Predicting the drivers of behavioral intention to use

Transcription

Computers in Human Behavior 36 (2014) 198–213
Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
Predicting the drivers of behavioral intention to use mobile learning:
A hybrid SEM-Neural Networks approach
Garry Wei-Han Tan a, Keng-Boon Ooi b, Lai-Ying Leong a, Binshan Lin c,⇑
a
Faculty of Business and Finance, Universiti Tunku Abdul Rahman, Kampar, Malaysia
Faculty of Business, Management and Accountancy, Linton University College, Malaysia
c
Department of Management and Marketing, Louisiana State University in Shreveport, Shreveport, LA 71115, USA
b
a r t i c l e
i n f o
Article history:
Available online 20 April 2014
Keywords:
Mobile learning (m-learning)
Technology Acceptance Model (TAM)
Structural Equation Modeling (SEM)
Artificial Neural Networks (ANN)
User behavior
a b s t r a c t
This study empirically investigates on the elements that affect the user’s intention to adopt mobile learning (m-learning) using a hybrid Structural Equation Modeling–Artificial Neural Networks (SEM–ANN)
approach. A feed-forward-back-propagation multi-layer perceptron ANN with the significant determinants from SEM as the input units and the Root Mean Square of Errors (RMSE) indicated that the ANN
achieved high prediction accuracy. All determinants are relevant and their normalized importance was
examined through sensitivity analysis. The explanation on new computer technologies acceptance have
been primarily based on the Technology Acceptance Model (TAM). Since TAM omits the psychological science constructs, the study address the weaknesses by incorporating two additional constructs, namely
the personal innovativeness in information technology (PIIT) and social influences (SI). Out of the 400
survey distributed to mobile users, 216 usable questionnaires were returned. The results uncovered that
the intention to adopt m-learning has significant relationship with TAM. The findings for PIIT, SI and the
control variables of age, gender and academic qualifications however show mixed results. The results
provide valuable information for mobile manufacturers, service providers, educational institutions and
governments when strategizing their adoption strategies. Additionally, from the perspective of an emerging market, the study has successfully extended TAM with psychological constructs.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
Learning has always been restrained to brick and mortar classroom and traditional books, for decades. However, a new wave of
learning has emerged with the increased capabilities of mobile
devices (m-devices) and the wide availability of various network
connectivity, e.g. UMTS (3G), HSPA (3G+), LTE (4G), WIMAX, and
WAP (Hu, Lu, & Tzeng, 2014). Sharples (2007) defined mobilelearning (m-learning) as the learning between the learners via
the technology of portability. Take for example, learning with the
integration of Personal Digital Assistant (PDA), Smartphones, iPod,
mobile telephones, laptops and tablet personal computer technologies (Berri, Benlamri, & Atif, 2006; Donnelly, 2009; Liu, 2009).
Similarly, Sharma and Kitchens (2004) considers m-learning as
the delivery of digitized e-contents through wireless phones
hooked into PDAs. The development of m-learning has not only
⇑ Corresponding author. Tel.: +1 318 797 5025; fax: +1 318 797 5127.
E-mail addresses: [email protected] (G.Wei-Han Tan), [email protected] (K.-B. Ooi), [email protected] (L.-Y. Leong), Binshan.Lin@lsus.
edu (B. Lin).
http://dx.doi.org/10.1016/j.chb.2014.03.052
0747-5632/Ó 2014 Elsevier Ltd. All rights reserved.
supports learning through a variety of settings, but also acts as
an enabler to learning at different location and time (Gil &
Pettersson, 2010). Based on a study by Ambient Insight (2010) in
United States (U.S.), the m-learning’s market in 2009 for products
and services was at U.S. $632.2 million dollars but the figure was
forecasted to reach U.S. $1.4 billion by 2014.
M-learning is distinctly different from electronic learning (elearning) as the latter requires an Internet access as well as wired
connection before any learning can take course. However, m-learning works on the wireless environment where m-devices are used.
Therefore, learning is no longer restricted to only having attending
classes. Mulliah (2006) commented that there are three advantages
of m-learning in the likes of convenience, collaboration and fun as
opposed to e-learning. In addition, m-learning devices are portable
and small in size, thus it is easy to carry around at one’s convenience (Schwiderski-Grosche & Knospe, 2002). As a result, acquiring knowledge is now at one’s fingertips. According to Attewell
(2005), there are several advantages of adopting m-learning. They
include enhancing an individual’s skills, providing them with
opportunities to learn new things independently, the ability to
determine weak or slow learners who require assistance, and
G.Wei-Han Tan et al. / Computers in Human Behavior 36 (2014) 198–213
encourage reluctant individuals to learn, resulting in improved
learners’ confidence. Thus, m-learning is a new education paradigm and is a preferred choice in higher education and life-long
learning of every country (Liu, 2009). However, the factors influencing the adoption of m-learning are still unclear despite the
rapid development of the current study as a new form of learning.
Scholars like Pozzi (2007) stressed that m-learning is only adopted
occasionally and in a supplemental manner. The sentiments were
echoed by Herrington and Herrington (2007), who claimed that
pedagogical use of m-devices is not widespread in higher educations. The statistic from the Malaysian Communications and Multimedia Commission (MCMC) confirmed that there are 29.6 millions
mobile phone subscribers in Malaysia (Malaysian Communications
and Multimedia Commission, 2010). In comparison with the numbers of subscribers, scholars like Wei, Marthandan, Chong, Ooi, and
Arumugam (2009), stressed that the number of m-learning users in
Malaysia still falls behind other developing countries. Further evidence from Wong and Hiew (2005) indicated that m-learning is
still very much at an early stage in Malaysia. The availability of different m-devices according to Corbeil and Valdes-Corbeil (2007)
does not indicate that students will adopt them for education
purposes.
While the development of m-learning have been frequently discussed, most past studies were carried out in countries such as Taiwan (Hwang, Wu, Zhuang, Kuo, & Huang, 2010), New Zealand (Lu &
Viehland, 2008), Macedonia (Fetaji & Fetaji, 2008), China (Liu, Li, &
Carlsson, 2010) and Thailand (Poonsri, 2008). M-learning studies
from a developing country perspective like Malaysia remains limited. Scholars studying m-learning primarily focused from the perspective of software/infrastructure for library services (Cummings,
Merrill, & Borrelli, 2010; Hahn, 2008; Walsh, 2009), higher education (Cook, Bradley, Lance, Smith, & Haynes, 2007; Fetaji & Fetaji,
2008), museum (Hsu, Ke, & Yang, 2006) and further education
(Savil-Smith, Attewell, & Stead, 2006). Interestingly, the driving
factors on the intention to adopt m-learning have remained unexplored. Only through understanding why consumers lack of motivation to adopt a certain information technology (IT) can we make
certain on the substantial return on investment (Magni, Taylor, &
Venkatesh, 2010; Rogoski, 2005). The study therefore empirically
creates a framework to explain on the factors that influence the
intention to adopt m-learning through the extension of Technology
Acceptance Model (TAM) with psychological science constructs. In
addition, the study also incorporates gender, age and academic
qualifications as control variables. The following is the structure
of the paper. In the following section, we present on the overview
of m-learning. Then, we present our research model, hypotheses
development and methodology of our study. In the final section,
the findings, conclusion, limitation and future research of m-learning adoption is discussed.
2. Literature review
2.1. An overview of mobile learning
Given that m-learning is a relatively new concept, it has been
defined in various ways by earlier studies (Lu & Viehland, 2008).
Attewell (2005) and Lu and Viehland (2008) defines m-learning
as a learning which is similar to e-learning. M-learning uses wireless transmission and m-devices such as smartphones, tablets,
multi-game devices and personal media players, instead of wired
connections or traditional personal computers. Similarly, Lehner
and Nosekabel (2002, p. 103) elaborated on m-learning definition
as ‘‘any service or facility that supplies a learner with general electronic information and educational content that aids in acquisition
of knowledge regardless of location and time’’. Therefore,
199
individuals can learn independently of time and space (Amaral,
2006). Due to the numerous advantages, m-learning has gained
popularity and many learning institutions are starting to adopt to
this technology (Koike, Akama, Chiba, Ishikawa, & Miura, 2005).
M-learning’s popularity is largely due to its low cost, as well as
allowing users to learn anytime and anywhere. In this study, mlearning refers to as the activities of learning with the usage of
m-devices such as a mobile phone/smart phone through wireless
communications among its users on a 365/24/7 basis.
2.2. Models of IT/IS adoption
Most models in predicting the acceptance of new technologies
were derived from scholars with diverse backgrounds. IT scholars
like Davis (1989) proposed the Technology Acceptance Model
(TAM), while psychologists scholars like Fishbein and Ajzen
(1975) and Ajzen (1991) concentrated on their research using Theory of Reason Action (TRA) and Theory of Planned Behavior (TPB).
Rogers (1995) a marketing scholar on the other hand proposed the
Diffusion of Innovation (DOI).
TPB is extended from TRA by adding an additional variable
namely ‘perceived behavior control’ (Ajzen, 1991) to increase predictive power. TPB suggests that the new added variable with subjective norm and the individual’s act of behavior can explained on
the behavioral achievement of an individual. Attitude refers to as a
person’s favorable or unfavorable feelings about performing the
behavior. On the other hand, subjective norms is defined as ‘one’s
beliefs whether others approve or disapprove in engaging an activity (Fusilier & Durlabhji, 2005), while the perception on the individual’s ability to perform a behavior explained on the concept
on perceived behavioral control (Ajzen & Madden, 1986).
TAM focuses on perceived ease of use (PEOU) and perceived
usefulness (PU) as the two prime purposes behind the intention
to adopt information systems (IS) (Davis, 1989) and evolved using
TRA. According to Agarwal and Karahanna (2000), O’Cass and
Fenech (2003) and Lee (2006), TAM has been successfully carried
out by IT scholars to forecast a wide variety of technology settings
such as websites, internet shopping and e-learning. While TAM is
useful in the explanation of users’ intention, the external variables
that impact the PU and PEOU were not completely discussed.
Scholars therefore suggested for TAM to be extended to provide
clearer understanding of users’ decisions to adopt a certain technology (Chong, 2013a; Legris, Ingham, & Collerette, 2003). TAM2
for example was proposed as an extension of TAM (Venkatesh &
Davis, 2000). In TAM2, however the attitude towards using was
omitted as it shows a weak predictor of either actual system usage
or behavioral intention (Venkatesh & Davis, 2000). DOI offers
insights into how an innovation among users is diffused over time
(Rogers, 1983). The model which is similar to TAM has been
adopted by researchers to explain on the diffusion of IT adoption.
Based on the relative time of adoption, the study lists five categories of adoption. They can be classified as late majority and laggards, innovators, early adopters early majority (Rogers, 1995).
The innovators are risk takers and thus more likely accept new
products and services. Gatignon and Robertson (1985) explained
that the innovators are highly educated, have higher income,
young, more socially mobile, have favorable attitudes towards risks
and shows greater social participation. Studies by Serenko (2008)
indicated that the user’s readiness for innovation adoption is
impacted by different personal traits. Individual with a higher
degree of personal innovativeness for example are anticipated to
be more confident on new technologies (Lewis, Agarwal, &
Sambamurthy, 2003). Additionally, the theory consists of perceived characteristics of innovation which could be used to verify
the adoption rate (Lu, Yao, & Yu, 2005). The elements are compatibility, relative advantage, trialability, complexity, and
200
Fig. 1. Research model of mobile learning adoption. Notes: PEOU = perceived ease of use; PU = perceived usefulness; PIIT = personal innovativeness in information
technology; SI = social influence. Adapted from Lu et al. (2005)
observability (Chong & Ooi, 2008; Rogers, 1995). Agarwal and
Prasad (1998) further added that the characteristics here are key
influence on the adoption behavior.
Deriving from the current studies, since each of the proposed
traditional frameworks have some restrictions, the research
decided to adopt TAM as the foundation model since Kuo and
Yen (2009) mentioned that the model is able to predict different
IT utilization. However, one major limitation is that the model only
considers two constructs in the overall estimation of technology
and did not include the psychological science perspective. Based
on the notion that consumer’s learning habits are often associated
with the surrounding and the characteristics of an individual (Tan,
Ooi, Chong, & Hew, 2014), the study decided to incorporate two
additional constructs namely personal innovativeness in information technology (PIIT) and social influences (SI) for the technology
under study with TAM. The study also decided to adopt usage
intention than actual intention in the TAM since usage intention
is related to the actual usage (Hill, Smith, & Mann, 1987). Similar
sentiment by Fishbein and Ajzen (1975) indicated the strong correlation between intention and actual usage, thus the adoption of
usage intention is reasonable in this context of study. The integrated model is believed to provide a clearer explanation on the
intention to adopt m-learning in Malaysia which currently at the
embryonic level (see Fig. 1).
case, the acceptance of IS is more likely if users find it is easy to
learn (Chong, 2013b; Pikkarainen, Pikkarainen, Karjaluoto, &
Pahnila, 2004). The PEOU of the application system does not necessary guarantee that the application is easy to use but a higher
adoption is likely if the devices is not perceived to be complex.
Kukulska-Hulme (2007) believes that the activity regarding mlearning leads to usability issues as they often takes place on
devices that are not compatible for education use. Likewise,
Maniar, Bennett, Hand, and Allan (2008) stressed that poor screen
resolution and small screen size may hamper m-learning adoption.
Gururajan, McDonald, Gururajan, and Genrich (2007) and Curran
and Huang (2008) both further described that the usage of mdevices can be limited by slow text input facilities, low and unreliable bandwidth, small storage capacity, slow CPU speed, limited
battery life and environment of user. Theoretically, user friendly
features should be incorporated into the m-devices or the overall
usage experience may be less than what have been desired. Tan,
Sim, Ooi, & Kongkiti, 2012) in their study on m-learning in Malaysia further stated that PEOU positively affects PU. The study was in
accordance with most IS literatures (Agarwal & Prasad, 1998). In
conclusion, PEOU of m-learning will greatly affects their attitude
to adopt m-learning as well as the impact of PEOU on PU and the
intention to adopt thus led us to the following hypotheses:
H1. PEOU has a significant and positive association with intention
to adopt m-learning.
3. Hypotheses development
3.1. Perceived ease of use (PEOU)
PEOU is the ‘‘degree to which an individual believes that using a
particular system would be free of physical and mental effort’’
(Davis, 1993, p. 477). PEOU has been recognized in predicting a
diverse acceptance of IT studies and is evidenced by its applicability in m-commerce (Lin & Wang, 2005; Luarn & Lin, 2005; Teo, Tan,
Sim, Ooi, & Kongkiti, 2012), online banking (Lu, Yu, Liu, & Yao,
2003; Tan, Chong, Ooi, & Chong, 2010; Wang, Lin, & Tang, 2003)
and wireless Internet (Lu et al., 2003; Shih & Fang, 2004). According to Rogers (1995) the adoption of innovation will be discouraged if there is complexity in the Information System (IS). In this
H2. PEOU has a significant and positive association with PU.
3.2. Perceived usefulness (PU)
As defined by Mathwick, Malhotra, and Rigdon (2001), PU is
described as the degree to which an individual perceives the adoption of a technology is useful in boosting his or her performance.
Similarly, Lopez-Nicolas, Molina-Castillo, and Bouwman (2008)
defined PU is the level whereby the belief of individuals in which
a system will able to assist a person to carry out a job easier,
quicker and good quality. Thus, PU plays a vital role in determining
the acceptance of innovation (Tan & Teo, 2000). Studies on past literatures have validated PU as an essential factor in the acceptance
of technology (Agarwal & Karahanna, 2000; Hong, Thong, Moon, &
Tam, 2006; Kim, Chuan, & Gupta, 2007; Lee, 2009; Sim, Tan, Ooi, &
Lee, 2011; Venkatesh & Morris, 2000). In fact, Yang (2005) discovered that PU has a greater effect over predicting consumers’ attitudes as compared to PEOU. Research work conducted in the
field of m-learning discovered that individuals will exploit mlearning if they find it useful. A study on learning English in Taiwan
by Tan and Liu (2004) revealed that when compared to the traditional learning method, mobile technology helps to enhance students’ motivation and interest. Similarly the study was also
supported by Hwang et al. (2010) in Taiwan. Learners are able to
access learning material related applications anywhere and anytime due to the portability of m-devices, thus helping learners to
meet their studying objectives according to their learning pace
(Barkhuus & Tashiro, 2010; Malek, Laroussi, & Derycke, 2006). In
terms of convenience, m-learning also helps to minimize input
overhead (Chao & Chen, 2009). Furthermore, students could share
assignments and improve collaboration in the classroom (Liu,
2009). In addition, m-learning according to Markett, Sanchez,
Weber, and Tangney (2006), allows better communication among
learner-learner and learner-instructor. Similar sentiments were
also echoed by Rosario, Giordano, Lucetti, Procissi, and Risi
(2006), whereby m-learning significantly improves the quality of
interaction among students and between students and teachers.
Chen, Chang, and Wang (2008) highlighted on the design of an
m-learning system which could assist students to observe different
birds outdoor, while teachers can send questions to them via the
devices. This led us to the following hypothesis.
H3. PU has a significant and positive association with the intention
to adopt m-learning.
3.3. Personal innovativeness in information technology (PIIT)
Rogers (1995) mentioned that innovative individuals are the
one who could handle uncertainty and also have a better intention to adopt new innovations in IT. On the other hand,
Agarwal and Prasad (1998) mentioned that the immediate influence of how an individual interpret IT is actually linked to them.
They further explained that PIIT is considered as a personality
construct since they only exist in selected individuals (Wood &
Swait, 2002). Lu et al. (2005) found that individual will bound
to have better perceptions on innovation with greater level of
PIIT. Citrin, Sprott, Silverman, and Stem (2000) explained that
personal innovativeness can be classified in two different categories in which is the domain-specific innovativeness and open-processing innovativeness. The open-processing innovativeness
touches on the prediction of general behavior of innovation adoptions such as individual’s intellectual, attitudinal characteristics
and perceptual (Joseph & Vyas, 1984) while the domain-specific
innovation is regarded as the tendency of individual seeking
knowledge on innovation adoption for a particular product
(Gatignon & Robertson, 1985). In general, when deciding the
results, personal innovativeness restrains the impacts of an individual’s decision on his or her perception of mobile adoption,
thus individuals with more innovative is anticipated to generate
a more positive thinking for new IT (Lopez-Nicolas et al., 2008).
According to Wood and Swait (2002), in examining the behavior
adoption of products and services, personal innovativeness has
also been applied. Additionally, numerous past IT literatures indicate that personal innovation as a construct is critical in the
understanding of behavioral intention (Crespo & Rodriguez,
2008), PEOU (Serenko, 2008) and PU (Lu, Liu, Yu, & Wang,
201
2008; Yang, 2005) of new IT adoptions. Lewis et al. (2003) conducted a survey on the staff members to determine the influence
on the individual, institutional and social basis on how they actually have the interaction with IT. The study shows that PIIT has
an important link with PU and PEOU (Lu et al., 2005). Based on
the personal innovativeness literatures, we conclude that individuals who have higher PIIT are risk-taker, thus developing a positive intention to adopt m-learning. Therefore, this lead to the
following three hypotheses:
H4. PIIT has a significant and positive association with PU of
m-learning.
H5. PIIT has a significant and positive association with PEOU of
m-learning.
H6. PIIT has a significant and positive association with intention to
adopt m-learning.
3.4. Social influences (SI)
SI is defined as ‘‘the degree to which an individual perceives
that important others believe he or she should use the new system’’ (Venkatesh, Morris, Davis, & Davis, 2003, p. 451). Although
the original TAM omitted SN, more research was called for in view
of the importance of SN (Davis, 1989). Studies so far have found
that SI has been considered essential in understanding the behavioral to adopt m-commerce (Hong, Thong, Moon, & Tam, 2008;
Khalifa & Cheng, 2002). According to Venkatesh and Davis
(2000), SI’s theoretical underpinnings come from TRA. TRA indicates that SI can be divided into three elements such as subjective
norm (SN), image, and voluntariness. SN is an individual belief if
others think that one should perform a specific behavior (Lu
et al., 2003). The example includes the opinions from a user’s
friends, relatives, family members, superiors, colleagues, peer students and instructors (Miller, Rainer, & Corley, 2003; Tan et al.,
2010). The rational goes that a user may be reluctant to engage a
certain behavior at first but when taking into considerations the
importance of what the referents think, a user may concord to
the behavior (Venkatesh & Morris, 2000). Fishbein and Ajzen
(1975) considered SN as the other’s mandate on an individual
whether to exhibit or not to exhibit a particular behavior. Choi,
Choi, Kim, and Yu (2003) highlighted that SN has the greatest
impact on behavior intention. Image on the other hand is defined
as the perception of improvement of an individual’s status with
the use of an innovation within the social group (Moore &
Benbasat, 1991). Individuals consider owning and using an mdevice such as a mobile phone as a fashion as it could enhance
their images (Lopez-Nicolas et al., 2008). Therefore, social image
is vital to most individuals as the use of m-devices could enhance
their social status in the society and also their self-importance
(Sarker & Wells, 2003). Venkatesh and Davis (2000) commented
that in TAM2, SN and image can influences the cognitive belief of
PU, in spite of the context. Similarly, TAM2 also suggested that
when there is voluntariness in the system adoption, the intention
to adopt is not affected directly by subjective norms (Lu et al.,
2005). The following hypotheses are proposed since there is a
strong empirical support for SI:
H7. SI in the form of subjective norm and image has a significant
and positive association with PU.
and positive association with PEOU.
202
and positive association with intention to adopt m-learning.
3.5. Control variables: gender, age and academic qualifications
Based on a study on mobile entertainment using a response rate
of 89.66 percent of students, Leong, Ooi, Chong, and Lin (2013)
found that age, gender and academic qualification have no confound on mobile entertainment adoption in Malaysia. Similarly in
another study on m-commerce adoption in China and Malaysia,
only age is found to be significant with Malaysian consumers while
gender and education level have no significant (Chong, Chan, & Ooi,
2012). Since the number of studies using gender, age and academic
qualifications have been sparsely limited, it would be interesting to
include all of the above as control variables in our proposed
research framework. We hypothesize that gender, age and academic qualifications will lead to different intention to adopt mlearning.
4. Research methodology
4.1. Sampling and data collection
In accessing the mentioned hypotheses, data was collected from
questionnaire surveyed to mobile users in a private university in
Malaysia. In view that the users consist of diverse groups with different characteristics, they are relevant in this context of m-learning study (Barkhuus & Tashiro, 2010). Moreover, according to
Leong, Hew, Ooi, and Lin (2011) and Yang (2005), younger age
users tend to be more ready to adopt to new technologies. Since
the sample consists of users with different ethnics, cultures and
religions from different parts in Malaysia, Leong, Hew, et al.
(2011) mentioned that the study represented the different multireligion and multi-ethnic of the Malaysia’s population. Thus, the
Malaysian context can be generalized from this finding. The study
adopted convenience sampling in the initial step. There are advantages of using convenience sampling in data collection. First, the
respondents in our convenience samples are most likely to adopt
m-learning due to their education background and influence of
friends and classmates (Yang, 2005). These groups of respondents
are knowledgeable on the topic on m-learning, in which they
would likely to adopt the current undertaken study to suit their
learning purposes. This approach is also consistent with Lee’s
(2006) study on e-learning adoption in Taiwan. Secondly, when
compared to other consumers in the mass market, the convenience
samples representatives have higher probability to be the first
group among consumers to adopt to m-learning. By collecting data
from them, it can help to ensure the validity of sample selection in
this study. With the approval of participating university, the survey
was distributed to various classes in one of the largest private universities in Malaysia. With an estimation of over 12,000 mobile
users, the samples would therefore consist of a balance representation from all of the fourteen states in Malaysia. The purpose of the
study and a short video presentation on what is m-learning (e.g.,
http://www.youtube.com/watch?v=EgGaEbQsKWY) was shown
to the mobile users before the survey was distributed. The technique is similar with m-learning researchers such as Liu et al.
(2010) and Cheon, Lee, Crooks, and Song (2012). In addition, users
were also briefed on the definition of m-learning in the questionnaire whereby the term refers to the extent to which users are willing to adopt m-devices with access to wireless communication
networks as a platform to support learning activities such as
365/24/7 access to learning resources (e.g., lecture materials,
examination papers, e-books) so that learning can take place anytime and anywhere or to participate in collaborative learning with
instructors and classmates inside and outside the classroom whenever appropriate. In order to avoid respond bias in the survey, they
were told not to fill in their names. Out of the 400 surveys handed
out over the period of two weeks, 220 were returned. Out of these
220 samples, 4 samples were discarded due to incompletion and
missing data, thus yielding a total of 216 usable questionnaires,
which cover a response rate of 54.0%. Please see Table 1 for the target respondents.
4.2. Variable measurement
4.2.1. Independent variables
The independent variables for the study were adopted upon the
scope and structure of existing and past literatures wherever possible. There are four independent variables in this study namely,
PU, PEOU, SI and PIIT. There are four to six questions in each of
these variables. Hence, a total of 20 items were developed from
the four factors to measure the intention to adopt m-learning. Likert 5-point scale was used to measure each of the questions with
anchors of ‘‘1’’ = strongly disagree, and ‘‘5’’ = strongly agree. Items
for PEOU were measured using the items developed based on the
study of Davis (1989). For PU and PIIT, we decided to adopt the
construct from those developed and validated by Lu et al. (2005).
SI on the other hand was measured using the four-scale originated
by Wei et al. (2009) for testing m-commerce.
4.2.2. Dependent variable
For measuring the intention toward the m-learning adoption,
four questions were originally developed from Davis (1989,
1993). In order to fit within m-learning context, the questions were
reworded and modified. Each of the questions were also measured
using the five-point Likert scale, whereby ‘‘1’’ represents as
strongly disagree, and ‘‘5’’ represents as strongly agree.
4.3. Data analysis
4.3.1. Sample profile
The demographic details are in the likes of age, the highest level
of academic qualification, gender and marital status are presented
in Table 1. The sample consists of, 38.0% males, and 62.0% females.
In terms of the their age, the results show that majority of the
respondents are between 21 and 25 years (62.5%) while only
0.5% are above 25 years old suggesting that most respondents are
relatively young. Regarding the education level, 63.5% of the
respondents had achieved at least an undergraduate degree or professional qualifications.
4.3.2. Testing of multivariate assumptions and construct validity
Adopting the Structural Equation Modeling (SEM), the research
model as in Fig. 1 was analyzed. In order to conduct the SEM, many
scholars such as Hair, Anderson, Tatham, and Black (1998), Liao,
Table 1
Demographic profile of respondents.
Variables
Frequency
Percentage (%)
Gender
Male
Female
82
134
38.0
62.0
Age
620 years old
21–25
26–30
>30 years old
79
135
1
1
36.6
62.5
0.5
0.5
Education
No College Degree
Diploma
Bachelor
Master
PhD
50
29
135
1
1
23.1
13.4
62.5
0.5
0.5
203
Table 2
Correlation analysis, criterion validity and nomological validity.
**
Correlation is significant at the 0.01 level (2-tailed). Gray cells indicate criterion and nomological validity.
Chen, and Yen (2007), Lin and Lee (2004, 2005), Llorens, Schaufeli,
Bakker, and Salanova (2007), Sit, Ooi, Lin, and Chong (2009) and
Teo, Cheah, et al. (2012), Wu, Chen, and Lin (2007), Yu, Lu, and
Liu (2010) have proposed a two-stage modeling process, whereby
the confirmatory factor analysis (CFA) need to be examined first
before testing the structure model. The analysis of the data was
performed in the following ways (Demirbag, Koh, Tatoglu, &
Zaim, 2006; Fotopoulos & Psomas, 2009).
The assumption of normality underlying the maximum likelihood should first be examined (Kuo, Wu, & Deng, 2009; Lu et al.,
2005; Tan et al., 2014) prior to data analysis. The normality of
the distribution was validated based on skewness and kurtosis as
well as histogram, normal P–P plots and Kolmogorov–Smirnov test
whereas the linearity and homoscedasticity were verified based on
R2 of the matrix scatter plot and scater plot of standardized residual and predicted values respectively. Finally, based on the values
of Variance Inflation Factor (VIF < 10) and Tolerance (>0.1) as well
as the Pearson’s product moment correlation coefficients of less
than 0.90 (Table 2), the problem of multi-collinearity has been
eliminated (Hew & Leong, 2011).
It is a prerequisite to ensure that construct validity was validated before further statistical analyses were to be conducted. To
ensure content validity, we have adapted items from previous
research and let these items reviewed by several academician
who are well-versed in IT adoption. Besides that, criterion and
nomological validity were verified through correlation analysis
(Table 2) whereby the significant level, magnitude and sign of
the correlation coefficient were consistent to the theoretical outcomes hypothesized based on the current literature. In the next
step, we carried out the common method variance as well as the
confirmatory factor analysis (CFA) to examine the measurement
model’s overall fit (Kuo et al., 2009).
4.3.3. Common method variance
In order to address on the common method variance (CMS), the
Harman’s single factor analysis was performed. The largest factor
based on the results contributed only 30.715% of the total variance.
As the percentage is less than 50%, no dominant factor emerged as
a single factor. According to Delerue and Lejeune (2010), due to
CMS, the data would not be significant.
4.3.4. Exploratory factor analysis
Exploratory Factor Analysis (EFA) and varimax rotation was
performed separately on both adoption factors and intention to
adopt m-learning in order to separate the dimensions of each factor (Leong, Ooi, Chong, & Lin, 2011; Leong, Ooi, et al., 2013). Table 3
shows the results of EFA. All the independent and dependent variables have a Cronbach’s alpha value ranges from 0.697 to 0.836 as
in Table 4, which according to Hair et al. (1998), the measurement
of the variables are above the acceptable threshold 0.60 and are
statistically significant at p < 0.01.
4.3.5. The measurement (CFA) model
In this stage the measurement model is examined, whereby
confirmatory factor analysis (CFA) was carried out to the factors
that affect m-learning adoption so as to determine the validity of
the construct. The measurement model based on the study consists
of 12 (i.e. after removing 8 items from the original battery due to
poor factor loadings during the EFA’s validation process) items
looking into the four factors affecting the diffusion of m-learning,
such as SI, PIIT, PU and PEOU. In order to determine the liability,
convergent validity and discriminatory validity of the measurement model, the measurement model is tested using CFA (Leong,
Hew, Ooi, & Lin, 2012). The convergent validity was tested by
examining the factor loadings and the relevant p-values. Kline
(1998) commented that the common rule of all indicator standardized loadings (k) should be significant and surpass 0.50 for acceptability. As summarized in Table 4, the results show that the kvalues for all items were significant and greater than 0.500, the
composite reliability (CR) of each factor was: SI = 0.774;
PIIT = 0.718; PU = 0.813, PEOU = 0.799 and BI = 0.766. As indicated
in Table 4, from the point of view of all cases, the scales are not
only within the satisfactory limits, but the composite reliability
of all latent constructs also surpassed the benchmark of 0.7 as proposed by Molina, Montes, and Ruiz-Moreno (2007). This implies
that the measurement is good. In addition, all the AVE of each factor was greater than 0.5, indicating good convergent validity and
reliability (Table 4).
For the correlations between two constructs we adopted the
confidence intervals (i.e. within two standard errors) to validate
discriminant validity. Based on Table 5, all confidence intervals
do not include the value one (Anderson & Gerbing, 1988). We
summed up that since discriminant validity has been statistically
proven; therefore all items in the research instrument are significantly unique.
As for this research study, the CFA model measure the goodness
of fit using six common measures (Kaynak, 2003; Lee, Ooi, Tan, &
Chong, 2010; Lin & Lee, 2005; Segars & Grover, 1998), such as
‘‘the ratio of v2 statistics to the degree of freedom’’ (df), ‘‘goodness-of-fit index’’ (GFI), ‘‘normed fit index’’ (NFI), ‘‘comparative
fit index’’ (CFI), ‘‘adjusted goodness-of-fit index’’ (AGFI), and the
‘‘root mean square error of approximation’’ (RMSEA). As showed
in Table 6, the observed normed v2 for this model was 1.091 (pvalue = 0.311 > 0.05). This value is less than 3 as confirmed by
Bagozzi and Yi (1988). The rest of the fit indices include the
GFI = 0.970; AGFI = 0.923; CFI = 0.996; NFI = 0.961 which exceeded
the suggested cut-off level of 0.9 (Bagozzi & Yi, 1988), whereas the
RMSEA = 0.021 is less than the cut-off level of 0.08 as recommended by Browne and Cudeck (1993). The conclusion of these
results has proved that the CFA model represent a well fit with
the data collected (Hair et al., 1998) (Note: Root Mean Squared
Residual (RMR) = 0.025). Table 6 shows the results of measurement
model.
4.3.6. The structural model
Table 7 showed the overall results of the SEM analysis. According to Anderson and Gerbing (1988), values greater than 0.90 are
desirable for GFI, CFI, AGFI and NFI while values less than 0.08
for RMSR and RMSEA are acceptable. As ascertained by the normed
Chi-square index (v2/df = 0.844; p-value = 0.849 > 0.05) together
204
Table 3
Summary of measurement results for factors influencing the adoption of m-learning.
Factors
Number of items
Factor loadings
Mean
S.D
Social influences (SI)
Personal innovativeness in information technology (PIIT)
Perceived of usefulness (PU)
Perceived ease of use (PEOU)
Behavioral intention to use mobile learning (BI)
3
5
4
5
4
0.666–0.852
0.673–0.768
0.747–0.859
0.693–0.811
0.600–0.811
3.469
3.743
3.713
3.563
3.796
0.754
0.730
0.714
0.699
0.695
Note: Three items from social influences (SI), namely S14, SI5 and SI6 were deleted due to low factor loadings.
with other indices (GFI = 0.965; AGFI = 0.931; CFI = 1.000;
NFI = 0.947; RMSEA = 0.000), the structural model fits well. As
can be seen in Table 7, the model fit indices went beyond their normal acceptable levels, recommending that the portrayed structural
model shows an acceptable model fit to the data (Bagozzi & Yi,
1988; Browne & Cudeck, 1993; Lin & Lee, 2005; Sit et al., 2009).
The significant causal associations supported in this research are
showed in Fig. 2.
4.3.7. Hypothesis testing
The hypotheses were examined using SEM method in this
hypotheses testing stage. The analysis results affirm the positive
relationship between PEOU, PU and intention to adopt m-learning
(BI) (H1 to H3). The result also imply positive associations between
PEOU, PU and consumer intention to adopt m-learning (BI); that is,
the higher consumer PEOU (b = 0.355, critical ratio = 3.608,
p < 0.01) is, the more positive their intention toward using mlearning is. There is also significant relationship between PEOU
and PU (b = 0.339, critical ratio = 4.297, p < 0.01). Moreover, the
association between PU and BI is significant (b = 0.553, critical
ratio = 3.821, p < 0.01). Previous TAM studies that examined and
evaluated the strong consistent relationships between PEOU, PU
and BI were supported with these findings (Davis, 1989; Yang,
2005). In comparison with other relationships (e.g. H1: PEOU ?
BI; H2: PEOU ? PU and H3: PU ? BI), PU has stronger impacts
on BI (b = 0.553, critical ratio = 3.821, p < 0.01) (see Table 8).
Hypotheses of H1, H2 and H3 were hence supported.
Onto the effects of PIIT (H4 to H6) on m-learning adoption decision, PIIT can predict consumers’ PEOU (H5). The associations
between these variables are also shown in Fig. 2 that provided b
values for each hypothesis. Thus, the hypothesis of H5 (PIIT ? PEOU) was supported.
The positive relationship between SI, and PU (b = 0.155, critical
ratio = 2.885, p < 0.01) was supported by the empirical data. However, the association between SI and PEOU is insignificant
(b = 0.061, critical ratio = 0.731, p > 0.05) and the relationship
between SI and BI is also insignificant (b = 0.068, critical
ratio = 1.124, p > 0.05). However, previous TAM research that
examines and evaluates the strong consistent associations
between SI, PEOU, PU and BI (Lu et al., 2005) was not supported
in these findings. Thus, H7 was supported (see Table 8) and the
hypotheses of H8 and H9 were not supported. The structural model
is able to explain 53.4% of the variance in BI. In order to assess the
effect size, we have calculated Cohen’s f2 statistics using the following equation.
f2 ¼
R2
1 R2
ð1Þ
The f2 = 0.399 indicated a large effect size (Cohen, 1988).
PIIT and PEOU were found to have significant indirect effects on BI.
Likewise, PIIT was found to have significant indirect effect on PU.
Nevertheless, there were no significant indirect effects of SI and
PEOU on PU as well as direct effect of SI on BI.
4.3.9. Mediating effects with Baron-Kenny’s technique
The mediating effects between the independent (IV), mediator
(M) and dependent (DV) variables were tested based on BaronKenny’s (1986) test statistics using the following equation.
ba bb
ffi
t ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2
ba sb þ b2b s2a þ s2a s2b
ð2Þ
where s is the standard errors and b is the coefficient of path.
The following criteria were used in examining the mediating
effects:
(a) the relationships in IV ? M, M ? DV and IV ? DV must be
significant
(b) in the IV + M ? DV model, full mediating effect occurs when
only M is significant but not IV and partial mediating effect
happens when both M and IV are significant.
As can be seen from Table 11, only PU partially mediates the
association between PEOU and BI. There were no significant mediation effects on other linkages due to non-fulfillment of the above
mentioned mediation criteria.
4.3.10. Artificial Neural Networks (ANN) and sensitivity analysis
ANN is ‘‘a massively parallel distributed processor made up of
simple processing units, which have a neural propensity for storing
experimental knowledge and making it available for use’’ (Haykin,
2001, p. 2). ANN is very similar to the human brain in the sense
that knowledge is gathered through learning or training process
and stored by ‘‘interneuron connection strengths known as synaptic weights’’ (IBM SPSS Neural Networks 22, 2013, p. 1). ANN
entails neurons or nodes in input, hidden and output layers with
connections strengths known as synaptic weights that are adjusted
through iterative learning process (Leong, Hew, Tan, & Ooi, 2013).
There are two types of learning process (Chen & Du, 2009):
(a) Supervised learning
Patterns of known inputs and outputs are exposed to the ANN.
The function of the supervised learner is to forecast the output
value of the function for any input which is valid after going
through a certain number of training examples. Supervised ANN
has been commonly deployed for solving problems of function
approximation and classification (Chen & Du, 2009).
(b) Unsupervised learning
4.3.8. Indirect effects
Based on the p-values of the two-tailed significance of the biascorrected percentile method with bootstrapping (Tables 9 and 10),
Patterns are fed to the ANN in feature values without a priori
output. Unsupervised ANN has been successfully utilized for data
205
mining and classification purposes (Chen & Du, 2009). Self-Organizing Map (SOM) and Adaptive Resonance Theory (ART) are two
popular examples of ANN using unsupervised learning.
The usefulness of ANN includes:
(a) Non linearity: The output from the computational neuron can
be linear or non linear. The ANN is formed by interconnection of non-linear neurons which by itself is non linear.
(b) Adaptive learning: ANN is able to identify the relationship
between the different examples being presented to it without requiring a prior model.
(c) Self-organization: ANN is capable of distributing knowledge
in the entire network structure.
(d) Fault tolerance: ANN is able to handle noise or variability and
even if any of the elements of the network fails, it does not
affect its functionality.
A feed forward-back propagation (FFBP) neural network is an
ANN which utilizes a supervised learning process with a feed forward algorithm for prediction and classification and is assumed
as an advanced multiple regression analysis (MRA) able of handling complex and non-linear relationships. In fact, ANN has been
used in various disciplines including marketing, retail, insurance,
telecommunications, operations management, banking and
finance (Smith & Gupta, 2000) but there is dearth in its use in IT
and IS studies (Shmueli & Koppius, 2010).
Due to the fact that SEM is only able to detect linear relationships, there are possibilities that it may over-simplify the complexities of the decision to adopt a technology. Therefore, in order to
address this slack, an SEM–ANN approach was taken since ANN
is capable of identifying both linear and non-linear relationships
with requiring any distribution assumptions such as normality, linearity or homoscedasticity (Leong, Hew, et al., 2013). Moreover,
ANN is more robust and can provide higher prediction accuracy
and has even out-performed other conventional regression techniques such as MRA, SEM or Multiple Discriminant Analysis
(MDA) (Morris, Greer, Hughes, & Clark, 2004). Anyway, due to
the ‘‘black-box’’ operational nature, ANN is not appropriate for
testing hypothesis of causal relationships (Lee, Leong, Hew, &
Ooi, 2013). Hence, to address this limitation, significant variables
or determinants from SEM are used as the input units for the
ANN. Hence in this study, three ANN models were deployed
(Fig. 3).
In order to avoid over-fitting, a ten-fold cross-validation was
engaged with 90% of the data used for training and the rest of
the hold-out data for testing purpose (Sim, Tan, Wong, Ooi, &
Table 4
Instrument reliability and validity.
Latent
constructs
Indicator
Standardized
loading (k)
Cronbach’s
alpha (a)
Composite
reliability (CR**)
Average variance
extracted (AVE*)
SI
I will use mobile learning if my colleagues use it (SI3)
Friend’s suggestion and recommendation will affect my decision
to use mobile learning (SI1)
Family/relatives have influence on my decision to use mobile
learning (SI2)
0.547
0.633
0.697
0.774
0.548
PIIT
I like to experiment with new ways of doing things (PIIT2)
I like to take a chance (PIIT3)
0.758
0.738
0.715
0.718
0.560
PU
Using mobile learning increases my productivity (PU2).
Using mobile learning would enhance my effectiveness in my
daily work (PU3)
Overall, I would find mobile learning to be advantageous (PU4)
Using mobile learning will enable me to accomplish the required
more quickly (PU1)
0.798
0.786
0.836
0.813
0.524
PEOU
Mobile learning is understandable and clear (PEOU4)
I find it easy to do what I want to do in mobile learning (PEOU3)
It would be easy for me to become skillful at using the mobile
learning system (PEOU5)
0.573
0.867
0.805
0.712
0.799
0.576
BI
I intend to increase my use of the mobile learning systems in the
future (BI3)
I believe my interest towards mobile learning will increase in the
future (BI4)
I would use mobile learning for my personal needs (BI1)
0.730
0.764
0.766
0.522
0.972
0.590
0.701
0.754
0.681
Notes:
*
AVE = Rki2/n (i = 1. . .n, k = standardized factor loadings, i = observed variables).
**
CR = (Rki)2/[(Rki)2 + Rdi)], (ki = standardized factor loadings, i = observed variables, di = error variance); Items PIIT1, PIIT4, PIIT5, PEOU1, PEOU2 and BI2 were deleted due
to poor standardized loadings. SI = social influences; PIIT = personal innovativeness in information technology; PU = perceived usefulness; PEOU = perceived ease of use;
BI = behavioral intention to use mobile learning.
Table 5
Confidence intervals (2 standard errors) for correlation estimate between constructs.
SI
PIIT
PEOU
PU
BI
SI
PIIT
PEOU
PU
BI
–
(0.050,
(0.195,
(0.014,
(0.048,
(0.041, 0.204)
–
(0.066, 0.384)
(0.025, 0.357)
(0.080, 0.276)
(0.113, 0.094)
(0.046, 0.270)
–
(0.069, 0.344)
(0.246, 0.525)
(0.007, 0.191)
(0.016, 0.233)
(0.064, 0.320)
–
(0.334, 0.593)
(0.023, 0.164)
(0.046, 0.160)
(0.203, 0.443)
(0.296, 0.526)
–
0.248)
0.162)
0.353)
0.342)
Notes: SI = social influences; PIIT = personal innovativeness in information technology; PU = perceived usefulness;
PEOU = perceived ease of use; BI = behavioral intention to use mobile learning.
206
Table 6
Measures of the model fit – measurement model.
Table 8
Hypothesis testing results.
Goodness of fit measures
Recommended
value
CFA model
(result)
Path
v2 test statistics/df
63.00a
P0.90a
P0.90a
P0.90a
P0.90a
60.08b
1.091
0.970
0.923
0.996
0.961
0.021
GFI (goodness-of-fit index)
AGFI (adjusted goodness-of-fit index)
CFI (comparative fit index)
NFI (normed fit index)
RMSEA (root mean square error of
approximation)
Sources:
a
Bagozzi and Yi (1988).
b
Browne and Cudeck (1993).
Table 7
Measures of the model fit – structural model.
Goodness of fit measures
Recommended
value
Structural model
(result)
v2 test statistics/df
63.00a
P0.90a
P0.90a
P0.90a
P0.90a
60.08b
0.844
0.965
0.931
1.000
0.947
0.000
GFI (goodness-of-fit index)
AGFI (adjusted goodness-of-fit index)
CFI (comparative fit index)
NFI (normed fit index)
RMSEA (root mean square error of
approximation)
Sources:
a
Anderson and Gerbing (1988).
b
Browne and Cudeck (1993).
Hew, 2013). A FFBP multi layer perceptron (MLP) in SPSS 21 with
sigmoid activation function for hidden and output layers was utilized. The number of hidden units was generated automatically
and the Root Mean Square of Errors (RMSE) was calculated
together with the normalized importance in the sensitivity analysis. Table 12 shows that the models are able to provide very accurate prediction based on the very small RMSE values.
The relevance of the variables is validated based on the number
of non-zero synaptic weights connected to the relevant hidden
units (Table 13). Hence, all factors are indeed relevant in predicting
the dependent variable. The normalized importance is the ratio of
Estimate
Std.
error
Critical
ratio
pValue
Remarks
A ? BI
0.022
0.019
1.146
0.252
G ? BI
0.015
0.016
0.974
0.330
AQ ? BI
PEOU ? BI
PEOU ? PU
PU ? BI
PIIT ? PU
0.067
0.355
0.339
0.553
0.125
0.029
0.098
0.079
0.145
0.082
2.339
3.608
4.297
3.821
1.525
0.019⁄
0.000⁄⁄
0.000⁄⁄
0.000⁄⁄
0.127
PIIT ? PEOU
PIIT ? BI
0.633
0.007
0.123
0.122
5.137
0.060
0.000⁄⁄
0.952
SI ? PU
SI ? PEOU
0.155
0.061
0.054
0.084
2.885
0.731
0.004⁄⁄
0.465
SI ? BI
0.068
0.060
1.124
0.261
Not
supported
Not
supported
Supported
Supported
Supported
Supported
Not
supported
Supported
Not
supported
Supported
Not
supported
Not
supported
Note:
PEOU = perceived ease of use; PU = perceived usefulness; PIIT = personal innovativeness in information technology.
SI = social influences; A = age; AQ = academic qualifications; G = gender;
The values in bold indicates the relationship is supported.
**
p < 0.01.
*
p < 0.05.
the relative importance of each variable with its highest relative
importance and expressed in percentage form (Table 14).
Since only significant linear factors from the SEM analysis were
used as the input units of the ANN models, only linear relationships
were detected. The relative strengths of the causal relationships
were examined based on the normalized importance in the sensitivity analysis. PEOU was found to be the key determinant in predicting PU followed by SI. In terms of BI prediction, PU constituted
the most effect followed by PEOU and AQ.
5. Discussion
The research paper aims to study on the consumer’s intention to
adopt m-learning in Malaysia. The proposed hypotheses deriving
Fig. 2. SEM path analysis results. Notes: PEOU = perceived ease of use; PU = perceived usefulness; PIIT = personal innovativeness in information technology; SI = social
influence; G = gender; A = age; AQ = academic qualifications; p < 0.01; p < 0.05.
207
Table 9
Indirect effects.
PU
BI
PIIT
SI
PEOU
.215
.413
.021
.119
.000
.187
Notes: SI = social influences; PIIT = personal innovativeness in information technology; PU = perceived usefulness.
The values in bold indicates the relationship is significant.
Table 10
Two-tailed significance (bias-corrected) percentile method.
PU
BI
PIIT
SI
PEOU
.003
.001
.482
.077
–
.006
Notes: SI = social influences; PIIT = personal innovativeness in information technology; PU = perceived usefulness.
The values in bold indicates the relationship is significant.
from TAM and psychological science constructs were tested using a
hybrid structural equation modeling–artificial neural networks
approach. The following subsections provide more detailed discussions on the findings.
5.1. Relationships between PU, PEOU and intention to adopt mobile
learning
PU has the highest normalized importance and also showed significant relationship with the users’ intention to adopt m-learning
in our study. The finding is align with prior researches on m-learning acceptance such as Viehland and Marshall (2005), Lu and
Viehland (2008) and Alzaza and Yaakub (2010a). Similarly, this
study also supports other researchers indifferent settings such as
online banking (Azam, 2007; Pikkarainen et al., 2004; Tan et al.,
2010), m-commerce adoption (Teo, Tan et al., 2012) and wireless
internet services (Lu et al., 2005) in which consumers are likely
to adopt if they find the particular system to be beneficial. Thus,
when compared to old fashion learning methods, if users can feel
the usefulness aspects of m-learning, this will lead to the increase
of adoption rate. For example, m-learning helps to improve the
quality of interaction among students, enhancing the learning process and the convenience of accessing information 365/24/7. The
study in Malaysia by Alzaza and Yaakub (2010b) revealed that students would like to adopt m-learning for obtaining exam results,
course registration, calendar, schedule services, library services,
campus facilities, etc. Service providers, universities or mobile
manufacturers, should highly concentrate on this elements of usefulness in their advertisement campaigns. The findings in this
study also show that PEOU with 91.3% normalized importance is
positively related to the acceptance of m-learning. The same
results collaborates was reached by Alzaza and Yaakub (2010a),
Huang, Lin, and Chuang (2007), Li, Qi, and Shu (2008) and Lu and
Viehland (2008) in which PEOU is seen as a key inhibitor on the
intention to adopt m-learning. In a study by Alzaza and Yaakub
(2010b) in Malaysia, they commented that the biggest limitation
adopting m-learning is the slow data exchange with network. As
m-learning is still very new in Malaysia, service providers and educational institutions can accelerate the adoption rate by considering the determinants like wireless bandwidth, screen size,
storage capacity, screen resolution, battery life span, convenience
to input data and designing learning software that are compatible
for mobile usage. Regarding the impact of PEOU over PU, the
results indicated that PEOU has the highest normalized importance
in comparison to SI which has 41.7% normalized importance and is
a significant predictor of system’s usefulness. The results were consistent with a past study on m-learning adoption using 401 students in Malaysia by Tan et al. (2012) whereby they stressed that
user-friendly interface will lead to more younger consumers in
accepting m-learning in which case when the device is easy to
use this will alter the perception of usefulness and thus likelihood
of acceptance. Therefore, mobile manufacturers should improve on
building m-devices that are compatible for m-learning purposes.
5.2. Relationships between PIIT, PU, PEOU and intention to adopt
mobile learning
PIIT shows a significant influence in predicting the ease of use of
m-learning. The study was supported by prior scholars such as
Yang (2005), Lu et al. (2005) and Parveen and Sulaiman (2008).
Consistent with many studies on personal innovativeness, individuals with greater level of PIIT have more courage and higher personality values and social-economy status, thus when
considering any technology adoption they are likely to develop
positive feeling towards PEOU as opposed to individuals with
lower PIIT. In this instance, during the introduction stage of mlearning, service provider should concentrate on the ease of use
of m-learning services to innovative users (Liu et al., 2010). The
absent of significant relationship between PIIT and PU is also con-
Table 11
Mediating effects.
a
Variable
Baron and Kenny test statistic (t)
p-Value
IV
M
DV
PEOU
PU
BI
2.731
0.003
PIIT
PU
BI
0.599 0.274 PIIT
PEOU
BI
3.235 0.001 BI
SI
SI
PU
PEOU
BI
0.637
0.506
0.262
0.306
Beta coefficients of mobile learning structural model
IV ? M
M ? DV
ba (sa)
bb (sb)
0.268
(0.085)
0.051
(0.081)
0.957
(0.159)
0.038
(0.057)
0.074
(0.092)
0.871
(0.152)
0.528
(0.144)
0.315
(0.081)
0.700
(0.181)
0.406
(0.106)
Mediating level
IV ? DV
IV + M ? DV (M controlled)
IV
M
0.314
0.416
0.692
Partial
0.010
0.011
0.599
None
0.231
0.194
0.303
None
0.027
0.034
0.146
0.075
0.775
None
0.415
None
Notes: IV = independent variable; M = mediator; DV = dependent variable; = not applicable because some relationships are not significant; p < 0.05; p < 0.01; p < 0.001;
Baron-Kenny test statistic (Baron & Kenny, 1986); standard error is shown in bracket.
SI = social influences; PIIT = personal innovativeness in information technology; PU = perceived usefulness; PEOU = perceived ease of use; BI = behavioral intention to use
mobile learning.
a
208
sistent with a study in China on the acceptance of wireless mobile
data services (Lu et al., 2008). While the users perceive that the use
of m-learning is easy to operate the adoption will not necessary
lead to better results in their education. On the relationship
between PIIT and intention, the results contradict with a study
by Crespo and Rodriguez (2008) and Liu et al. (2010) but corroborate the results by Lu et al. (2005). As the majority of the respondents are bachelor degree holder, the decision to adopt mlearning is not based on braveness or curiosity but on rationality
and usefulness of m-learning (Wong, Lee, Lim, Chua, & Tan, 2012).
5.3. Relationships between SI, PU, PEOU and intention to adopt mobile
learning
SI with normalized importance of 41.7% was found to have significant relationship in predicting PU. The finding corroborates
result by Sim, Kong, Lee, Tan, and Teo (2012) in Malaysia, whereby
SI in the form of subjective norm and image influence the evaluation of whether a certain technology devices is easy to use. SI however was found to have no significant relationship with PEOU. The
finding is contrary with the study by Lu et al. (2005). One possible
reason is that the since most mobile users are young, they have the
necessary skills to use m-devices and therefore the influence of
friends and family members are not needed in the navigation mdevices for learning purposes. On the relationship between SI
and the intention, the findings suggest that social factors do not
exert any influence on the intention to adopt m-learning. The study
was similar with Lu et al. (2005) but opposed by research conducted by Venkatesh et al. (2003), Teo and Pok (2003), Hsu, Lu,
and Hsu (2007) and Lopez-Nicolas et al. (2008). With 62.5% of
respondents holds a bachelor degree, the users are not easily influenced by social status or social pressure from friends, family or colleagues and their decision on whether to adopt m-learning is based
purely on rationality and logic thinking.
novel perspective in examining the key determinants of m-learning acceptance while enriching and closing up the knowledge gaps
that exist in the current body of knowledge. The research model is
able to explicate 53.4% of the variance in behavioral intention to
use m-learning. The Cohen’s f-square effect size of 0.399 also indicates that the findings are highly significant. Surely, the findings
from this hybrid approach will further enhance the current literature on m-learning acceptance. Besides that the rigorous and
advanced statistical techniques such as Harman’s single factor
analysis, criterion and nomological validity analysis, discriminant
validity analysis with confidence interval with two standard errors,
indirect effects with two-tailed significance of the bias-corrected
percentile method using bootstrapping, neural network sensitivity
Model A
5.4. Control variables
The study indicates that there were no confounding effects of
age and gender in the intention to adopt m-learning. Therefore, a
one solution marketing strategies can be adopted irrespective of
age and gender of users. Academic qualifications (i.e. normalized
importance = 31.5%) however revealed that there is significant
impact on the intention. Thus, practitioners may need to take this
into consideration in their business strategies.
6. Implications
Among a few theoretical contributions of this study are as followed. The study extended previous researches conducted in other
developing countries. This study therefore is able to provide
greater insights from the Malaysia’s perspective in understanding
the intention to adopt m-learning from the multi-cultural and
multi-religion perspective. Consistent with many past IT studies,
PU and PEOU were found to be significant in adopting new technology and in this case m-learning thus helping to validate the
model from the emerging market perspective. The study has also
added two new constructs namely SI and PIIT and the control variables of gender, age and academic qualifications to the original
TAM. We believe that the extended TAM provides better contributions on the adoption of m-learning than TAM alone. Methodologically, the use of a hybrid SEM–ANN approach has contributed to
the understanding of m-learning acceptance specifically and IT
adoption generally as this approach is a balanced method of complementing the linear and non-compensatory SEM model with the
non-linear and non-compensatory ANN model. This has provided a
Model B
Model C
Fig. 3. Examples of three ANN models.
209
Table 12
RMSE values for training and testing routines.
Artificial neural networks
Model C
Input neurons: PEOU, PU, AQ
Output neuron: BI
Model B
Input neurons: PEOU, SI
Output neuron: PU
Model A
Input neuron: PIIT
Output neuron: PEOU
Training
Testing
Training
Testing
Training
Testing
ANN1
ANN2
ANN3
ANN4
ANN5
ANN6
ANN7
ANN8
ANN9
ANN10
0.1261
0.1263
0.1236
0.1316
0.1257
0.1209
0.1221
0.1209
0.1259
0.1203
0.1146
0.0989
0.1393
0.1301
0.0975
0.1402
0.1387
0.1403
0.1027
0.1479
0.1103
0.1206
0.1058
0.1066
0.1114
0.1048
0.1053
0.1187
0.1088
0.1084
0.1063
0.0890
0.0937
0.0684
0.1143
0.1281
0.1367
0.1032
0.1033
0.0994
0.1036
0.0996
0.1120
0.0966
0.1207
0.1035
0.0969
0.1069
0.1023
0.0976
0.1172
0.1232
0.1140
0.0992
0.1090
0.1151
0.1087
0.0900
0.0608
0.0784
Mean RMSE
Standard deviation
0.1243
0.0035
0.1250
0.0196
0.1101
0.0055
0.1042
0.0194
0.1040
0.0076
0.1016
0.0197
Notes: PIIT = Personal innovativeness in information technology; SI = social influence; PEOU = perceived ease of use; PU = perceived usefulness; AQ = academic qualifications;
Table 13
Relevance of variables based on non-zero synaptic weight with hidden neurons.
Model
Predictor variable
A
PIIT
B
PEOU
SI
C
PEOU
PU
AQ
ANN1
p
ANN2
p
ANN3
p
ANN4
p
ANN5
p
ANN6
p
ANN7
p
ANN8
p
ANN9
p
ANN10
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
p
Note: Dependent variable = BI (behavioral intention to use mobile learning); PIIT = personal innovativeness in information technology; SI = social influence; PEOU = perceived
p
ease of use; PU = perceived usefulness; AQ = academic qualifications;
indicates at least one non-zero synaptic weight was connected to the hidden neurons.
Table 14
Neural networks sensitivity analysis.
Model A
Output neuron: PEOU
Relative importance
Model B
Output neuron: PU
Relative importance
Model C
Output neuron: BI
Relative importance
PIIT
PEOU
SI
PEOU
PU
AQ
ANN1
ANN2
ANN3
ANN4
ANN5
ANN6
ANN7
ANN8
ANN9
ANN10
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.617
0.885
0.709
0.712
0.640
0.635
0.671
0.969
0.575
0.642
0.383
0.115
0.291
0.288
0.360
0.365
0.329
0.031
0.425
0.358
0.360
0.393
0.411
0.415
0.417
0.425
0.409
0.535
0.413
0.320
0.521
0.456
0.476
0.461
0.257
0.468
0.429
0.430
0.468
0.522
0.119
0.151
0.113
0.125
0.326
0.107
0.162
0.035
0.119
0.158
Average relative importance
Normalized importance (%)
1.000
100.0
0.706
100.0
0.295
41.7
0.410
91.3
0.449
100.0
0.142
31.5
Notes: Dependent variable = BI (behavioral intention to use mobile learning); PIIT = personal innovativeness in information technology; SI = social influence; PEOU = perceived ease of use; PU = perceived usefulness; AQ = academic qualifications.
analysis and Baron–Kenny’s analysis on mediating effects may contribute as a point-of-reference for other researchers.
On the practical contributions of this study, mobile manufacturers, service providers, educational institutions and even governments may want to apply the suggested strategies from this
study to increase the number of users. Considering that PU, PEOU,
SI and PIIT are critical in this study, they should be considered so
that users will adopt m-learning in their daily life. As PU has signif-
icant effect in influencing the intention to adopt m-learning,
governments and service providers should concentrate on the usefulness of m-learning in their advertising campaigns. The promotion and advertisement can emphasize on the benefits of mlearning when traveling (Geddes, 2004) or for distance education
and lifelong learning (Jin, 2009). Educational institutions can also
develop contents which might perceive as valuable to students
such as to check examination results, course registration, etc. when
210
compared to other traditional learning methods. Likewise since
PEOU is significant towards both BI and PU it is also important
for mobile manufacturers to consider designs m-devices which is
compatible for m-learning purposes, so that users would require
less effort in operating the devices. For example, mobile manufacturers can focus on producing clearer display screen or improving
on the navigation keys. Emphasis should also be given by educational institutions to develop user friendly-applications such as
those of with richer presentation and better reading layout since
most students learn-on-the-fly. Users will develop positive feeling
on the usefulness of m-devices if there are no usability issues or
when users do not require a lot of effort in using m-learning, thus
will adopt m-learning. On the other hand since SI also plays a critical role in influencing PU, government and educational institutions can play their part by promoting at websites frequented by
youths such as Twitter, YouTube, Facebook and Friendster. They
can promote on the usefulness of the products to users who will
then disseminate to their friends and family members. This would
lead to the perception of usefulness among potential users. Lastly,
knowing that PIIT has significant effect on PEOU, service providers
and mobile manufacturers should cast their advertisement in different market segments based on the difficulty levels of using mlearning so that they can satisfy different needs and wants thus
raising the level of m-learning acceptance.
7. Limitation and future studies
The research has following limitations, which should be dealt
with for future research. Firstly, the demographic profiles under
investigation are a group of young mobile users within a university. The imbalance samples cannot represent the views held by
different segments in Malaysia since different results could be
obtained using different age group. For example, Venkataram and
Price (1990) indicated that older consumers have lower innovativeness index thus, may not be significant for this study. Therefore, it might be helpful if future research can examine different
age group and provide comparison. Secondly, we collected data
in Malaysia and thus precluding the generalization of the findings
to other countries. To more fully reflect on the intention to adopt
m-learning, we recommend a cross-country comparison studies
for future research. Thirdly, the research only focused on education
sector, thus it cannot be viewed as a total representation of the
total adoption of m-learning in Malaysia. The research should be
repeated to in other industries such as finance, insurance, manufacturing and banking to find out whether the result is similar.
Fourthly, it is not possible to consider all factors in the model
under investigation. Other factors involving technology adoption
that were excluded, such as government supports, price of devices,
perceived enjoyment and so on, may be important from the Malaysian context. Huang, Rau, Salvendy, Gaoa, and Zhou (2011) concluded that the perception of risks influence the adoption of IT.
Given that, future studies may include the additional variables to
better understanding the intention to adopt m-learning. Lastly, in
terms of PEOU, the study only captured whether m-learning is perceived to be easy or complicated to use without considering the
real challenges present when adopting m-devices for learning such
as the size of keyboards, the ambient light, the screen size and etc.
Therefore, the next stage to advance this model is to examine on
the determinants of PEOU in m-learning such as the m-devices
design and issues related to usability.
8. Conclusion
To sum up, the purpose of this research was to investigate on
the factors that influence the intention to adopt m-learning in
Malaysia using a hybrid SEM-ANN approach. To this end, we found
that TAM significantly influences the intention to adopt
m-learning. However the results for PIIT, SI and the moderating
variables of age, gender and academic shows mixed results thus
requires further investigation. The results provide valuable information and advice not only to the mobile manufacturers, service
providers, and educational institutions but also to governments
especially when advancing the development of marketing services
and when implementing business strategies.
Acknowledgement
This is a revised and extended version of a paper co-authored by
Tan, G.W.H. et al. (2013) presented at the 2013 Conference on
Medical Innovation and Computing Service (MICS), Tainan, Taiwan,
3rd–4th August 2013.
References
Agarwal, R., & Karahanna, E. (2000). Time flies when you’re having fun: Cognitive
absorption and beliefs about information technology usage. Management
Information Systems Quarterly, 24(4), 665–694.
Agarwal, R., & Prasad, J. (1998). A conceptual and operational definition of personal
innovativeness in the domain of information technology. Information Systems
Research, 9(2), 204–215.
Ajzen, I. (1991). The theory of planned behavior. Organisational Behaviour and
Human Decision Processes, 50(2), 179–211.
Ajzen, I., & Madden, T. J. (1986). Prediction of goal-directed behavior: Attitudes,
intentions, and perceived behavioral control. Journal of Experimental Social
Psychology, 22, 453–474.
Alzaza, N. S., & Yaakub, A. R. (2010a). Student’s mobile information prototype
(SMIP) for the higher education environment. In Proceedings knowledge
management fifth international conference 2010, University Utara Malaysia.
Alzaza, N. S., & Yaakub, A. R. (2010b). Students’ awareness and requirements of
mobile learning services among Malaysian students in the higher education
environment. In Proceedings knowledge management 5th international conference,
University Utara Malaysia.
Amaral, A. M. (2006). Learning and adapting to mobile technologies: An intraorganizational assessment. IEEE International Conference on Management of
Innovation and Technology.
Ambient
Insight
(2010).
<http://www.ambientinsight.com/Reports/
MobileLearning.aspx>. Accessed 01.08.10.
Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A
review and recommended two-step approach. Psychological Bulletin, 103(3),
411–423.
Attewell, J. (2005). Mobile learning: Reaching hard-to-reach learners and bridging the
digital device, methods and technologies for learning. Southampton: WIT Press.
Azam, M. S. (2007). Internet adoption and usage in Bangladesh. Japanese Journal of
Administrative Science, 20(1), 43–54.
Bagozzi, R. P., & Yi, Y. (1988). On the evaluation of structural equation model. Journal
of Academy of Marketing Science, 16(1), 74–94.
Barkhuus, L., & Tashiro, J. (2010). Student socialization in the age of Facebook. In
Paper presented in the 28th international conference on human factors in
computing systems (CHI 2010) conference on human factors in computing
systems (pp. 133–142), Atlanta, Georgia, USA, 10–15 April 2010.
Baron, N. S., & Kenny, D. A. (1986). The moderator–mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical
considerations. Journal of Personality and Social Psychology, 51, 1173–1182.
Berri, J., Benlamri, R., & Atif, Y. (2006). Ontology-based framework for contextaware mobile learning. International workshop context-aware mobile
computing’06, Vancouver, Canada.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A.
Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136–162).
Newbury Park, CA: Sage.
Chao, P. Y., & Chen, G. D. (2009). Augmenting paper-based learning with mobile
phones. Interacting with Computers, 21(3), 173–185.
Chen, G. D., Chang, C. K., & Wang, C. Y. (2008). Ubiquitous learning website: Scaffold
learners by mobile devices with information-aware techniques. Computers &
Education, 50(1), 77–90.
Chen, W. S., & Du, Y. K. (2009). Using neural networks and data mining techniques
for the financial distress prediction model. Expert Systems with Applications,
36(2), 4075–4086.
Cheon, J., Lee, S., Crooks, S. M., & Song, J. (2012). An investigation of mobile learning
readiness in higher education based on the theory of planned behavior.
Computers & Education, 59(3), 1054–1064.
Choi, H., Choi, M. S., Kim, J. W., & Yu, H. S. (2003). An empirical study on the
adoption of information appliances with the focus on interactive TV. Telematics
and Informatics, 20(2), 161–183.
Chong, A. Y. L. (2013a). Predicting m-commerce adoption determinants: A neural
network approach. Expert Systems with Applications, 40(2), 523–530.
Chong, A. Y. L. (2013b). A two-staged SEM-Neural network approach for
understanding and predicting the determinants of m-commerce adoption.
Expert Systems with Applications, 40(4), 1240–1247.
Chong, A. Y. L., Chan, F. T. S., & Ooi, K. B. (2012). Predicting consumers’ decision to
adopt mobile commerce: Cross country empirical examination between China
and Malaysia. Decision Support Systems, 53(1), 34–43.
Chong, A. Y. L., & Ooi, K. B. (2008). Collaborative commerce in supply chain
management: A study of adoption status in Malaysian electrical and electronic
industry. Journal of Applied Sciences, 8(21), 3836–3844.
Citrin, A. V., Sprott, D. E., Silverman, S. N., & Stem, D. E. Jr., (2000). Adoption of
internet shopping: The role of consumer innovativeness. Industrial Management
and Data System, 100(7), 294–300.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Lawrence Erlbaum.
Cook, J., Bradley, C., Lance, J., Smith, C., & Haynes, R. (2007). Generating learning
contexts with mobile devices. In N. Pachler (Eds.), Mobile learning: Towards a
research agenda. WLE Centre Occasional Papers in Work-based Learning 1 (pp.
55–74), WLE Centre, London.
Corbeil, J. R., & Valdes-Corbeil, M. E. (2007). Are you ready for mobile learning?
Educause Quarterly, 30(2), 51–58.
Crespo, A. H., & Rodriguez, I. A. R. D. B. (2008). Explaining B2C e-commerce
acceptance: An integrative model based on the framework by Gatignon and
Robertson. Interacting with Computers, 20(2), 212–224.
Cummings, J., Merrill, A., & Borrelli, S. (2010). The use of handheld mobile
devices: Their impact and implications for library services. Library Hi Tech,
28(1), 22–40.
Curran, K., & Huang, W. (2008). A multilingual mobile university information portal
designed to cater for visiting overseas students. Interactive Technology and Smart
Education, 5(1), 4–28.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use and user acceptance
of Information Technology. Management Information Systems Quarterly, 13(3),
319–340.
Davis, F. D. (1993). User acceptance of information technology: System
characteristics, user perceptions and behavioral impacts. International Journal
of Man-Machine Studies, 38(3), 475–487.
Delerue, H., & Lejeune, A. (2010). Job mobility restriction mechanisms and
appropriability in organizations: The mediating role of secrecy and lead time.
Technovation, 30(5–6), 359–366.
Demirbag, M., Koh, S. C. L., Tatoglu, E., & Zaim, S. (2006). TQM and market
orientation’s impact on SMEs’ performance. Industrial Management & Data
Systems, 106(8), 1206–1228.
Donnelly, K. (2009). Learning on the move: How m-learning could transform
training and development. Development and Learning in Organizations, 23(4),
8–11.
Fetaji, M., & Fetaji, B. (2008). Universities go mobile – Case study experiment in
using mobile devices. In Proceedings of the ITI, 2008, 30th international conference
on information technology interfaces, Croatia.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intentions and behaviour: An
introduction to theory and research. Reading: Addison-Wesley.
Fotopoulos, C. B., & Psomas, E. L. (2009). The impact of ‘‘soft’’ and ‘‘hard’’ TQM
elements on quality management results. International Journal of Quality &
Reliability Management, 26(2), 150–163.
Fusilier, M., & Durlabhji, S. (2005). An exploration of student internet use in India.
Campus-wide Information Systems, 22(4), 233–246.
Gatignon, H., & Robertson, T. S. (1985). A propositional inventory for new diffusion
research. Journal of Consumer Research, 11, 849–867.
Geddes, S. J. (2004). Mobile Learning in the 21st Century: Benefit for Learners,
Knowledge Tree e-Journal. <http://knowledgetree.flexiblelearning.net.au/
edition06/download/Geddes.pdf>. Accessed 01.08.10.
Gil, D., & Pettersson, O. (2010). Providing flexibility in learning activities systems by
exploiting the multiple roles of mobile devices. Sixth IEEE international
conference on wireless, mobile and ubiquitous technologies in education.
Gururajan, R., McDonald, J., Gururajan, V., & Genrich, R. (2007). An exploratory
qualitative study to determine factors influencing the adoption of mobile
learning for tertiary education. The International Journal of Technology,
Knowledge and Society, 2(6), 33–40.
Hahn, J. (2008). Mobile learning for the twenty-first century librarian. Reference
Services Review, 36(3), 272–288.
Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data
analysis (5th ed.). Englewood Cliffs, NJ: Prentice Hall.
Haykin, S. (2001). Neural networks: A comprehensive foundation. Upper Saddle River,
NJ: Prentice Hall.
Herrington, A., & Herrington, J. (2007). Authentic mobile learning in higher education.
<http://www.aare.edu.au/07pap/her07131.pdf> Accessed 01.09.10.
Hew, T. S., & Leong, L. Y. (2011). An empirical analysis of Malaysian pre-university
students’ ICT competency gender differences. International Journal of Network
and Mobile Technologies, 2(1), 15–29.
Hill, T., Smith, N. D., & Mann, M. F. (1987). Role of efficacy expectations in predicting
the decision to use advanced technologies: A case of computers. Journal of
Applied Psychology, 72(2), 307–318.
Hong, S. J., Thong, J. Y. L., Moon, J. Y., & Tam, K. Y. (2006). Understanding continued
information technology usage behavior: A comparison of three models in the
context of mobile internet. Decision Support Systems, 42(3), 1819–1832.
Hong, S. J., Thong, J. Y. L., Moon, J. Y., & Tam, K. Y. (2008). Understanding the
behavior of mobile data services consumers. Information Systems Frontier, 10(4),
431–445.
211
Hsu, T. Y., Ke, H. R., & Yang, W. P. (2006). Knowledge-based mobile learning
framework for museums. The Electronic Library, 24(5), 635–648.
Hsu, C. L., Lu, H. P., & Hsu, H. H. (2007). Adoption of the mobile internet: An
empirical study of multimedia message service (MMS). Omega, 35(6), 715–726.
Hu, S. K., Lu, M. T., & Tzeng, G. H. (2014). Exploring smart phone improvements
based on a hybrid MCDM model. Expert Systems with Applications, 41(9),
4401–4413.
Huang, J. H., Lin, Y. R., & Chuang, S. T. (2007). Elucidating user behavior of mobile
learning: A perspective of the extended Technology Acceptance Model. The
Electronic Library Emerald Group Publishing Limited, 25(5), 586–599.
Huang, D. L., Rau, P. L. R., Salvendy, G., Gaoa, F., & Zhou, J. (2011). Factors affecting
perception of information security and their impacts on IT adoption and
security practices. International Journal of Human-Computer Studies, 69(12),
870–883.
Hwang, G. J., Wu, P. H., Zhuang, Y. Y., Kuo, W. L., & Huang, Y. M. (2010). An
investigation on students’ cognitive load and learning achievements for
participating in a local culture mobile learning activity. In Sixth IEEE
international conference on wireless, mobile and ubiquitous technologies in
education.
IBM SPSS Neural Networks 22, IBM Corporation 1989–2013.
Jin, Y. (2009). Research of one mobile learning system. International conference on
wireless networks and information systems.
Joseph, A., & Vyas, S. J. (1984). Concurrent validity of a measure of innovative
cognitive style. Journal of the Academy of Marketing Sciences, 12(2), 159–175.
Kaynak, H. (2003). The relationship between total quality management practices
and their effects on firm performance. Journal of Operations Management, 21(4),
405–435.
Khalifa, M., & Cheng, S. (2002). Adoption of mobile commerce: Role of exposure. In
Proceedings of the 35th annual international management, Hawaii.
Kim, H. W., Chuan, H. C., & Gupta, S. (2007). Value-based adoption of mobile
internet: An empirical investigation. Decision Support Systems, 43(1), 111–126.
Kline, R. B. (1998). Principles and practices of Structural Equation Modeling. New York,
NY: Guilford Press.
Koike, H., Akama, K., Chiba, M., Ishikawa, T., & Miura, K. (2005). Developing an elearning system which enhances students’ academic motivation. ACM Special
Interest Group on University and College Computing Services (SIGUCCS) Fall 2005
Conference.
Kukulska-Hulme, A. (2007). Mobile usability in educational contexts: What have we
learnt? The International Review of Research in Open and Distance Learning, 8(2),
1–16.
Kuo, Y. F., Wu, C. M., & Deng, W. J. (2009). The relationships among service quality,
perceived value, customer satisfaction, and post-purchase intention in mobile
value-added services. Computers in Human Behavior, 25(4), 887–896.
Kuo, Y. F., & Yen, S. N. (2009). Towards an understanding of the behavioral intention
to use 3G mobile value-added services. Computers in Human Behavior, 25(1),
103–110.
Lee, Y. C. (2006). An empirical investigation into factors influencing the adoption of
an e-learning system. Online Information Review, 30(5), 517–541.
Lee, M. C. (2009). Predicting and explaining the adoption of online trading: An
empirical study in Taiwan. Decision Support Systems, 47(2), 133–142.
Lee, V. H., Leong, L. Y., Hew, T. S., & Ooi, K. B. (2013). Knowledge management: A key
determinant in advancing technological innovation? Journal of Knowledge
Management, 17(6), 848–872.
Lee, V. H., Ooi, K. B., Tan, B. I., & Chong, A. Y. L. (2010). A structural analysis of the
relationship between TQM practices and product innovation. Asian Journal of
Technology Innovation, 18(1), 73–96.
Legris, P., Ingham, J., & Collerette, P. (2003). Why do people use information
technology? A critical review of the technology acceptance model. Information
& Management, 40(3), 191–204.
Lehner, F., & Nosekabel, H. (2002). The role of mobile devices in e-learning – First
experience with an e-learning environment. IEEE international workshop on
wireless and mobile technologies in education. Los Alamitos, CA: IEEE Computer
Society, pp. 103–106.
Leong, L. Y., Hew, T. S., Ooi, K. B., & Lin, B. (2011). Influence of gender and English
proficiency on Facebook mobile adoption. International Journal of Mobile
Communications, 9(5), 495–521.
Leong, L. Y., Hew, T. S., Ooi, K. B., & Lin, B. (2012). The determinants of customer
loyalty in Malaysian mobile telecommunication services: A structural analysis.
International Journal of Services, Economics and Management, 4(3), 209–235.
Leong, L. Y., Hew, T. S., Tan, G. W. H., & Ooi, K. B. (2013). Predicting the determinants
of the NFC-enabled mobile credit card acceptance: A neural networks approach.
Expert Systems with Applications, 40(14), 5604–5620.
Leong, L. Y., Ooi, K. B., Chong, A. Y. L., & Lin, B. (2011). Influence of individual
characteristics, perceived usefulness and ease of use on mobile entertainment
adoption. International Journal Mobile Communications, 9(4), 359–382.
Leong, L. Y., Ooi, K. B., Chong, A. Y. L., & Lin, B. (2013). Modeling the stimulators of
the behavioral intention to use mobile entertainment: Does gender really
matter? Computers in Human Behavior, 29(5), 2109–2121.
Lewis, W., Agarwal, R., & Sambamurthy, V. (2003). Sources of influence on beliefs
about Information Technology use: An empirical study of knowledge workers.
Management Information Systems Quarterly, 27(4), 657–679.
Li, Y. Q., Qi, J. Y., & Shu, H. Y. (2008). Review of relationships among variables in
TAM. Tsinghua Science and Technology, 13(3), 273–278.
Liao, C., Chen, J. L., & Yen, D. C. (2007). Theory of planning behavior (TPB) and
customer satisfaction in the continued use of e-service: An integrated model.
Computers in Human Behavior, 23(6), 2804–2822.
212
Lin, H. H., & Wang, Y. S. (2005). Predicting consumer intention to use mobile
commerce in Taiwan. In Proceeding of the International Conference on Mobile
Business (ICMB’05), Sydney, Australia.
Lin, H. F., & Lee, G. G. (2004). Perceptions of senior managers toward knowledgesharing behavior. Management Decision, 42(1), 108–125.
Lin, H. F., & Lee, G. G. (2005). Impact of organizational learning and knowledge
management factors on e-business adoption. Management Decision, 43(2),
171–188.
Liu, X. (2009). Applied research on the mobile learning in foreign language. Second
international conference on education technology and training.
Liu, Y., Li, H., & Carlsson, C. (2010). Factors driving the adoption of m-learning: An
empirical study. Computers & Education, 55(3), 1211–1219.
Llorens, S., Schaufeli, W., Bakker, A., & Salanova, M. (2007). Does a positive gain
spiral of resources, efficacy beliefs and engagement exist? Computers in Human
Behavior, 23(1), 825–841.
Lopez-Nicolas, A., Molina-Castillo, F. J., & Bouwman, H. (2008). An assessment of
advanced mobile services acceptance: Contributions from TAM and diffusion
theory models. Information & Management, 45(6), 359–364.
Lu, X., & Viehland, D. (2008). Factors influencing the adoption of mobile learning.
19th Australasian conference on information systems.
Lu, J., Liu, C., Yu, C. S., & Wang, K. (2008). Determinants of accepting wireless mobile
data services in China. Information & Management, 45(1), 52–64.
Lu, J., Yao, J. E., & Yu, C. S. (2005). Personal innovativeness, social influences and
adoption of wireless Internet services via mobile technology. Journal of Strategic
Information Systems, 14(3), 245–268.
Lu, J., Yu, C. S., Liu, C., & Yao, J. E. (2003). Technology acceptance model for wireless
internet. Internet Research, 13(3), 206–222.
Luarn, P., & Lin, H. H. (2005). Toward an understanding of the behavioral intention
to use mobile banking. Computer in Human Behaviour, 21(6), 873–891.
Magni, M., Taylor, M. S., & Venkatesh, V. (2010). ‘To play or not to play’: A crosstemporal investigation using hedonic and instrumental perspectives to explain
user intentions to explore a technology. International Journal of Human–
Computer Studies, 68(9), 572–588.
Malaysian Communications and Multimedia Commission (2010). Statistics and
achievements
<http://www.skmm.gov.my/
index.php?c=public&v=art_view&art_id=36,2010> Accessed 01.11.10.
Malek, J., Laroussi, M., & Derycke, A. (2006). A middleware for adapting context to
mobile and collaborative learning. In Proceedings of the fourth annual IEEE
international conference on Pervasive Computing and Communications Workshops
(PERCOMW’06).
Maniar, N., Bennett, E., Hand, S., & Allan, G. (2008). The effect of mobile phone
screen size on video based learning. Journal of Software, 3(4), 51–61.
Markett, A., Sanchez, I. A., Weber, S., & Tangney, B. (2006). Using short message
service to encourage interactivity in the classroom. Computer and Education,
46(3), 280–293.
Mathwick, A., Malhotra, N. K., & Rigdon, E. (2001). The effect of dynamic retail
experiences of value: An internet and catalog comparison. Journal of Retailing,
78(1), 51–60.
Miller, M. D., Rainer, R. K., & Corley, J. K. (2003). Predictors of engagement and
participation in an on-line course. Online Journal of Distance Learning
Administration, 6(1), 1–13.
Molina, L. M., Montes, J. L., & Ruiz-Moreno, A. (2007). Relationship between quality
management practices and knowledge transfer. Journal of Operations
Management, 25(3), 682–701.
Moore, G., & Benbasat, I. (1991). Development of an instrument to measure the
perceptions of adopting an information technology innovation. Information
Systems Research, 2(3), 192–222.
Morris, S. A., Greer, T. H., Hughes, C., & Clark, W. J. (2004). Prediction of CASE
adoption: A neural network approach. Industrial Management & Data Systems,
104(2), 129–135.
Mulliah, A. (2006). Mobile Collaborative Peer Learning in an Apprenticeship
Context. Unpublished Master thesis, University of Alberta, Edmonton, Alberta,
Canada.
O’Cass, A., & Fenech, T. (2003). Web retailing adoption: Exploring the nature of
internet users’ web retailing adoption. Journal of Retailing and Consumer Services,
10(2), 81–94.
Parveen, A., & Sulaiman, A. (2008). Technology complexity, personal innovativeness
and intention to use wireless internet using mobile devices in Malaysia.
International Review of Business Research Papers, 4(5), 1–10.
Pikkarainen, T., Pikkarainen, K., Karjaluoto, H., & Pahnila, S. (2004). Consumer
acceptance of online banking: An extension of the Technology Acceptance
Model. Internet Research, 14(3), 224–235.
Poonsri, V. (2008). Mobile learning: Major challenges for engineering education.
38th ASEE/IEEE frontiers in education conference, Saratoga Springs, NY.
Pozzi, A. (2007). The impact of m-learning in school contexts: An ‘‘inclusive’’
perspective. Lecture Notes in Computer Science, 4556, 748–755.
Rogers, E. M. (1983). Diffusion of innovations. New York: The Free Press.
Rogers, E. M. (1995). Diffusion of innovations (4th ed.). New York: The Free Press.
Rogoski, R. R. (2005). Wireless by design. Health Management Technology, 26(1), 1–7.
Rosario, G. G., Giordano, S., Lucetti, S., Procissi, G., & Risi, G. (2006). A pilot e-learning
experience on telecommunications: Towards the e-campus concept. In
Proceedings of the Advanced International Conference on Telecommunications
and International Conference on Internet and Web Applications and Services (AICT/
ICIW).
Sarker, S., & Wells, P. J. (2003). Understanding: Mobile handheld device use and
adoption. Communications of the ACM, 46(12), 35–40.
Savil-Smith, A., Attewell, J., & Stead, G. (2006). Mobile learning in practice: Piloting a
mobile learning teachers’ toolkit in Further Education Colleges. London: Learning
and Skills Network.
Schwiderski-Grosche, S., & Knospe, H. (2002). Secure mobile commerce. Special Issue
of the IEE Electronics and Communication Engineering Journal on Security for
Mobility, 14(5), 228–238.
Segars, A. H., & Grover, V. (1998). Strategic information systems planning success:
An investigation of the construct and its measurement. MIS Quarterly, 22(2),
139–163.
Serenko, A. (2008). A model of user adoption of interface agents for email
notification. Interacting with Computers, 20(4–5), 461–472.
Sharma, S. K., & Kitchens, F. L. (2004). Web services architecture for m-learning.
Electronic Journal on e-Learning, 2(1), 203–216.
Sharples, M. (2007). Big issues in mobile learning: Report of a workshop by the
Kaleidoscope Network of Excellence Mobile Learning Initiative. Nottingham, UK:
University of Nottingham, Learning Sciences Research Institute.
Shih, Y. Y., & Fang, K. (2004). The use of a decomposed theory of planned behaviour
to study internet banking in Taiwan. Internet Research, 14(3), 213–223.
Shmueli, G., & Koppius, O. R. (2010). Predictive analytics in information systems
research. MIS Quarterly, 35(3), 553–572.
Sim, J. J., Tan, G. W. H., Wong, J. C. J., Ooi, K. B., & Hew, T. S. (2013). Understanding
and predicting the motivators of mobile music acceptance-A multi-stage MRAArtificial Neural Network approach. Telematics and informatics, <http://
dx.doi.org/10.1016/j.tele.2013.11.005>.
Sim, J. J., Kong, F. M., Lee, V. H., Tan, G. W. H., & Teo, A. C. (2012). Determining factors
affecting broadband services adoption? An empirical analysis of Malaysian
consumers. International Journal of Services, Economics and Management, 4(3),
236–251.
Sim, J. J., Tan, G. W. H., Ooi, K. B., & Lee, V. H. (2011). Exploring the individual
characteristics on the adoption of broadband: An empirical analysis.
International Journal of Network and Mobile Technologies, 2(1), 1–14.
Sit, W. Y., Ooi, K. B., Lin, B., & Chong, A. Y. L. (2009). TQM and customer satisfaction
in Malaysia’s service sector. Industrial Management & Data Systems, 109(7),
957–975.
Smith, K. A., & Gupta, J. N. D. (2000). Neural networks in business: Techniques and
applications for the operations researcher. Computers & Operations Research,
27(2000), 1023–1044.
Tan, H. T., & Liu, T. Y. (2004). The mobile-based interactive learning environment
(mobile) and a case study for assisting elementary school English learning. In
Proceedings of the IEEE International Conference of Advanced Learning Technologies
(ICALT’ 04).
Tan, G. W. H., Chong, C. K., Ooi, K. B., & Chong, A. Y. L. (2010). The adoption of online
banking in Malaysia: An empirical analysis. International Journal of Business and
Management Science, 3(2), 169–193.
Tan, G. W. H., Ooi, K. B., Chong, S. C., & Hew, T. S. (2014). NFC mobile credit card: The
next frontier of mobile payment? Telematics and Informatics, 31(2), 292–307.
Tan, G. W. H., Sim, J. J., Ooi, K. B., & Kongkiti, P. (2012). Determinants of mobile
learning adoption: An empirical analysis. Journal of Computer Information
Systems, 52(3), 82–91.
Tan, M., & Teo, T. S. H. (2000). Factors influencing the adoption of internet banking.
Journal of the Association for Information Systems, 1(1), 1–42.
Teo, A. C., Cheah, C. M., Leong, L. Y., Hew, T. S., & Shum, Y. L. (2012). What matters
most in mobile payment acceptance? A structural analysis. International Journal
of Network and Mobile Technologies, 3(3), 49–69.
Teo, T. S. H., & Pok, S. H. (2003). Adoption of WAP-enabled mobile phones among
internet users. Omega, 31(6), 483–498.
Teo, A. C., Tan, G. W. H., Cheah, C. M., Ooi, K. B., & Yew, K. T. (2012). Can the
demographic and subjective norms influence the adoption of mobile banking?
International Journal of Mobile Communications, 10(6), 578–597.
Venkataram, M. P., & Price, L. L. (1990). Differentiating between cognitive and
sensory innovativeness: Concepts, measurement, and implications. Journal of
Business Research, 20(1), 293–315.
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the Technology
Acceptance Model: Four longitudinal field studies. Management Science, 46(2),
186–204.
Venkatesh, V., & Morris, M. G. (2000). Why don’t men ever stop to ask for
directions? Gender, social influence, and their role in technology acceptance
and usage behavior. MIS Quarterly, 24(1), 115–139.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of
information technology: Toward a unified view. Management Information
Systems Quarterly, 27(3), 425–478.
Viehland, A., & Marshall, J. (2005). The effectiveness of PDAs for enhancing
collaboration in m-learning. In Proceedings of the 11th Americas conference on
information systems, Omaha, Nebraska.
Walsh, A. (2009). Text messaging (SMS) and libraries. Library Hi Tech News, 26(8),
9–11.
Wang, Y. S., Lin, Y. M., & Tang, T. I. (2003). Determinants of user acceptance of
internet banking: An empirical study. International Journal of Service Industry
Management, 14(5), 501–519.
Wei, T. T., Marthandan, G., Chong, A. Y. L., Ooi, K. B., & Arumugam, S. (2009). What
drives Malaysian m-commerce adoption: An empirical analysis. Industrial
Management & Data Systems, 109(3), 370–388.
Wong, C. C., & Hiew, P. L. (2005). Factors influencing the adoption of mobile
entertainment: Empirical evidence from a Malaysian survey. In IEEE
Computer Society International Conference on Mobile Business (ICMB’05), Los
Alamitos, CA.
Wong, C. H., Lee, H. S., Lim, Y. H., Chua, B. H., & Tan, G. W. H. (2012). Predicting the
consumers’ intention to adopt mobile-shopping: An emerging market
perspective. International Journal of Network and Mobile Technologies, 3(3),
24–39.
Wood, S. L., & Swait, J. (2002). Psychological indicators of innovation adoption:
Cross-classification based on need for cognition and need for change. Journal of
Consumer Psychology, 12(1), 1–13.
213
Wu, J. H., Chen, Y. C., & Lin, L. M. (2007). Empirical evaluation of the revised
end user computing acceptance model. Computers in Human Behavior, 23(1),
162–174.
Yang, K. C. C. (2005). Exploring factors affecting the adoption of mobile commerce in
Singapore. Telematics and Informatics, 22(3), 257–277.
Yu, T. K., Lu, L. C., & Liu, T. F. (2010). Exploring factors that influence knowledge
sharing behavior via weblogs. Computers in Human Behavior, 26(1), 32–41.

Predicting the drivers of behavioral intention to use

Transcription

Similar documents

www.nycacc.org - Animal Care Centers of NYC

Deborra-lee Furness launches #AdoptChange

The Netherlands Nutrition Centre

Adoption - Rackcdn.com

Helping Classmates Understand Adoption

Small Shelter, Big Adoption Increases: “The Little Shelter

PREPARE FOR YOUR AIRBRUSH TAN: AFTER

Adoption Services - Children`s Home Society of Florida

Al Roker to Cook Up Excitement for Adoption in Cleveland

Our Lady of the Mountains