{Insert Table 3 about here}
Our initial regression is the choice to use the Internet for shopping. The results
corroborate extant literature by illustrating that males and the higher educated are more likely
4
The empirical findings of this study have several limitations. These refer to the chosen geographic context,
the chosen predictors for modelling, the cross-sectional nature of the survey, and the possible non-response
bias considering that sample respondents may not represent the working population.
5
See Churchill and Iacobucci (2002) for a description of the principal components approach.
8
to use the Internet for shopping. Moreover, the results show that greater enjoyment
(significant at the 10 percent level) and greater personal capacity both enhance the likelihood
that an individual will use the Internet for shopping, while greater perceived risks negatively
influence the likelihood that a person will use the Internet for shopping, all given that they are
already Internet users. The positive influences observed, and especially on enjoyment, are in
line with empirical findings that emotions influence positively usage behaviour (Martin et al.,
2008; Shamdasani et al., 2008; Watson and Spence, 2007; Wood and Moreau, 2006).
Different factors are important in enhancing the likelihood that an Internet user will
use the Internet for banking. Greater enjoyment and greater ease of use both enhance the
likelihood that an Internet user will also use the Internet for banking; conversely, lower
perceptions of usefulness and a lack of trial will both diminish this likelihood. Although the
core application is the same, usage of subsequent innovations may be influenced by different
factors and therefore proceed to estimate a sequential model.
Conditional marginal effects
Set within a bivariate probit regression approach, it is possible to examine our proposition
that the equations should be estimated in sequence. Accordingly, Table 3 also presents two
sets of marginal effects estimates corresponding to those which are not based on a sequential
approach (i.e. unconditional) and those that are based on a sequential approach (i.e.
conditional). Several points are worthy of emphasis here. First, there is a slight disagreement
on which variables enhance the likelihood that an Internet user will use Internet banking as
the significant coefficients corresponding to education and time-energy are not significant at
tradition level of acceptance within the unconditional set up. These findings suggest that
time-energy and higher education affect the likelihood of using the Internet for banking and
not only the decision to use the Internet for shopping, as would be inferred under the
unconditional approach. Perhaps these results are reflecting the possibility that having higher
education increases the speed of learning new innovations and therefore increases the
likelihood that an Internet user will more quickly adopt Internet banking facilities. The
greater coefficient estimates for the time-energy variable when the structure is conditional
also emphasises the increasing importance of speed when using the Internet for banking
transactions.
Second, when the variables have been identified as being highly statistically
significant, the coefficients estimates corresponding to the condition approach are typically
much larger than for the unconditional approach. This suggests the standard unconditional
approach leads to coefficient estimates that are biased towards zero (i.e. having no effect),
and therefore underestimates their impact on the likelihood of using the Internet for banking.
The clearest example of this bias corresponds to having post-graduate education, where the
marginal effect is more than four-times greater under the conditional approach relative to the
unconditional approach. A lack of trail has a much greater hindering effect on using the
Internet for banking under the condition approach, being almost 50 percent greater than
expected under the unconditional approach.
Third, the empirical observation that there is a negative influence of usefulness on
Internet banking use offers new evidence on Internet banking usage behaviour; indeed our
empirical estimates suggest that usefulness is the most important lever for policy formation,
and emphasising the usefulness of Internet banking could lead to the greatest amount of new
Internet bankers. Moreover, the negative effects of usefulness on Internet banking adoption
rates contrasts with existing research on the influence of usefulness on customer intentions
and usage toward technology-based service offerings (TAM model and extensions: Gilbert et
9
al., 2004; Hernandez et al., 2009; McKechnie et al., 2006; Ozdemir and Trott, 2009). In a
study comparing Internet banking acceptance to older self-service technologies for banking,
such as the ATM and phone, Curran and Meuter (2005) found that ease of use and usefulness
are not important predictors of Internet banking, although there are important for the adoption
of ATM. They concluded that for an innovation to be successful, i.e. reach the majority of
potential customers, it must be both useful and easy to use. Thus, the low adoption rates of
Internet banking may be associated with the negative influence of usefulness. Considering
that usefulness is similar to relative advantage and that relative advantage has been found to
be one of the strongest predictors of an innovation’s adoption rate (Rogers, 2003), the
negative effect of usefulness on Internet banking propensity may further explain low adoption
rates. Another explanation for the negative influence of usefulness on Internet banking is that
Internet shoppers may perceive Internet banking to be less useful when compared to
alternative channels, such as the ATM, and therefore not interesting for further consideration.
Fourth, a lack of trial of Internet banking is found to be strongly associated with non-
usage behaviour, which could be difficult to circumvent as trialability of Internet banking
cannot readily be experienced before adoption. In our study, although Internet banking users
are positively influenced by ease of use, enjoyment, and to some extent value time-energy
savings, usefulness and lack of trial explain non-use. Thus, Internet shoppers may not be fully
aware of the usefulness aspects of Internet banking use. This is supported by the negative
influence of lack of trial, i.e. less or no opportunities to understand how it works do not add
knowledge on the usefulness aspects.
It may also be the case that customers use a combination of banking methods other
than the online. The literature supports a multi-channel integration (Coelho and Easingwood,
2003; Zeithaml et al., 2009). Earlier results indicate that consumers have a preference for a
mix of delivery channels rather than exclusive reliance upon anyone single channel
(Howcroft et al., 2002), and new delivery channels tend to complement rather than replace
the existing ones (Hughes, 2006). Finally, it is interesting in the results of this study that
interactivity, and any perceived risks, security and privacy concerns do not influence
adoption rates. This also contrasts empirical research on the preceding influences. It may be
the focus of extant work on pre-adoption behaviour, compared to the method adopted in this
study examining usage behaviour, that explains their non-importance.
5. Conclusion and implications
This study examined empirically the conditional link between Internet shopping and Internet
banking in order to identify key factors influencing Internet banking adoption rate. The focus
was on the behaviour of Internet users that adopt and use Internet shopping first in order to
decide whether to use the Internet for their banking needs.
The application of bivariate probit regression analyses revealed that for those with
Internet shopping experience, the probability of using Internet banking services is positively
affected by customer perceptions on enjoyment, ease of use, and time-energy savings, and
negatively influenced by usefulness and lack of trial. The results of this study support the
presence of a conditional link, and suggest Internet shopping and Internet banking should be
examined sequentially. Empirical research usually examines diffusion theories within the
context of one technological innovation. The results of this study strongly support the
proposition that diffusion research should examine new technological innovations
sequentially when they are based on a core application and the adoption of the first would
probably lead or not to the adoption and use of the second.
10
The above findings suggest further research inquiry and practical implications for
service providers. Future research could further apply the sequential modelling approach in
similar service contexts (e.g., smart phone, personal assistance shopping) and in other
geographic areas exhibiting different adoption rates. This is particularly applicable when
technological innovations are introduced, or have been introduced, sequentially in the market,
and there is a lack of an accurate sample frame. Usage behaviour could be examined further
to identify the most relevant predictors and confirm the key influences revealed by this study.
As the chosen innovations are in the market for some time now, different factors might be
more important influences to their adoption rates. This would facilitate a better understanding
of non-usage behaviour too. In addition, longitudinal studies based on panel data could
provide better indication of the future direction of the conditional link and the most important
influencing factors.
Finally, the evidence provided in this research study on the conditional link between
Internet shopping and Internet banking suggests that even though the basis of a technological
interface is the computer, new services developed through this can have different perceived
characteristics. Thus, banks should be aware that Internet banking use presents unique
characteristics compared with the other alternative channels as well as with related services.
Bank customers that use the Internet for shopping but not for banking should be informed for
the usefulness aspects of online banking, as well as be given the opportunity to experience
how it works, given the perceived lack of trial. In particular, the usefulness aspects should be
examined further through a comparison of the alternative methods for banking. It may be that
some customers find Internet banking to be less useful when compared to other method(s), or
they may not be fully aware of its benefits. Existing users of Internet banking facilities value
the enjoyment aspects, so service providers should make the experience enjoyable so as to
sustain positive emotions. As a result, good knowledge of the behaviour of users would
facilitate their effort to understand those likely to be new users and incorporate this
understanding in their marketing strategy to develop a more targeted communications policy
that would generate useful customer feedback.
References
Bass, F.M. (1969). A new product growth model for consumer durables. Marketing Science, 15, 215-227
Berry, L.L., Seiders, K. and Grewal, D. (2002). Understanding service convenience. Journal of Marketing, 66
(3), 1-17.
Bitner,M.J., Ostrom, A.L. and Meuter, M.L. (2002). Implementing successful self-service technologies.
Academy of Management Executive, 16 (4), 96-108.
Black, N., Lockett, A., Winklhofer, H. and Ennew, C. (2001). The adoption of Internet financial services: a
qualitative study. International Journal of Retail and Distribution Management 29(8), 390-398
Bobbitt, M. and Dabholkar, P.A. (2001). Integrating attitudinal theories to understand and predict use of
technology-based self-service: the internet as an illustration. International Journal of Service Industry
Management, 12 (5), 423-450.
Churchill, G.A. and Iacobucci, D. (2002). Marketing Research: Methodological Foundations, 8th ed. Orlando,
FL: Harcourt College Publishers.
Coelho, F. and Easingwood, C. (2003). Multiple channel structures in financial services: A framework. Journal
of Financial Services Marketing, 8 (1), 22-34.
Cunningham, L., Gerlach, J., Harper, M. and Young, C. (2005). Perceived risk and the consumer buying
process: Internet airline reservations. International Journal of Service Industry Management, 16 (4), 357-372
Curran, J.M. and Meuter, M.L. (2005). Self-service technology adoption: comparing three technologies. Journal
of Services Marketing, 19 (2), 103-113.
Dabholkar, P.A. (1996). Consumer evaluations of new technology-based self-service options: An investigation
of alternative models of service quality. International Journal of Research in Marketing, 13 (1), 29-51.
11
Dabholkar, P. and Bagozzi, R. (2002). An attitudinal model of technology-based self-service: Moderating
effects of consumer traits and situational factors. Journal of the Academy of Marketing Science 30(3), 184-
201
Davis, F.D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology.
MIS Quarterly, 13 (3), 319-340.
Davis, F.D., Bagozzi, R.P. and Warshaw, P.R. (1992). Extrinsic and intrinsic motivation to use computers in the
workplace. Journal of Applied Social Psychology, 22 (14), 1109-1130.
Durkin, M. (2007). Understanding registration influences for electronic banking. International Review of Retail,
Distribution and Consumer Research. 17(3), 219-231
Ellen, P., Bearden, W. and Sharma, S. (1991). Resistance to technological innovations: an examination of the
role of self-efficacy and performance satisfaction. Journal of the Academy of Marketing Science 19(4), 297-
307
Elliott, K.M. and Hall, M.C. (2005). Assessing consumers’ propensity to embrace self-service technologies: Are
there gender differences? The Marketing Management Journal, 15 (2), 98-107.
Fitzsimmons, J.A. and Fitzsimmons, M.J. (2011). Service Management: Operations, Strategy, Information
Technology. 7
th
ed. New York, NY: McGraw-Hill/Irwin.
Fourt, L.A. and Woodlock, J.W. (1960). Early prediction of market success for grocery products. Journal of
Marketing. 25 (October), 31-38.
Gan, C., Clemes, M., Limsombunchai, V. and Weng, A. (2006). A logit analysis of electronic banking in New
Zealand. International Journal of Bank Marketing. 24 (6), 360-383.
Gatignon, H. and Robertson, T.S. (1989). Technology diffusion: an empirical test of competitive effects.
Journal of Marketing, 53 (1), 35-49.
Gilbert, D., Balestrini, P. and Littelboy, D. (2004). Barriers and benefits in the adoption of e-government.
International Journal of Public Sector Management 17(4) 286-301
Greene, W.H. (2008). Econometric Analysis, Upper Saddle River, NJ: Prentice-Hall.
Hernandez, B., Jimenez, J. and Martin, M.J. (2009). Adoption vs acceptance of e-commerce: two different
decisions. European Journal of Marketing, 43 (9/10), 1232-1245.
Hernandez, J.M.C. and Mazzon, J.A. (2007). Adoption of internet banking: proposition and implementation of
an integrated methodology approach. International Journal of Bank Marketing, 25(2), 72-88.
Howcroft, B., Hamilton, R. and Hewer, P. (2002). Consumer attitude and the usage and adoption of home-based
banking in the UK. International Journal of Bank Marketing. 20 (3), 111-121.
Hughes, T. (2006). New channels/old channels – Customer management and multi-channels. European Journal
of Marketing, 40 (1/2), 113-129.
Jauda, S., Trocchia, P.J. and Gwinner, K.P. (2002). Consumer perceptions of Internet retail service quality.
International Journal of Service Industry Management 13 (5), 412-431.
Jun, M. and Cai, S. (2001). The key determinants of Internet banking service quality: A content analysis.
International Journal of Bank Marketing, 19 (7), 276-291.
Lassar, W.M., Manolis, C. and Lassar, S. (2005). The relationship between consumer innovativeness, personal
characteristics, and online banking adoption. International Journal of Bank Marketing, 23 (2), 176-199.
Lee, H-J, Cho, H.J., Xu, W. and Fairhurst, A. (2010). The influence of consumer traits and demographics on
intention to use retail self-service checkouts. Marketing Intelligence and Planning, 28 (1), 46-58.
Lilien, G.L., Kotler, P. and Moorthy, K.S. (1992). Marketing Models. Englewood Cliffs, NJ: Prentice-Hall.
Littler, D. (2001). The idiosyncrasies of new product adoption for technologically innovative offerings. The
Marketing Review. 2, 3-19.
Lockett, A. and Littler, D. (1997). The adoption of direct banking services. Journal of Marketing Management,
13 (8), 791-811.
Mahajan, V., Muller, E. and Bass, F.M. (1990). New product diffusion models in marketing: A review and
directions for research. Journal of Marketing. 54 (January), 1-26.
Mahajan, V., Muller, E. and Wind, Y. (2000). New Product Diffusion Models, Boston, MA: Kluwer Academic.
Makarem, S.C., Mudambi, S.M. and Podoshen, J.S. (2009). Satisfaction in technology-enabled service
encounters. Journal of Services Marketing, 23 (3), 134-144.
Mansfield, E. (1961). Technical change and rate of imitation. Econometrics, 29 (October), 741-766.
Martin, D., O’Neill, M., Hubbard, S. and Palmer, A. (2008). The role of emotion in explaining consumer
satisfaction and future behavioural intention. Journal of Services Marketing, 22 (3), 224-236.
McKechnie, S., Winklhofer, H. and Ennew, C. (2006). Applying the Technology Acceptance Model to online
retailing of financial services, International Journal of Retail & Distribution Management 34(4/5), 388-410
Meuter, M.L., Ostrom, A.L., Bitner, M.J. and Roundtree, R. (2003). The influence of technology anxiety on
consumer use and experiences with self-service technologies. Journal of Business Research, 56, 899-906.
Moore, G.C. (1991). Crossing the Chasm, New York, NY: Harper Business.
12
Moore, G.C. and Benbasat, I. (1991). Development of an instrument to measure the perceptions of adopting an
information technology innovation. Information Systems Research, 2 (3), 192-222.
Ozdemir, S. and Trott, P. (2009). Exploring the adoption of a service innovation: A study of Internet banking
adopters and non-adopters. Journal of Financial Services Marketing, 13 (4), 284-299.
Parasuraman, A. (2000). Technology readiness index (TRI): a multiple-item scale to measure readiness to
embrace new technology. Journal of Service Research, 2 (4), 307-320.
Patsiotis, A.G., Hughes, T. and Webber, D.J. (2012). Adopters and non-adopters of internet banking: a
segmentation study. International Journal of Bank Marketing, 30 (1), 20-42.
Robertson, T. (1971). Innovative Behaviour and Communication, New York, NY: Holt, Reinhart and Winston.
Rogers, E.M. (2003). Diffusion of Innovations, 5th ed. New York, NY: Free Press.
Sawhney, M., Verona, G. and Prandelli, E. (2005). Collaborating to create: the Internet as a platform for
customer engagement in product innovation. Journal of Interactive Marketing, 19 (4), 4-17.
Shamdasani, P., Mukherjee, A. and Malhotra, N. (2008). Antecedents and consequences of service quality in
consumer evaluation of self-service Internet technologies. The Service Industries Journal, 28 (1), 117-138.
Shi, W., Shambare, N. and Wang, J. (2008). The adoption of internet banking: An institutional theory
perspective. Journal of Financial Services Marketing. 12 (4), 272-286.
Simon, F. and Usunier, J. (2007). Cognitive, demographic, and situational determinants of service customer
preference for personnel-in-contact over self-service technology. International Journal of Research in
Marketing, 24, 163-173.
Tapp, A. and Hughes, T. (2004). New technology and the changing role of marketing. Marketing Intelligence
and Planning. 22 (3), 284-296.
Varadarajan, P.R. and Yadav, M.S. (2002). Marketing strategy and the Internet: An organizing framework.
Journal of the Academy of Marketing Science 30(4) 296-312.
Veryzer, R.W., Jr. (1998a). Key factors affecting customer evaluation of discontinuous new products. Journal of
Product Innovation Management, 15, 136-150.
Veryzer, R.W., Jr. (1998b). Discontinuous innovation and the new product development process. Journal of
Product Innovation Management, 15, 304-321.
Walker, R.H. and Johnson, L.W. (2006). Why consumers use and do not use technology-enabled services.
Journal of Services Marketing, 20 (2), 125-135.
Walker, R.H., Craig-Lees, M., Hecker, R. and Francis, H. (2002). Technology-enabled service delivery: An
investigation of reasons affecting customer adoption and rejection. International Journal of Service Industry
Management, 13 (1), 91-106.
Wang, G., Dou, W. and Zhou, N. (2008). Consumption attitudes and adoption of new consumer products: A
contingency approach. European Journal of Marketing. 42 (1/2), 238-254.
Watson, L. and Spence, M.T. (2007). Causes and consequences of emotions on consumer behaviour – A review
and integrative cognitive appraisal theory. European Journal of Marketing, 41 (5/6), 487-511.
Wood, S.L. and Moreau, C.P. (2006). From fear to loathing? How emotion influences the evaluation and early
use of innovations. Journal of Marketing, 70 (3), 44-57.
Yadav, M.S. and Varadarajan, R.R. (2005a). Understanding product migration to the electronic marketplace: A
conceptual framework. Journal of Retailing, 81 (2), 125-140.
Yadav, M.S. and Varadarajan, R.R. (2005b). Interactivity in the electronic marketplace: an exposition of the
concept and implications for research. Journal of the Academy of Marketing Science, 33 (4), 585-603.
Yousafzai, S., Pallister, J. and Foxall, G. (2009). Multi-dimensional role of trust in Internet banking adoption.
The Service Industries Journal, 29 (5), 591-605.
Zeithaml, V.A., Bitner, M.J. and Gremler, D.D. (2009). Services Marketing, 5th ed. New York, NY: McGraw-
Hill/Irwin.
Zeithaml, V.A., Parasuraman, A. and Malhotra, A. (2002). Service quality delivery through websites: A critical
review of extant knowledge. Journal of the Academy of Marketing Science, 30 (4), 362-375.
13
Figure 1: Conceptual framework
Internet Use
No Internet
Shopping Use
Internet Shopping
Use
No Internet
Banking Use
Internet Banking
Use
Predictors
- Perceived
Innovation
Attributes
- Demographic
Characteristics
14
1
I use the Internet for shopping I use the Internet for Banking
Yes = 42.13 %
Yes = 0.00 %
Yes = 65.93 %
No = 57.87 %
No = 100.00 %
No = 34.07 %
0.00 %
38.15 %
34.07 %
27.78 %
Final
probabilities
Figure 2: Tree diagram
15
Table 1. Description of non-principal component variables
Variables Description
Dependents:
- Internet shopping use:
- Internet banking use:
Binary = 1 if someone uses; 0 non-use
Binary = 1 if someone uses; 0 non-use
Independent
Individual characteristics:
Gender (binary)
Age (years)
Education
1 if male (47%); 0 if female (53%)
18-25 (26.3%); 26-35 (41.5%); 36-45 (20%); 46-55
(7.8%); 56-65 (4.4%)
School (3.7%); College (20%); University (43%);
Postgraduate (33.3%)
1
Table 2: Results of principal components analysis
Enjoyment Time and energy Ease of use Usefulness Interactivity Personal capacity Perceived risks
Privacy
concerns
Security
concerns
Lack of trial
Loadings 7a, b, c 8g1, 2, 3, 4 8h, i, j 8g5, e, f 9 (all 33) 10b, c, d, e 11a, b, c 11d1, 2, 3 11d4, 5, 6 11e, f, g, h, i, j
Eigenvalue 2.263 2.594 1.483 1.753 13.343 2.474 2.265 2.421 2.348 4.022
Proportion (%) 0.754 0.649 0.494 0.584 0.404 0.619 0.755 0.807 0.783 0.670
1
Table 3: Estimates of probit regressions
Unrelated bivariate probit Unconditional marginal effects Conditional marginal effects
Coef. Std. error p dy/dx Std. error p dy/dx Std. error p
Internet shopper
Male
0.627
0.193
0.001
Age: 18-25 0.110 0.228 0.630
Age: 26-35 Control variable
Age: 36-45 0.1600 0.269 0.552
Age: 46-55 -0.057 0.401 0.887
Age: 56-65 0.106 0.452 0.815
Education School 0.4500 0.460 0.328
Education College 0.189 0.234 0.420
Education Degree Control variable
Education Post-grad 0.969 0.262 0.000
Enjoyment 0.128 0.074 0.082
Time and energy -0.035 0.079 0.659
Ease of use 0.124 0.096 0.199
Usefulness -0.134 0.095 0.160
Interactivity 0.035 0.032 0.273
Personal capacity 0.272 0.075 0.000
Perceive risks -0.309 0.100 0.002
Privacy 0.190 0.116 0.102
Security -0.137 0.102 0.177
Lack of trial 0.040 0.065 0.538
constant -0.165 0.203 0.417
Internet banker
Male 0.589 0.226 0.009 0.168 0.065 0.009 0.148 0.083 0.075
Age: 18-25 -0.156 0.291 0.591 -0.043 0.077 0.577 -0.077 0.109 0.480
Age: 26-35 Control variable Control variable Control variable
Age: 36-45 0.776 0.292 0.008 0.256 0.106 0.016 0.285 0.099 0.004
Age: 46-55 0.144 0.475 0.761 0.043 0.148 0.772 0.065 0.187 0.729
Age: 56-65 1.002 0.494 0.042 0.360 0.192 0.060 0.382 0.170 0.025
Education School -0.653 0.678 0.335 -0.338 0.098 0.160 -0.319 0.263 0.224
Education College -0.135 0.295 0.647 -0.037 0.077 0.634 -0.079 0.108 0.464
Education Degree Control variable Control variable Control variable
Education Post-grad -0.164 0.252 0.514 -0.045 0.068 0.504 -0.196 0.090 0.029
Enjoyment 0.291 0.093 0.002 0.082 0.026 0.002 0.098 0.036 0.006
Time Energy 0.137 0.093 0.140 0.039 0.026 0.138 0.059 0.033 0.078
Ease Of Use 0.225 0.110 0.041 0.064 0.031 0.038 0.072 0.039 0.066
Usefulness -0.316 0.101 0.002 -0.089 0.029 0.002 -0.107 0.036 0.003
Interactivity 0.004 0.037 0.909 -0.001 0.010 0.909 -0.003 0.013 0.817
Personal Capacity 0.127 0.095 0.180 0.036 0.027 0.177 0.013 0.035 0.699
Perceived Risks -0.109 0.103 0.292 -0.031 0.029 0.291 -0.001 0.040 0.975
Privacy 0.051 0.110 0.640 0.014 0.031 0.640 -0.005 0.042 0.898
Security -0.025 0.099 0.798 -0.007 0.028 0.798 0.009 0.037 0.816
Lack Of Trial -0.231 0.073 0.002 -0.065 0.022 0.003 -0.097 0.030 0.001
constant -1.154 0.234 0.000
Rho
Log Likelihood
1
-191.55
5.21e-09
Notes: N=259; bold font highlights statistical significance at the 10% confidence level. Likelihood-ratio test of
rho=0, chi
2
(1) = 43.2469, Prob > chi
2
= 0.0000.
doc_118858653.pdf
This paper presents an empirical investigation into the factors that shape the propensity to use the Internet for shopping and banking through application of bivariate probit regression techniques to data sourced from a survey of 259 respondents in Athens, Greece.
1
Internet shopping and Internet banking in sequence
Athanasios G. Patsiotis
a
, Tim Hughes
b
and Don J. Webber
c
a
Department of Marketing, Deree College, American College of Greece, Athens, Greece
b
Department of Business and Management, University of the West of England, Bristol, UK
c
Department of Accounting, Economics and Finance, University of the West of England, Bristol, UK
Abstract
This paper presents an empirical investigation into the factors that shape the
propensity to use the Internet for shopping and banking through application of
bivariate probit regression techniques to data sourced from a survey of 259
respondents in Athens, Greece. Based on the observation that Internet banking
usage typically requires familiarity with Internet shopping, we estimate the
marginal effects of the determinants of Internet banking use conditioned on the
determinants of Internet shopping use.
Our results suggest that not controlling for this conditioning will bias
estimates and could lead to incorrect policy-recommendations. For instance,
personal capacity is found to be an important determinant of the propensity to use
Internet banking in a non-sequential approach but it is found to have no
significant effect after conditioning. In particular, our results suggest that policy-
makers should emphasise usefulness attributes of computer-based innovations
when attempting to increase the use of the Internet for banking by people who
already use the Internet for shopping.
Keywords: Internet banking; Internet shopping; adoption rate; bivariate probit regression;
conditional marginal effects
Address for correspondence: Dr Athanasios Patsiotis, Department of Marketing, Deree
College, American College of Greece, Athens, Greece. Email: [email protected]
2
1. Introduction
The Internet represents a huge source of information that can be organized and retrieved in
many different ways based on individual users needs (Mahajan et al., 2000). It facilitates
communication and shopping through computer-mediated environments and it is a market
where a large variety of new technologies and interdependent products are introduced
(Mahajan et al., 2000; Varadarajan and Yadav, 2002). It can also be a discontinuous
innovation process and lead to new product developments. Following the introduction and
acceptance of Internet shopping, new technological interfaces developed by banks, such as
Internet banking, are innovative delivery and communication channels where new products
and services are introduced. These innovations have facilitated interaction and the building of
relationships between banks and their customers (Tapp and Hughes, 2004).
New technologies and especially the developments of self-service technologies
present several challenges for banks in terms of their customer relationships. Banks that offer
Internet banking services can benefit from lower costs due to the utilization of less human
and physical resources and the potential of economies of scale in bank operations (Shi et al.,
2008). Consumers’ transferring their decision making processes from traditional offline to
online can engender cost and time savings benefits (Shi et al., 2008) at the expense of various
risks (Durkin, 2007). Consequently banks need to alter their operations and internal and
external communication media, and such major changes can encounter resistance.
1
A number of articles present investigations of the determinants of Internet use for a
variety of services. Such studies typically examine either one Internet service in isolation or
assume away structure or order between Internet services. This paper purports that there is a
sequence of Internet usage choices, with consumers first becoming familiar with the Internet
for their shopping experience and, once proficiency in this area has been achieved, consumers
will then consider using the Internet for banking services. Based on the idea of a conditional
and sequential link between Internet shopping and Internet banking we proceed to examine
empirically the factors that influence the rate of Internet banking adoption. Our results
strongly support the assumption of association between Internet banking and Internet
shopping, and once sequencing has been integrated into the modelling approach we identify
potential conflicting results and important policy levers.
The paper is structured as follows. The next section presents a review of extant
literature and our conceptual model. Section 3 details the data and the modelling approach.
Section 4 is a discussion of findings and implications, and Section 5 concludes.
2. Theoretical background
Innovations can be defined in terms of the amount of behavioural change necessary to use the
innovation effectively; they can be classified along a continuum from the least to the most
disruptive which is related to the extent to which the innovation is functionally new
(Robertson, 1971). Technological innovations can be discontinuous (Moore, 1991), are
typically viewed as being rooted in new information and computer-based technologies, and
can disrupt existing patterns of behaviour (Fitzsimmons and Fitzsimmons, 2011; Littler,
1
For example, customers do not always welcome technology or they may increasingly use a combination of
banking methods. The theoretical literature on diffusion research of such technological innovations is well-
developed but it lacks empirical evidence on non-adopter behaviour and focuses mainly on mental behaviour
(Hernandez et al., 2009). There is limited evidence on the possible differences between pre-adoption and
usage behaviour (Hernandez and Mazzon, 2007).
3
2001; Veryzer, 1998b). Their usage can involve a very high degree of technological
uncertainty, a sequence of innovations based on the core application, a longer development
process, and a greater distance from the end user in terms of customer familiarity with the
innovation and the time it takes to evolve (Veryzer, 1998a, 1998b). Thus, technological
innovations can require a change in the behaviour of potential adopters and the development
of new skills.
The introduction of the core application of the Internet (i.e. linking computers in
networks) has spawned a sequence of Internet based innovations that have facilitated
communication and shopping. Such innovations include Internet shopping and Internet
banking that required major behavioural changes by potential adopters in their business and
personal relationships. Internet shopping is now widespread and typically includes books,
cosmetics, consumer durables and service operations, such as retailing, entertainment and
travel. Internet banking is also available and requires a more sophisticated application of
Internet technologies to satisfy consumers’ banking needs, and it differs from other self-
service innovations in financial services since it requires major changes in behaviour (i.e. a
completely new way of consumer banking).
An order of succession
There may be a logical dependence of engagement with innovations, and usage of Internet
shopping and Internet banking may be a prime example. This paper purports a sequence of
events whereby individuals make a series of decisions:
1) Decide (consciously or otherwise) to use the Internet;
2) After some familiarity of use with the Internet has been achieved, a next step in using
the Internet is for shopping;
3) After some familiarity of use with the Internet for shopping has been achieved, a next
step in using the Internet is for banking.
Decision (1) is beyond the scope of this paper. This paper examines empirically the factors
that contribute to decisions (2) and (3) for a sample of Internet users.
The decisions above are each associated with different, albeit potentially sequential
innovations that require interaction with technological interfaces and necessitate a degree of
behavioural change by potential adopters. The conceptual framework for this study, shown in
Figure 1, illustrates that someone must adopt the Internet first, and then must adopt Internet
shopping as a prerequisite to deciding whether to adopt Internet banking. The ordering
purported in Figure 1 corresponds to both the sequence of introduction of the above
technological interfaces (shopping and banking) and the degree of the behavioural change
required by potential adopters to use these functional innovations. As a result, there is a
sequence of two dichotomous decisions addressing the respective conditional link:
{Insert Figure 1 about here}
Factors influencing rates of adoption
The rate of adoption refers to the frequency of new users of the innovation out of its market
potential (Rogers, 2003). Diffusion research has explicitly considered the communication
process for the diffusion of a technological innovation using mathematical models (e.g. Bass,
1969; Fourt and Woodlock, 1960; Lilien et al., 1992; Mansfield, 1961). Following these
4
classic works, a number of models have been developed to capture other dynamics of the
innovation diffusion process, such as the influence of the marketing mix on new product
diffusion (Mahajan et al., 1990, 2000).
The Bass model and its revised forms have been used in marketing for forecasting
innovation diffusion (i.e. the lifecycle dynamics of a new product) in retail service and
consumer durable goods, among others (Mahajan et al., 1990). However, some of the
assumptions underlying the Bass model have been questioned, such as that market potential
remains constant over time and that adoption is an individual decision (Lilien et al., 1992).
Moreover, to explain consumer acceptance diffusion research has focused on i) the perceived
attributes of an innovation and ii) the potential adopter’s characteristics. The empirical
literature emphasises the importance of gender, age and income differences (e.g. Gan et al.,
2006) but offers little consistency on the importance of other individual characteristics in
explaining adoption of an innovation (e.g. Wang et al., 2008), and this inconsistency may be
based on the variety of research contexts, the nature of the innovation, the sample’s
representation of the target population and the geographic context. The research also
indicates that the perceived attributes of an innovation are stronger predictors of adoption
than the personal characteristics of potential adopters (Gatignon and Robertson, 1989;
Lockett and Littler, 1997; Moore and Benbasat, 1991; Rogers, 2003). Although the perceived
attributes of an innovation can be used to explain adoption rates, researchers have devoted
little effort in examining how these factors affect adoption rates (Rogers, 2003).
Within the context of Internet-based innovations, the literature has examined
empirically the factors that influence consumer attitudes and their effects on intentions
towards adopting these services. Many of these studies are based on the works of Rogers
(2003) (innovation characteristics), Davis (1989) and Davis et al., (1992) (technology
acceptance model - TAM), Parasuraman (2000) (technology readiness index - TRI),
Dabholkar (1996) (service quality) and extensions and combinations of these theories. In
addition, the constructs of perceived risk (Bobbitt and Dabholkar, 2001; Cunningham et al.,
2005; Curran and Meuter, 2005), interactivity (to understand future buyer-seller activities in
the electronic marketplace) (Sawhney et al., 2005; Varadarajan and Yadav, 2002; Yadav and
Varadarajan, 2005a; Yadav and Varadarajan, 2005b), human interaction (Gilbert et al., 2004;
Makarem et al., 2009; Simon and Usunier, 2007), perceived ability/capacity (Bitner et al.,
2002; Ellen et al., 1991; Walker et al., 2002), time and energy savings (Walker and Johnson,
2006), aspects of service convenience (Berry et al., 2002) and consumer demographics and
personal characteristics (Dabholkar and Bagozzi, 2002; Elliott and Hall, 2005; Lee et al.,
2010) have all been found to be important factors in explaining adoption.
The above studies and related theories could be examined in order to identify what
influences the initial adoption of innovations (Mahajan et al., 2000) but equally there is space
to examine post-adoption or usage behaviour (Hernandez et al., 2009; Mahajan et al., 2000),
which is particularly important for service innovations that are used frequently after adoption.
The construct of an e-service quality is an important factor contributing to repeat purchases
from websites (Zeithaml et al., 2002) and the measurement of this construct must rest on
perception data which is based on the use (adopters) or non-use (non-adopters) of the Internet
for shopping. The most frequently found e-service quality perception factors are: reliability,
privacy/security, site design, ease of use, responsiveness, control, accessibility, speed of
delivery, enjoyment and accuracy (Bobbitt and Dalholkar, 2001; Jauda et al., 2002; Jun and
Cai, 2001; Shamdasani et al., 2008; Zeithaml et al., 2002).
There are some overlaps between aspects and constructs with reliability, privacy and
security concerns being aspects of trust (Yousafzai et al., 2009) and with reliability also being
an aspect of perceived risk (Curran and Meuter, 2005). Davis’s constructs of usefulness and
5
ease of use are similar to Rogers’ constructs of perceived relative advantage and complexity,
respectively, while aspects of relative advantage and convenience are related. Rogers’
compatibility construct is broadly defined and captures knowledge aspects, such as personal
ability/capacity and cultural values.
2
There is limited empirical evidence on non-adopter behaviour, which is particularly
important for situations where adopters of the core application, such as the Internet, have not
adopted or have adopted only some of its applications. The next section presents an empirical
development of the adoption model to test for and explain a sequential link between adoption
rates of Internet shopping and adoption rates of Internet banking, and hence to fill this gap in
the literature.
3. Data and method
This study draws on a data set collected via questionnaires distributed to organizations in
Athens, Greece, and has been described in detail in Patsiotis et al. (2012). That study
employed categorical data to capture respondent demographics and 7-point Likert scales to
measure respondent perceptions on characteristics of innovation. Descriptions of the
variables used in this study are presented in Table 1.
{Insert Table 1 about here}
Modelling approach
Our proposition is that there is a clear sequence of decisions relating to the adoption of
computer-based innovation technologies: first the Internet user must make the decision to use
the Internet for shopping (Yes = 1; No = 0) and then make a similar decision to use the
Internet for banking (Yes = 1; No = 0). These two dichotomous choices are traditionally
modelled separately as dependent variables in regression; we continue to do this but we
model them simultaneously initially and then sequentially.
An alternative way to think about these questions is to consider it as a (potential)
sample selection issue: if these are sequential decisions then individuals who are not Internet
shoppers are much less likely to consider the Internet for banking. Hence any efforts to
identify the contributory effect of explanatory variables on the decision to use the Internet for
banking services may produce biased results, as part of the sample are not considering using
the Internet for banking, as they have not first engaged enough in shopping on the Internet.
Therefore, estimates of the effect of explanatory variables on Internet banking should be
conditional on the factors that influence adoption of the Internet for shopping.
An appropriate method to employ in this instance is the bivariate probit regression
approach and conditional marginal effects can be obtained where P(Internet Banker = 1 |
Internet Shopper = 1). Given marginal effect estimates of this conditional probability, it
2
Some of these constructs may not be relevant to usage behaviour. For instance, technology readiness factors
focus on consumer traits and personal orientation, such as innovativeness, but do not necessarily predict
technology usage (Meuter et al., 2003). Human interaction may be an inhibitor to Internet shopping and
Internet banking and is related more to pre-adoption behaviour. As people use a combination of shopping
modes, Internet facilities may only serve as a substitute channel for users and hence it is less likely for
Rogers’ observability to be an important factor (Black et al., 2001; Lassar et al., 2005). Conversely, Rogers’
trialability construct (the lack of trial opportunities to start using Internet banking) may be an important
inhibitor to adoption.
6
would be possible to identify whether respondent demographics and perceived attributes
contribute either to the decision to participate in Internet banking or to the decision to
participate in Internet shopping, or to both decisions.
We adopt the formal model for estimating the probabilities according to Greene
(2003). Let
i
y
1
be a latent variable that denotes the probability that an individual is an
Internet banker, which is dependent on a range of contributory factors,
i
X
1
. Also let
i
X
2
be a
latent variable that denotes the probability that the individual is an Internet shopper, where
this is also dependent upon a range of factors,
i
X
2
. The model is represented as follows:
i i i
X y
1 1 1 1
c | + =
i i i
X y
2 2 2 2
c | + =
where the values for
i
y
1
are observable and related to the following binary dependent
variables, on the basis of the following conditions:
0 , 1
1
> =
i i
y if banker Internet 0 , 0
1
s =
i i
y if banker Internet
and
0 , 1
2
> =
i i
y if shopper Internet 0 , 0
2
s =
i i
y if shopper Internet
where 1 =
i
shopper Internet denotes that the individual is an Internet shopper, and
1 =
i
banker Internet denotes that the individual is an Internet banker. The errors ) , (
2 1 i i
c c are
assumed to have the standard bivariate normal distribution, with ) ( 1 ) (
2 1 i i
V V c c = = and
µ c c = ) , (
2 1 i i
Cov . Thus the individual’s probability of being an Internet banker can be written
as:
) ( banker Internet P ) 1 , 1 ( = = =
i i
shopper Internet banker Internet P
) , (
2 2 1 1 i i i i
x X x X P < < =
i i
x x
i i
dz dz z z
i i
2 1 2 1
2 1
) ; , ( 2
} }
· ÷ · ÷
= µ |
) ; , (
2 2 1 1
µ | |
i i
X X F =
where F denotes the bivariate standard normal distribution function with correlation
coefficient µ .
3
The bivariate probit model has full observability if
i
banker Internet and
i
shopper Internet are both observed in terms of all their four possible combinations [i)
‘Internet banker
i
= 1, Internet shopper
i
= 1’, ii) ‘Internet banker
i
= 1, Internet shopper
i
= 0’,
iii) ‘Internet banker
i
= 0, Internet shopper
i
= 1’, and iv) ‘Internet banker
i
= 0, Internet
shopper
i
= 0’]. Category (ii) is always equal to zero in our sample. If there is a clear sequence
3
Greene (2003) shows that the density function is given by:
2 / 1 2 ) 1 /( ) 2 )( 2 / 1 (
2
) 1 ( 2 /
2
2 1
2
2
2
1
µ t |
µ µ
÷ =
÷ ÷ + ÷
i i i i
x x x x
e .
7
of decisions in adopting Internet shopping and then Internet banking then it is appropriate to
deem this to be a naturally constrained complete set of observations and effectively provides
us with full observability in our data. It is known that full observability leads to the most
efficient estimates (Ashford and Sowden, 1970; Zellner and Lee, 1965).
The following section presents the results from application of the bivariate probit
method to the case explained above, and specifically will discuss the conditional marginal
effects obtained from P(Internet Banker
i
= 1 | Internet Shopper
i
= 1).
4. Results
A sequential structure
Figure 2 shows the structure of our purported sequential structure. It shows that
approximately one-third of Internet users do not use the Internet for shopping. More
importantly for our research, it illustrates that no one uses the Internet for banking if they do
not use the Internet for shopping. In our sample, less than 30 percent of respondent use the
Internet for banking, and this represents more than 40 percent of the respondents who use the
Internet for shopping. Figure 2 corroborates the perspective that the decisions to use the
Internet for shopping and for banking may be sequential as it is in line with the proposition
that using the Internet for shopping is a prerequisite for using the Internet for banking.
{Insert Figure 2 about here}
We draw our data from Patsiotis et al.’s (2012) study which captures the
multidimensional nature of each construct, but inclusion of all dimensions in our regression
approach would produce excessive multicollinearity and potentially incorrect coefficient
estimates due to included variable bias.
4
To circumvent these problems we apply principal
component analysis to each construct as a data reduction technique. We retain each
construct’s principal component and use these as regressors into our bivariate probit
analysis.
5
Table 2 presents descriptions of these principal components.
{Insert Table 2 about here}
Bivariate regression results
Our regression results are presented in Table 3. They indicate there is a very strong
association between the choices to use the Internet for shopping and to use the Internet for
banking; this is illustrated by the highly statistically significant result of the log-likelihood of
rho (chi
2
= 43.2469, p < 0.000).
{Insert Table 3 about here}
Our initial regression is the choice to use the Internet for shopping. The results
corroborate extant literature by illustrating that males and the higher educated are more likely
4
The empirical findings of this study have several limitations. These refer to the chosen geographic context,
the chosen predictors for modelling, the cross-sectional nature of the survey, and the possible non-response
bias considering that sample respondents may not represent the working population.
5
See Churchill and Iacobucci (2002) for a description of the principal components approach.
8
to use the Internet for shopping. Moreover, the results show that greater enjoyment
(significant at the 10 percent level) and greater personal capacity both enhance the likelihood
that an individual will use the Internet for shopping, while greater perceived risks negatively
influence the likelihood that a person will use the Internet for shopping, all given that they are
already Internet users. The positive influences observed, and especially on enjoyment, are in
line with empirical findings that emotions influence positively usage behaviour (Martin et al.,
2008; Shamdasani et al., 2008; Watson and Spence, 2007; Wood and Moreau, 2006).
Different factors are important in enhancing the likelihood that an Internet user will
use the Internet for banking. Greater enjoyment and greater ease of use both enhance the
likelihood that an Internet user will also use the Internet for banking; conversely, lower
perceptions of usefulness and a lack of trial will both diminish this likelihood. Although the
core application is the same, usage of subsequent innovations may be influenced by different
factors and therefore proceed to estimate a sequential model.
Conditional marginal effects
Set within a bivariate probit regression approach, it is possible to examine our proposition
that the equations should be estimated in sequence. Accordingly, Table 3 also presents two
sets of marginal effects estimates corresponding to those which are not based on a sequential
approach (i.e. unconditional) and those that are based on a sequential approach (i.e.
conditional). Several points are worthy of emphasis here. First, there is a slight disagreement
on which variables enhance the likelihood that an Internet user will use Internet banking as
the significant coefficients corresponding to education and time-energy are not significant at
tradition level of acceptance within the unconditional set up. These findings suggest that
time-energy and higher education affect the likelihood of using the Internet for banking and
not only the decision to use the Internet for shopping, as would be inferred under the
unconditional approach. Perhaps these results are reflecting the possibility that having higher
education increases the speed of learning new innovations and therefore increases the
likelihood that an Internet user will more quickly adopt Internet banking facilities. The
greater coefficient estimates for the time-energy variable when the structure is conditional
also emphasises the increasing importance of speed when using the Internet for banking
transactions.
Second, when the variables have been identified as being highly statistically
significant, the coefficients estimates corresponding to the condition approach are typically
much larger than for the unconditional approach. This suggests the standard unconditional
approach leads to coefficient estimates that are biased towards zero (i.e. having no effect),
and therefore underestimates their impact on the likelihood of using the Internet for banking.
The clearest example of this bias corresponds to having post-graduate education, where the
marginal effect is more than four-times greater under the conditional approach relative to the
unconditional approach. A lack of trail has a much greater hindering effect on using the
Internet for banking under the condition approach, being almost 50 percent greater than
expected under the unconditional approach.
Third, the empirical observation that there is a negative influence of usefulness on
Internet banking use offers new evidence on Internet banking usage behaviour; indeed our
empirical estimates suggest that usefulness is the most important lever for policy formation,
and emphasising the usefulness of Internet banking could lead to the greatest amount of new
Internet bankers. Moreover, the negative effects of usefulness on Internet banking adoption
rates contrasts with existing research on the influence of usefulness on customer intentions
and usage toward technology-based service offerings (TAM model and extensions: Gilbert et
9
al., 2004; Hernandez et al., 2009; McKechnie et al., 2006; Ozdemir and Trott, 2009). In a
study comparing Internet banking acceptance to older self-service technologies for banking,
such as the ATM and phone, Curran and Meuter (2005) found that ease of use and usefulness
are not important predictors of Internet banking, although there are important for the adoption
of ATM. They concluded that for an innovation to be successful, i.e. reach the majority of
potential customers, it must be both useful and easy to use. Thus, the low adoption rates of
Internet banking may be associated with the negative influence of usefulness. Considering
that usefulness is similar to relative advantage and that relative advantage has been found to
be one of the strongest predictors of an innovation’s adoption rate (Rogers, 2003), the
negative effect of usefulness on Internet banking propensity may further explain low adoption
rates. Another explanation for the negative influence of usefulness on Internet banking is that
Internet shoppers may perceive Internet banking to be less useful when compared to
alternative channels, such as the ATM, and therefore not interesting for further consideration.
Fourth, a lack of trial of Internet banking is found to be strongly associated with non-
usage behaviour, which could be difficult to circumvent as trialability of Internet banking
cannot readily be experienced before adoption. In our study, although Internet banking users
are positively influenced by ease of use, enjoyment, and to some extent value time-energy
savings, usefulness and lack of trial explain non-use. Thus, Internet shoppers may not be fully
aware of the usefulness aspects of Internet banking use. This is supported by the negative
influence of lack of trial, i.e. less or no opportunities to understand how it works do not add
knowledge on the usefulness aspects.
It may also be the case that customers use a combination of banking methods other
than the online. The literature supports a multi-channel integration (Coelho and Easingwood,
2003; Zeithaml et al., 2009). Earlier results indicate that consumers have a preference for a
mix of delivery channels rather than exclusive reliance upon anyone single channel
(Howcroft et al., 2002), and new delivery channels tend to complement rather than replace
the existing ones (Hughes, 2006). Finally, it is interesting in the results of this study that
interactivity, and any perceived risks, security and privacy concerns do not influence
adoption rates. This also contrasts empirical research on the preceding influences. It may be
the focus of extant work on pre-adoption behaviour, compared to the method adopted in this
study examining usage behaviour, that explains their non-importance.
5. Conclusion and implications
This study examined empirically the conditional link between Internet shopping and Internet
banking in order to identify key factors influencing Internet banking adoption rate. The focus
was on the behaviour of Internet users that adopt and use Internet shopping first in order to
decide whether to use the Internet for their banking needs.
The application of bivariate probit regression analyses revealed that for those with
Internet shopping experience, the probability of using Internet banking services is positively
affected by customer perceptions on enjoyment, ease of use, and time-energy savings, and
negatively influenced by usefulness and lack of trial. The results of this study support the
presence of a conditional link, and suggest Internet shopping and Internet banking should be
examined sequentially. Empirical research usually examines diffusion theories within the
context of one technological innovation. The results of this study strongly support the
proposition that diffusion research should examine new technological innovations
sequentially when they are based on a core application and the adoption of the first would
probably lead or not to the adoption and use of the second.
10
The above findings suggest further research inquiry and practical implications for
service providers. Future research could further apply the sequential modelling approach in
similar service contexts (e.g., smart phone, personal assistance shopping) and in other
geographic areas exhibiting different adoption rates. This is particularly applicable when
technological innovations are introduced, or have been introduced, sequentially in the market,
and there is a lack of an accurate sample frame. Usage behaviour could be examined further
to identify the most relevant predictors and confirm the key influences revealed by this study.
As the chosen innovations are in the market for some time now, different factors might be
more important influences to their adoption rates. This would facilitate a better understanding
of non-usage behaviour too. In addition, longitudinal studies based on panel data could
provide better indication of the future direction of the conditional link and the most important
influencing factors.
Finally, the evidence provided in this research study on the conditional link between
Internet shopping and Internet banking suggests that even though the basis of a technological
interface is the computer, new services developed through this can have different perceived
characteristics. Thus, banks should be aware that Internet banking use presents unique
characteristics compared with the other alternative channels as well as with related services.
Bank customers that use the Internet for shopping but not for banking should be informed for
the usefulness aspects of online banking, as well as be given the opportunity to experience
how it works, given the perceived lack of trial. In particular, the usefulness aspects should be
examined further through a comparison of the alternative methods for banking. It may be that
some customers find Internet banking to be less useful when compared to other method(s), or
they may not be fully aware of its benefits. Existing users of Internet banking facilities value
the enjoyment aspects, so service providers should make the experience enjoyable so as to
sustain positive emotions. As a result, good knowledge of the behaviour of users would
facilitate their effort to understand those likely to be new users and incorporate this
understanding in their marketing strategy to develop a more targeted communications policy
that would generate useful customer feedback.
References
Bass, F.M. (1969). A new product growth model for consumer durables. Marketing Science, 15, 215-227
Berry, L.L., Seiders, K. and Grewal, D. (2002). Understanding service convenience. Journal of Marketing, 66
(3), 1-17.
Bitner,M.J., Ostrom, A.L. and Meuter, M.L. (2002). Implementing successful self-service technologies.
Academy of Management Executive, 16 (4), 96-108.
Black, N., Lockett, A., Winklhofer, H. and Ennew, C. (2001). The adoption of Internet financial services: a
qualitative study. International Journal of Retail and Distribution Management 29(8), 390-398
Bobbitt, M. and Dabholkar, P.A. (2001). Integrating attitudinal theories to understand and predict use of
technology-based self-service: the internet as an illustration. International Journal of Service Industry
Management, 12 (5), 423-450.
Churchill, G.A. and Iacobucci, D. (2002). Marketing Research: Methodological Foundations, 8th ed. Orlando,
FL: Harcourt College Publishers.
Coelho, F. and Easingwood, C. (2003). Multiple channel structures in financial services: A framework. Journal
of Financial Services Marketing, 8 (1), 22-34.
Cunningham, L., Gerlach, J., Harper, M. and Young, C. (2005). Perceived risk and the consumer buying
process: Internet airline reservations. International Journal of Service Industry Management, 16 (4), 357-372
Curran, J.M. and Meuter, M.L. (2005). Self-service technology adoption: comparing three technologies. Journal
of Services Marketing, 19 (2), 103-113.
Dabholkar, P.A. (1996). Consumer evaluations of new technology-based self-service options: An investigation
of alternative models of service quality. International Journal of Research in Marketing, 13 (1), 29-51.
11
Dabholkar, P. and Bagozzi, R. (2002). An attitudinal model of technology-based self-service: Moderating
effects of consumer traits and situational factors. Journal of the Academy of Marketing Science 30(3), 184-
201
Davis, F.D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology.
MIS Quarterly, 13 (3), 319-340.
Davis, F.D., Bagozzi, R.P. and Warshaw, P.R. (1992). Extrinsic and intrinsic motivation to use computers in the
workplace. Journal of Applied Social Psychology, 22 (14), 1109-1130.
Durkin, M. (2007). Understanding registration influences for electronic banking. International Review of Retail,
Distribution and Consumer Research. 17(3), 219-231
Ellen, P., Bearden, W. and Sharma, S. (1991). Resistance to technological innovations: an examination of the
role of self-efficacy and performance satisfaction. Journal of the Academy of Marketing Science 19(4), 297-
307
Elliott, K.M. and Hall, M.C. (2005). Assessing consumers’ propensity to embrace self-service technologies: Are
there gender differences? The Marketing Management Journal, 15 (2), 98-107.
Fitzsimmons, J.A. and Fitzsimmons, M.J. (2011). Service Management: Operations, Strategy, Information
Technology. 7
th
ed. New York, NY: McGraw-Hill/Irwin.
Fourt, L.A. and Woodlock, J.W. (1960). Early prediction of market success for grocery products. Journal of
Marketing. 25 (October), 31-38.
Gan, C., Clemes, M., Limsombunchai, V. and Weng, A. (2006). A logit analysis of electronic banking in New
Zealand. International Journal of Bank Marketing. 24 (6), 360-383.
Gatignon, H. and Robertson, T.S. (1989). Technology diffusion: an empirical test of competitive effects.
Journal of Marketing, 53 (1), 35-49.
Gilbert, D., Balestrini, P. and Littelboy, D. (2004). Barriers and benefits in the adoption of e-government.
International Journal of Public Sector Management 17(4) 286-301
Greene, W.H. (2008). Econometric Analysis, Upper Saddle River, NJ: Prentice-Hall.
Hernandez, B., Jimenez, J. and Martin, M.J. (2009). Adoption vs acceptance of e-commerce: two different
decisions. European Journal of Marketing, 43 (9/10), 1232-1245.
Hernandez, J.M.C. and Mazzon, J.A. (2007). Adoption of internet banking: proposition and implementation of
an integrated methodology approach. International Journal of Bank Marketing, 25(2), 72-88.
Howcroft, B., Hamilton, R. and Hewer, P. (2002). Consumer attitude and the usage and adoption of home-based
banking in the UK. International Journal of Bank Marketing. 20 (3), 111-121.
Hughes, T. (2006). New channels/old channels – Customer management and multi-channels. European Journal
of Marketing, 40 (1/2), 113-129.
Jauda, S., Trocchia, P.J. and Gwinner, K.P. (2002). Consumer perceptions of Internet retail service quality.
International Journal of Service Industry Management 13 (5), 412-431.
Jun, M. and Cai, S. (2001). The key determinants of Internet banking service quality: A content analysis.
International Journal of Bank Marketing, 19 (7), 276-291.
Lassar, W.M., Manolis, C. and Lassar, S. (2005). The relationship between consumer innovativeness, personal
characteristics, and online banking adoption. International Journal of Bank Marketing, 23 (2), 176-199.
Lee, H-J, Cho, H.J., Xu, W. and Fairhurst, A. (2010). The influence of consumer traits and demographics on
intention to use retail self-service checkouts. Marketing Intelligence and Planning, 28 (1), 46-58.
Lilien, G.L., Kotler, P. and Moorthy, K.S. (1992). Marketing Models. Englewood Cliffs, NJ: Prentice-Hall.
Littler, D. (2001). The idiosyncrasies of new product adoption for technologically innovative offerings. The
Marketing Review. 2, 3-19.
Lockett, A. and Littler, D. (1997). The adoption of direct banking services. Journal of Marketing Management,
13 (8), 791-811.
Mahajan, V., Muller, E. and Bass, F.M. (1990). New product diffusion models in marketing: A review and
directions for research. Journal of Marketing. 54 (January), 1-26.
Mahajan, V., Muller, E. and Wind, Y. (2000). New Product Diffusion Models, Boston, MA: Kluwer Academic.
Makarem, S.C., Mudambi, S.M. and Podoshen, J.S. (2009). Satisfaction in technology-enabled service
encounters. Journal of Services Marketing, 23 (3), 134-144.
Mansfield, E. (1961). Technical change and rate of imitation. Econometrics, 29 (October), 741-766.
Martin, D., O’Neill, M., Hubbard, S. and Palmer, A. (2008). The role of emotion in explaining consumer
satisfaction and future behavioural intention. Journal of Services Marketing, 22 (3), 224-236.
McKechnie, S., Winklhofer, H. and Ennew, C. (2006). Applying the Technology Acceptance Model to online
retailing of financial services, International Journal of Retail & Distribution Management 34(4/5), 388-410
Meuter, M.L., Ostrom, A.L., Bitner, M.J. and Roundtree, R. (2003). The influence of technology anxiety on
consumer use and experiences with self-service technologies. Journal of Business Research, 56, 899-906.
Moore, G.C. (1991). Crossing the Chasm, New York, NY: Harper Business.
12
Moore, G.C. and Benbasat, I. (1991). Development of an instrument to measure the perceptions of adopting an
information technology innovation. Information Systems Research, 2 (3), 192-222.
Ozdemir, S. and Trott, P. (2009). Exploring the adoption of a service innovation: A study of Internet banking
adopters and non-adopters. Journal of Financial Services Marketing, 13 (4), 284-299.
Parasuraman, A. (2000). Technology readiness index (TRI): a multiple-item scale to measure readiness to
embrace new technology. Journal of Service Research, 2 (4), 307-320.
Patsiotis, A.G., Hughes, T. and Webber, D.J. (2012). Adopters and non-adopters of internet banking: a
segmentation study. International Journal of Bank Marketing, 30 (1), 20-42.
Robertson, T. (1971). Innovative Behaviour and Communication, New York, NY: Holt, Reinhart and Winston.
Rogers, E.M. (2003). Diffusion of Innovations, 5th ed. New York, NY: Free Press.
Sawhney, M., Verona, G. and Prandelli, E. (2005). Collaborating to create: the Internet as a platform for
customer engagement in product innovation. Journal of Interactive Marketing, 19 (4), 4-17.
Shamdasani, P., Mukherjee, A. and Malhotra, N. (2008). Antecedents and consequences of service quality in
consumer evaluation of self-service Internet technologies. The Service Industries Journal, 28 (1), 117-138.
Shi, W., Shambare, N. and Wang, J. (2008). The adoption of internet banking: An institutional theory
perspective. Journal of Financial Services Marketing. 12 (4), 272-286.
Simon, F. and Usunier, J. (2007). Cognitive, demographic, and situational determinants of service customer
preference for personnel-in-contact over self-service technology. International Journal of Research in
Marketing, 24, 163-173.
Tapp, A. and Hughes, T. (2004). New technology and the changing role of marketing. Marketing Intelligence
and Planning. 22 (3), 284-296.
Varadarajan, P.R. and Yadav, M.S. (2002). Marketing strategy and the Internet: An organizing framework.
Journal of the Academy of Marketing Science 30(4) 296-312.
Veryzer, R.W., Jr. (1998a). Key factors affecting customer evaluation of discontinuous new products. Journal of
Product Innovation Management, 15, 136-150.
Veryzer, R.W., Jr. (1998b). Discontinuous innovation and the new product development process. Journal of
Product Innovation Management, 15, 304-321.
Walker, R.H. and Johnson, L.W. (2006). Why consumers use and do not use technology-enabled services.
Journal of Services Marketing, 20 (2), 125-135.
Walker, R.H., Craig-Lees, M., Hecker, R. and Francis, H. (2002). Technology-enabled service delivery: An
investigation of reasons affecting customer adoption and rejection. International Journal of Service Industry
Management, 13 (1), 91-106.
Wang, G., Dou, W. and Zhou, N. (2008). Consumption attitudes and adoption of new consumer products: A
contingency approach. European Journal of Marketing. 42 (1/2), 238-254.
Watson, L. and Spence, M.T. (2007). Causes and consequences of emotions on consumer behaviour – A review
and integrative cognitive appraisal theory. European Journal of Marketing, 41 (5/6), 487-511.
Wood, S.L. and Moreau, C.P. (2006). From fear to loathing? How emotion influences the evaluation and early
use of innovations. Journal of Marketing, 70 (3), 44-57.
Yadav, M.S. and Varadarajan, R.R. (2005a). Understanding product migration to the electronic marketplace: A
conceptual framework. Journal of Retailing, 81 (2), 125-140.
Yadav, M.S. and Varadarajan, R.R. (2005b). Interactivity in the electronic marketplace: an exposition of the
concept and implications for research. Journal of the Academy of Marketing Science, 33 (4), 585-603.
Yousafzai, S., Pallister, J. and Foxall, G. (2009). Multi-dimensional role of trust in Internet banking adoption.
The Service Industries Journal, 29 (5), 591-605.
Zeithaml, V.A., Bitner, M.J. and Gremler, D.D. (2009). Services Marketing, 5th ed. New York, NY: McGraw-
Hill/Irwin.
Zeithaml, V.A., Parasuraman, A. and Malhotra, A. (2002). Service quality delivery through websites: A critical
review of extant knowledge. Journal of the Academy of Marketing Science, 30 (4), 362-375.
13
Figure 1: Conceptual framework
Internet Use
No Internet
Shopping Use
Internet Shopping
Use
No Internet
Banking Use
Internet Banking
Use
Predictors
- Perceived
Innovation
Attributes
- Demographic
Characteristics
14
1
I use the Internet for shopping I use the Internet for Banking
Yes = 42.13 %
Yes = 0.00 %
Yes = 65.93 %
No = 57.87 %
No = 100.00 %
No = 34.07 %
0.00 %
38.15 %
34.07 %
27.78 %
Final
probabilities
Figure 2: Tree diagram
15
Table 1. Description of non-principal component variables
Variables Description
Dependents:
- Internet shopping use:
- Internet banking use:
Binary = 1 if someone uses; 0 non-use
Binary = 1 if someone uses; 0 non-use
Independent
Individual characteristics:
Gender (binary)
Age (years)
Education
1 if male (47%); 0 if female (53%)
18-25 (26.3%); 26-35 (41.5%); 36-45 (20%); 46-55
(7.8%); 56-65 (4.4%)
School (3.7%); College (20%); University (43%);
Postgraduate (33.3%)
1
Table 2: Results of principal components analysis
Enjoyment Time and energy Ease of use Usefulness Interactivity Personal capacity Perceived risks
Privacy
concerns
Security
concerns
Lack of trial
Loadings 7a, b, c 8g1, 2, 3, 4 8h, i, j 8g5, e, f 9 (all 33) 10b, c, d, e 11a, b, c 11d1, 2, 3 11d4, 5, 6 11e, f, g, h, i, j
Eigenvalue 2.263 2.594 1.483 1.753 13.343 2.474 2.265 2.421 2.348 4.022
Proportion (%) 0.754 0.649 0.494 0.584 0.404 0.619 0.755 0.807 0.783 0.670
1
Table 3: Estimates of probit regressions
Unrelated bivariate probit Unconditional marginal effects Conditional marginal effects
Coef. Std. error p dy/dx Std. error p dy/dx Std. error p
Internet shopper
Male
0.627
0.193
0.001
Age: 18-25 0.110 0.228 0.630
Age: 26-35 Control variable
Age: 36-45 0.1600 0.269 0.552
Age: 46-55 -0.057 0.401 0.887
Age: 56-65 0.106 0.452 0.815
Education School 0.4500 0.460 0.328
Education College 0.189 0.234 0.420
Education Degree Control variable
Education Post-grad 0.969 0.262 0.000
Enjoyment 0.128 0.074 0.082
Time and energy -0.035 0.079 0.659
Ease of use 0.124 0.096 0.199
Usefulness -0.134 0.095 0.160
Interactivity 0.035 0.032 0.273
Personal capacity 0.272 0.075 0.000
Perceive risks -0.309 0.100 0.002
Privacy 0.190 0.116 0.102
Security -0.137 0.102 0.177
Lack of trial 0.040 0.065 0.538
constant -0.165 0.203 0.417
Internet banker
Male 0.589 0.226 0.009 0.168 0.065 0.009 0.148 0.083 0.075
Age: 18-25 -0.156 0.291 0.591 -0.043 0.077 0.577 -0.077 0.109 0.480
Age: 26-35 Control variable Control variable Control variable
Age: 36-45 0.776 0.292 0.008 0.256 0.106 0.016 0.285 0.099 0.004
Age: 46-55 0.144 0.475 0.761 0.043 0.148 0.772 0.065 0.187 0.729
Age: 56-65 1.002 0.494 0.042 0.360 0.192 0.060 0.382 0.170 0.025
Education School -0.653 0.678 0.335 -0.338 0.098 0.160 -0.319 0.263 0.224
Education College -0.135 0.295 0.647 -0.037 0.077 0.634 -0.079 0.108 0.464
Education Degree Control variable Control variable Control variable
Education Post-grad -0.164 0.252 0.514 -0.045 0.068 0.504 -0.196 0.090 0.029
Enjoyment 0.291 0.093 0.002 0.082 0.026 0.002 0.098 0.036 0.006
Time Energy 0.137 0.093 0.140 0.039 0.026 0.138 0.059 0.033 0.078
Ease Of Use 0.225 0.110 0.041 0.064 0.031 0.038 0.072 0.039 0.066
Usefulness -0.316 0.101 0.002 -0.089 0.029 0.002 -0.107 0.036 0.003
Interactivity 0.004 0.037 0.909 -0.001 0.010 0.909 -0.003 0.013 0.817
Personal Capacity 0.127 0.095 0.180 0.036 0.027 0.177 0.013 0.035 0.699
Perceived Risks -0.109 0.103 0.292 -0.031 0.029 0.291 -0.001 0.040 0.975
Privacy 0.051 0.110 0.640 0.014 0.031 0.640 -0.005 0.042 0.898
Security -0.025 0.099 0.798 -0.007 0.028 0.798 0.009 0.037 0.816
Lack Of Trial -0.231 0.073 0.002 -0.065 0.022 0.003 -0.097 0.030 0.001
constant -1.154 0.234 0.000
Rho
Log Likelihood
1
-191.55
5.21e-09
Notes: N=259; bold font highlights statistical significance at the 10% confidence level. Likelihood-ratio test of
rho=0, chi
2
(1) = 43.2469, Prob > chi
2
= 0.0000.
doc_118858653.pdf