Acceptance models of enterprise resource planning systems


Nov 9, 2013 (4 years and 8 months ago)


Acceptance models of enterprise resource planning systems

Levi Shaul
Information Systems Research Department at the Bar
Ilan University (Corresponding author).

Doron Tauber
, Information Systems Research Department at the Bar
Ilan University.


Levi Shaul

Bar Ilan University

Ramat Gan 52900



Fax: 972



Dr. Doron Tauber

Bar Ilan University

Ramat Gan 52900


Phone: 972

Fax: 972


Acceptance models of enterprise resource planning systems


Organizations perceive Enterprise Resource Planning (ERP) as a vital tool for organizational competition as it integrates
dispersed organizational systems, and enables flawless transactions

and production. The importance of ERP systems has
been increasingly recognized by organizations of all kinds. Nevertheless the implementation of such systems has proved
to be difficult, in that it demands considerable resources for long periods of time. T
his study has several goals: (1) review
the literature on information systems acceptance models in terms of prospective individual adoption, (2) empirically
compare eight prominent models and their extensions to ERP systems, (3) examine the relationships a
mong fundamental
constructors, (4) examine the effect of moderators on these relationships including age, gender and experience and (5)
formulate a model that integrates elements across these eight models and best describes the acceptance of ERP systems.

: Enterprise Resource Planning, ERP, Acceptance Models, Moderating variables, Information systems.



Organizations consider ERP to be its backbone and a vital tool for organizational excellence because it integrates
varied organizational systems, and enables flawless transactions and production (Al
Mashari et al. 2003,
Koh et al.

et al. 2007).An ERP system can reduce costs, and thus lead to greater effectiveness and a better
competitive edge in terms of improved strategic initiatives and responsiveness to cus
tomers (O'Leary 2000, Sandoe et al.
2001, Rashid et al. 2002, Bharadwaj et al. 2007
, Ge & VoB 2009
). Enterprise system software constitutes a multi
dollar industry that produces components to support a variety of business functions (Chellappa

af 2010). IT
investments have grown to be the largest category of capital expenditures in United States
based businesses over the past
decade (Ranganathan

Brown 2006). Implementing an ERP system is different from implementing a traditional
software devel
opment system since it is not “built to order” but rather bought “as is”, and is transaction driven rather than
centric in its focus, with different levels of adaptability (Basu
Kumar 2002). Although ERP has been depicted
as a panacea in both th
e literature and in practice, there are many reports of difficulties in implementing ERP systems

(Ram et al. 2013)
. Chang (2004) reported that (a) 90% of ERP implementations are delivered late or are over
(b) enterprise initiatives show a 67% fai
l rate in achieving corporate goals and outcomes are considered negative or
unsuccessful, (c) more than 40% of all large
scale projects fail. Furthermore, ERP projects also fail because of errors in
managing leadership (42%), organizational and cultural (2
7%), human and people (23%), technology and other
dimensions (8%) (Waters 2006).

This study has several goals: (1) review the literature on information systems acceptance models in terms of
prospective individual adoption, (2) empirically compare eight pr
ominent models and their extensions in the field of ERP
systems (Table 1), (3) examine the relationships among fundamental constructors, (4) examine the effect of moderators
on these relationships including age, gender and experience and (5) formulate a mo
del that integrates elements across
these eight models that best captures the steps toward acceptance of ERP systems.






Technology acceptance model

Davis, 1989



a revised model of TAM

Venkatesh & Davis, 2000



Unified theory of acceptance and use of technology

Venkatesh et al., 2003



Task technology fit model

Goodhue & Thompson, 1995



愠combin敤 mod敬

䑩ahaw C p瑲ongI 1999



䑩afus楯n of fnnov慴楯n mod敬

䵯or攠C B敮b慳慴Ⱐ 1991

Table 1

eight prominent technology acceptance models


Literature review

Along with increasing investments in new technologies, their acceptance has become a frequently studied topic in the
field of information systems. In the last two decades acceptance models have been proposed, tested, refined, extended
and unified. Previous

studies have presented a variety of theoretical models to support successful ERP adoption and
implementation (Calisir

Calisir, 2004). Studies on acceptance in the field of information systems reflect two
mainstreams of research (Venkatesh et al. 2003).

Each of these which has made an important and unique contribution to
the literature, although as noted by Lin et al. (2007) most empirical studies of technology acceptance models have been
limited to the technology acceptance
related issues of individua
l users.

One stream examines the individual psychological characteristics that influence technology acceptance, and use
intention or usage as a dependent variable (Compeau

Higgins 1995b; Davis et al. 1989). This type of approach is valid
for almost any

technology. Although developed within the IS field, it nevertheless does not consider the specific
characteristics of software and makes no distinction between software, hardware and services of the IT departments

& McLean

2003). Thus although



perceptions are

differentiated the












are included (Bhattacherjee

Sanford 2006).

The second
stream examines implementation success through the fit of the

technology either overall in terms of its technological
characteristics or at the organizational level (Goodhue

Thompson 1995, Autry et al. 2010). This stream explicitly
considers the attributes of information and systems which produce information such

as data









and reliability (Moore

Benbasat 1991, Delone &

1992, 2003).

Among the theoretical models within the first stream, the technology acceptance model (TAM) developed by Davi
(1989) appears to be the most widely used by technology researchers and managers because
of its empirical support (
et al.
. The TAM model draws on the theory of reasoned action (TRA) developed by Fishbein and

Ajzen (1975)

and is based on the h
ypothesis that technology acceptance and use can be explained in terms of the individual's internal
and perceived beliefs of technology usefulness, ease of use and intentions (Davis 1989). The TAM model can be applied
to predict future technology use by ex
amining data from the time that the technology was introduced. The TAM has
given rise to two subsequent models. TAM2, developed by Venkatesh
Davis (2000) preserves the core philosophy of
the model but incorporates additional theoretical constructs spanni
ng social influence processes to reflect the impact on
an individual deciding to adopt or reject a new system. The UTAUT model refines how the determinants of intention and
behavior evolve over time and emphasizes that most of the key relationships in the
model are moderated (e.g. age,
gender, experience) to respond to the interest in workplace environments to create equitable settings for women and men
of all ages (Venkatesh et al. 2003).

Higgins (1995b) extended one of the most influential theories of human behavior, Social Cognitive
Theory (SCT) to the context of technology utilization. SCT, developed by Bandura (1986) defines human behavior as an
interaction of personal factors, behavior

and the environment. SCT posits that learning will most likely occur if there is a
close identification between the observer and the model (i.e. the individual who is imitated) and if the observer also has a
good deal of self
efficacy. Bandura (1986) argu
ed that an individual's self
efficacy beliefs affect behavior and function as
an important set of proximal determinants of human motivation and action which operate on action through affective
intervening processes. These include motivational process (peop
le are more likely to expend more effort and persist


Computer self

efficacy model

Compeau & Higgins 1995




䑥汯n攠and 䵣L敡n I匠獵捣ess mod敬

䑥汯n攠 & 䵣L敡n, 2003

longer in a task) and cognitive process (people are more likely to take a wider picture of a task and be encouraged by
obstacles to greater effort when performing the task).

Several models draw on const
ructs from both streams of research. Diffusion of Innovation (DOI) theory views
innovation as communicated through certain channels over time and within a particular social system (Rogers, 1995).
The rate of adoption of innovations is influenced by five fa
ctors: relative advantage (i.e. usefulness), complexity (i.e. ease
of use), compatibility, trainability and observability (Rogers, 1995).

Benbasat (1991), working in an IS context,
expanded on the Rogers' factors to generate eight factors: voluntari
ness, relative advantage, compatibility, image, ease of
use, result demonstrability, visibility and trialability which all impact the adoption of IT.

Since the early applications of
DOI to IS research, the theory has been applied and adapted in numerous w

However, research has consistently
found that technical compatibility, technical complexity, and relative advantage are important antecedents to the adoption
of innovations (Bradford
Florin, 2003; Crum et. al., 1996) all of which have led

to a gene
ralized and simpler model.

Strong (1998) adapted key models of information technology (IT) utilization behavior from the MIS
literature (TAM and TTF models) to suggest a combined model that delivers more explanatory power than either model
. The result is an extension of TAM to include a Task
technology fit (TTF) construct. Models that integrate
constructs from both streams of research have greater explanatory power. They argued that research using the integrated
models should lead to a bet
ter understanding of choices concerning the use of IT. Each of these combined models
provides a much needed theoretical basis for exploring the factors that explain software utilization and its links with user

& McLean

(1992) defined f
our antecedents of user acceptance and organizational benefits: system quality,
information quality, user satisfaction and user intention to use the technology. DeLone
& McLean

(2003) suggested that
use and intention to use are alternatives in their model,

and that intention to use may be worthwhile in the context of
mandatory usage such as ERP systems. Most researchers agree with DeLone &
’s (2003) argument that service
quality, when properly measured, should be added to system quality and informatio
n quality as predictors of user
satisfaction and user intention to use the technology (Wang
Liao 2006).

These models have contributed to our understanding of user technology acceptance factors and their relationships.
The acceptance models in the field o
f information systems are based on different (and partially overlapping) sets of
dependent and independent constructs. Nevertheless they also present two limitations: their relatively low explanatory
power and inconsistent influences of the factors across
studies (Sun
Zhang 2006).



Beyond the empirical comparison of these known acceptance models as described above, this research also aims to
explore the effect of key individual user differences on the main relationships among core constructs.
& Prasad

explored the effect of individual user differences on technology acceptance. They found that each of these
moderators was fully mediated by core constructs, implying that simpler models could be constructed that exclude

differences. However, different studies have shown that core constructs do not fully mediate the effects of key
individual user differences (Burton
Hubona 2006; Venkatesh et al. 2003; Morris
Venkatesh 2000; Karahanna
et al. 1999; Taylor
& Todd

995a). Burton
Hubona (2006) found consistent proof of relationships between
users’ characteristics and IT in the literature

They argued that there are several justifications for key individual user
differences including the fact that older users t
end to resist change and may be less able to appreciate or understand it.
They therefore perceive new IT as less useful, and find it more difficult to learn and use unfamiliar technology even if
they are willing to adopt a new IT. In addition, these author
s' view most user behavior as non
cognitive and claim that
core constructs cannot fully mediate individual differences associated with user habits.

everal key individual user differences have been found to be significant in acceptance models in the c
ontext of
information systems. This study incorporates: age, gender and experience, the three best documented individual user
differences to examine the key relationships among fundamental constructors system in both mandatory and voluntary
settings (Yi et

al. 2006, Burton
Jones & Hubona

2005, Morris
& Venkatesh

2000). It deliberately neglects other
individual user differences because of either irrelevance to the field of ERP systems (i.e. voluntariness, since ERP is
perceived to be associated with mandato
ry usage) or inconsistent findings the field of information systems (i.e. level of

Most of the models investigated in this study aim to measure potential user's attitudes toward adopting an information
technology (Moore
Benbasat 1991, Davis 1
989, Venkatesh

Davis 2000, Dishaw
Strong 1998, Venkatesh et al.
2003). Therefore, intention to use an information technology is a prominent dependent variable in most models.
However, the CSE model developed by Compeau
Higgins (1995b), used actual us
age as a dependent variable. Here
we examine the predictive validity of all models in the context of intention to enable a comparison of the models.
However, the intention construct in many technology acceptance studies has been measured via voluntary orie
statements of usage such as "I intend", "I plan'' or "I predict". Nah et al. (2004) claimed that these measures are
inappropriate to assess acceptance of mandatory technologies such as ERP systems.
Chang et al. (2008) argue that
lthough the use of
ERP systems

may not be voluntary, the understanding of system adoption from the user’s perspective
is useful in helping

the organizations prepare their employees to face new challenges and learn how to make good use of

Seymour et al. (2007)

suggested that this dependent variable should be redubbed the 'symbolic adoption'
variable, to describe potential adopters' mental acceptance of mandatory information technology in a better way. Based
on these the models and literature review, a number of

hypotheses were formulated to identify antecedents of symbolic
adoption (Table 2). These hypotheses are refined to include the moderating variables that have been acknowledged as
having an effect on the relationships between the independent variables and
symbolic adoption.


Research methodology


Data Collection

The authors developed eight structured questionnaires, one for each model. The instruments were adapted from
measures developed throughout the model development and from instruments validated in
previous quantitative studies
of a similar nature as listed in Table 3.



Source for validated instruments



Davis 1989; Davis et al. 1989



Venkatesh & Davis 2000



Venkatesh et al. 2003



Goodhue & Thompson 1995



Dishaw & Strong 1999, Goodhue & Thompson 1995



Moore & Benbasat 1991




Higgins 1995b, Compeau et al. 1999




McLean 2003, Iivari 2005, Ifinedo
Nahar 2007

Table 3

Source for validated instruments

questionnaire consisted of two components. The first component was demographic questions about the
respondents and the extent to which they used the ERP system. This questionnaire was administered at a certain point in
time and therefore a question on prio
r experience in ERP systems was added to enable an analysis of the impact of
experience on adoption. The second component consisted of the items measuring the core constructs that were defined in
the models. A five point Likert
type scale was used where 1=
strongly disagree to 5=strongly agree. The full
questionnaires are not shown due to space constraints.

Each questionnaire was referred by approximately 100
respondents. The questionnaires were mailed, from September 2010 to December 2011 and returned by ap
800 respondents (eight questionnaires in overall

one for each model) in the Mediterranean region working in SMEs in
which an ERP system was implemented.

Several constructs are common across models. For example, previous studies have indicate
d that performance
expectancy (defined in the UTAUT and CSE model) and relative advantage (defined in the DOI model) constructs are
similar (Compeau
Higgins 1995b

Davis et al. 1989, Moore
& Benbasat 1991,
Plouffe et al. 2001, Venkatesh et al.
2003). The
refore, to enhance the explanatory power of the following analyses, constructs that were common across
models were measured in the same manner to enlarge the data sample. Thus, for example, the analysis of the TAM model
that was returned by approximately 1
00 respondents could be measured on a sample size of approximately 500
respondents because the 'perceived usefulness' and 'perceive ease of use' constructs are common across five models


Reliability analysis

A reliability analysis determines the extent to which the measurements resulting from an analysis are the result of
characteristics of the features being measured. A reliability analysis also evaluates the internal consistency of the
measurement items grou
ped under the core constructs in the models. In most cases and in this research as well, the
available variables were only the observed variables and therefore this method is purely theoretical. As a result, we used
an internal consistency method that is c
losely associated with reliability analysis and enables an empirical analysis of
measurement reliability.

Internal consistency was measured by Cronbach’s Alpha. High communality values for all sub factors indicate that
the total amount of variance that an

original factor shares with all other factors is high. Hair et al. (1995) indicated that
the lowest acceptable value ranges between 0.60 and 0.70 whereas Nunnally (1978) and Fornell
Larcker (1981)
recommended a Cronbach's Alpha limit of 0.70 for reasona
bly high reliability.

The measurement model estimations for the models, based on the internal consistency reliability (ICR) analysis,
showed similar internal consistency values, means and standard deviations for both the entire questionnaire and the set o
reduced measurement items. In addition, the square roots of the shared variance between the constructs and their
measurement items were higher than the correlations across constructs, supporting convergent and discriminant validity.
The results of the me
asurement model estimations for both cases are not shown here due to space considerations.


Multicollinearity analysis

Unlike reflective measurement items where multicollinearity between construct items is desirable as illustrated by a
high Cronbach’s alpha

or internal consistency scores, excessive multicollinearity in formative constructs can destabilize
the model. If measures are highly correlated, it may suggest that multiple indicators are tapping into the same aspect of
the construct (Diamantopoulos
iguaw 2006). Therefore, to ensure that multicollinearity was not present,
multicollinearity analysis was performed using the variance inflation factor statistic (VIF). Although general statistics
theory posits that multicollinearity occurs if the VIF value

is higher than 10, the authors tested multicollinearity for a
strict VIF threshold of 3.3 out of model destabilization considerations (Diamantopoulos
Siguaw 2006).


Hierarchical regression

Cronbach (1987) suggests that interaction effects should be evalu
ated by stepwise hierarchical regression. Prior to the
hierarchical regression an additive transformation on the predictor variables should be performed. The transformation for
a given predictor involves subtracting the mean of the predictor variable from
each individual's raw score on that
predictor, thus forming deviation scores. To eliminate the effect of multicollinearity of variables, the interaction term
was formed by multiplying the two centered variables together (Aiken
West 1991). Thus, such a t
ransformation will
yield low correlations between the product term and the component parts of the term. This is desirable, because it
decreases the probability of computational errors (Jaccard et al. 1990).

In the first step, we entered the independent
riables into the regression model to verify the main effects of the independent variables. Then, in a separate step, the
product of the independent variables, which represents the moderator effect, was entered. This stepwise hierarchical
approach provides
an unambiguous test of moderator effects (Aiken
West 1991). Furthermore, to determine the nature
of this interaction, we performed a simple slopes analysis (Aiken & West 1991).

Past studies have used this technique for
determining the influence of potent
ial moderator variables (Stone & Hollenbeck 1989)



The variance explained by the models, without the inclusion of the moderating variables, was relatively modest, as
presented in Table 12. In addition, the variance explained by the models after th
e inclusion of the moderating variables
increased across all models. However, the variance explained by the models, in the field of ERP systems, in an absolute
manner, even after the inclusion of moderating variables, increased only slightly and at best o
nly accounts for 41% of the
variance. The models show a 29% increase in explained variance (on average) whereas the CSE model shows the
highest percentage of increase in explained variance after including the moderating variables (45%) but nevertheless
ows the least explained variance in both cases (before and after the inclusion of moderating variables

15% and 21%
respectively). The D&M model does not include the influence of any moderating variables and therefore was analyzed
for the influence of core

constructs alone.




% change




% change




































Table 4

Variance explained by the models before and after including moderating variables

With regard to TAM model and its extensions (i.e. TAM2 and UTAUT) the findings indicate that newer versions
increased the amount of explained variance of the previous model
both before including the moderating variables (i.e.
TAM explains 24%, TAM2: 25% and UTAUT: 29%) and after (i.e. TAM explains 31%, TAM2: 35% and UTAUT:
37%). In addition, three models

DOI, the combined model (TAM+TTF) and UTAUT model

showed the highest

explained variance in both cases. These three models, in contrast to the other models, are not focused solely on the
individual perspective but include organizational and management dimensions in addition to the individual dimensions.
Brown et al. (2002)
found that using TAM to evaluate ERP acceptance provided a limited explanation of end
behavior, attitudes and perceptions towards the system, and thus delivers misleading recommendations for organizations.

In addition, UTAUT is considered an improv
ement over the TAM extension models when evaluating end
acceptance of ERP systems because it makes it possible to consider the mandatory nature of ERP systems. An implicit
assumption of earlier technology acceptance models (i.e. TAM, TAM2) is that use
rs of the information systems have
some level of choice with regard to the extent that they use the technolog
y (Amaoko
Gyampah & Salam, 2004, Nah et al.

2004). Furthermore, the UTAUT model incorporates a facilitating conditions construct which is defined a
s the objective
factors, such as the provision of support for users, in the environment that makes an application easy to use. The DOI
model is based on a diffusion process developed by Rogers (1962) which is defined as a communicative process rather
an individually focused process. Thus, the DOI model introduces variables related to the organizational aspects such
as result demonstrability, trialability and visibility within the organization. In this sense, the DOI model is considered an

improvement o
ver previous models when evaluating end
user acceptance of ERP systems.

It is important to emphasize that most of the key relationships in the models were moderated. Gender, which has
received more attention in the literature, was found to be a key
moderating influence. User prior experience in complex
IT settings, such as ERP systems, was the second key moderating variable. According to Venkatesh et al. (2003) another
moderating variable, age, has received little attention in the technology acceptan
ce research literature. Our findings
indicate that in the context of complex IT settings, age emerges as an important moderator of key relationships in the








Medium Support (
4 of 7






Week Support (
1 of 7 positive)






Strong Support (
7 of 7 positive)




Not Supported


Strong Support (
5 of 5 positive)






Support (
5 of 5 positive)






Strong Support (
3 of 4 positive)




Strong Support (
4 of 4 positive)




Hypotheses results

The perceived usefulness, performance expectancy, relative advantage and task
fit constructs were
acknowledged by previous studies as similar (Calisir et al. 2009, Venkatesh et al. 2003). These constructs, in this study,
were not found to be si
gnificant within all models. This finding corroborates a few studies in the field of ERP (Seymour
et al. 2007) but is inconsistent with most general information systems acceptance research. This result is nevertheless is
very significant in that it shows t
hat in a complex technology implementation environment such as ERP implementation,
unlike less complex environments, the perceived usefulness of the technology is perhaps less important than its ease of
use. Many organizations are committed to a “vanilla”
implementation to avoid ERP software modifications and business
process re
engineering in particular to align best business standards for a successful ERP implementation (Al
2007, Finney
Corbett 2007, El
Sawah et al. 2008). Consequently, potenti
al adopters are less troubled by how to
execute old processes in the new system because of the obligation to run new business processes based on best practice
that are already well implemented in the ERP system with minimal changes needed. Thus, managerial

attempts that have
focused on enhancing the perceived usefulness of the ERP system will be less worthwhile than the managerial attempts
focused on enhancing the perceived ease of use.

In addition, in cases where these similar constructs were found to be
significant, they were not found to be the strongest predictor of user symbolic adoption by contrast to several studies.
These results perhaps suggest that perceived usefulness has lower explanatory power in comparison to other constructs in
the context o
f complex IT settings.

Contrary to predictions and in contrast to previous studies, the results indicate, that the influence of
on symbolic adoption was not moderated by age or gender. Venkatesh et al. (2003) posited that since men
to be highly task
oriented, performance expectancy centered on task accomplishment is likely to be especially important
to men because of socialization processes. In addition, they argued that research on age differences indicates that younger
users m
ay place more importance on extrinsic rewards. However, in the case of ERP systems the latter may be perceived
as rich in functionality and beyond the needs of the reasonable user (Yi

et al.
2006). Therefore, users' main concern may
be the extent to which
the ERP system is easy to use rather than the extent to which the system is useful. Thus, the
present study reveals that age and gender differences do not play a role in ERPs contexts with regard to the perceived
usefulness construct.

Another frequent hyp
othesis concerns the potential moderating effect of experience. According to Castaneda et al.
(2007) user beliefs are the key perceptions driving IT usage and may change with time as users gain experience. It was
found that the effect of perceived usefulne
ss on user symbolic adoption increases with increasing experience. One
explanation may be related to training programs. Users' training is important not only for acquiring skills but also enables
adjustment to changes created by the implementation of an ER
P system and allows potential adopters to get firsthand
experience and explore the ERP system (Amoako
Salam, 2004, Aldwani 2001, Brown et al. 2002).
Experienced users evaluate a system in a more in
depth way and hence may consider perceived usefu
lness to a greater
extent than inexperienced ones (Jasperson et al. 2005).

In this study, and consistent with most previous studies, perceived ease of use, as formulated by different constructs
(e.g. effort expectancy), was found to be a significant pred
ictor of user symbolic adoption. Furthermore, in the context of
moderating factors, and consistent with previous research (e.g., Agarwal
Prasad 1997, 1998; Davis et al. 1989;
Thompson et al. 1991, 1994, Morris
Venkatesh 2000), less experienced younger
woman ascribed more importance to
ease of use aspects than men, as they tend to gain efficacy over time. Age differences have been associated with growing
difficulty in processing complex stimuli and allocating attention to information on the job (Venkates
h et al .2003). Scott
Walczak (2009) suggested that ERP users in organizations with diverse ages often find ERP training challenging,
despite their work experience. In addition, it was found that women may place more importance on ease of use aspects
n men because of individual perceptions related to gender roles. Thus, age, gender and experience differences exist in
the context of ERPs.

Consistent with most previous studies in mandatory settings, the results showed for all models that the social
uence construct is a significant predictor of symbolic adoption. In addition and in line with previous research, the
social influence effect on symbolic adoption of ERP system was moderated by: 1) age because affiliation requirements
increase with age, 2
) gender because women tend to be more sensitive to others’ opinions and 3) experience, in
mandatory settings, because in the early stages of individual experience social issues impact the technology and its roles
but eroding over time and eventually beco
me non
significant with sustained usage (Venkatesh

Davis 2000, Morris
Venkatesh 2000, Venkatesh et al. 2003). Thus, these moderating variables simultaneously influence the social influence
intention relationship not only in a simple technology environm
ent but in a complex technology environment as well.

The facilitating conditions construct, in the context of information systems, is associated with the provision of IT
support. Venkatesh (2000) argued that effort expectancy fully mediates the effect of

facilitating conditions on intention
because facilitating condition issues (e.g. support) are largely captured within the effort expectancy construct which taps
the ease with which that tool can be applied. Thus in the context of complex IT settings, such

as an ERP system, these
constructs may not share similar themes since the support given to users may not be good enough to satisfy users and
deliver an ease of use experience. The current results show that in complex IT settings such as an ERP system, t
construct is not fully mediated by effort expectancy and influences symbolic adoption considerably. In addition and
consistent with previous studies, this study shows that the effect of facilitating conditions on symbolic adoption increases
with exper
ience in that users gradually find multiple avenues for help and support. Age also has an effect since older
users attach more importance to receiving help and assistance on the job which is more strongly emphasized in the
context of a complex IT because
of the increasing cognitive and physical limitations associated with age (Morris
Venkatesh 2000, Venkatesh et al. 2003).

Self efficacy and anxiety constructs emerged as significant direct determinants of intention. McIlroy et al. (2001)
found that the ma

female gap in computer anxiety, which initially showed women to be more anxious, is slightly
declining but still persists in the USA.
In addition, a
lthough affect was found to be a significant determinant of user
symbolic adoption, previous research ha
s shown that affect, associated with intention to use, is fully mediated by
performance and effort expectancy (Venkatesh et al. 2003).

Rogers (1995) related compatibility with existing values, belief, past experiences and the needs of potential adopters.

Since the early applications of DOI to IS research, this theory has been applied and adapted in numerous ways.

studies defined compatibility as the extent to which the innovation is perceived to be consistent with the potential
adopters' existing

values, previous experience and needs. Other studies defined it in terms of technical compatibility with
regard solely to hardware and software issues (Bradford
Florin 2003). Nevertheless, studies have consistently found
that technical compatibility is
an important antecedent to the adoption of innovations (Bradford
Florin, 2003).
However, in terms of ERP packages, compatibility, from a standards perspective, may be broader.

Iivari (2005) found that system quality emerged as more significant than infor
mation quality, presumably because of
the mandatory nature of analyzing the system for acceptance. The present study is consistent with Iivari's (2005) study.
Since an ERP system is used on a daily basis in organizations, it is natural that the information

output is timely.
However, Zhang et al. (2004) argued that the variables of information quality and system quality from the D&M model
should be modified to take the specific conditions of a large mature off


shelf ERP package into account. First, in
environment of an ERP system, the integrity of raw input data affects others users who operates the different modules.
Second, ERP system packages have been developed for many years and used in many sites, which enables the packages
to be very mature a
nd reliable. In addition, this study showed that service quality is a significant predictor of symbolic


Enterprise resource planning acceptance model

A major paradigm in psychology and marketing argues that affect (defined as an umbrella for a se
t of more specific
mental processes including emotions, moods, and attitudes) and cognition (referring to more specific mental processes
are separate and partially independent systems (Zajonc, 1984). Most models or theories in IS focus on the cognitive and

behavioral aspects of human decision
making processes and on individual reactions to using technologies in
organizations (Sun
Zhang 2006).

The basic idea in the model proposed below is that a user's symbolic adoption of an information system in
IT settings is influenced by cognitive reactions and technical features that are considered separate and partially
independent systems. The hypothesis is that these two components together determine the user's final symbolic adoption.

We drew on th
e analysis above to identify several key constructs and key moderators to make up the main dimensions
of the model (see Figure 9
. The model is based on the incorporation of the main constructs defined in previous research
in the field of information syste
ms that are thought to be significant in the field of ERP systems, as described in Table 13.

With regard to ERP systems we assumed that the facilitating condition construct is very similar to the service quality
construct in terms of the extent to which an

individual believes that an organizational and technical infrastructure exists
to support use of the system. In addition, task
fit and compatibility are very similar constructs. The
compatibility construct incorporates items that tap the fit be
tween all aspects of an individual’s work and the use of the
system in the organization (Venkatesh et al. 2003). These aspects are covered by three constructs in the new model: 1)
perceived usefulness, defined by the degree to which a person believes that
using an IS system will enhance his job
performance, 2) level of integration, which influences job performance beyond users' initial perception and 3) offset from
standard, which can increase job performance, and its counterpart, hazard system quality.

In this study, as in previous work, the CSE model was analyzed for the effect of these constructs on users' willingness
to use the system (dropping the ease of use construct). According to Venkatesh et al. (2003) self
efficacy and anxiety are
theorized not

to be direct determinants of intention. Previous research has shown that self
efficacy and anxiety are
conceptually and empirically distinct from perceived ease of use and yet are fully mediated by perceived ease of use in
explaining intention to use and
thus were modeled as indirect determinants of user symbolic adoption. Therefore, the
suggested model ignores the self
efficacy and anxiety construct although they were found significant.


Service Quality

The Service Quality construct is defined as the overa
ll support delivered by the service provider, and applies
regardless of whether this support is delivered by the IS department, a new organizational unit, or outsourced (Delone

2003). Support of users by the service provider is often measured by

the assurance, responsiveness,
reliability, and empathy of the support organization (Petter


2009). The inclusion of service quality in the
updated DeLone
& McLean

(2003) model reflects IS functions or IS organizations rather than IS applications,

to reflect
the importance of service and support in successful information system (Iivari 2005, Wu
Wang 2006). It was added
because the changing nature of IS called for a measure to assess service quality when evaluating IS acceptance (Petter

09). Lin et al. (2006) argued that system quality and information quality may be the most important quality
dimensions whereas service quality may be the most important factor for measuring the overall success of the IS
department. Therefore, service quali
ty was not considered in their study, because their focus was to measure the success
of ERP systems rather than the IS department. However, researchers believe that service quality is an important element
in information system success (Landrum
Prybutok 2

et al. 2008
). Although a claim could be made that
service quality is merely a subset of the system quality, the changes in the role of IS over the last decade argue for a
separate variable (Delone
& McLean

2003). Chien
Tsaur (2007) argued

that service quality needs to be included to
measure service
level aspects since system quality focuses more on technology
level measures.
k et al. (2008)
found empirical evidence for a significant causal relationship between service quality and c
onstructs related to users'
satisfaction and intention to use.


Level of Integration

Organizations perceive ERP as a vital tool for organizational competition as it integrates dispersed organizational
systems and enables flawle
ss transactions and produc
tion (
Koh et al.
. ERP vendors traditionally offered a single
ERP system

Huang et al. 2003
. ERP systems suffered from limitations in coping with integration challenges dealing
with changing requirements. However, companies preferred to implement an

ERP suite from one vendor that
incorporated stand
alone point solutions (that once filled functionality gaps in older ERP releases) to achieve higher
levels of integration and improve customer relationships and the supply chain's overall efficiency (Huang

et al. 2003
Tchokogue et al. 2005
). However, although most companies still follow the single source approach, a significant number
of firms employ a strategy of “best of breed” ERP to maintain or create a competitive advantage (Shaul
Tauber, 2013).
ERP vendors begun to acquired products or develop their own functionality that was either comparable or better than
many of the "best of breed" applications, and hence enabled companies to maintain or create a competitive advantage
based o
n unique business processes, rather than adopting the same business processes which would leave no firm with an
advantage (Bradley 2008). In recent years, integration has prompted leading investments due to the functionality gap and
the need to extend and
integrate the ERP system to other enterprises or "best of breed" applications (Jacobson et al.
2007). Integration was ranked as one of the leading investments for 2003, and well over 80% of U.S. companies
budgeted for some type of integration in

d roughly one
third of U.S. companies defined application integration
as one of their top three I
T investments in 2003 (Caruso 2003)
. ERP license revenue remained steady as companies
continued their efforts to broadly deploy core applications and then adde
d complementary functionalities in later phases.
Today a greater effort is being made to integrate more mobile devices with the ERP system. ERP vendors are working to
extend ERP to these devices along with users’ other business applications. The technical
stakes of the ERP concern
integration: this has involved hardware, applications, networking, supply chains and has covered more functions and
roles including decision making, stakeholders' relationships, standardization, transparency, globalization, etc.
Akkermans et al. 2003, Lim et al. 2005, Botta
Genoulaz et al. 2005


Offset from standard

An ERP system is radically different from traditional systems development (Dezdar
Sulaiman, 2009). ERP
systems are based on industry best practices, and are intend
ed to be deployed

as is, thus offering organizations
configuration options that allow them to incorporate their own business rules. However, there are often functionality gaps
remaining even after the configuration is complete between the best practices pr
ocesses implemented within the ERP
system and the organization's pre
implementation business processes, and organizations often suffer from poor fit
between the ERP system and the organization. Organizations can avoid major misfits by applying two differe
nt strategies
to better match the delivered ERP functionality: technical customization such as rewriting part of the delivered
functionality within the ERP system, or interfacing to an external system, which is the most invasive, or finally business
s reengineer
ing (
Fryling 2010

Customization potentially leads to more software process customization, more cycles of re
implementation and an
increase in testing activities, complexity, resources and a longer project schedule which can slow down the pro
ject and
generating risky bugs in both present and in future maintenance. ERP vendors provide upgrades to guarantee support for
the system o 'fix' outstanding ‘bugs’, current best
practices or design weaknesses (
Agerfalk et al. 2009, Shaul


To avoid ERP software modifications and its consequences many organizations are committed

to a “vanilla”
implementation (
Mudimigh 2007, Finney
Corbett 2007
. However, ERP vendors have a rather different view of
customization than the adopting organi
zations, in that most vendors consider customiza
tion to be an evolving process

Strong 2004



Regardless of the significance of the relationships between factors in the regression model, these relationships may
not apply to large enterp
rises since the respondents' experience relates to SMEs operating in the
local market.
unlike LEs, face much greater constraints in terms of the resources that can be committed to all stages of information
gathering, although the complexity and amoun
t of IT functionality and integration requirements are often
similar (Chan
et al. 2012, Shaul & Tauber 2011)
. As a result, SMEs are forced to make implementation compromises according to
resource constraints, which increase the risks inherent to

the implem
entation process (
Sun et al. 2005
. In addition
differences in the scope of implementation in general as well as organizational, technological and environmental factors
make it difficult to present a generalized perspective on implementation

(Koh & Saad 20
. Finally this study was
conducted with limited samples across different models and therefore, for practical analytical reasons, the authors
operationalized each of the core constructs in the models by using the highest
loading items from each of the re
scales as recommended by Nunnally
Bernstein (1994).



The primary purpose of this paper was to synthesize the current state of the art with respect to users' symbolic
adoption of information technologies in complex IT settings such as

ERPs. It reviewed the literature on the main
information system acceptance models and their extensions, and empirically compared them as regards ERP systems

Each of these models makes important and unique contributions to the literature on user acceptanc
e of IT. It also
examined the effect of key moderators on these relationships (i.e. age, gender and experience) were also examined.

The findings are consistent with previous research in less complex IT settings, with regard to the interaction between
key moderators and core construct in complex IT settings such as ERPs. For instance, in implementing enterprise
systems such as ERP syste
ms, PEOU was found to be a significant predictor of user symbolic adoption within each
model and less experienced users place more importance on ease of use r than experienced users as they tend to gain
efficacy over time. However, the findings also show t
hat complex IT settings are unique in a certain sense. Contrary to
initial hypotheses, and in contrast to previous studies, the influence of the perceived usefulness (defined in TAM, TAM2
and TAM+TTF models), performance expectancy (defined in the UTAUT a
nd CSE model) and relative advantage
(defined in the DOI model) on user symbolic adoption of an ERP system is not moderated by age and gender but rather
by experience. In addition, these constructs were found to be unstable across the different studies, t
hus implying that
further examination is needed. Complex IT settings such as ERP systems are rich in functionalities beyond the needs of
the average user. Therefore, users' main concern may be the extent to which the ERP system is easy to use rather than t
extent to which the system is useful.


Future research

The acceptance of complex information technology such as ERPs is still affected by intangibles; hence future work
on adoption is critical. As shown in the review of the literature, recent efforts to

develop technology acceptance models
have mostly focused on two dimensions: enriching or extending the model from theoretical perspectives and empirically
further validating the performance of the models with various innovations in different environments.

Although studies have made great progress and the variance explained by several models are respectable in terms of
behavioral research, further work should attempt to identify and test additional boundary conditions of the model to
provide an even richer
understanding of technology adoption and usage behavior. In particular more attention should be
paid to investigating the influence of broad organizational, managerial, technological, operational and environmental
variables. The influence of other moderati
ng variables such as organization size, education level, orientation (e.g.
technological, business), level of management, private vs. public sector and developing countries vs. developed countries
also deserve work. A closer examination of the role moderat
ing variables and their psychological and organizational
basis could also shed light on their moderating role.


garwal, R., Prasad, J,

Are individual differences germane to the acceptance of new information technologies?.
Decision Science
s, 30(2), 361


Agerfalk, P.
., Fitzgerald, B., Slaughter, S,

2009. Flexible and distributed information systems development: state of
the art and research challenges. Information Systems Research, 20(3), 317


Aiken, L.S., West, S.G,

1991. Multiple

regression: testing and interpreting interactions. Newbury Park, CA: Sage.

Akkermans, H.A., Bogerd, P., Yucesanc, E., Van
hove, L.N,

2003. The impact of ERP on supply chain
management: exploratory findings from a European Del
phi study. European Jou
rnal of
Operational Research,
146(2), 284


Mudimigh, A.S,

2007. The role and impact of business process management in enterprise systems implementation.
Business Process Management Journal, 13(6), 866

Gyampah, K., Salam, A.F,

2004. An ext
ension of the Technology Acceptance Model in an ERP
Implementation Environment. Information & Management, 41(6), 731


Bienstock, C

Royne, M
Sherrell, D

Stafford, T
, 2008.

An expanded model of logistics service

Incorporating logistics information technology
International Journal of Production Economics

cherjee, A., Sanford, C,

2006. Influence processes for information

technology acceptance: An elaboration
likelihood model

. MI
S Quarterly, 30(4), 805

az, V., Millet, P.A., Grabot, B,

2005. A survey on the recent research literature on ERP systems.
Computers in Industry, 56(6), 510


Bradley, J,

2008. Management based critical success factors in the implementation of Enterprise Resource Planning
systems. International Journal of Accounting Information Systems, 9(3), 175

Brown, S.A., Massey, A.P., Mo
Weiss, M.M., Burkman, J.R,

2002. Do I

Really Have To? User Acceptance of
Mandated Technology. European Journal of Information Systems, 11(4), 283


Jones, A., Hubona, G.S,

2005. Individual Differences and Usage Behaviour: Revisiting a Technology
Acceptance Model Assumption. The DATA

BASE for Advances in Information Systems, 36(2), 58


Calisir, F., Calisir, F,

2004. The relation of interface usability characteristics, perceived usefulness, and perceived ease
of use to end
user satisfaction with enterprise resource planning (ERP) sy
stems. Computers in Human Behaviour,
20(4), 505


Caruso, D,

2003. The World Class Challenge: Six Critical Issues Midmarket Manufacturers Must Address. AMR
Research, Boston, MA

J.A., Munoz
Leiva, F., Luque, T,

2007. Web Acceptance Model (WAM)
: Moderating effects of user
experience. Information & Management, 44(4), 384


Chan, F

Chong, A
.Y.L., Zhou, L., 2012

An empirical investigation of factors affecting e
collaboration diffusion in
International Journal of Production Economics

, M.
heung, W.,

, C.
, J.H.

2008. Understanding ERP system adoption from the user’s
perspective. International Journal of Production Economics, 113(2), 928

Chien, S.W., Tsaur, S.M.

2007. Investigating the success of ERP systems: Case studies in three
Taiwanese high
industries. Computers in Industry, 58 (8
9), 783

793 .

Compeau, D.R.,
Higgins, C.A,

1995a. Application of Social Cognitive Theory to Training for Computer Skills.
Information Systems Research, 6(2),118

au, D.R., Higgins, C.A,

1995b. Computer Self
Efficacy: Development of a Measure and Initial Test. MIS
Quarterly, 19(2), 189

Compeau, D.R., Higgins, C.A.,
Huff, S,

1999. Social Cognitive Theory and Individual Reactions to Computing
Technology: A Longitudinal Study. MIS Quart
erly, 23(2), 145

Cronbach, L., 1987
. Statistical tests for moderator variables: Flaws in analysis recently proposed. Psychological
Bulletin, 102, 414

Davis, F.D,

1989. Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Informati
on Technology. MIS
Quarterly, 13(3), 319


DeLone, W.H.,
, E.R.

1992. Information systems success: the quest for the dependable variable. Information
Systems Research, 3(1), 60


DeLone, W.H.,
McLean, E.R,

2003. The DeLone and

model of i
nformation systems success: a ten
update. Journal of Management Information Systems, 19(4), 9


Dezdar, S., Suliman, A.

2009. Successful enterprise resource planning implementation: taxonomy of critical factors.
Industrial Management & Data Systems
, 109(8), 1037

Sawah, S., Tharwat, A.A.E.F., Rasmy, M.H.

2008. A quantitative model to predict the Egyptian ERP
implementation success index. Business Process Management Journal, 14(3), 288


Finney, S., Corbett, M.

2007. ERP implementation:

a compilation and analysis of critical success factors. Business
Process Management Journal, 13(3), 329

Fryling, M,

2010. Estimating the impact of enterprise resource planning project management decisions on post
implementation maintenance costs: a case study using simulation modeling.
Enterprise Information Systems,


Ge, L

, S
, 2009.

ERP Application in China: An Overview
International Journal of Production

Huang, A.
, Yen, D.C., Chou, D.C., Xu., Y.,

Corporate applications integration: Challenges, opportunities

implementation strategies. Journal of Business and Management, 9(2), 137

Iivari, J.

2005. An empirical test of the DeLone

model of information system success.
Database, 36(2), 8

Jacobson, S., Shepherd, J., D’aquila, M., Car
ter, K.

The ERP Market Sizing Report, 2006

2011. AMR Research,
Boston, MA.

Jasperson, J., Cater, P.E., Zmud, R.W.

2005. A comprehensive conceptualization of postadoptive behavior associated
with information technology enabled work systems. MIS Quarterly, 29 (3), 525


Karahanna, E., Straub, D.W., Chervany, N.L.

1999. Information technology adoption across time: a cross
comparison of pre
adoption and post
adoption beliefs. MIS Quarterly, 23 (2), 183


Koh, S.C.

, S.M.

2006. Managing uncertainty in ERP
controlled manufacturing environments in
International Journal of Production Economics, 101(1), 109

Koh, S.C.L., Gunasekaran, A.,

Rajkumar, D.

ERP II: The involvement, benefits and impediments of
collaborative information sharing
. International Journal of Production Economics
, 113(1), 245

Landrum, H., Prybutok, V.R.

2004. A service quality and success model for the information service industry. European
Journal of Operational Research, 156(3), 628

Lee, D.H., Lee, S.M., Olson, D.L., Chung, S.H.

2010. The effect of organizational support on ERP implementation.
ndustrial Management & Data Systems, 110(2), 269

im, E.T.K.,
an, S.L,
an, C.W.

2005. Managing user acceptance towards enterprise resource planning (ERP)
systems: understanding the dissonance between user expectations and managerial policies. Europ
ean Journal of
Information Systems, 14(2), 135


Lin, H.Y., Hsu. P.Y., Ting, P.H.

2006. ERP Systems Success: An Integration of IS Success Model and Balanced
Scorecard. Journal of Research and Practice in Information Technology, 38(3), 215

Luo, W.,

Strong, D.M.

2004. A Framework for Evaluating ERP Implementation Choices. IEEE Transactions on
Engineering Management, 51(3), 322

Morris, M.G., Venkatesh. V.,

2000. Age Differences in Technology Adoption Decisions: Implications for a Changing
Work F
Personnel Psychology, 53(2), 375


Nah, F.F., Tan, X., Teh, S.H.

An Empirical Investigation on End
Users’ Acceptance of Enterprise Systems.
Information Resource Management Journal, 17(3), 32


Nunnaly, J.

1978. Psychometric Theory. New York: McGrawHill

Petter, S.,
, E.R.

2009. A meta
analytic assessment of the DeLone and

IS success model: An
examination of IS success at the individual level. Information & Management, 46(3), 159

Ram, J.,

Corkindale, D., Wu, M.L.,

Implementation critical success factors (CSFs) for ERP: Do they contribute
to implementation success and p
implementation performance?.
International Journal of Production Economics
144(1), 157

Rogers, E.M.

Diffusion of Innovations. New York: Free Press.

Shaul, L., Tauber, D.,

2011. CSFs along ERP life
cycle in SMEs: a field study. Industrial Management & Data
Systems, 112(3), 360

Shaul, L., Tauber, D.

2013. Critical Success Factors in Enterprise Resour
ce Planning Systems: Review of the Last
Decade. ACM Computing Surveys,
45(4), 481

Stone, E.F., Hollenbeck, J.R.

1989. Clarifying some controversial issues surrounding statistical procedures for
detecting moderator variables: empirical evidence and re
lated evidence. Journal of Applied Psychology, 74(1), 3

Sun, H., Zhang, P.

2006. The role of affect in IS research: A critical survey and a research model. In: Zhang, P.,
Galletta, D. (Eds). Series of Advances in Management Information Systems, M.E. S
harpe Publisher, 295

Sun, A

Yazdani, A

Overend, J


for enterprise resource planning (ERP) system
implementations based on
critical success factors (CSFs)
. International Journal of Production Economics
, 98(

e, A., Bareil, C., Duguay, C.
, 2005
. Key lessons from t
he implementation of an ERP

Pratt &
Canada. International Journal of Production Economics, 95
(2), 151

Venkatesh, V., Davis, F.D.

1996. A model of the antecedents of perceived ease of use: development and test, Decision
Sciences, 27(3), 451


Venkatesh, V., Davis, F.D.

2000. A Theoretical Extension of the Technology Acceptance Model:
Four Longitudinal
Field Studies.
Management S
cience, 45(2), 186

Venkatesh. V., Morris,

M.G., Davis, G.B.,

Davis, F.D.

2003. User Acceptance of Information Technology: Towards a
Unified View. MIS Quarterly, 27(3), 425


Wu, J.H., Wang, Y.M.

2006. Measuring KMS success: A respecification of t
he DeLone and
’s model.
Information & Management, 43(6), 728


Yi, Y., Wu, Z., Tung, L.L.

2006. How Individual Differences Influence Technology Usage Behaviour? Towards an
Integrated Framework. Journal of Computer Information Systems, 46(2), 52


Zajonc, R.B. 1984. On the primacy of affect. American Psychologist, 39(2), 117

Zhang, Z.,
Lee, M.K.O., Huang, P., Zhang, L., Huang, X.

2005. A framework of ERP systems implementation success
in China: An empirical study. International Journal of Production Economics, 98(1), 56

Zviran, M., Pliskin, N., Levin, R.

2005. Measuring User Satisf
action and Perceived Usefulness in the ERP context.
Journal of Computer Information Systems, 45(3), 43