The Propensity of E-commerce Usage: the Influencing Variables

Or

The Propensity of E-commerce Usage: an Empirical Analysis of the

Determinants
Abstract

The following paper addresses online consumer behavior in purchase decision

making. Research - in the form of various psychological and sociological studies -
has, to date provided many theoretical models to explain how consumers utilize new
technologies. These models represent a point of reference and are intended to provide
explanation for the propensity of particular classes of consumers to use specific
applications. Further, some of the proposed models have been partially modified and
applied to cases of online purchasing. The intention of this study is to utilize these
existing theories, not simply adapt them to the phenomenon of e-commerce, but to
question and therefore change fundamental variables. The main variables are derived
from models in the literature which have been tested together using a statistical
methodology not previously applied to this type of problem: the CART procedure.

1. Introduction

In recent years academics have been highly interested in the Internet, given its
expected ability to profoundly influence the nature of business, markets and the
economy in general (Biswas and Krishnan, 2004). In a 1995 study, authors Weiner
and Brown proposed a connection between the Internet and a number of potential
social changes relating to demographic shifts of population, the growing needs of the
individual in the optimization of time, the growing body of information available in
relation to the purchase of products, and the resulting emergence of the Internet as
market space. Indeed, to date, there has been a real proliferation of commercial Web
sites that offer products or services purchased directly online. According to a 2008
report out of B2C, the School of Management of Milan Polytechnic, the total
purchases made via online business sites by Italian consumers in 2008 exceeded 6
billion Euros, an increase of more than 20% compared to the previous year, which
had recorded 23% sales growth. The sale of products on the Internet is a departure
from traditional distribution channels which leverage innovations related to
information and communication technology (ICT). It is possible to study online
purchasing behavior and thereby the propensity of consumers to use the new
technology of ecommerce.

The acceptance of technology by consumers has been widely investigated in order to

understand which factors encourage the adoption and, therefore, the use of new
technologies. The development of the Web and the activities of e-commerce have led
to the adaptation of the main theories of analysis to fit the specific situation of Internet
commerce (Shin 2007).

The first task of this paper is to investigate the foundations of the theoretical models
so far proposed to explain the phenomenon of acceptance of technology. Following
this is an exploration of the core constructs themselves. This will allow for combining
the empirical research designed to study the phenomenon of e-commerce with a new
methodological approach which looks at technology acceptance and which has not
been applied to traditional business models.

In the study, the intention is to exploit the previous theories, not simply adapting them
to the phenomenon of e-commerce, as has been in past research, but to put into
question and, therefore, to bring about a change in key variables. The determinants of
the phenomenon in this study were derived from the model prediction of acceptance
of technology and were tested using a statistical methodology previously never
applied to this type of problem, i.e. the CART procedure.

Therefore the study has provided for the assessment of variables derived from
different models, leading to a more detailed and accurate picture of the behavior of
online purchasing, the results of which are outlined below. Finally, a discussion also
ensues regarding prospects for the development of future research, as well as
considering the limitations and implications of similar applications of these models in
business management.

2. The theoretical models

As previously stated, the use of e-commerce by consumers may be considered and,

therefore, studied as a phenomenon of acceptance of technology. Since the initial
adoption of new technologies (for example, personal computers, spreadsheets, and
emails) can be very challenging for new users, there is considerable uncertainty
regarding the success of their continued use. Indeed, there is in practice a high rate of
failure of initiatives to introduce new information and communication technologies.
Therefore, a study of the determinants of acceptance seems appropriate, not only for
academic purposes, but also to ascertain the usefulness of such determinants and the
role they might play in the effective implementation of improved processes via the
introduction of new technologies by business firms.

Since the early ‘70s, numerous models have been proposed in order to understand and
explain determinants of acceptance in the use of information technologies cancel.
Some theories examine the relationship between the attitudes, perceptions and beliefs
of the users of a technology and the level of use of the technology itself. Among these
could include the Theory of Reasoned Actions (TRA), (Fishbein and Ajzen 1975,
Ajzen and Fishbein, 1980), the Theory of Planned Behavior (TPB), (Ajzen, 1985:
Ajzen and Madden, 1986), Model of PC Utilization (MPCU) (Thompson, Higgins &
Howell 1991; Triandis 1977), the Technology Acceptance Model (TAM), (Davis,
1989; Davis et al., 1989), the Motivation Model (MM) (Davis, Bagozzi & Warshaw
1992), the Unified Theory of Acceptance and Use of Technology (UTAUT),
(Venkatesh et al., 2003), the Innovation Diffusion Theory (IDT) (Rogers 1995), the
Combined TAM and TPB (Taylor & Todd 1995),and the Social Cognitive Theory
(Bandura 1986; Compeau & Higgins 1995a; Compeau & Higgins 1995b; Compeau,
Higgins & Huff 1999).
Theory of Reasoned Actions (TRA) was developed through research in social
cognition and the subsequent Theory of Planned Behavior. The TRA explains almost
every type of human behavior based on the beliefs and intentions of individuals. The
motivation to comply is the greater or lesser propensity of individuals to adapt to the
expectations of reference groups. Individuals may be more or less inclined or biased
to such expectations. It is possible to distinguish subjects at high-sensitivity rules,
meaning subjects are sensitive and strongly influenced by expectations of others, and
also by low-sensitivity rules, referring to subjects who are the least influenced by the
expectations of others. The TRA has produced a considerable number of contributions
in various fields of marketing and influence on behavior. Further, it has demonstrated
predictive power provided that the behavior is under the individual decision power,
which falls within the scope of possible actions and practically concrete actions in
which the person has room to maneuver. The TRA assumes that a motivated person
could not act if they receive little control over external environmental factors which
affect the ability to act. Combined with the idea that the effort appears beyond the
perceived scope, the client fails to act even when tempted by the promise of any
perceived benefit the purchase may produce.

To overcome the limitations of the Theory of Reasoned Action, the authors have
proposed a second model, the Theory of Planned Behavior (TPB). This model was
developed to predict behaviors in which the subject does not have the complete
decision power due to exclusion via some internal or external barrier. The TPB
considers the impact of the subject’s attitude toward the intention to use a certain
technology with respect to consumption, subjective norms and perceived behavioral
control. The biggest difference between TRA and TPB is the addition of perceived
behavioral control variables. These theories suggest that even a difficult task can be
attempted if the perception of the possibility to succeed is high. Conversely, even
simple actions will not be implemented, if the individual perceives barriers (internal
or external) that can stop him or her. The TPB has helped to clarify the importance of
the beliefs of the subject as the main determinants of his behavior. Therefore, it can be
reasoned that there exists a determinant to behavior rooted in subjective belief, not
objective reality.

According to TAM, which indeed has its roots in the TRA theory, people acquire
attitudes and, consequently, certain levels of intention with respect to the use of new
technology before ever making an effort to use it. The model suggests that the
perceived ease of use and usefulness of a technology are good predictors of the
attitude and intention of potential users who may then choose to utilize (or not to
utilize) that technology based on those initial perceptions. The addition of external
agents can influence the use of technology by changing the perception of ease of use
and usefulness of the system introduced. Several studies (Adams et al., 1992; Davis et
al., 1989; Hendrickson et al., 1993; Segars and Grover, 1993; Subramanian, 1994;
Szajna, 1994) confirmed the validity of the TAM model to explain the phenomenon
of acceptance of new technologies in various fields, including the banking sector
(Adamson and Shine, 2003), writing programs and word processing (Davis et al.,
1989), applications relating to spreadsheets (Mathieson , 1991), use of web browsers
(Morris and Dillon, 1997), the Internet (Koufaris, 2002), mobile commerce (Bruner
and Kumar, 2005), online games (Hsu and Lu, 2004), email (Huang et al., 2003),
Lotus notes (Li et al., 2004), telemedicine (Chau and Hu, 2001 Hu et al., 1999), and
so forth.

The first version of TAM provides that attitude affects the intention to use the system,
where, in a later version, called TAM2 (Venkatesh and Davis, 2000), the two
variables "attitude" and "intention to use" have been merged in a single construct.
Venkatesh & Davis (2000) explained how perceived usefulness and usage intentions,
taking into account social influences and cognitive processes, affect technology
acceptance. Their research found that social influence processes, as well as cognitive
instrumental processes, are all determinants in acceptance of new technologies.
TAM2 re-introduces subjective norm as a key determinant of behavioral intention to
use, bringing it back in line with TRA.

The original model has repeatedly been the subject of studies designed to extend the
search for variables that may influence the main constructs, perceived ease of use, and
perceived usefulness. For example, Venkatesh and Davis (1996) introduced the
concept of a perceived self that acts on the perception of ease of use before a
difference of objective usability can be determined when actually using it. Taylor &
Todd (1995a) investigated the possibility of an augmented TAM that allowed for an
experience construct to account for the behavior of inexperienced users. Also the ease
of use would suffer the effects of control (perceived self in the use of computers and
facilitating conditions), the intrinsic motivation (enjoyment in using the computer),
and emotions (anxiety in using computers) (Venkatesh, 2000).

In many cases, the TAM has been borrowed from the workplace and implemented in
the study of e-commerce. Incorporating aspects that are both hedonistic and utilitarian
(Hackbarth, Grover, & Yi, 2003; Koufaris, 2002; Moon & Kim, 2001; Venkatesh,
2000). Some researchers have therefore suggested the introduction of additional
factors than the original model such as trust, e-trust (Jarvenpaa et al., 2000, Gefen and
Straub, 2003; Pavlou, 2003), and the level of perceived playfulness or enjoyment
(Van der Heijden and Verhagen, 2004).

In particular, given the open and global nature of the Internet as an infrastructure for
commercial transactions, trust makes for a crucial factor in e-commerce (Hoffman et
al., 1999). Confidence in the seller and in the Internet media is a determinant of
consumer perception of safety. The trust in the Web retailer has empirically shown to
influence the intention to use the merchant site for e-commerce purposes (Gefen,
Karahanna, Straub, 2003; Jarvenpaa, Tractinsky, Saarinen and Vitale, 1999;
Jarvenpaa, Tractinsky, & Vitale, 2000). Trust in the Web retail environment is the
"willingness of the consumer to rely on the Web retailer when there is vulnerability
for the consumer"). Prior studies incorporated trust into TAM in different ways. Trust
was used as an antecedent of ease of use (Pavlou, 2003), usefulness (Dalberg et al.,
2003; Pavlou, 2003), attitude (Chen and Tan, 2004; Suh and Han, 2002) and
behavioral intention (Gefen and Straub, 2003; Pavlou, 2003; Suh and Han, 2002).
Moreover the trust-enhanced technology acceptance model was developed (Dahlberg
et al., 2003).

Likewise, the perceived potential for fun is a factor that impacts the use of new
technologies in general (Bruner and Kumar, 2005; Lee et al., 2005). Incorporating
hedonistic aspects of the technology adoption Davis, Bagozzi and Warshaw (1992)
have extended the original TAM model to introduce the perception of entertainment
as a factor influencing the adoption of technology. As part of ecommerce the two
strands in the literature, one being utilitarian and the other, hedonism, were analyzed
by Scarpi, Dall'Olmo Manaresi and Riley (2006).

Unified Theory of Acceptance and Use of Technology (UTAUT), (Venkatesh et al.,

2003), in order to explain user intentions to use an information system and subsequent
usage behavior, has empirically compared eight existing models that earlier research
had employed to explain information systems usage behavior. The theory holds that
four key constructs (performance expectancy, effort expectancy, social influence, and
facilitating conditions) are direct determinants of usage intention and behavior.
Gender, age, experience, and voluntariness of use are posited to mediate the impact of
the four key constructs on usage intention and behavior.

3. Research methodology

3.1. The hypothesis of research

When the user interfaces with a new technology, you can identify a number of factors
which influence the decision to use this technology in addition to the mode and time
of use. In order to conduct this analysis, first it was necessary to identify the variables
that constitute the various models and theories outlined above. The concordance or
discrepancies between the definitions of these variables showed similarities and
overlaps between the different models taken for examination. The factors or
constructs used in this study were measured by using items already validated in
previous research and have been adapted to the case of e-commerce. In addition to
these constructs, additional information was collected on gender, age, educational
qualifications and employment.

In particular, the factors or constructs used in the analysis are the following:

- Facilitating conditions;
- Perceived self efficacy;
- Experience;
- The perceived usefulness;
- The perceived ease of use;
- Subjective norms;
- Playfulness;
- E-trust;
- The intention to buy online.

Facilitating conditions

Conditions defined as facilitant factors also support and represent the perception that
the individual has the technical and organizational infrastructure required to support
the use of the information system (Venkatesh et al., 2003). This construct has strong
similarities with the perceived behavioral control in TPB (Ajzen and Fishbein, 1980).
In studies, systems related to Web facilitant conditions are translated into items
related to ease of connection to the Network.
The first hypothesis (H1) is therefore that the presence of a technical infrastructure
that allows easy connection to Internet has a positive influence on the intention of
buying online.

Perceived self efficacy

The perceived self was theorized by Bandura (1977) in the field of social cognitive
theory. According to Bandura, perceived self is the conviction of the individual that
they will be able to obtain the desired result with success. The self-efficacy is thus
divided by outcome expectancy, which refers to the probability estimated by
individuals that a certain behavior will lead to certain outcomes. The Theory of
Planned Behavior can be applied to perceived self via the perceived behavioral
control variable—that is, the individual’s perceived ease or difficulty in succeeding in
a certain behavior (Ajzen, 1988).
Considerable evidence from previous research suggests that perceived self efficacy
has an influence on the perceived probability of the behavior to have positive results;
therefore the second hypothesis (H2) of research is that greater self-efficacy perceived
by the users of the site would lead to a higher level of intention to purchase online.

Experience

The construct consists of the level of knowledge of the technology in question

acquired through use and practice. Experience facilitates the acquisition of
information and increasing individual knowledge, which can alter initial perceptions
(Venkatesh and Davis, 2000, Min and Galle, 2003).
The third hypothesis (H3) that is tested is the past experience in online buying, the
same should have a positive influence on online purchase intention.

Perceived usefulness
Perceived usefulness was defined by Davis (1989) as the degree to which a person
believes that using a particular system would enhance his or her job performance. The
items used in this work were taken from Davis (1989) and Davis et al., (1989) and
adapted to the case of electronic commerce.
The fourth hypothesis (H4) of research is that the perception of the usefulness of the
website causes the increase in online purchase intention.

Perceived ease of use

Perceived ease of use represents the level of physical and mental effort that users
expect to expend while considering use of the technology or the perception of the
degree to which the use of a particular system would be free from effort (Davis,
1989). Again, the items used were adapted from Davis (1989) and Davis et al. (1989).
The fifth hypothesis (H5) focuses on the perceived ease of use, so if the e-commerce
Website is perceived as easy to use the intention to purchase online will grow.

Playfulness

The hedonism of an action reflects its potential in terms of emotional impact and fun.
Customer satisfaction comes not only from the extrinsic reason for the product or
service purchased, but also from personal and emotional response connected to the
pleasure of it. This is true for both offline and online shopping (Jarvenpaa and Todd
1997; O'Cass and Fenech 2003; Shang et al., 2005). The items used were derived
from the study conducted by Davis et al. (1992).
The sixth hypothesis (H6) concerns the subjective norms, which indicate that the
purchase intention of the customers would be greater if the people who they believe
are relevant, agree to the purchase.

Subjective norms

Subjective norms are the rules by which operates the subjective motivation of
individuals to act consistently with the views of the individuals’ peer and social
group. This concept is taken into account in many models traditionally used to study
the adoption of e-commerce, such as the Theory of Reasoned and Action and the
Theory of Planned Behavior, although not in the TAM model. Bandura (1977) argues
that individuals learn and adopt the behaviors that operate within their social groups.
Social norms and networks of interpersonal communication have a decisive role in the
adoption of new technologies. In the present study the measurement of subjective
norms was already validated by Taylor and Todd (1995).
The seventh hypothesis (H7) of this research is that the perception of the ability of the
site to entertain and emotionally involve will increase the purchase intent.

E-trust
In different areas in which it was studied (i.e. organizational behavior, sociology and
social psychology), the concept of perceived safety has different, often even abstract
and complex meanings. In terms of the acceptance of technology, Chervany
McKnight (2001) defined safety as the perceived belief that the new technology is
reliable and credible. Pavlou (2003), in the context of business to consumer
ecommerce, has referred to trust as "the belief that allows consumers to willingly
become vulnerable to Web retailers after having taken the retailers' characteristic into
consideration."
Confidence in both the seller and in the Internet is a determinant of consumer
perception of safety. The perceived safety, then, is measured relative to both the seller
online and compared to the security of the e-commerce system used.
The last hypothesis (H8) of research is that the reliability of the seller and the online
purchase in other words the perceived safety influences positively the intention to
purchase.

Intention of use (intention to buy online)

Intention of use, as considered in this analysis is the dependent variable. It is

predictive of the actual rate of adoption of technology as well as of the intention of
the consumer to use the site to make an online purchase. This item was adopted from
Davis (1989) and Davis et al., (1989).

As already stated, from TAM2 (Venkatesh and Davis, 2000), the constructs "attitude"
and "intention" have been consolidated into a single variable, i.e. intention to use the
system. This factor incorporates the construct of "attitude about the conduct" in the
TRA (Ajzen and Fishbein, 1970, 1975) and TPB (Ajzen, 1988).

Figure 1 - The causal relationships between constructs

-- Insert Figure 1 Here --

3.2. The survey

The questionnaire administered to users of the site consists of 22 questions. Most of

the items included in the questionnaire, including the one on the dependent variable,
are measured with a likert scale using 6 levels1 ranging from "strongly agree" to

1
It has been chosen to use a 6 point likert scale even though previous research used 5 or 7 point likert scale
for two main reasons. First the aim was to use a response scale with an even number of responses and no
middle neutral or undecided choice. Removing the explicit midpoint the respondent is forced to decide
(Kalton & Schuman, 1982). Kalton, Roberts, and Holt (1980) pointed out that when investigators decide
against offering an explicit middle alternative, they usually assume that the middle category consists largely
of responses from those who lean toward one or the other polar alternatives, albeit perhaps with little
"strongly disagree." Some items are of a binary type. Some variables used were
analyzed as quantitative variables, others as ordinal and nominal.

The questionnaire was administered online to users of certain e-commerce sites.

There were 24 such sites used here, all of which sell products within the different
sectors of the consumer market. The participation of these companies was made
possible thanks to the collaboration of AICEL (Italian Association of Electronic
Commerce), an association which brings together a variety of Italian merchants
operating both within Italy and abroad.

The responses received numbered 1365, of which 1178 were considered useful for
analysis. Using the statistical methodology described below, cases with missing data
were not excluded. Only those with a missing response rate exceeding 50% were
excluded.

Four types of explanatory variables have been used in order to identify the profile of
the respondents related to gender, age, level of training and profession. Below (Table
1) are given the characteristics of the components of the sample.

Table 1 - The profile of respondents

-- Insert Table 1 Here --

Source: Research Data

3.2.1. Methodology of statistical analysis

To analyze the data collected, a technique of data mining (CART) has been used in
order to achieve an effective response to the particular requirements of the process of
data analysis.

While traditional statistical analysis was designed primarily to test specific research
hypotheses by default and then be applied to the data collected, data mining can be
understood as a secondary use of data in order to pursue other objectives.

intensity. I aimed to decrease the central tendencies and increase the variance in the responses. Bishop
(1987) states that people who select a middle response alternative when it is offered would not necessarily
answer the question in the same way as other respondents forced to choose sides on an issue.
Moreover the same likert response scale was used for all items. According to Rea and Parker (1997), this
ensures that the numbers have the same meaning across the items.
The characteristic that distinguishes this technique from other traditional statistical
procedures is the relative ease with which new insights can be gained. Moreover, this
technique has the ability to handle heterogeneous data, not just figures on which it is
necessary to apply strong assumptions on their distribution, but larger data sets, with
no or few assumptions about it.

Specifically, the technology used is called Classification and Regression Trees

(CART), a non-parametric methodology that communicates with the user by means of
increasing the figure to ease the interpretation of results. This technique, introduced
by Leo Breiman, Jerome Friedman, Richard Olshen and Charles Stone in 1984, shows
the results as a reversed tree diagram and has proved a very flexible. The methods of
classification and regression tree-based on a non-parametric approach, have proven,
since their introduction, a useful analytical process of knowledge discovery and
supervised learning based on collected data.

In the present work it was decided to use that methodology in one of the techniques of
sample stratification, since it does not require any assumption on the target variable
and even with regard to relations between the variables in question would require
specific assumptions. All this makes the technique useful to assess the constructs
proposed by previous models without taking into account the relationships between
variables in the same proposals. In addition, the algorithm is able to automatically
select the most significant variables.

Further, the tree structure allows the output to CART to treat large data sets and very
complex data structures, which permits the use of continuous independent variables,
categorical or continuous and categorical together. The CART is, therefore, a
procedure that can be used to analyze data both qualitative (classification) and
quantitative (regression). Being a non-parametric procedure is not necessary to excuse
a normal or other assumptions regarding the statistical distribution of data. The CART
can also prosecute cases with missing data.

The advantages of this technique are also due to the ease of interpretation of results
(interpretability) with the same accuracy (accuracy). The trees, when used in the
particular case of complex data with many independent variables, may provide an
error rate significantly lower than that produced by the usual parametric procedures
such as discriminant analysis or logistic regression.
4. Results

Starting from these considerations, the aim in this study was to create a hierarchical
segmentation of the population using the above technique of data mining, the CART.
The terminal nodes, or the tree’s final determinations, contained sets of observations
forming very homogeneous, but different, classes.
In this way the sampling is carried out according to objective, in a sense, foresight:
the variables by which it is first segmented into clusters are, in fact, explanatory of the
phenomenon in question.

The final tree obtained by analysis includes only those independent variables that are
predictive of the dependent variable. The other independent predictor variables did
not affect the final result. Even in this respect the CART differs from other traditional
statistical procedures such as regression.

Data analysis was conducted twice, first considering only the constructs, in order to
support the hypothesis and the second including all the items that make up the
constructs. This secondary analysis was conducted in order to obtain more practical
information on the relevant features for the online customer.

In the first case (Fig. 2) the tree considers the dichotomous dependent variable and
ordinal on the intention to purchase (the likelihood that future purchases will be made
on the same site) and constructs previously described. The variables were considered
as quantitative, so the end nodes represent instead of the classes of individuals, the
expected values, i.e. the average of the dependent variables (intention to buy).

Firstly, it seems worthwhile to note how the tree has generated interesting results and,
already on the first split of the first variable, divides rather sharply the respondents
have a strong propensity to buy (final leaf value nearest to 1) and the respondents with
a lower propensity to purchase (final leaf value nearest to 8).

It is useful, at this point, to comment most significant leaves that showed values
closer to unity (strongly agree) and therefore relate to those who are more willing to
purchase and on the opposite leaves that have generated values closer to six (strongly
disagree) and representing individuals less willing to purchase online.

The first leaf presents a value of 1.195 (average purchase intent of responses), the
subject strongly agrees with the intention to purchase online. These subjects also
strongly agree (1) or agree (2) to the perception of playfulness of the reference site,
they found the site compelling and impacting on an emotional level. In addition they
strongly agree (1) with the e-trust perception. That is to say that users of a site deemed
safe and playful show a very high probability of making an online purchase.

The second interesting leaf presented an average value of 1.300. Again this leaf
represents the respondents who have high propensity to shop online on the site. In this
case the playfulness variable appears twice, people who perceive the site as very
playful (1), even if the perception e-trust ranged from agree to strongly disagree (2 to
6) have a strong propensity to buy. Therefore, those respondents strongly agree with
the playfulness of the site regardless of how they perceive the security of the site.
Instead, considering the leaves that have higher values and hence corresponding to the
lower propensity to purchase, it is possible to evaluate the predictors of poor intention
to acquire. Note that these leaves are both on the right side because, as stated above,
the first predictor variable (the playfulness) is already able to clearly separate the
cluster with higher and lower propensity to buy.

The right branch also has more nodes because it concerns subjects who assigned to
the playfulness construct a 3,4,5,6 level of response (from partially agree to strongly
disagree) and takes into account other variables. The third interesting leaf has a value
of 4.833 on average, and includes those with a very low propensity to purchase
(disagreement with the intention to purchase). These people reported a level of
perception of the playfulness of more than 2.5 on the scale (from partially agree to
strongly disagree), of the usefulness of more than 3.5 (therefore from tend to
disagree to strongly disagree) and a level of perceived e-trust of more than 3.5
(therefore from tend to disagree to strongly disagree). In this leaf it is possible to find
respondents who considered the site not very playful, useless and unsafe and are less
likely to purchase on the site.

The fourth leaf has an average level of 3.600 therefore it is still in the low propensity
to purchase area, although it is higher than the previous case. The variables that lead
to this condition are the same of the previous leaf commented, with a higher
perception of the usefulness of the site, above 2.5 (therefore from agree to strongly
disagree). Respondents reported perception of the playfulness at more than 2.5 (from
partially agree to strongly disagree), they quite agree with the utility (3) but just like
before they consider the site not safe (4, 5, 6). Perhaps these people have a low
propensity to buy assuming it to be useful but not safe.

Consequently, the variables that affect most negatively on the intention of buying are:
• the lack of playfulness;
• the lack of perceived usefulness;
• the lack of e-trust in the site.

The survey results confirm that, in addition to utilitarian aspects of online use, the
perception of the usefulness of the system, very important seem to be the hedonistic
aspects related to the higher or lower emotional commitment. In the first instance, the
online purchase intention is influenced by the playfulness construct, followed by the
perceived security and perceived usefulness. The same constructs were also repeated
in subsequent nodes with values of more restrictive response, which confirms their
importance.

Figure 2 - with an official tree and only the constructs

-- Insert Figure 2 Here --

Source: Research Data

The constructs that affect the intention to purchase are:

- playfulness (seen two times);

- e-trust (seen several times);
- perceived usefulness (seen two times).

The hypothesis H7 (playfulness), H4 (perceived usefulness), H8 (e-trust) seem to be

supported while the others refer to variables that are not so relevant.

Note in particular that the perceived ease of use, as in other studies previously cited,
seems to be a factor not influencing the level of use of electronic commerce.

The second tree (Figure 3) considers the same ordinal dependent variable that is the
intention to purchase. The variables are considered as quantitative. In this case, the
items of the survey were used without considering the constructs. The width of the set
being compared to traditional stores can be seen as the most significant variable in
explanations of the phenomenon of intent to purchase a product or service online.

Figure 3 - Shaft with ordinal dependent and only the items

-- Insert Figure 3 Here --

Source: Research Data

The items that mostly affect the intention to purchase are:

- the width of the set being compared to traditional stores (seen two times);
- security of payments (seen several times);
- the convenience of prices;
- the cost of shipping;
- the preference of the specific site in relation to others;
- the frequency of purchases made online.

5. Discussion and implications

Still, the results to date offer some interesting food for thought for managers as well
as academics involved in Web sites development. They should help to understand
user behavior in Web-based online retailing. The results showed that playfulness
plays an important role in enhancing user behavioral intention to buy in a site. This
study presumes that websites with a playful potential can influence online consumers’
shopping motivations and entice them to modify their original shopping
predispositions by providing them with attractive and enhanced interactive features
and controls, thus generating a positive attitude towards products and services offered
by the retailer. Consequently, consumer choice in taking part in e-commerce certainly
has an emotional determinant.

Moreover, perceived usefulness related to the online price policies and online product
policies, the perceived safety, and therefore the trust in online commerce sites, seems
to correlate with a higher purchase intention than the elements of usability or ease of
use in the development of an e-commerce marketing project and web design. These
should also be taken into account, but probably.

E-commerce success determinants are used by researchers who wish to build models
and enhance our deep understanding of the factors that influence the variables. It is
thus imperative for technology acceptance success models to be both valid and
precise enough to be useful for developing practical guidelines.

From a practical standpoint, systems that are judged successful become a part of the
organizational decision-making framework and or part of the organization's system
structure and processes. As the competition in e-commerce is intensified, it becomes
more important for online retailers to understand the antecedents of consumer
acceptance of online shopping. Because of this, it is already evident that ecommerce
sites, even at the functional and graphic design stage, should consider elements that
make it possible to increase the potential for participation, engagement, interaction
and fun. In conclusion, the big importance given to the perceived usefulness suggests
that the marketing managers plan carefully their policies of price and product on the
Net.

6. Limitations and future research lines

Of course an inherent limitation of this research lies in the fact that the current study
has been restricted only to the users of the participating sites. This means that users
had likely already developed a preference for the site they used, even if the data on
past purchases on the site itself is not too high. While we have mainly focused on
consumer factors in online shopping research, there are other factors related to the
product/service system, and to the vendor that could be important determinants of
consumer acceptance of online shopping.
If the results suggested here are confirmed by further experiments, new models can be
constructed to explain the choices made by individuals regarding ecommerce. We
could, in fact, study the relationships between variables, omitting those which have
been eliminated as a result of this analysis. There is much more to be learned about
consumer online shopping acceptance through rigorous empirical studies. In the
future, the same type of study could also be conducted and the differences observed
when different types of online products and services are offered. Indeed, we believe
that the variables used will lead to different results in different areas. In addition, a
meta analysis to handle similar and contradictory results of surveyed studies could be
carried out in future.

Finally, it is of interest to conduct longitudinal studies to examine online buying

behavior as users of online shopping technology progress from "beginners" to the
more experienced. For example, Cheung et al. (2005) suggested a progression of
sophistication: intention to buy on the Internet, adoption (first Internet purchase), and
continuance (repeat Internet purchase). Moreover it would be preferable to conduct
the study using actual online purchases as opposed to self reported intention to buy.

Bibliography

A.H. Segars, V. Grover (1993). “Re-examining perceived ease of use and usefulness:
a confirmatory factor analysis”, MIS Quarterly, Vol. 17 No. 4, pp. 325-343.

Adams, D. A., Nelson, R. R., & Todd, P. A. (1992). “Perceived usefulness, ease of
use, and usage of information technology: A replication” MIS Quarterly, Vol. 16:
227-247.

Adamson, I. Shine J. (2003). “Extending the new technology acceptance model to

measure the end user information systems satisfaction in a mandatory environment: A
bank’s treasury” Technology Analysis and Strategic Management, Vol. 15 No.4, pp.
441-55.

Ajzen, I. (1985). “From intentions to actions: A theory of planned behavior” in Khul

J., Beckmann J. (Eds.), Action control: Frem cognition to behavior, Springer –
Verlag, New York, pp. 11-39.

Ajzen, I., Fishbein, M. (1980). Understanding Attitudes Predicting Social Behavior,

Prentice-Hall, Englewood Cliffs, NJ.

Ajzen, I., Madden T.J. (1986). “Prediction of goal-directed behavior: attitudes,

intentions, and perceived behavioral control” Journal of experimental Social
Psychology, Vol. 22, pp. 453-474.
Bagozzi, R. P. (2007). “The legacy of the technology acceptance model and a
proposal for a paradigm shift”, Journal of the Association for Information Systems,
Vol. 8 No 4, pp. 244-254.

Bishop G.F. (1987). “Experiments with the Middle Response Alternative in Survey
Questions”, Public Opinion Quarterly, 1987, 51, pp. 220-232.

Biswas, A., Krishnan, R. (2004). “The Internet’s impact on marketing” Special

edition of JBR Marketing on the Web – behavioral, strategy, and practices and public
policy, Journal of Business Research, Vol. 57, pp.681-684.

Breiman L., Friedman J., Olshen R. e Stone C. (1984). Classification and Regression
Trees, Chapman & Hall, New York - London.

Bruner G.C., Kumar A. (2005). “Explaining consumer acceptance of handheld

Internet devices”, Journal of Business Research, Vol. 58 No. 5, pp.553-558.

Chau, P. Y. K., and Hu, P. J. (2001). "Information Technology Acceptance by

Individual Professionals: A Model Comparison Approach," Decision Sciences, Vol.
32 No. 4, pp. 699-719.

Chen K., Tarn, J.M., Han, B.T. (2004). "Internet dependency: its impact on online
behavioral patterns in e-Commerce", Human Systems Management, Vol. 23 No.1, pp.
49-58.

Chen L.D., Gillenson M.L., Sherrell D.L. (2002). “Enticing online consumers: an
extended technology acceptance perspective”, Information and Management, Vol. 39
No. 8, pp. 705-719.

Chen, L. Tan, J. (2004). “Technology adaptation in e-commerce: key determinants of

virtual stores acceptance”, European Management Journal Vol. 22 No.1, pp.74-86.

Davis F. D. (1989). Perceived Usefulness, “Perceived Ease of Use, and User

Acceptance of Information Technology”, MIS Quarterly, Vol. 13 No. 3, pp. 319-339.

Davis F. D., Bagozzi, R. P. e Warshaw, P. R. (1989). “User Acceptance of Computer

Technology: A Comparison of Two Theoretical Models”, Management Science, Vol.
35, No. 8, pp. 982-1002.

Davis F. D., Bagozzi, R. P. e Warshaw, P. R. (1992). “Extrinsic and intrinsic

motivation to use computers in the workplace”, Journal of Applied Social Psychology,
Vol. 22, pp. 1111-1132-
Fishbein M., Ajzen, I. (1975). Belief, attitude, intention, and behavior: An
introduction to theory and research. Reading, Mass.; Don Mills, Ontario: Addison-
Wesley Pub. Co.

Gefen D., Straub D. (2003). “Managing user trust in BtoC e-services”, E-service
Journal, Vol. 2 No2, pp. 263-288.

Hendrickson, A.R., Massey, P.D., & Cronan, T.P. (1993). “On the test-retest
reliability of perceived ease of use scales”, MIS Quarterly, Vol. 17, pp. 227-230.

Hoffman D.L., Novak T.P, Peralta M.A. (1999). “Building consumer trust online”
Communications of the ACM, Vol. 42 No. 4, pp- 80-85.

Hsu, C.L., Lu H.P. (2004). “Why do people play on-line games? An extended TAM
with social influences and flow experience”, Information & Management, Vol. 41 No.
7, pp. 853-68.

Hu, P., Chau P., Sheng O., and Tam K. (1999). “Examining the technology
acceptance model using physician acceptance of telemedicine technology”, Journal of
Management Information Systems Vol. 16 No. 2, pp. 91-113.

Jarvenpaa S.L., Todd P.A. (1997). “Consumer reactions to electronic shopping on the
World Wide Web”, International Journal of Electronic Commerce Vol. 2 No. 1.

Jarvenpaa S.L., Tractinsky N., Vitale M. (2000). “Consumer trust in an Internet store”
Information Technology and Management, Vol. 112, pp. 45-71.

Kalton G. and Schuman H., (1982). “The effect of the question on survey responses:
A review”, Journal of the Royal Statistical Society, A 145, pp. 42-73.

Kalton G., Roberts J., and Holt D. (1980). “The effects of offering a middle response.
Option with opinion questions”, The Statistician, 29, pp. 65-79.

King, W. R. ,He, J. (2006). “A meta-analysis of the technology acceptance model”

Information and Management, Vol. 43, pp. 740-755.

Koufaris (2002). “Applying the technology acceptance model and flow theory
consumer behavior”, Information Systems Research, Vol. 13 No. 2, pp.205-223.

Lee M.K.O., Cheung C.M.K. and Chen Z. (2005). “Acceptance of Internet-based

learning medium: the role of extrinsic and intrinsic motivation”, Information
Management, Vol. 42 No. 8, pp. 1095-1104.
Legris, P., Ingham, J., and Collerette, P. (2003). “Why do people use information
technology? A critical review of the technology acceptance model”, Information
Management, Vol. 40 No. 3, pp.191-204.

Li D., Lou H. , Day J. and Coombs G. (2004). “The effect of affiliation motivation on
the intention to use groupware in an MBA program”, Journal of Computer
Information Systems, Vol. 44 No.3, pp. 1-8.

Mathieson K. (1991). “Predicting User Intentions: Comparing the Technology

Acceptance Model with the Theory of Planned Behavior”, Information Systems
Research, Vol. 2 No. 3 p. 173.

Min H., Galle W. (2003). “E-shopping: profiles of adopters and non-adopters”,

Industrial Marketing Management, Vol. 32 No 3, pp.227-233.

O’Cass A., Fenech T. (2003). “Web retailing adoption: exploring the nature of
internet users Web retailing behavior”, Journal of Retailing and Consumer Services,
Vol. 10: pp.81-94.

Pavlou P.A. (2003). “Consumer acceptance of electronic E-commerce: integrating

trust and risk with the technology acceptance model”, International Journal of
Electronic Commerce, Vol. 73, pp. 69-103.

Pontiggia, A. (2001). I modelli di impiego, in Pontiggia, A. (a cura di): Impiego

efficiente delle tecnologie dell’informazione. Apprendimento e cambiamento
organizzativo. EGEA, Milano.

Rea and Parker (1997). Designing and Conducting Survey Research: A

Comprehensive Guide, Jossey-Bass Publishers, San Francisco, CA.

Report (2008). School of Management, Politecnico di Milano.

Scarpi D., Dall'Olmo Riley F., Manaresi A. (2006). “Le ricerche sui comportamenti di
acquisto on line: L'orientamento edonico e utilitarista”, Mercati e competitività, Vol.
3.

Shang R.A., Chen Y.C. Shen L. (2005). “Extrinsic versus intrinsic motivations for
consumers to shop online” Information Management, Vol. 42 No. 3, pp.401-413.

Shin, D.H. (2007). “User acceptance of mobile Internet: implication for convergence
technologies” Interacting with Computers, Vol. 19, pp.472-483.

Szajna B. (1994). “Software evaluation and choice: predictive evaluation of the

Technology Acceptance Instrument” MIS Quarterly, Vol. 18 No. 3, pp.319-324.
Taylor S., Todd P.A. (1995). “Understanding information technology usage: a test of
competing models” Information Systems Research, Vol. 6 No. 2, pp.144-176.

Townsend A.M., DeMarie S.M., and Hendrickson A.R. (2001), “Desktop Video
Conferencing in Virtual Workgroups: Anticipation, System Evaluation, and
Performance” Information Systems Journal. Vol. 11, pp.213-227.

Van der Heijden He Verhagen T., (2004). Online store image: conceptual foundations
and empirical measurement: Information Management Vol. 41 No. 5: pp.609-17.

Venkatesh V., Morris M. G., Davis G. B. e Davis F. D. ( 2003). ”User Acceptance of

Information Technology: Towards a Unified View” MIS Quarterly, Vol. 27 No. 3,
pp.425-478.

Venkatesh V., Davis F. D., (2000). “A theoretical extension of the technology

acceptance model: four longitudinal field studies”, Management Sciences, Vol. 46
No. 2, pp. 186-204.

Vijayasarathy L.R. (2004). “Predicting Consumer Intentions to Use Online Shopping:

The Case for an Augmented Technology Acceptance Model”, Information
Management, Vol. 41 No. 6, pp. 747-762.

Weiner E., Brown A. (1995). “The new marketplace”, The Futurist, Vol. 29 No. 3,
pp.12-16.
Figure 1 - The causal relationships between constructs

Facilitating conditions

Perceived self efficacy

Experience

Perceived usefulness
Intention to buy
online
Perceived ease of use

Subjective norms

Playfulness

E-trust
Figure 2 - with an official tree and only the constructs

Construct: Playfulness <2,5

Construct: Perceived
Construct: E-Trust < 1,5 usefulness < 3,5

1,195 Construct:
Playfulness < 1,5
Construct: E-Trust < 3.5

1,300 1,920
3,430 4,833

Construct: Perceived
usefulness < 2.5

2,135 Construct: E-Trust < 3.5

2,615 3,600

Source: Research Data

Figure 3 - Shaft with ordinal dependent and only the items

The width of the set being compared to

traditional stores < 2,5

Security of payments < 1,5 The convenience of prices < 3,5

Security of payments < 2,5

1,192 1,887
The width of the set being
2,041 compared to traditional stores <1,5

1,944 2,772

The cost of shipping < 4,5

2,524
The preference of the specific site in
relation to others< 4.5

The frequency of purchases 4,778

made online < 3,5

2,682
Security of payments < 3.5

3,311 4,429

Source: Research Data

Table 1 - The profile of respondents

Gender Women 44.90

Men 55.10
Age Less than 30 years 23.20
Between 30 and 49
54.90
years
Over 50 years 21.90
Level of training elementary-school 6.90
High school 54.70
University 27.28
Post-graduate
10.19
training
Employment status Student 10.83
Employee 38.43
Manager 5.20
Freelancer 14.83
Entrepreneur 7.22
Military 1.80
Pensioner 5.94
Unemployed 4.14
Other 11.25

Source: Research Data

Autobiographical note

Michelle Bonera
Assistant Professor at the Business Economics Department
of Università degli Studi di Brescia
Contrada Santa Chiara, 50 - 25122 Brescia
mbonera@eco.unibs.it
Mobile phone: +393332341484
Tel. 030/2988.551-2-3-4
Fax 030/295814

At present she is an Assistant Professor framed in the scientific group of Business Administration at the “Università
degli Studi di Brescia” and she holds the chair of “Tourism Marketing” and “Competitive and market analysis”.
After her Ph. D. in Business Administration at the Brescia University, she has begun her activity at the same University.
Here she has collaborated to the teachings of “Marketing”, “Communication Management” and “Communication Mix”.
Her research activity focuses on Web marketing and Tourism Marketing.