How Consumers Assess Products Features?

: A Case Study of Product

Features of Smartphone

WEE-KHENG TAN*, YI-DER YEH, SHIN-JIA CHEN, YU-CHENG LIN, CHIA-YU KUO
Department of Information & Electronic Commerce
Kainan University
No. 1, Kainan Road, Luchu, Taoyuan County
TAIWAN
*tanwk@mail.knu.edu.tw http://www.knu.edu.tw


Abstract: - Assessing product features is often a multiple criteria decision-making problem. Consumers may
mentally link related product features to simplify the decision making process. This study tests the above idea
by using the case study of consumers assessing the product features of smartphones. Our analysis revealed that
this is the case and bi-directional relationship may exist between product features. This has implications for
new technology features such as NFC and its associated application (e-wallet). New product features need to
entrench themselves as a collective part of the product by locking (linking) themselves to other existing and
more established features. Other implications arising from the analysis are also discussed.

Key-Words: - smartphone, product features, multiple criteria decision-making problem, DEMATEL

1 Introduction
Consumer decision making is always a central
theme in marketing research. Studies (e.g. [1]) have
shown that consumers follow a two-stage strategy to
making choices. Consumers include some brands
which they are aware of (the awareness set) in their
consideration set and make comparison within the
latter set to finalize their decision. Having a product
in the consideration set is often a precondition for
final purchase choice [2]. A products inclusion may
depend on its features. Products with features absent
in other competitors can increase its appeal [3].
However, there is also argument that new features
do not always improve product evaluation [4].
Many studies aim to find product features that
customers feel important or desirable. After these
features have been identified, suggestions on how to
improve and market the products can then be made.
Quite often, studies such as those using the Analytic
Hierarchy Process (AHP) method, assume that the
features are independent of one another. However,
everyday experience suggests otherwise.
Real-life decision making is often multiple
criteria decision-making (MCDM) problem. Human
can make poor choice if the decision is complex and
involves many attributes [5]. This study suggests
that consumers may mentally link related product
features to simplify the decision making process. In
other word, customers perception of some of the
product features is built and based on a collection of
perceptions of other inter-linked features. This idea
is not far-fetched as feedback loops exist in many
causal processes [6]. Hence, the recursive model,
which is frequently used in many system adoption
studies, may be restrictive for our purpose.
Pursuing the above ideas further and getting the
mental map of consumers is not just a matter of
academic interest. It also has important practical
implications. It helps in designing and marketing
next generation products, in specific, allowing
manufacturers and marketers to focus more efforts
on the cause features (those which affect others)
than the effect features (those affected by others).
This study uses a case study of consumers
assessing the smartphone features to shed light on
the above issues. This case study is appropriate for
theoretical and practical reasons. Its nature
contributes theoretically to the lack of research on
the relationship between feature preferences [4],
especially in the context of MCDM problem.
Consumers consider many product features
simultaneously when choosing smartphones. Hence,
it is a MCDM problem. Given the array of complex
features, it is unreasonable to treat these features as
independent of one another. Hence, consumers
simplifying their decision making process by
associating some of the product features perception
to a collective set of perceptions of other inter-
linked features is possible. The smartphone industry
is also of great practical importance. Consumers are
adopting smartphones at a remarkable pace and see
smartphones as personal expressions of their
lifestyle [7]. Smartphones have short life-cycles.
Competition between manufacturers is fierce and
Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 131
new smartphone features are introduced at a
dazzling pace to attract buyers. Nokia and Apple
introduced Near Field Communication (NFC) and
Siri software respectively in their latest models.
This study uses the Decision Making Trial and
Evaluation Laboratory (DEMATEL) method [8] as
the analytical tool. A MCDM method, DEMATEL
can discover the cause and effect relations between
factors and allow one to visualize its structural
model [9]. Since the case study is a MCDM problem
and the objective is to get the consumers
assessment mental picture of how some smartphone
features are built on the collective perception of
other features, it is appropriate to apply this method
to construct the cause-effect structural model.
This paper is structured as follows: relevant
literatures are first presented, follows by research
method, results, discussion and conclusion.


2 Literature Review
Many academic researchers and practitioners have
investigated issues related to smartphone usage.
Study [10] found that personal innovativeness,
perceived enjoyment and smartphone satisfaction
positively impact compulsive usage of smartphones.
Determining which smartphone features are
attractive and important in purchase decisions is
also a common research area. Using two criteria
(user-related criteria and product-related criteria),
[11] used the MCDM approach to evaluate mobile
phone options in respect to the users preference
order. The product-related criteria comprises: basic
requirements (e.g. reasonable price, standard part
used), physical characteristics (e.g. weight,
dimension) and technical features (e.g. talk time,
safety standards). The sub-criteria of user-related
criteria are: functionality (ease of use), brand choice
and customer excitement (e.g. games, ringing
tones).
Study [12] investigated diversity in smartphone
usage along four dimensions: user interactions,
application use, network traffic and energy drain.
Using AHP and Data Envelopment Analysis
method, [13] compared smartphones using 19
attributes: dimensions, weights, standby time, talk
time, memory, ROM, expansion slot, Infrared,
Bluetooth, WiFi, GPRS, WAP, Java application,
MP3, types of messaging, digital camera resolution,
screen resolution, colour display and price.
Users overall satisfaction with smartphone is
higher for devices that are of a specific size and
weight and equipped with the latest technological
advancements [14]. The same report also mentioned
that 42% of the owners received free mobile phone
when subscribing to a wireless service (the price
plan), and touch screen-only smartphones generate
higher satisfaction with ease of operation than
QWERTY-only based devices.
Based on their needs and interests, to enhance
user experience [14], and partly for enjoyment [15],
smartphone users install applications of their liking.
Two-thirds of them download games and social
networking applications, more than one-half use
entertainment-oriented applications, while 52%
download travel software [14]. Hence, there is huge
market opportunity for mobile software and
contents.
A successful smartphone brand should engage
users in an emotional experience [16]. In that aspect,
Apple has created a very strong marketing strategy
for its iPhone [17], and ranks highest among
manufacturers of smartphones in customer
satisfaction, followed by HTC [14].
Smartphone manufacturers often introduce new
features to attract buyers. One recent move is to add
NFC to the phone (such as the recent Nokia launch).
NFC allows for wireless exchange of data between
devices within a radius of about 10cm. NFC can be
used as e-wallet [18], enabling mobile micro-
payments at physical stores and transport services.


3 Research Method - DEMATEL
DEMATEL is very suitable for studying and
analysing complicated and intertwined problems. It
classifies factors of the problem into cause group
(factors that have more effect on others) and effect
group (factors that receive more influence from
others). It also identifies the interdependence among
factors. Using digraphs, the structure of complicated
causal relationships can be visualized through an
intelligible cause-effect structural model [19].
DEMATELs procedure is as follows:
Step 1 - With n factors and H respondents, each
respondent states the degree a factor i affects factor j
with score ranging from 0 as no influence to 4 as
very high influence. Using nxn answer matrix X
k

with 1 k H, the initial direct relation matrix A is:

=
=
H
k
k
ij ij
X
H
a
1
1

(1)

Step 2 - Normalised initial direct-relation matrix
D is obtained by D = A/s where:

=

= =

n
j
n
i
ij
n j
ij
n i
a a s
1 1
1 1
max , max max

(2)

Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 132
Step 3 - A continuous decrease of the indirect
effects of problem along the powers of matrix D, for
example, D
2
, D
3
, , D

guarantees convergent
solutions to the matrix inversion similar to an
absorbing Markov chain matrix. The n x n total
relation matrix T is obtained as in (3) as m
approaches and I is an n x n identity matrix.

T = D + D
2
+ + D
m
= D(I D)
-1

(3)

Step 4 - The sum of rows r and the sum of
columns c of matrix T is obtained as in (4).

[ ] [ ]
'
1
1
'
1
1
1
1
,
xn
n
i
ij
xn
j
nx
n
j
ij nx i
t c C t r R

= =

= =

= =

(4)

The sum r
i
+c
i
represents the total effects both
given and received by factor i. The difference r
i
-c
i
is
the net effect factor i contributes to the problem. If
the difference r
i
-c
i
is positive, factor i is a net causer.
If r
i
-c
i
is negative, factor i is a net receiver. Finally,
a threshold value is set to draw the influence map.


4 Survey and Analysis

4.1 Smartphone Features and Survey Design
Using literature review and product specification
sheets provided by manufacturers, a preliminary list
of common features was obtained. Discussion with
experts and postgraduate students was conducted to
obtain a more manageable list of eleven features.
NFC and its application, e-wallet, were included as
they have received attention recently and appear in
recent product launch. Brief but concise description
of these features were designed, and using
comments from the same group of experts and
postgraduates, ambiguous wordings were re-phrased
for better clarify. The final features are as below:

- price-related features: smartphone price (A1),
price plan offered by telecoms operators (A2);
- technology-related features: operating system
(OS) (B1), availability of NFC (B2), data
transmission method, such as Bluetooth (B3);
- design-related features: screen display such as
width and touch-screen (C1), body design such
as shape, weight and dimension (C2);
- application-related features: availability of e-
wallets (D1), APPs (D2);
- image-related features: brand (E1), fashion or
trendy to use (E2).

The questionnaire comprises three parts. The
first part asked the respondents, on 5-point Likert
scale (1 indicating totally do not understand or
very unimportant to 5 indicating understand very
well or very important) to indicate the extent
they understand each feature and how important
they take into account these features when selecting
smartphones. The second part is the DEMATEL
portion with an 11 by 11 matrix with 110 entries
that need filling up. The respondents were asked to
state the degree a factor i affects factor j through a
score of 0 to 4 with 0 indicating no influence and
4 indicating very high influence. The third part of
the survey asked for basic demographics.
DEMATEL survey is tedious, difficult to
understand and time consuming to complete.
Getting input is more challenging for this study
because the respondents are the much more
numerous general users instead of experts of the
field (which are the subjects in many DEMATEL
studies). Hence, posting survey online to get data
was not adopted. The more time-consuming face-to-
face interview method was used to guide the
respondents to increase survey accuracy.
Convenience sampling method was used. One-
hundred twenty-one (121) valid returns were
received. There was a good mix of respondents
where 72 (59.5%) of the respondents were male,
116 (95.9%) received tertiary education, 89 (73.6%)
were 20-40 year old, 79 (65.3%) were smartphone
users, 24 and 21 of them used HTC smartphone and
iPhone respectively. Among the 42 non-smartphone
users, 19 of them wished to have iPhone, follows by
13 who wished to use HTC smartphone.


4.2 Importance-Understanding Analysis
Table 1 summarizes the average understanding and
importance level of each feature.

Table 1. Mean and S.D. in Degree of Understanding
and Importance of Smartphone Features
Understanding Importance
Mean S.D. Mean S.D.
A1: Price 3.49 1.05 3.85 1.00
A2: Price Plan 3.36 1.08 4.11 0.93
B1: OS 3.36 0.98 3.61 0.99
B2: NFC 2.63 1.14 3.19 0.89
B3: Transmission 3.27 1.11 3.67 0.92
C1: Screen 3.60 0.92 3.98 0.83
C2: Body Design 3.55 0.99 3.96 0.87
D1: e-Wallet 2.94 0.98 2.91 0.98
D2: APPs 3.42 1.12 3.92 0.96
E1: Brand 3.73 0.83 3.68 0.85
E2: Fashion 3.59 0.86 3.34 0.94
Average 3.358 3.656
Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 133

The top 3 features which respondents understood
most, in order of importance, are brand (3.73),
screen (3.60) and fashion (3.59). The 3 least
understood features are NFC (2.63), e-wallet (2.94)
and transmission (3.27). The top 3 most important
features are price plan (4.11), screen (3.98) and
body design (3.96). The 3 least important features
are e-wallet (2.91), NFC (3.19) and fashion (3.34).
The overall understanding and importance mean
are 3.358 and 3.656 respectively. The importance-
understanding matrix grid is shown in Fig. 1 with
two overall means as the dividing lines for the 4
quadrants. NFC and e-wallet features are in the less
important-least understood quadrant. Transmission
is the only feature in the more important-least
understood quadrant. OS and fashion features are
in the less importance-most understood quadrant.
The other 6 features are in the most important-most
understood quadrant.
















Fig. 1: Importance-understanding matrix grid


4.3 DEMATEL Analysis
Total relation matrix T was obtained using (3). The
sum of rows r and sum of columns c of T were then
obtained using (4). The sum r
i
+ c
i
and difference r
i
-
c
i
were calculated and shown as Table 2.

Table 2. r+c and r-c
r + c r - c
Value Rank Value Rank
A1: Price 18.60 1 -0.13 (4)
A2: Price Plan 17.39 4 -0.14 (2)
B1: OS 17.55 3 0.19 4
B2: NFC 15.56 9 0.45 2
B3: Transmission 15.27 10 -0.02 (6)
C1: Screen 17.04 7 -0.14 (2)
C2: Body Design 16.49 8 -0.13 (4)
D1: e-Wallet 13.40 11 0.08 5
D2: APPs 17.24 6 -0.93 (1)
E1: Brand 17.97 2 0.22 3
E2: Fashion 17.25 5 0.54 1
The 3 factors with the highest r
i
+ c
i
value are
price, brand and OS. The top 3 net causer factors (r
i
c
i
>0) are fashion, NFC and brand. The top 3 net
receiver factors (r
i
c
i
<0) are APPs, screen and
price plan. The influence map was obtained and for
better clarify shown separately as Fig. 2 and 3.
Fig. 2: Influence map (bi-directional relationship)

Fig. 3: Influence map (uni-directional relationship)


5 Discussion and Conclusion
As shown in Section 4, consumers assessment of
the product features of smartphone is a MCDM
problem. The influence maps also show that
consumers mentally link features to simplify the
decision making process. Hence, customers
perception of some of the features is built and based
on a collection of perceptions of other inter-linked
features. For example, consumers do not view price
in isolation, their perception of price is closely
associated and inter-linked to other features such as
OS, screen and brand. The results thus justify our
intention to abandon recursive model in this study.
13 14 15 16 17 18 19
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
r - c
r + c
e-Wallet
Transmission
NFC
Body
APP
Price
Brand
OS
Fashion
Screen
Price Plan
13 14 15 16 17 18 19
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
r - c
r + c
e-Wallet
Transmission
NFC
Body
APP
Price
Brand
OS
Fashion
Screen
Price Plan r - c
r + c
e-Wallet
Transmission
NFC
Body
APP
Price
Brand
OS
Fashion
Screen
Price Plan

13 14 15 16 17 18 19
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
r - c
r + c
e-Wallet
Transmission
NFC
Body
APP
Price
Brand
OS
Fashion
Screen
Price
Plan
13 14 15 16 17 18 19
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
r - c
r + c
e-Wallet
Transmission
NFC
Body
APP
Price
Brand
OS
Fashion
Screen
Price
Plan

2.6 2.8 3 3.2 3.4 3.6 3.8 4
2.8
3
3.2
3.4
3.6
3.8
4
4.2
NFC
e-Wallet
Fashion
Transmission
OS
Brand
Screen
Body
Price Plan
Price
APP
Understanding
Importance
2.6 2.8 3 3.2 3.4 3.6 3.8 4
2.8
3
3.2
3.4
3.6
3.8
4
4.2
NFC
e-Wallet
Fashion
Transmission
OS
Brand
Screen
Body
Price Plan
Price
APP
Understanding
Importance
Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 134
Generally, consumers do not view technology-
related feature criteria (OS, data transmission and
NFC) as important as other feature criteria. They
also do not understand NFC and data transmission
as well as OS. Such a perception has an impact on
the much poorer performance of NFC and
transmission methods in terms of total effects given
and received by these two factors. It also causes
NFC and transmission methods to become
standalone features in the influence maps.
Consumers are not able to associate these two
features with other smartphone features. These
observations have several implications. Rapid
introduction of advanced and new technologies
(such as NFC) in the smartphones do not always
win the hearts of the consumers. Consumers must
have some basic awareness of the technologies
(similar to the concept of awareness set) before
they are willing to consider them (consideration
set), buy in, and tying them to other features. The
often discussion of Android and iOS in mass media
and in laymans terms has increased consumers
awareness of OS and enabled it to play a more
important role than the other two technology-related
features in the mental map of customers (the
influence map). Hence, OS has a bi-directional
relationship with price and brand, and influences
APPs uni-directionally.
NFC, being closely associated with e-wallet,
provides further insight on the problems associated
with introducing less-known technology as a
smartphone feature. Its related application, e-wallet,
is the second least understood and least important
feature. NFC and e-wallet are in the less important-
least understood quadrant. Both of them are also
standalone features in the influence map. Even
though academia and engineers often associate NFC
with e-wallet, consumers simply dont think there is
any relationship between them. NFC and e-wallet
are also unable to entrench themselves as an
essential and collective part of the smartphone by
locking (linking) themselves to other existing and
more established features. The NFC-e-wallet pair
provides a useful message to designers and
marketers that it is not enough to introduce new
technologies and their associated applications.
When they are being introduced, manufacturers and
marketers must ascertain whether the consumers
know about them. If it is not so, there should be
enough promotion before and immediately after the
launch to increase customers understanding and
appreciation of their usefulness and importance.
They can design their messages so that consumers
can associate them with other integral part of
smartphone features such as brand and fashion.
Otherwise, customers will not be able to appreciate
the good intent of manufacturers and marketers in
enhancing the functions of smartphones and making
the phone more useful.
Except for the lead users and early majority users
who hotly pursue the latest technology, majority of
the consumers focus on how smartphone adds value
in ease of use and usefulness. APPs feature is the
fourth most important features after price plan,
screen display and body design. The influence map
also shows that APPs feature is the strongest net
receiver feature. Price, price plan, OS, brand and
fashion influence perception of APPs. In other word,
consumers use these five features to provide an
indication of APPs feature. Hence, manufacturers
and marketers do not need to allocate too much
resources and effort on promoting the huge and rich
collection of APPs to purchasers. Instead, they can
focus on features such as brand. Consumers will
logically use them to provide an indication of the
richness of APPs. Obviously, such kind of
mechanism is absent between NFC and e-wallet.
Fashion is the strongest net effect feature.
Fashion influences price, price plan and APPs. This
result may be a surprise to some observers since
fashion is expected to be the outcome of other
features. This outcome is a strong indication of the
success of smartphone industry and bear witness to
smartphones being a personal expression of users
lifestyle. Many consumers now take fashion as a
given. This study also shows that the industry is
brand-driven. Fashion and brand are two sides of the
same coin, evidenced by their bi-directional
relationship. Brand also shares bi-directional
relationship with OS and price and affects screen,
price plan and APPs uni-directionally.
The price of smartphone is highly inter-related
with other features (price plan, OS, screen, brand).
Price is influenced by fashion and in turn influences
APPs uni-directionally. Hence, customers do not
view price in isolation when assessing whether the
smartphone is worth its value. The mental map of
the consumers also takes into account features
closely associated with price. Hence, as long as the
manufacturers and marketers manage related
features carefully and are reasonable in their pricing,
it is easier to justify higher price by inducing
consumers to link and justify price with other
important features. Our interviews with some
respondents reveal that some of them even associate
high price with quality smartphone features.
On limitations and future research directions, this
study considers only one case study. Hence,
generalization of result may be limited. This study
suggests that other consumer decision-making
Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 135
scenario can be considered. This study also shows
that new product features need to entrench
themselves as a collective part of the product by
locking themselves to other existing and more
established features. This issue can be studied
further. Another limitation is the sample size. The
reason is DEMATEL survey is not easy to fill up
and this study has to put in lot of effort to encourage
participation. Despite this limitation, this study still
provides useful ideas on how consumers assess
product features.
To conclude, using consumers assessing
smartphone features as the case study, this study
shows that in MCDM decision-making problem,
consumers may mentally link product features to
simplify the decision making process. Bi-directional
relationship may also exist between product features,
thereby imposing a limit to the applicability of
recursive model. Since customers perceive some
features as inter-linked, manufacturers may group
them together in the product R&D process.
Marketers also need to take into account this issue
for better marketing outcome.


References:
[1] W.W. Moe, An Empirical Two-Stage Choice
Model with Varying Decision Rules Applied to
Internet Clickstream Data, Journal of
Marketing Research, Vol. 43, 2006, pp. 680-
692.
[2] J.H. Roberts & J.M. Lattin, Development and
Testing of a Model of Consideration Set
Composition, Journal of Marketing Research,
Vol. 28, 1991, pp. 429-440.
[3] J. Goldenberg, R. Horowitz, A. Levav & D.
Mazursky, Finding your Innovation Sweet Spot,
Harvard Business Review, Vol. 81, No. 3, 2003,
pp. 120-129.
[4] M. Haverila, Mobile Phone Feature
Preferences, Customer Satisfaction and
Repurchase Intent among Male Users,
Australasian Marketing Journal, Vol. 19, No.
4, 2011, pp. 238-246.
[5] J. Doyle & R. Green, Strategic Choice and Data
Envelopment Analysis: Comparing Computers
across Many Attributes, Journal of Information
Technology, Vol. 9, 1994, pp. 61-69.
[6] R.B. Kline, Principles and Practice of
Structural Equation Modeling (2nd ed.),
Guildford Publications, New York, 2005.
[7] M. Castells, M. Fernandez-Ardevol, J.L. Qiu &
A. Sey, Mobile Communication and Society: A
Global Perspective, The MIT Press, Cambridge,
M.A., 2006.
[8] E. Fontela & A. Gabus, The DEMATEL
Observer, DEMATEL 1976 Report, Battelle
Geneva Research Center, Switzerland, 1976.
[9] W.W. Wu, Choosing Knowledge Management
Strategies by Using a Combined ANP and
DEMATEL Approach, Expert Systems with
Applications, Vol. 35, No. 3, 2008, pp. 828-
835.
[10] B.W. Park & K.C. Lee, The Effect of Users
Characteristics and Experiential Factors on the
Compulsive Usage of the Smartphone, in: T.H.
Kim, H. Adeli, R.J. Robles & M. Balitanas
(Eds.), Ubiquitous Computing and Multimedia
Applications - Communications in Computer
and Information Science 2011, Vol. 151, 2011,
pp. 438-446.
[11] G. Isiklar & G. Buyukozkan, Using a Multi-
criteria Decision Making Approach to Evaluate
Mobile Phone Alternatives, Computer
Standards & Interfaces, Vol. 29, No. 2, 2007,
pp. 265-274.
[12] H. Falaki, R. Mahajan, S. Kandula, D.
Lymberopoulos, R. Govindan & D. Estrin,
Diversity in Smartphone Usage, MobiSys '10
Proceedings of the 8th International
Conference on Mobile Systems, Applications,
and Services, 2010.
[13] L.P. Tan & A. Mustafa, Incorporating AHP in
DEA Analysis for Smartphone, Proceedings of
the 2nd IMT-GT Regional Conference on
Mathematics, Statistics and Applications, 2006.
[14] J.D. Power and Associates, The Right Blend of
Design and Technology is Critical to Creating
an Exceptional User Experience with
Smartphones and Traditional Mobile Devices,
J.D. Power and Associates, 2011.
[15] H. Nysveen, P. Pedersen & H. Thorbjornsen,
Intentions to Use Mobile Services: Antecedents
and Cross-service Comparisons, Journal of the
Academy of Marketing Science, Vol. 33, No. 3,
2005, pp. 330-346.
[16] P.M.A. Desmet, C.J. Overbeeke & S.J.E.T. Tax,
Designing Products with Added Emotional
Value: Development and Application of an
Approach for Research through Design, The
Design Journal, Vol. 4, 2001, pp. 32-47.
[17] K. Mickalowski, M. Mickelson & J. Keltgen
(2008). Apple's iPhone Launch: A Case Study
in Effective Marketing, The Business Review
Cambridge, Vol. 9, No. 2, 2008, 283-288.
[18] H.H. Lee & S.E. Lee, Internet vs Mobile
Services: Comparisons of Gender and Ethnicity,
Journal of Research in Interactive Marketing,
Vol. 4, No. 4, 2010, pp. 346-375.
Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 136
[19] G.H. Tzeng, C.H. Chiang & C.W. Li,
Evaluating Intertwined Effects in e-learning
Programs: A Novel Hybrid MCDM Model
Based on Factor Analysis and DEMATEL,
Expert Systems with Applications, Vol. 32, No.
4, 2007, pp. 1028-1044.

Recent Researches in Applied Mathematics and Economics
ISBN: 978-1-61804-076-3 137