”nag... 5r ......
H - 11.?

2...

 

... . .
.... rib... .
... 5..

2 3
. \lMﬁavﬁ’
i... ...-

...1- 3 f

2-..... g

 

£339$§ﬁ ...,maﬁagnmwgﬁﬂ , ,
. . . x :15: ,
, _ g? $3.... $1.».-. 5 ,
.. .. a?...sﬁs.%%§§é§. ...? . sum".

....\...u!_

\./
. /

105(53i77’

LIBRARIES
MICHIGAN STATE UNIVERSITY
EAST LANSING, MICH 48824-1048

This is to certify that the
dissertation entitled

SOCIAL INFLUENCE AND TASK-TECHNOLOGY FIT

IN TECHNOLOGY ACCEPTANCE:

AN EMPIRICAL STUDY OF USERS’ PERCEPTION AND

Ph.D.

USE INTENT IN M-COMMERCE

presented by

ZOO—HYUN CHAE

has been accepted towards fulfillment
of the requirements for the

degree in Mass Media

 

 

- Major Prdfesscy’é Signature
g/ﬁ/v f

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

..-.----.--.---—.-------.-.—----.--»—-a--~-----—-—--‘—---------—--->----.-------.-.-.-..

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

ms -‘2 ii “Zane

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 aEIRC/DatoDueJndd-p. 15

SOCIAL INFLUENCE AND TASK-TECHNOLOGY FIT
IN TECHNOLOGY ACCEPTANCE:
AN EMPIRICAL STUDY OF USERS’ PERCEPTION AND USE INTENT IN
M-COMMERCE

By

Zoo-Hyun Chae

A DISSERTATION

Submitted to
Michigan State University
in partial ﬁilﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Telecommunication
Information Studies and Media

2005

ABSTRACT
SOCIAL INFLUENCE AND TASK-TECHNOLOGY FIT
IN TECHNOLOGY ACCEPTANCE:
AN EMPIRICAL STUDY OF USERS’ PERCEPTION AND USE INTENT IN
M-COMMERCE
By

Zoo-Hyun Chae

During the last few years, high-speed 3G mobile technologies have been
proliferating rapidly, signaling prospects for the mobile Internet. However, such great
expectation is only an anticipation of the marketers in mobile telecommunication industry,
and while the projected numbers are not fully guaranteed to be attained, there still exist
controversies regarding the future of mobile Internet and m-commerce. After an extensive
review of the m-commerce literature, we found that little has been studied on the factors
that drive the users’ acceptance of this new technology. This dissertation speciﬁcally
examines the main factors that drive technology acceptance by testing the effects of TTF
and Social Influence on the TAM model’s variables in the context of m-commerce.

An experimental setting was conceived to artiﬁcially create a 3x2
(positive/neutral/negative social inﬂuence by good/poor task-technology ﬁt) between-
subjects factorial design, in which subjects were randomly assigned to one of six
conditions. A total of 210 college students participated in the study. Data analysis results
indicated that there are signiﬁcant main effects of Social Inﬂuence and TTF, as well as
interaction effects, on users’ perceptions and intention to use.

Based on these ﬁndings the followings could be inferred: First, positive social

inﬂuence was crucial in forming positive perception on a new technology and high

intention to use it, and when combined with appropriate goodness of ﬁt between task and
technology, the effect on use intent would maximize. Second, although task-technology
ﬁt impacts the users’ success of trials in ﬁnding music titles, it does not directly affect
their perceived ease of use and perceived usefulness, but rather has an indirect effect
through perceived TTF. Third, perceived ease of use has almost no effect on perceived
usefulness and use intent, but instead, perceived usefulness was found to signiﬁcantly
associate with users’ willingness to use a new technology. These ﬁndings suggest that as
much as TTF plays an important role in shaping and changing users’ perception on new
mobile technologies, social inﬂuence is equally, if not more, inﬂuential when it comes to

the decisions of these new users toward an intention to actually use them.

Com/right by
ZOO-HYUN CHAE

2005

DEDICATION

To my beautiful wife, Yoon
my beloved son, Sean

and my unborn child

ACKNOWLEDGEMENT

I would like to express gratitude to all professors in my guidance and dissertation
committee, Dr. Charles Steinﬁeld, Dr. Cheri Speier, Dr. Hairong Li, Dr. Pamela Whitten,
and Dr. Dan J. Kim, for providing their priceless devotion and care during the course of
my graduate journey. I would like to give special thanks to Dr. Charles Steinﬁeld, who
has been my graduate academic adviser, as well as a tough critic, sage counsel, and
reliable friend since I came in the Master’s Program. I would also like to acknowledge Dr.

Cheri Speier’s invaluable comments and insightful remarks that shaped this dissertation.

I am grateful to my colleagues in the Mass Media Ph.D. Program for all their
help and support during my doctoral studies. In particular, I would like to give thanks to
Bertalan David for providing me with great help in programming the experimental Web
site, and to Jesse Page, who spared his time training and instructing the subjects of this
experiment during the data collection period. I am also grateful to Doowhang Lee and his
wife, Jungsun Ahn, for offering invaluable assistance and support in analyzing the data

gathered.

Finally, I could never thank enough my family, for their continued love and
encouragement. In particular, I am deeply appreciative of my parents for being silent
supporters, and my wife for giving me conﬁdence whenever I was discouraged,
sacriﬁcing her time whenever I was rushed, and guiding me forward whenever I went

astray. This dissertation could not have been completed without her.

vi

TABLE OF CONTENTS

LIST OF TABLES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.......... x
LIST OF FIGURES ____________________ _ xi
CHAPTER 1
INTRODUCTION 1
Mobile Internet: A Brief Summary of the Current Situation 1
M-Commerce and Digital Entertainment Goods 2
Technology Acceptance, Social Inﬂuence, and Task-Technology F it_ 5
Overview of the Dissertation and Research Questions ___9
CHAPTER 2
REVIEW OF THEORETICAL AND EMPIRICAL LITERATURE 12
M-commerce Web Navigation 12
Technology Acceptance Model (TAM) 15
Previous TAM Studies _______ 18
Task-Technology Fit (TTF) Theory 21
Previous TTF Studies 22
Social Inﬂuence Model (SIM) 25
Social Inﬂuence through Word-Of-Mouth Messages - - 26
TAM and Social Inﬂuence 27
Process Research Model____ 30
Chapter Summan' .......... 32
CHAPTER 3
METHODS AND PROCEDURES .............................. 36
Research Design ........................ 36
Stimuli 37
Task-Technology Fit Stimuli 37
Social Inﬂuence Stimuli 40
Sampling 43
Procedure 43
Measurement: 45
Multiple Item Measures ,,,,,,,,,,,,,, 46
Data Collection Instruments __________ _47
Perceived Task-Technology Fit 48
Perceived Ease of Use 49
Perceived Usefulness _ ____ 50
Use Intent 50
Success of Trials .......... 5]
Reliability Test__ ......... 52

 

 

 

vii

Exploratory Factor Analysis (EPA) 53

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Conﬁrmatory Factor Analysis (CFA) - 58

Other Measures 62

Demographic Data 62

Prior Experience 62

' Enjoyment 63

Trainer’s Communication Style 63

Self Efﬁcacy 64

Familiarity with the Task Object 64

Manipulation Check 65

Chapter Summary _66
CHAPTER 4

RESULTS 67

Descriptive Statistics 67

Hypothesis Testing _____ 69

Analysis of Variance (ANOVA) 69

Main Effect of TTF on Success of Trials 70

Main Effect of TTF on Perceived TTF ______ 70

Main Effect of TTF on Perceived Ease of Use 71

Main Effect of TTF on Perceived Usefulness 71

Main Effect of TTF on Use Intention 72

Main Effect of Social Inﬂuence on Perceived TTF 72

Main Effect of Social Inﬂuence on Perceived Ease of Use ,,,,,,,,,,,,,,,,,, 7 3

Main Effect of Social Inﬂuence on Perceived Usefulness ,,,,,,,,,,,,,,,,,,,, 73

Main Effect of Social Inﬂuence on Use Intent 74

Multivariate Analysis of Variance (MANOVA) 82

 

Interaction Effect of TTF and Social Inﬂuence on Perceived TTF ...... 82
Interaction Effect of TTF and Social Inﬂuence on Perceived EOU ,,,,, 83
Interaction Effect of TTF and Social Inﬂuence on Perceived

 

 

 

 

 

 

 

 

 

 

 

Usefulness 84

Interaction Effect of TTF and Social Inﬂuence on Use Intention ________ 84

Repeated Measures 86
Structural Equation Model Testing 90

The Effects of the Moderating Variable 91
Lagrangian Multiplier (LM) Test 91

Likelihood Ratio (LR) Test 93

Chapter Summary 97

CHAPTER 5

DISCUSSION 103
Experimental Study _ _ _ 103
Repeated Measures _ 104
Task-Technology Fit in Mobile Commerce _ 105
Effect of TTF on Success of Trials 105

 

Effect of TTF on Perception and Intention to Use 106

viii

Technology Acceptance Factors: Implications of the Process Model Test
Results

 

Implications of'th—euMSEM Results,

 

Effect of Perceived Ease of Use on Perceived Usefulness
Effect of Perception on Use Intent
Social Inﬂuence in Technology Acceptance

 

Main Effects of Social Inﬂuence

 

Interaction Effects between Social Inﬂuence and Task-Technology-FiL"

Moderating Effect of Social Inﬂuence
Chapter Summary

CHAPTER 6
SUMMARY AND CONCLUSIONS

 

Summary of the Study .......
Managerial Implications____
Limitations and Directions for Future Studies

 

Conclusion

 

APPENDIX 1: QUESTIONNAIRES

 

 

APPENDIX 2: TRAINER’S SCRIPT

 

 

APPENDIX 3: HANDOUTS FOR TASK DESCRIPTION

 

 

APPENDIX 5: CONSENT FORMW

APPENDIX 6: PICTURE SAMPLES OF THE NAVIGATIONAL SYSTEMS

 

REFERENCES

 

 

ix

107
108
108
109
110
110
111
112
113

115
115
119
119
123
125
132
134
135
137

139

141

LIST OF TABLES

Table 1 — Questionnaire Item

Table 2 — Rotated Component Matrix of Perceptive Items at Trial 1

 

Table 3 — Rotated Component Matrix of Perceptive Items at Trial 2

 

Table 4 — Rotated Component Matrix at Trial 1 (after elimination of problematic

items) _- -

_57

Table 5 — Rotated Component Matrix at Trial 2 (after elimination of problematic

items)

57

 

Table 6 —- Fit Indices for CFA Models of Cognitive Measures at Trial 1 and Trial 2,,"60

Table 7 — A Comparison of Cronbach’s Alphas between the Initial and Selected

Items
Table 8 - Manipulation Check of Trainer Credibility

Table 9 — Study Participants’ Descriptive Statistics Summary

75

 

Table 10 — ANOVA Summary

Table 11 — Summary of Interaction Effect Tests from MANOVA

85

 

Table 12 — MANOVA Results for Repeated Measures Effects of Trials ,,,,,,,,,,,,,,

Table 13 — Goodness-of-ﬁt Summary for the Three-Group Structural Equation
Models

........... 89

101

 

Table 14 — Summary of Path Coefﬁcients in Trial 1 and Trial 2 Models ,,,,,,,,,,,,,

Table 15 — LM and LR Tests’ Results Summary"

......... 101

102

 

Figure l —
Figure 2 —
Figure 3 —
Figure 4 -
Figure 5 —
Figure 6 —
Figure 7 —
Figure 8 —

Figure 9 -

Figure 10 —
Figure 11 —
Figure 12 —
Figure 13 —
Figure 14 —
Figure 15 —
Figure 16 -
Figure 17 —
Figure 18 —
Figure 19 —
Figure 20 —

Figure 21 —

LIST OF FIGURES

 

Theory of Reasoned Action

 

 

 

 

 

 

 

Navigation

 

 

 

 

 

Time Spent at Trial 1

 

Time Spent at Trial 2

 

Perceived TTF at Trial 1

 

Perceived TTF at Trial 2

 

 

 

Perceived EOU at Trial 2.-

 

 

 

Horizontal-Depth versus Vertical-Depth ..................................... 15
_____ 17

Technology Acceptance Model _ ____ '17
The Technology-to-Performance Chain "22
Model of Task-Technology Fit and User Evaluations _____________ 2 2
Proposed Variance Model of the Study ,,,,,,,,,,,,,,, 34
Proposed Process Model of the Study 35
3x2 Between-Subjects Factorial Design 37
An Example of Horizontal-Depth Navigation and Interactive Web 41
CFA Model for Perceptive Measures at Trial 1 _________________________ 59
CFA Model for Perceptive Measures at Trial 2 ,,,,,, 59
Number of Web Pages Browsed at Trial 1 __________ 76
Number of Web Pages Browsed at Trial 2 ,,,,,,,,,,,,,,,,,,,,,,,,,,, 76
.......... 77

_____________ 77
.............................................. 78

- 78

.... 79

- . 79

Perceived Usefulness at Trial 1 _80
Perceived Usefulness at Trial 2 __________ 8O

 

xi

Figure 22 —
Figure 23 —
Figure 24 —
Figure 25 —
Figure 26 —
Figure 27 —
Figure 28 —

Figure 29 —

Figure 30 —

 

 

 

Use Intention at Trial 1 ,,,,,,,,,,,,,,,, 81
Use Intention at Trial 2 ____=81
Repeated Measure Effect of Trials and TTF on Perceived TTF ,,,,,,,,,,,,,,,,, 88
Repeated Measure Effect of Trials and TTF on Perceived EOU ,,,,,,,,,,,,,,,, 88

Repeated Measure Effect of Trials and TTF on Perceived Usefulness ,,,,,, 88
Repeated Measure Effect of Trials and SI on Perceived Usefulness ,,,,,,,,,, 89
Repeated Measure Effect of Trials and SI on Use Intent 89

MSEM Results for Positively (P), Neutrally (C) and Negatively (N)
Inﬂuenced Groups at Trial 1 99

MSEM Results for Positively (P), Neutrally (C) and Negatively (N)
Inﬂuenced Groups at Trial 2 100

xii

CHAPTER I

INTRODUCTION

Mobile Internet: A Brief Summary of the Current Situation

Technological advances in wireless and mobile communications has made it
possible to use the Internet on small handheld devices such as personal digital assistants
(PDAs) and cellular phones. The current most widely used mobile technologies are the
26, or second generation, standards, which include GSM (Global System for Mobile),
CDMA (Code Division Multiple Access) and TDMA (Time Division Multiple Access),
and dominate mobile communications in more than 160 countries (Schone, 2004).
However, 3G (third generation) wireless technologies are increasingly moving into the
spotlight as telecommunication carriers gradually shift from 2G to 3G infrastructure. 3G
is the collective term used for several engineering proposals to make wireless networks
more data-capable than 1G analog and 2G digital cellular systems. Wideband CDMA
(W-CDMA, also known as Universal Mobile Telecommunications System, or UMTS),
CDMAZOOO, and Time Division Synchronous CDMA (TD-SCDMA), are among the
leading 30 standards. Today, most of the European countries have already awarded the
W-CDMA, allowing for a data transmission rates of up to 2 Mbps (Buellingen and
Woerter, 2004). Japan started even earlier than any other countries with their NTT
DoCoMo’s mobile Internet services based on their own 3G standard, Freedom of Mobile
Multimedia Access (F OMA, an adaptation of W—CDMA) (Nadel, 2002). In South Korea,

where the number of hi gh-speed mobile Internet subscribers relative to the total number

of mobile subscribers reached 51.1% in 2002, the introduction of the 3G mobile
technology, CDMA2000, also gained great popularity (Minges, 2005).

However, in the United States, the development of the 3G cellular technology
was stunted by the competition of various technological standards, and most carriers have
shown more interest in intermediate and less ambitious 2.5G technologies.
Telecommunication carriers have been reluctant to roll out 3G networks because of the
growing popularity of the high-speed wireless Internet access from laptop computers
(Landler, 2002). Yet, many experts forecast a prosperous future when it comes to mobile
Internet. It is expected that the number of worldwide subscribers to 3G would jump to 70
million by the end of the year 2005 from about 16 million at the end of 2004 (Racanelli,
2005). Also predicted is that by the year 2007, more than 60 percent of Americans will be
using some form of wireless data (Young, 2002), with a revenue of $58.4 billion for the

US. mobile commerce sector (Betts, 2001).

M-Commerce and Digital Entertainment Goods

Mobile commerce (hereafter m-commerce) is, literally, “mobile electronic
commerce,” as such businesses can take advantage of the wireless technologies and
leverage the existing Internet infrastructure (Frolick, 2004). Thus, m-commerce activities
encompass both content delivery (notiﬁcation and reporting) and transactions (purchasing
and data entry) on mobile devices (Leung and Antypas, 2001). Among all issues
concerning m-commerce services, our research speciﬁcally examines the main factors

that drive users to purchase digital entertainment goods on a portable mobile device.

Digital goods, or digital products, are knowledge-based information, or
information goods, that can be digitized and transferred over networks. Information
goods include a wide variety of products mostly ﬁrst produced in digital format and
either later printed on paper or used in their original digital formats (Choi et al., 1997).
All recordable and digitally transferable products, such as music, software or news
information are, therefore, digital goods. (Royer and Van Der Velden, 2003). Thus,
digital goods are materially different from other tangible goods, mainly from the fact that
they are shapeless and have economic lives of their own; once born, they need not to be
reborn (McKenzie, 2003). Digital entertainment goods is a narrower term that represents
music titles, picture and video ﬁles, and game software that can be downloaded on a
terminal for the purpose of enjOyment. Recently, among various entertainment contents,
game is considered to be one of the main features driving the market (Brodsky, 2002),
while multimedia services such as MPEG video streaming are expected to unlock new
revenue streams for the cellular operators (Chevillat and Schott, 2003). Also, the digital
music delivery market has gone legitimate and is growing in its own right with the
successful launch of various platforms, such as Napster, iTunes and Sony Connect. In
Europe, iTunes alone has sold 5 million tracks in the ﬁrst ten weeks of it going live, and
T-Mobile and 02 have also started the mobile music bandwagon rolling with proprietary
mobile music services on the basis that downloads over the Internet will soon extend
across other delivery platforms (Jones, 2004).

Considering that most mobile Internet users are young in age, there is a high
probability that entertainment contents will become prevalent in the m-commerce market.

An exemplary case can be illustrated by NTT’s DoCoMo data service. When N'IT ﬁrst

started deploying their systems in 1997 under the name of DOPa packet transmission, the
target market was focused on business customers. Later, by the time NTT upgraded their
networks in 1999 and changed the service name to iMode wireless Internet, the company
reassessed the target market and decided to concentrate instead on the young mass
consumers (Fransman, 2002). Today, a majority of NTT’s DoCoMo subscribers is
consisted of young generation users especially interested in entertainment contents of the
mobile Internet (Okazaki, 2004). Also, upon estimating South Korean consumers’
preferences for the major CDMA2000 services, Kim (2005) found that two contents,
video telephony and multimedia mobile Internet applications were highly valuated among
subscribers, placing these services among the most anticipated mobile Internet
applications of the future.

In light of such trends, cellular phones are becoming increasingly intelligent,
especially with entertainment functions embedded in them. A recent appearance of
camera phones that integrated digital cameras in cellular phones has shown great success
as they turned out to be highly popular worldwide with a market boost from tens of
millions already in use to an expected number of hundreds of millions in the following
years (Crockett and Reinhardt, 2003; Drucker, 2003; Lewis, 2003). Moreover, with the
development of highly efﬁcient digital audio and video compression technology, such as
MP3 and MPEG-4, the next generation cellular phones are anticipated to fully support
audio- and video-on-demand (Fairlie, 2001; Dunn, 2002; Takahashi, 2002). In Japan,
mobile phones are already being used as portable entertainment players, cameras,
membership and loyalty cards, guidebooks, maps, tickets, watches, and devices for

accessing everything from new to corporate databases (McClelland, 2004). The critical

question lies on whether or not such digital entertainment goods provided wirelessly on
handheld devices will enjoy worldwide proliferation, because the user interface on a
mobile terminal, nevertheless, is less comfortable and less convenient for the user to
control and to interact with the mobile services than it is with conventional computers
(Figge, 2004). In today’s technology-driven situation, customer usability and their future
demands seem to be of only subordinate interest. However, the possibilities for proﬁtable
applications and services for success can only be secured with the knowledge of user
needs. From this perspective, detailed examination of human nature, and understanding
of consumers’ cognitive and behavioral reactions is essential to gain knowledge about
future market potentials (Gerstheimer and Lupp, 2004). Consequently, user acceptance
and innovation diffusion have been two recurring issues in the discipline of
telecommunication and information systems. Our study investigates and examines the
critical factors that speciﬁcally instigate potential m-commerce users to positively
perceive and accept this new technology. The theoretical basis of our arguments is

presented in the following paragraphs.

Technology Acceptance, Social Influence, and Task-Technology Fit

Ever since mankind has been prospering, countless innovations took place in the
form of new technologies that were either highly praised and adopted or neglected and
rejected by the users of their era. Scholars have been pondering over a logical explanation
of the success and failure of innovative ideas to identify the main drivers that cause the
acceptance of novelties. In social science, there exist several theories that discuss the

acceptance of new technologies by the social members. The Diffusion of Innovation

theory (Rogers, 1983) provides a description of the antecedents that inﬂuence the
adoption of innovations, and discusses the process of adoption. The Theory of Reasoned
Action (Ajzen and Fishbein, 1975; 1980) presents a detailed model that links beliefs,
attitude, subjective norm, intention, and behavior, while the Theory of Planned Behavior
(Ajzen, 1985; 1989), adds a new antecedent, the perceived behavioral control, to the
Theory of Reasoned Action. The Motivational Model (Davis et al., 1992) also provides
two key constructs, extrinsic motivation and intrinsic motivation, that lead to user
acceptance of a technology. Although these theories have effectively and
comprehensively demonstrated social and psychological explanations for acceptance and
use of different artifacts, they do not provide parsimonious approaches in testing and
assessing the variables. The Technology Acceptance Model (TAM) and the Social
Inﬂuence Model (SIM) offer simpliﬁed descriptions while maintaining the integrity of
the constructs in explaining technology acceptance. In addition, the Task-Technology Fit
(TTF) model sheds the light on the mechanism that stimulates utilization and enhances
performance in using a new technology.

The Technology Acceptance Model (Davis, 1989) examines the key variables in
the acceptance procedure of a new information technology. According to the theory, the
intention to accept a new technology is determined by two beliefs, one being the
perceived usefulness of using the technology, and the other, the perceived ease of use.
Although the model has been redeﬁned by numerous follow-up studies, which added
further variables that lead to acceptance, the main precursors of acceptance, which
include the perceived usefulness and perceived ease of use, have mostly remained

inherent to the user.

The Social Inﬂuence Model of Technology Use (Fulk et al., 1990) states that
individual’s media perceptions and use are, in part, socially constructed. Media properties
are posited to be subjective, that is, inﬂuenced to some degree by attitudes, statements,
and behaviors of others. Such social factors are the individual’s internalization of the
reference group’s subjective culture (Thompson et al., 1991). Therefore, although
applicability and convenience of a medium are still considered to inﬂuence how
individuals perceive and make use of it, these features are but a part of an equation that
determines media perceptions and adoption.

The Task-Technology F it model (Goodhue and Thompson, 1995) claims that the
ﬁt among task requirements, individual abilities, and the functionality and interface of the
technology determine how well the users’ needs are matched by the system, and as such
is reﬂected on users’ evaluative beliefs of task-technology ﬁt. The model indicates that
performance is affected by the level of correspondence between task and technology, and
utilization is also associated with the ﬁt level and performance. TTF has been broadly
applied to any situation in which individuals use technology to accomplish speciﬁc tasks
or fulﬁll certain needs.

This study approaches the issue of new technology acceptance by integrating the
variables from the TTF and SIM in the TAM theory. Indeed, the process of acceptance of
a new technology is not only inﬂuenced by some Objective features or users’ perception
on usability, but also by socially shaped predispositions. Therefore, we contend that the
level of goodness of ﬁt between task and technology, and social inﬂuence as well as
objective features of a technology needs to be equally examined to verify how an

innovative technology is evaluated and accepted among its new users.

The suggestion that people are inﬂuenced by the collective expression of others is
not new. In an article that examined the effect of poll results on other voters, Allport
(1940) stated that:

“We tend to behave as we see others behave, or, if we are feeling or acting

in a given direction of our own accord, we tend to be swayed the more

strongly if we see or hear a large number of others acting in the same

direction... based more upon social inﬂuence and suggestion than upon

rational decision.”(p.250)

Also, Harvey Leibenstein (1950) described the impact of the “bandwagon effect” on
product demand curves. According to Leibenstein, the bandwagon effect occurs when
people’s evaluation of a good increases aﬁer they observed others consuming the same
product.

Social inﬂuence is especially applicable in the acceptance procedure of new
communication technologies. Numerous studies demonstrated that new users intend to
rely on social relations when deciding on the use of a new and unfamiliar device. Schmitz
and F ulk (1991) empirically tested the SIM model in an actual setting through surveys,
and found signiﬁcant and direct effects of social inﬂuence from both co-workers and
supervisors on the actual individual use of a new media. Another study by Kraut et al.
(1998), which investigated the association between the amount of electronic network use
and virtualization of a ﬁrm, showed that the use of interpersonal relationships for
coordination was positively related to greater network use. While the literature of social
inﬂuence as a key determinant of technology use will be discussed more in detail in later

chapters, it is worth mentioning that only a small number of extant studies empirically

tested the variable in accordance with the TAM model. This dissertation explores the
extent of social inﬂuence as well as technological ﬁt’s effects on users’ success of trials,

perceptions, and intention in using a new technology, which is in our study, m-commerce.

Overview of the Dissertation and Research Questions

Most previous m-commerce studies have remained exploratory by focusing on
exploiting the main values of the wireless Internet (Clarke III, 2001; Anckar and D'Incau,
2002; Lee et al., 2002), discovering key applications (Koh and Kim, 2000;
Balasubramanian et al., 2002; Giaglis, 2003), presenting business strategies (Alanen and
Autio, 2003; Barnes, 2003), and predicting the success of the medium (Hammond, 2001;
Rowello, 2001; Siau et al., 2001). To date, only a handful amount of research examined
the major factors that affect the acceptance of m-commerce, and little, if any, pursued the
study from a theoretical perspective. Extant studies found that such factors, including
perceived risk, complexity, compatibility (Kleijnen et al., 2003), privacy, reliability, cost,
download time, usability (Coursaris et al., 2003), convenience (Magura, 2003) and
Simplicity (Haque, 2003) mostly inﬂuenced the adoption of mobile services.

The goal of this dissertation is to ﬁrrther develop the theory of TAM by
integrating with key factors from the TTF and SIM models, and by exploring them in the
m-commerce context to advance theoretical as well as practical implications. Also, this
study provides a uniﬁed and parsimonious classiﬁcation of the main drivers of m-
commerce acceptance. Experiments are conducted to obtain control of the key variables
and to observe causality. The results successfully contribute to the literature of TAM by

clarifying the key factors for new technology acceptance, as well as to the literature of m-

commerce by presenting a framework to assess the role of social inﬂuence and
technological support in terms of consumers’ perception and intention to purchase digital

goods. The research questions for this study can be stated as follows:

General Research Question:
What are the main factors inﬂuencing the individual acceptance of mobile Internet
services in mobile commerce dealing with entertainment applications such as music ﬁle

downloads?

Speciﬁc Research Question 1:
How does task-technology ﬁt of the navigational system in mobile commerce inﬂuence
users’ success of trials, perception, and intention of using mobile Internet services for

entertainment purposes?

Speciﬁc Research Question 2:
How does social inﬂuence of significant other’ word-of-mouth inﬂuence potential users’

perception and intention to use mobile commerce for entertainment purposes?

Speciﬁc Research Question 3:

How does users ’ perception aﬂect use intention in the context of m-commerce dealing

with entertainment services?

10

This dissertation is organized into six chapters. This introductory chapter
discusses the need to elaborate the theoretical framework of the TAM model by
emphasizing the importance of social inﬂuence and task-technology ﬁt as major
antecedents, and presents m-commerce applications as a context of new technology. The
next chapter provides an in-depth review of the theoretical and empirical literature on
TAM, TTF, SIM theories and m-commerce Web navigation usability, and deﬁnes the key
concepts along with the hypotheses tested in this study. Chapter 3 discusses the methods
and procedures of data collection as well as the rationale for the experimental study used
to examine the causal relations between the variables. Chapter 4 presents the results
obtained from statistical computation and data examination employed throughout the
preparation and analyses of the collected data. Final discussion of the outcome and
ﬁndings is provided in Chapter 5, and Chapter 6 concludes the dissertation by stating the
implications of the study, limitations stemming from conceptual caveats and drawbacks

associated with the research design, and closing remarks.

ll

CHAPTER 2

REVIEW OF THEORETICAL AND EMPIRICAL LITERATURE

This chapter examines the theoretical and empirical literature of the TAM model,
SIM model, and TTF theory. Also, the context of m-commerce is deﬁned, and concepts
pertaining to this dissertation are described. The hypotheses to be tested in the next

chapters will follow the discussions.

M-commerce Web Navigation

M-commerce relies on the usage of wireless devices, such as cellular phones or
PDAs, from which commercial transactions are made through instantaneous application
Of information technology or transmission of data streams in places that were previously
inaccessible (Raisinghani, 2002). Leading m-commerce services for consumers can be
categorized into different sets of application that include travel, ticketing, banking, stock
trading, news and sports, gambling, game and entertainment, and shopping (May, 2001).

However, the usability of these services heavily depends on the mobile device’s
screen size. “Usability” can be deﬁned as the quality of a system with respect to ease of
learning, ease of use, and user satisfaction (Rosson and Carroll, 2002). Although
technological developments have enhanced the connection speed and image quality, the
limited size of the small screens on cell phones or PDAs has impeded the display
capacity of regular Web pages on these handsets (Tarasewich, 2003). Consequently,

screen Size affects usability. A study by Jones et al. (1999) demonstrated that users in

12

“small screen” (640x480 pixels) environment showed lower effectiveness in completing
search and retrieval tasks than those with a “large screen” (1074x768 pixels).

As an important building block for enabling Web site navigability on wireless
devices, the Wireless Application Protocol (WAP) has now become a globally-accepted
standard (Ralph and Aghvami, 2001 ), but is required a new technique for user interface
design to minimize user interaction during input and output (F rick, 2000). Chittaro and
Cin (2001) tested the usability of the WAP interface with novice users. The subjects
evaluated four different types of interface, (1) a navigation using links (2) a list of links,
(3) action screens and (4) selection screens, and reported better performance and higher
perceived ease of use in the ﬁrst two types of interface, that is, the link-enabled
navigations. The study conﬁrmed the fact that the hyperlink function is as effective in
mobile commerce as it is in conventional Internet navigation.

The usability of small handheld devices also depends on the hierarchical menu
structure. The “hierarchy” of menus in mobile Internet refers to the arrangement and
display method of Web page procedure. Henneman and Rouse (1984) argued that one of
the most critical characteristics that must be considered in the hierarchical menu design is
the trade-off between depth and breadth of menu structure. “Depth” can be deﬁned as the
number of levels in the hierarchy, while “breadth” can be explained by the number of
options per menu panel (Paap and Cooke, 1997). Figure 1 depicts the difference between
. the horizontal-depth navigation and vertical-depth navigation. In essence, the horizontal-
depth, or depth, refers to the number of sub-menus, whereas the vertical depth, or breadth,

refers to the length of the list on a single menu. The following statement explains the

13

steps that a user needs to take in both horizontal- and vertical-depth navigations to reach
a target:

“The user starts searching information from the top level (point a) and

moves down to a sub-level (b) until encountering a list of street names. If

there are n street names on the list, there may be several ways to present

the whole list on the screen, such as providing n street names within one

page, or dividing the list into several pages. In the former case, the user

has to ﬁnd a certain street (list k) by scrolling down the long list. In the

latter case, the user has to move from the ﬁrst page to the third page

containing ‘list k’ (b -» b1 —-> b2), then on page b2 (s)he needs to go down

to ‘c ’ to reach the ‘Target’ information.” (Chae and Kim, 2003)
Through an experimental study, Chae and Kim (2003) demonstrated that as depth
increased, the performance of users using mobile devices decreased. They also found that
user satisfaction was signiﬁcantly affected by depth. However, the results showed higher
user preference of four-depth design over one- and six-depth designs.

Concurring with Chae and Kim’s study, the current research applies the
horizontal-depth design as a type of navigation in m-commerce. Additionally, an
interactive navigation will also be considered. The interactivity has been validated to be
an efﬁcient method of information retrieval (Benoit, 2002; Lin et al., 2003), as well as a
driver of online consumers’ intention to purchase and revisit (Kuk and Yeung, 2002).
Although there exist numerous ways to implement interactivity, for the purpose of
executing the ﬁle searching and retrieving function, a search box was created and

employed in the study. On this search box, users are told to type the name of the ﬁle to be

14

found. A ‘search’ button is displayed next to the search box so that when users click on it,
the screen shows a list of ﬁles matching the letters typed in the box. The users are then

able to choose and click on a ﬁle name hyperlink in the window.

F igyre I : Horizontal-Depth versus Vertical-Depth (Adopted from Chae and Kim, 2003)

 

 

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a l paging l
a
— -. i s i a
' List 2 : b2 ............... ) b3 ............... ’ i b 4 ............... )
I LISI 3 : . ' :
' . l l l l g
I i r I r
l E l i List Ir ll l
scrolling ; g c :
List Ir - L
List it
— b’ Target
............... ’ : Horizontal depth
(dotted line arrow)
c
: Vertical depth
(solid line arrow)
Target

Technology Acceptance Model (TAM)
The Technology Acceptance Model (TAM) (Davis, 1989; Adams et al., 1992)

suggests that two main factors, perceived usefulness and perceived ease of use of a new

15

technology, lead to its acceptance and actual usage. According to TAM, the intention to
voluntarily accept, that is to use, a new technology is determined by two beliefs dealing
with (1) the perceived usefulness of using the technology and (2) the perceived ease of
use of it. Perceived usefulness is a measure of the individual’s subjective assessment of
the utility offered by the new technology in a speciﬁc task-oriented context, and
perceived ease of use is and indicator of the cognitive effort needed to learn and to utilize
the new technology.

“The goal of TAM is to provide an explanation of the determinants of

computer acceptance that is general, capable of explaining user behaviour

across a broad range of end-user computing technologies and user

populations, while at the same time being both parsimonious and

theoretically justiﬁed.” (Davis et al. 1989, p.985)

A key purpose of TAM, therefore, is to provide a basis for tracing the impact of
internal factors on intentions and usage. TAM is formulated in an attempt to achieve
these goals by identifying a small number of fundamental variables suggested by
previous research dealing with cognitive and affective determinants of IT acceptance (Al-
Gahtani and King, 1999).

TAM’S theoretical foundation comes from the Theory of Reasoned Action
(TRA) by Ajzen and F ishbein (1975; 1980; Sheppard et al., 1988). TRA assumes that
behavior is determined by intention, which is, consecutively, inﬂuenced by attitude
toward the behavior and subjective norm. Attitude is shaped by beliefs (perceived

consequences of the act) and evaluations of outcomes, while subjective norm is affected

16

by normative beliefs (perceived signiﬁcant others’ preferences) and motivation to comply

(see Figure 2 below).

F tare 2: Theory of Reasoned Action (Fishbein and Ajzen, 1975)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Beliefs and Attitude toward
Evaluations Behavior
Behavioral Actual
Intention Behavior
Normative Beliefs and Subjective
Motivation to Comply Norm

 

 

 

 

 

 

TRA’S belief—attitude—intention-behavior relationship yielded to the formation of
TAM, which applies a more parsimonious approach of framing the exogenous variables.
Perceived usefulness and perceived ease of use are the two drivers originally deemed to

be the most important factors in acceptance of a new technology (see Figure 3 below).

F igure 3: Technology Acceptance Model (Davis et al., 1989)

Perceived
Usefulness :

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Extemal Attitude Behavioral Actual
Variables toward Using Intention Use
Perceived
Ease of Use

 

 

17

The causal link between perceived ease of use and perceived usefulness is not
only argued in Davis’ model but also conﬁrmed by numerous studies. Both Mathieson
(1991) and Szajna (1996) showed that case of use account for a signiﬁcant amount of the
variance of perceived usefulness. Similar ﬁndings have been reported by Thompson et al.
(1991), Straub et al. (1995), Igbaria et al. (1997), and Gefen et al. (2003). They suggest
that, while perceived ease of use and perceived usefulness inﬂuence use intention and

usage, perceived ease of use is an antecedent of perceived usefulness as well.

Previous TAM Studies

For more than a decade, a large number of studies have contributed to the
literature of TAM. While a majority of the literature introduced new variables by
combining TAM with other models, some other portion tested the model in different
contexts and compared the model with other theories to assess the main factors of
technology acceptance.

Mathieson (1991) compared TAM with the Theory of Planned Behavior (TPB)
in an experimental study that employed using new computer programs. He found that
both theories predicted intention to use fairly well, but TAM showed a slight empirical
advantage. Taylor and Todd (1995) tested a model that combined TAM and TPB, called
the decomposed TPB, and found that it provided better explanation of the variance in
usage than TAM. The same authors elaborated their ﬁndings and presented a new model,
Augmented TAM, encompassing a new variable: perceived behavioral control (Taylor
and Todd, 1995). In another study, Mathieson et al. (2001) also used TPB variables to

form a model of Extended TAM with an added predictor of perceived user resources.

18

According to the authors, the “perceived resources” refers to the extent an individual
believes that he or she has the personal and organizational resources needed to use an
information system. They found empirical evidence that supported the Extended TAM’s
contention that perceived resources affect an individual’s intention to use an information
system.

TAM has also been tested in the context of electronic commerce. Geffen et al. r:
(2003) advanced a new model of technology acceptance by incorporating the trust factor

in TAM. This Trust and TAM determinants were tested on consumers’ shopping

 

experiences through electronic vendors’ Web sites. The results of the study showed that L"
consumer trust is as important to online consumers as TAM’s antecedents. Similar
models of trust in TAM studies were conducted by other scholars. Pavlou (2001)
successfully incorporated two additional factors, trust and perceived risk, in testing
consumers’ intention to adopt e—commerce. Also, Dahlberg et al. (2003) implemented a
new model of Trust Enhanced TAM, and found that trust was an important predictor of
intent to use in the mobile payment context.

While the initial TAM underwent a series of model extension in numerous
studies, a more recent trend involves the elaboration of TAM model. Venkatesh et al.
(2002) proposed an integrated model of TAM and Motivational Model, where user
acceptance enablers were observed as predictors of perceived ease of use, perceived
usefulness, and intrinsic motivations. This integrated model was tested in an employee-
training environment through longitudinal data collection and analysis. The results
supported the importance of intrinsic motivational factors in the process of technology

acceptance in organizations. Another elaborate model of TAM by Venkatesh and Davis

19

(2000) was TAM2, a new version that includes subjective norm as one of the main
determinants. In this new model, the subjective norm, which was a key predictor of
behavioral intention in TRA, is deﬁned as a “person’s perception that most people who

are important to him think he should or should not perform the behavior in question”

 

(Fishbein and Ajzen, 1975). As the construct represents a crucial aspect of social

inﬂuence, the implication of TAM2 clearly indicates that societal factors are equally

:7

important in the acceptance process of a new technology.

To date, two comprehensive studies provide a general and detailed literature

 

review of TAM. Lee et al. (2003) explicitly organized most of extant TAM studies by L
chronological progress and ﬁndings, and presented a discussion on the model’s
limitations through interviews with major researchers in the ﬁeld. Venkatesh et al. (2003)
provided as well an almost complete list of TAM literature. However, the authors
formulated a Uniﬁed Theory of Acceptance and Use of Technology (U TAUT) through
revision of eight salient models of extended and elaborated TAM. Empirical evaluation
of the new model gave the authors validated results that supported the integrity of
UTAUT.

Concurring with the TAM theory, it is predicted that the perception of the new
technology’s ease of use will Show a positive relationship with the perceived usefulness,
hence the following hypothesis:

H] .' Perceived ease of use will be positively related to perceived usefulness.

20

Also, since the TAM model theorizes a positive effect of these perceptions on
use intention, we hypothesized that perceived ease of use and perceived usefulness will
positively inﬂuence the subjects’ willingness to reuse the new technology:

H2: Perceived ease of use will have a positive effect on use intention.

 

H3: Perceived usefulness will have a positive effect on use intention.

Task-Technology Fit (TTF) Theory 1'

The Task-Technology Fit (TTF) Theory (Goodhue, 1995; Goodhue and

 

Thompson, 1995) states that the correspondence between information systems b,
functionality and task requirements leads to positive user evaluations, higher utilization,
and positive performance impacts. The essence of TTF comes from Technology-to-
Perforrnance Chain (TPC), as depicted in Figure 4, which asserts that for an information
technology to have a positive impact on individual performance, the technology must be
a good ﬁt with the task. According to Goodhue (1995), in the context of information
systems research, technology refers to computer systems and user support services, while
tasks are deﬁned as the actions carried out by individuals in turning inputs into outputs.
Also, the model of TTF emphasizes the importance of user evaluations. Users will give
evaluations based on the extent to which systems meet their needs and abilities. Figure 5
represents the model of Task-Technology Fit and User Evaluations. The heart of the TTF
model is the assumption that information systems give value by being instrumental in
some tasks and that users will reﬂect this in their evaluations of the systems (Goodhue,
1998). Therefore, the evaluations from users are the key determinants of ﬁt between task

and technology.

21

Figure 4: The Technology-to-Performance Chain (Goodhue and Thompson, 1995)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Task
Characteristics
\ TaSk-
Technology Fit \
Technology / Performance
Characteristics Impacts
Utilization /
Figure 5: Model of Task-Technology Fit and User Evaluations (Goodhue, 1995) F
User Evaluations
(Surrogate for TTF) R
,",~’ ,’,’( II ‘
TaSk I III/1”,”! \ \
’ ' ’l ,’ Performance
Technology
. . Task-Technology Fit
Indrvrdual (TTF)
Previous T T F Studies

TTF shares some important similarity with organizational structural contingency
theories. One of the key questions of contingency theories is how task technology
inﬂuences the information-processing requirements of an organizational unit (Keller,
1994). These theories involve ﬁt measures of tasks at the organizational level, that is an
organization’s structure and its organizational context in which the technology is used
(Van de Ven and Drazin, 1985). For instance, information processing has been described

to be an integral concept in models that tested how organizations can match different

22

......” ...:-.——.. .-_..—-w~ ~—

 

technologies to the design and structure Of units in order to achieve high unit performance
(Tushman and Nadler, 1978; Daﬁ and Lengel, 1984; Cohen and Levinthal, 1990).
However, TTF is more concerned about individual level ﬁt of task and
technology. TTF can be deﬁned as “the degree to which technology assists an individual
in performing his or her portfolio of tasks” (Goodhue, 1997). Indeed, most extant TTF
studies examined individual users’ performance and perceptions of different IT use in
various contexts. Goodhue (1995) tested the TTF model in individual work-related
computer use, and found that users can successﬁrlly evaluate task-technology ﬁt by
accurately reﬂecting differences in the underlying systems and services provided to them.
Also, Goodhue and Thompson (1995) explored the effects of task-technology ﬁt on
performance and concluded that for an information technology to have a positive impact
on individual performance, it must match the individual needs and task characteristics.
Ferratt and Vlahos (1998) investigated the managerial decision making support of
computer-based information systems from the perspective of TTF, and found that
managers value information systems highly for resource allocation, evaluating
alternatives, identifying problems, ranking alternatives, and short-term decision making.
Some group level studies of TTF were also pursued. Massey et al. (Massey et al.,
2001) explicitly observed how cultural variability affects the difﬁculty with which global
virtual teams communicate, and suggested that establishing norms of behavior regarding
communication task-technology will enhance the performance of virtual teams separated
by space, time, and culture. Several scholars also tested the effectiveness of Group
Support Systems (GSS). Zigurs and Buckland (1998) examined the relationship between

attributes of group task complexity and group performance, while Shirani et al. (Shirani

23

 

et al., 1999) compared the quality of works performed by groups with synchronous

versus asynchronous communication technology. Murthy and Kerr (2000) discovered that

group members performed signiﬁcantly better when communicating face-to-face for

problem-solving tasks and when communicating via the G88 for the idea generation tasks.
The theory of TTF has been tested in conjunction with other theories as well.

Lim and Benbasat (2000) conducted a TTF model test in accordance with the information

 

richness theory. They developed and successfully tested their Task-Representation Fit

Model that explains how the multimedia representation is perceived as more useful than

 

the text-based representation in the context of individual decision makers utilizing 5,.
organizational data. In other papers, Dishaw and Strong (1999), Dishaw et al. (2002), and
Klopping and McKinney (2004) extended the model by integrating TTF with TAM
constructs. The introduction of the TTF variable in TAM as an antecedent of perceived
ease of use and perceived usefulness is worth emphasizing since it explores the ﬁt
between task and technology as a determinant of users’ perceptions. This argument
concurs with that of Mathieson and Keil (1998) who claimed that the perceived ease of
use can be a function of task-technology ﬁt as does the perceived usefulness. Indeed, the
level of task-technology ﬁt is quite likely to affect user’s perceptions and intention to use
a new technology. Pertaining to the TTF constructs and predicted effects of TTF on TAM,
the following hypotheses are presented:
H4: There will be a main effect of TTF on users’ success of trials, perceptions and
use intention of the new technology.
H40: There will be a main effect of TTF on users’ success of trials.

H4b: There will be a main effect of TTF on user’s perception of TTF.

24

H4c: There will be a main effect of TTF on perceived ease of use.
H4d: There will be a main effect of TTF on perceived usefulness.

H4e: There will be a main effect of TTF on use intention.

Social Inﬂuence Model (SIM)

The term of social inﬂuence is highly comprehensive and complex. Social
inﬂuence can be deﬁned as “the way in which one or more people alter the attitudes or
the behavior of others” (Peterson, 1997). Therefore social inﬂuence encompasses every
aspect of social exchange between two or more individuals sharing one or more common
social boundaries. The social exchange theory (Homans, 1958; Ekeh, 1974; Blau, 1987)
indicates that individuals involve in interpersonal and mutual interactions for the sake of
reward, but the beneﬁts contained in social exchange do not have an exact price in terms
of a single quantitative medium of exchange, and the nature of the return cannot be
bargained about. Such interactions among social members trigger collective conscience,
because each individual seeks to ﬁt him(her)self into the structuralism, or normative rule
that regulates behavior, of the society (Durkheim, 1958; Levi-Strauss, 1969). By the same
token, an inexperienced individual who Observes a new technology and gets exposed to
his (her) peers’ opinions about the quality, efﬁciency, and their willingness to use might
as well share similar beliefs and intention toward that novelty.

The Social Inﬂuence Model of Technology Use (F ulk et al., 1990; Schmitz and
F ulk, 1991) suggests that media choice and use behavior occurs in a very social world,

where the users’ perceptions are subjective and socially constructed. Thus, although these

25

 

 

 

 

media are evaluated in some degree by objective features, they are also substantially

determined by the attitudes, statements, and behaviors of coworkers.

Social Inﬂuence through Word-Of-Mouth Messages

The word-of-mouth (WOM) messages expressed in face-to-face communication
have been considered as highly persuasive means for molding and changing individual
opinion, especially when the communicator is a peer or someone who share similar

interests in similar environments. Indeed, peers oﬁen exchange ideas through expressing

 

overt statements or by hinting social cues. Riesman (1952) argued that American society
is moving toward an orientation in which both the values and styles of behavior are
learned not from the predecessors or ancestors, but from peers such as tutors, friends,
colleagues, and family members. Studies of individual technology use have also
examined the effect Of WOM messages, which turned out to be an important factor
inﬂuencing the decision making process. Meuter et al. (2003) explored the consumers’
usage pattern of self-service technologies and found that those with high levels of
technology anxiety were more likely to participate in positive WOM interactions with
other consumers who had had an initially satisfying technology experience. In another
study that tested the Social Judgment Analysis (SJA) to identify the information
judgment preferences held by professional groups (Steﬂ-Mabry, 2003), the WOM
messages were found to be one of the most important factors for user satisfaction of
information. Therefore, it is posited that such WOM messages expressed by peers play an
important role as socially signiﬁcant inﬂuence regarding the decision making of an

individual in shaping ﬁrst impressions toward a newly introduced technology.

26

TAM and Social Inﬂuence

While in the initial TAM, only two perceptual factors were suggested to affect
technology acceptance, several scholars have been suggesting that social inﬂuence Should
equally be reconsidered as a main determinant of the model. As presented previously, a
study on TAM2 corroborated that the subjective norm is a crucial social inﬂuence

variable that needs to be incorporated in the model. “Subjective norms” has been

 

considered to best represent social inﬂuence, especially in workplaces, and was validated

by a number of elaborated TAM studies (Jasperson et al., 1999; Malhotra and Galletta, l’

 

1999; Venkatesh and Morris, 2000; Lee et al., 2001). In a more recent study, Hsu and Lu Ly
(2004) reported that social norms, attitude, and ﬂow experience explain about 80% of
online game playing. Nevertheless, since the concept of subjective norm implies a
situation where the use of a certain technology is more ore less compulsive, that is, in a
mandatory context, other researchers attempted to measure social inﬂuence in a more
voluntary context. Thompson et al. (1991) tested the effect of social factors on personal
computing, and conﬁrmed that there exist positive relations between perceived inﬂuence
and actual utilization.

Since most social inﬂuence studies employed a self-reported or indirect data to
estimate the level of social inﬂuence on the research participants, there is a need to call
for a more delicate method of manipulating the construct to evaluate its directionality and
magnitude. An interesting experimental research pursued by Galletta et al. (1995)
effectively controlled and manipulated social factors in soﬁware learning environment by
introducing confederates in their experiment site. In this study, the trained “fake”

participants joined the “rea ” subjects and acted as being knowledgeable about the

27

software in question during the experimental sessions. Furthermore, through word-of-
mouth communication, they expressed either favorable or unfavorable opinions to the
other subjects. Results of the study revealed that social inﬂuence through word-on-mouth
communication was a signiﬁcant determinant of attitudes, behavior, and performance.
Although the research design and data collection procedure of this dissertation
will be described with more details in the next chapter, it is worth mentioning that the
current study will take Galletta et al.’s approach as an exemplary designing method and
implement the use of confederates to induce social inﬂuence through prearranged WOM

messages. It is posited that regardless of the actual TTF of the new medium, positively

 

inﬂuenced subjects will show positive perceptions and positive intention to use, while
negatively inﬂuenced subjects will exhibit less favorable perceptions and use intent.
Neutrally inﬂuenced subjects are also predicted to show lower degree of perceptions and
use intent than the positive group’s subjects, but higher scores than the negatively
inﬂuenced ones. Therefore, in an identical level of TTF situation, each of the three
conditions’ perceptive and intentional outcomes regarding the new technology is likely to
vary according to the direction of social inﬂuence. The hypotheses concerning these main
effects of social inﬂuence are stated as follows:
H5 .' There will be a main effect of social inﬂuence on users’ perceptions and use
intention of the new technology.

H5a: There will be a main effect of social inﬂuence on perceived TTF.

H5b: There will be a main effect of social inﬂuence on perceived ease of use.

H5c: There will be a main effect of social inﬂuence on perceived usefulness.

H5d: There will be a main effect of social inﬂuence on use intention.

28

Since we hypothesized that TTF and social inﬂuence form the key external
variables for the TAM theory that are considered to inﬂuence the perceived ease of use,
perceived usefulness and use intent, it is also highly likely that these two factors interact
with each other to emphasize the effects on users’ perceptions and intention. In other

words, social inﬂuence may as well magnify the effect of TTF on the individuals’

 

perceptions and use intention of a new technology. Thus the hypotheses concerning the
interaction effects of social inﬂuence and TTF can be stated as follows:

H6: There will be an interaction effect between social inﬂuence and TTF on users’

 

perceptions and use intent. L
H6d: There will be an interaction effect between social inﬂuence and task-
technology ﬁt on users’ perception of TTF.
H6b: There will be an interaction effect between social inﬂuence and task-
technology ﬁt on perceived ease of use.
H6c: There will be an interaction effect between social inﬂuence and task-
technology ﬁt on perceived usefulness.
H6d: There will be an interaction effect between social inﬂuence and task-

technology ﬁt on use intent.

These relationships are summarized in Figure 6 as a variance research model that
classiﬁes the variables into three categories: manipulations, cognitive variables, and
conative variables. Manipulations include the independent variables that directly affect
the cognitive and conative reactions, while the cognitive variables consist of all the

perceptive measures regarding TTF, EOU, and usefulness. Conative variables comprise

29

the outcomes concerning the actual success of trials, and individual willingness to further

use the system tested.

Process Research Model
The predictions derived from TTF, SIM, and TAM models can also be stated in

terms of a process chain, where the relationships among the variables are explained
through causal links. Such a process model is an extension of the TAM model, as well as
an alternative way to consider the associations among the social inﬂuence factor, TTF
factor, and other technology acceptance drivers. Our process model places social
inﬂuence as a moderating variable based on the assumption that it could intensify the
effect of TTF on perceptions. Also, because TTF is a dichotomous variable, a surrogate
for TTF, namely success of trials, was introduced as an independent variable that
represents a key external factor in TAM model. As previously described, TTF theorizes
that the level of ﬁt between task and technology affects utilization (Goodhue and
Thompson, 1995), which allowed us to formulate the following hypotheses:

H7: The degree of a trial’s success in completing a task with a new technology will

be positively associated with the users’ perception of TTF.
H8: Success of trials will be positively related to perceived ease-of-use.

H9: Success of trials will be positively related to perceived usefulness.

Also, perceived TTF, which is one of the many dimensions Of TTF evaluation, is

hypothesized to have a positive relation with other perceptive variables of TAM. We

employed the term “perceived TTF” instead of “TTF evaluation” because the latter

30

encompasses numerous types of cognitive measures, such as quality, authorization, ease
of use, production timeliness, system reliability, relationship with users, and locatability
(Goodhue, 1998). Since the TAM variables overlap with the dimensions of ease of use
and relationship with users, implementing all of these measures would be redundant and
recursive to the proposed model. Thus the locatability dimension was chosen to represent
the concept of perceived TTF. The hypotheses regarding perceived TTF are stated as
follows:

H10: Users’ perception of TTF will be positively associated with perceived ease-of-

use.
HI I .' Users’ perception of TTF will be positively associated with perceived

usefulness.

As was the case with the variance model, the effects of perceived ease of use and
perceived usefulness on use intent will also be tested in this process model. Pertaining to
the TAM literature (Davis, 1989), the following statements are the same hypotheses as
presented previously.

HI : Perceived ease of use will be positively related to perceived usefulness.
H2: Perceived ease of use will have a positive effect on use intention.

H3: Perceived usefulness will have a positive effect on use intention.

Since social inﬂuence is expected to interact with TTF in this variance model,

instead of viewing it as another independent variable, we believe that it contains a trait

for moderating the effect of TTF on the perceptive outcomes. Therefore, it is

31

 

 

hypothesized that social inﬂuence will affect the relationship between success of trials

and the three perception-related variables.

H12:

H13:

H14:

The effect of the trials’ success on perceived TTF will be moderated by social
inﬂuence, such that the relationship will be stronger for positively inﬂuenced
users, moderate for neutrally inﬂuenced users, and weaker for negatively
inﬂuenced users.

The effect of the trials’ success on perceived ease-of-use will be moderated by
social inﬂuence, such that the relationship will be stronger for positively
inﬂuenced users, moderate for neutrally inﬂuenced users, and weaker for
negatively inﬂuenced users.

The effect of the trials’ success on perceived usefulness will be moderated by
social inﬂuence, such that the relationship will be stronger for positively
inﬂuenced users, moderate for neutrally inﬂuenced users, and weaker for

negatively inﬂuenced users.

Chapter Summary

The Technology Acceptance Model (TAM) has been receiving great attention

since its ﬁrst appearance. Numerous studies contributed to the model’s extension and

modiﬁcation, including those that employed social factors. Since the Social Inﬂuence

Model (SIM) literature had already been discussing new media use behaviors, while the

Task-Technology Fit (TTF) model attempted to account for the factors enhancing

utilization and performance of new technology users, the current study combines these

32

three models in assessing the antecedents that trigger new technology acceptance in a
mobile commerce context. Figure 6 illustrates a variance research model, while Figure 7
depicts a process research model showing the relationships between the key variables in
process chains. All arrows denote the hypotheses to be tested in the current research. The
following chapter discusses the empirical portion of this dissertation by precisely
describing the methodological procedures and explaining the measures taken for data

analyses and model testing.

33

 

 

 

\

 

/

E85 33

 

\

 

 

 

/

ﬂat-H
MO 3826

835.8;

95360

       

 

-—-—.—i—-—‘—-—-—

 

38533 33083

30m woZoBom
FE. 33083

 

K ”mcomﬂooeom

 

 

 

._.— —-—-—---— —.—.—.—.—

  

OOGOSEGH
38m

J

 

 

 

 

\

 

«VI

SEBEBS
méhmmob

 

F

_
_
L
_
.
_
_
_
_

33m 2: me $82 SHEEN, 88qu “gm

em

 

Eotﬁamwae:

j

zwoﬁqnwo-Hixmwh

 

K

34

83:85
33

N1

m3

mama—£33

 

33683
_ _I
_I
o _ I
83 mo Ommm
vozooeom

on

E mo
cosmooLom

wI

 
  

ﬂat-_-

 

  
  

 

4 mm \7 mo $825

SE SE 3:

 

 

 

 

8:265 _mmoom

 

 

 

33m 05 MO 382 339$ 388$ ”NM-am

35

CHAPTER 3

METHODS AND PROCEDURES

This chapter describes the research methods and procedures used in data
collection process for the current study. As indicated in previous chapters, since the
Objective of the research is to test the causal relationships among variables from the TAM,
TTF, and SIM model, data were collected in an experimental setting to control for
unwanted intervening factors and observe the relations among the independent, mediating,

moderating, and dependent variables.

Research Design

There exist two treatments in this research: (1) three different social inﬂuences
through word-of-mouth messages (positive, negative, and neutral — control), and (2) two
contrasting ﬁts of technology with tasks (good ﬁt and poor ﬁt). A three-by-two (3x2)
between-subjects factorial design was employed to test the hypotheses, with task-
technology ﬁt as the independent variable, and social inﬂuence as the moderating
variable. A total number of 210 subjects participated in the study from which 37
participants were assigned in positive/ good condition, 37 in positive/poor condition, 32 in
neutral/good condition, 32 in neutral/poor condition, 36 in negative/good condition, and
36 in negative/poor condition. Figure 8 illustrates the conditions and the number of

participants for each treatment group.

36

Figure 8: 3x2 Between-Subjects Factorial Design

Social Inﬂuence

 

 

 

Positive Control Negative
138.19. Good Fit 37 32 36
Technology
F_it Poor Fit 37 32 36

 

 

 

 

 

The fact that there was a limited number of PDAs and that each participant
needed to be individually trained and socially inﬂuenced, without the presence of other
participants, made the data collection procedure require no more than one participating
subject at a time during the experiment. Therefore every subject was individually and
independently assigned to a treatment, tested, and asked to complete the questionnaires at

the location of experiment.

Stimuli

As previously mentioned, we tested the effects of two treatments, task-
technology ﬁt and social inﬂuence, on users’ success of trail, perceptions, and use intent
in an experimental setting. The following paragraphs thoroughly explain how the stimuli

were controlled to differentiate one treatment from another.

Task- Technology Fit Stimuli

In order to test the effect of task-technology ﬁt on perception and technology
acceptance, we built two similar Web sites that pretended selling MP3 format music ﬁles
as products. Music titles were chosen from the researcher’s personal collection of

compact discs (CDs), and converted into MP3 ﬁle format to make it transferable through

37

 

electronic networks. These Web sites were temporarily created to represent seemingly
commercial MP3 online vendors that enable music search and direct download from the
Site. However, the two Web vendors employed different interface for searching and
ﬁnding music ﬁles, and participants were asked to perform speciﬁc tasks, i.e. ﬁnd and
download speciﬁc songs, that either ﬁtted well or poorly with the interface. No MP3 ﬁles
were actually offered to the participants, and all ﬁles were discarded from the system
when the entire experiment was over. The optical resolution of the Web pages were
narrow enough (320 horizontal pixels) to be represented on PDA screens. Also, both sites
contained the same list of music ﬁles downloadable on the handheld devices.

The ﬁrst Web site, which was named MP3 Topia. Com, featured an interactive
system, where a user was ﬁrst asked to type a part or the entire name of the artist, leading
to a matching list of names. Then, when a name was clicked, the screen displayed another
search window asking the user to type the title of the song. The user was able to choose
and click any title from the list whenever (s)he sees the one (s)he was seeking. The next
Web page displayed all information of the song, with a button that let the user conﬁrm
the download as a ﬁnal step. During the download process, a status bar would appear and
indicate how much the MP3 ﬁle was being downloaded on the device. Finally, the last
page would display a message verifying that the ﬁle was successfully transmitted on the
device.

The second Web site, called MP3Kingdom. Com, featured horizontal-depth Web
navigation, that is, a menu-by-menu navigation technique. For instance, if a user was
searching for the song “Yesterday” by “The Beatles” from the album “Helpl”, (s)he

would ﬁrst require to click on the “Rock” genre, choose “Beatles” in the list of artists,

38

choose “Help!” in the list of albums, and then choose “Yesterday” in the list of songs.
The conﬁrming, downloading, and verifying pages were analogous to those of
MP3 Topia. com. Both Web sites were created on an Internet server, and the contents were
wirelessly uploaded to the PDA through 802.11b wireless LAN (Wi-Fi) connection.
Figure 9 illustrates the navigation processes for MP3Kingdom and MP3 Topia. Actual
picture samples of the two systems are in Appendix 6

Task-technology ﬁt was controlled by providing subjects either with tasks that
matched the navigational characteristic of the Web site or with tasks that did not exactly
correspond to the navigational characteristic of the Web site. More speciﬁcally, subjects
in the good ﬁt group were required to ﬁnd and download any ﬁve MP3 songs of a
particular album from a speciﬁc artist from the horizontal-depth navigation Web site,
MP3Kingdom. com. Later, they were once again asked to ﬁnd and download a speciﬁc
song from a speciﬁc artist using the interactive navigation Web site, MP3 T opia. com. On
the other hand, subjects in the poor ﬁt group were asked to seek and download any ﬁve
songs from a speciﬁc artist’s album by using the interactive navigation, MP3 T opia, and
later ﬁnd a speciﬁc song using the horizontal-depth Web site, MP3Kingdom. Since the
interactive Web navigation (MP3 T opia) does not provide any information of the album
until a song is chosen, the user would be compelled to tap any song in the list and verify
if the song is from the album required to ﬁnd. If not, she would go back to the song
search page, and choose other songs repeatedly until ﬁve songs are found. By the same
token, to search a speciﬁc song in the horizontal-depth Web navigation (MP3Kingdom),

the user would need to look into all albums of an artist to ﬁnally ﬁnd and download the

39

required song. Such arbitrary manipulations of Web site interface and task assignment

has allowed us to effectively manipulate task-technology ﬁt in mobile Internet use.

Social Inﬂuence Stimuli

Social inﬂuence was controlled through word-of-mouth messages expressed by
the experiment trainer, who, at the same time, was a confederate for the experiment. Prior
to perform the actual tasks, all subjects were required to go through a training session in a
separate location where a trainer instructed them on using the PDA, connecting
wirelessly to the online system, and searching and ﬁnding the songs from the Web sites.
During the instruction, the trainer casually but clearly conveyed his opinion about using
wireless Internet on mobile devices and downloading MP3 ﬁles. The subjects were
ignorant that his expressions had been planned and scripted prior hand, and were
arbitrarily exposed to either positive or negative messages. Those who were randomly
assigned to the positive social inﬂuence group were exposed to positive expressions,
while those who were randomly assigned to negative social inﬂuence group heard
negative expressions coming from the trainer. For the consistency of social inﬂuence, the
same trainer continually instructed all subjects and expressed unvarying positive word-of-
mouth messages to every positively inﬂuenced group subjects or the same negative
messages to each negatively inﬂuenced group subjects. As scripted, the trainer spoke the
following messages to the positively and negatively inﬂuenced participants. The entire

script for the trainer on instructing and inﬂuencing the subjects can be found in Appendix

2.

40

Figure 9: An Example of Horizontal-Depth Navigation and Interactive Web Navigation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Horizontal-Depth Interactive
-W,....~-v-“T“'”—‘ WI)" “ ... —- -- _....~ _.——-——-"'""_“" LL" " " "‘
MP3Kingdom.Com MP3Topia.Com
Welcome to MP3 Topia Type in your artist’s name
Search for:
Search for:
Blues Latin Music
Broadway New A e
_ ...—L
Gos el _Rjjg
Classical ﬂu
County Rock
‘ International Soundtrack
k Jazz Other ’- \

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Rock at MP3Topla Rock at MP3Topla a
Browse by Artist’s Name “A” Artists: Search Songfl'itle
A 'B‘ 9 Aaron Johnson Search for:
Q E E ‘—
G H | Abacus
— — - Abdomen .
Ill T1 (L) _’ QC/DIC t ——Back In blackI
_ _ _ ooe era ors
§ I u Adam Rickitt
V w x Adema
V f _ Adicts
_ — Aerosmith
Afﬁnig ’v
ACIDC a ACIDC “Back In Black”
Back In Black
Browse by Album:
- . Do you want to download and
I .
Back in Black On "“5 A bum purchase:
Bonﬁre
—— - ll Bell
Flick of the Switch .. 5:03, ,0 ﬁn." __ Artist: AC/DC
Highway to Hell _ ave a Drink on Me Song: Back In Black
In This WOI'Id aaCk in BIaCk Album: BaCk In BIaCk
Wh——————-° Made Wh° Give the Dog a Bone
Rock n Roll Arnt Norse Download Now
Shake 3 Le
K7

 

4l

 

 

Positive Inﬂuence: “Isn’t this cool? Of course, you might be able to
download songs for free by using Kazaa or Morpheus, but that is now
considered illegal, right? And you don’t actually need to synchronize
with your computer to put your music on your PDA. I wonder if they are
planning to commercialize this type of MP3 service, because if they do,
and if I have the device, I’ll deﬁnitely use it. I heard there exist similar
services abroad, like in Japan, Australia, and Europe. They can actually
download songs on cell phones and use them as MP3 players! You may
know that using P2P ﬁle sharing programs can actually penalize the
individual users since it’s now legally punishable by law.”

Negative Inﬂuence: “I’m doing the training sessions because I was asked
to, but personally, I don’t think this will work commercially. You can buy
MP3 players for a fraction of the price of a PDA, and you can download
songs for free through any P2P ﬁle sharing program, such as Kazaa or
Morpheus. I heard that there exist such services in overseas, like in Japan
and Europe, but it seems that they are not quite popular. Downloading is
fast here because we’re using a wireless LAN, but if we were using a
cellular technology, the process will take so much more time. I guess this

will end up being useless.”
Our assumption on the effects of such word-Of-mouth messages is that those who

were exposed to negative expressions would be sufficiently socially inﬂuenced in a

negative manner, while those exposed to positive expressions would be amply positively

42

affected. As for the control group subjects, instead of going through a trainer for the
training session, they were directly led to the experiment administrator to receive basic
instructions without any affective statements, and immediately begin the trials at the

same location.

Sampling

Samples were drawn from a population of college undergraduate students
enrolled in telecommunication, communication, or advertising classes at a major
Midwestern university. Subjects were gathered through announcements in the classrooms
during a period of two academic semesters in the year 2004. To recruit the students, we
contacted the instructors to obtain permission to come to their class and ask for
participation. When permission was granted, we advertised, without telling the real
purpose, the subject matter of the experiment, the location for the trials, and informed
that a Web page had been posted, on which those who would be interested can Sign up
their names and e-mail addresses, and choose a time slot desired. The instructors have
agreed on offering extra credits for the course to the participating students in order to
encourage participation. As a result, a total of 210 students were recruited and randomly

assigned to one of the six treatment group conditions.

Procedure
Upon entering the experimental site, each participant ﬁrst completed a consent
form of voluntary participation and then received basic instructions on using a PDA for

Web search. While performing the training session, the trainer casually expressed

43

negative or positive opinion on using such a new technology for searching, purchasing
and downloading MP3 ﬁles on the mobile devices. To prevent any potential suspicion
that might arise among the subjects, the training area was located in a separate room
situated at a certain distance from the experimental site. The trainer was a Caucasian
male graduate student who could speak to the participants with a seemingly expert
impression. After the training session, the subject went to the experiment site, and
following a strict protocol, he or she was asked to ﬁnd and download MP3 ﬁles from the
Web sites on the PDA. There were a total of two tasks to complete, each of which
consisting of two subtasks; two using a horizontal-depth navigation Web site
(MP3Kingdom), and two on the interactive Web site (MP3 Topia). Each task was
separately described on a paper Sheet so that the participants could not know in advance
the subsequent tasks (see appendix 3). Every participant was allowed to freely listen to
the music or browse on the Web page when a task was over. The search time, number of
retraced Web pages, and number of correct downloads were automatically and
objectively monitored and collected on a database computer ﬁle. After each set of tasks, a
short questionnaire designed to record the subjects’ evaluation, perceptions, and purchase
intention was provided to measure their perceptive responses regarding searching and
purchasing MP3 ﬁles on mobile devices. Thus, a total of two questionnaires were handed
out to the participants. After ending the experimental session, the subjects were debriefed
and dismissed. Each trial, from training to completing the questionnaires, lasted between
20 to 30 minutes, depending on the condition to which the student was assigned, and the

personal ability of using an browsing the Web on a PDA.

44

Measurement

Measurement is the ﬁrst building block of science, and this fact applies to both
physical and social sciences. However, in social science, not all measures can be
observed systematically. Cognitive reactions, such as affection (attitude), belief, or
impression (perception) need to be measured subjectively and in some degrees
equivocally from the individual’s point of view. Yet, this characteristic of measurement is
by no means unique to social science, and thus it must be recognized that all
measurements are arbitrary at base (Babbie, 1998). Therefore, in social science, where
most phenomena are assessed with hypothetical constructs not directly observable, such
variables must be measured through inference (Emmert, 1989).

Most hypotheses of this study, except for those concerning success of trials, deal
with perceptive measures of evaluation, perception, and intention. An individual’s
perception is a function of his past history, his state of the moment he is viewing the
stimulus and the value of the object to the individual (Schiff, 1970). Attitudes
conceptually differ from other such states of readiness in that they predispose people to
- respond in a preferential manner (Emmert, 1989), they are implicit processes having
reciprocally antagonistic properties and varying in intensity (Osgood et al., 1970), and
they tend to be kept more internally consistent without occurring at the spur of the
moment (Schiff, 1970). Because our study involves introducing and varying stimuli by
applying experimental treatments to the subjects, the concept of perception would
correspond better than attitude to the cognitive reactions measured in regard to the device

that has been tested.

45

Although there exist numerous scale construction methods, for the sake of
measuring these perceptive variables, seven-point Likert scale questionnaire items were
created. The Likert method is based on the assumption that the overall score based on
responses to the many items that seem to reﬂect the variable under consideration provides
a reasonably good measure of the variable (Babbie, 1998). While other scaling methods,
such as Bogardus Social Distance, Thurstone, or Guttrnan prove to be equally, if not
more, effective measures of perceptive responses, the process of item creation may
present a tremendous cost in terms of time and effort to the researchers. Also, the
Semantic Differential Scaling was exempted from scale creation. This method is
considered to better approach attitude measurement since it represents a general

procedure for assessing affective responses (Heise, 1970).

Multiple Item Measures

The process of operationalization of the constructs yielded to a multiple-item
measure of the variables to be tested. Although single-item measures should not be
judged as being less effective than multiple-item scales in providing reliable measure of
relatively complex constructs (L00, 2002), in social science, a consensus is established
that the latter provides elaborate reﬂection of the rather abstract perceptive constructs. In
fact, when studying the TTF effect on utilization, Goodhue and Thompson (1995)
questioned 662 subjects about using information systems and obtained high reliability
estimates for each factor in the TTF dimension. For instance, they found Cronbach’s
alpha values of .84 for quality, .75 for locatability, .74 for ease of use/training, and .88 for

relationship with users. Each factor had at least two items indicating the construct to be

46

measured. Also, in a TAM study, Davis et al. (1989) showed that a two-item behavioral
intention scale obtained a Cronbach alpha reliability of .84 and .90 at time 1 and time 2 in
their study, while a four-item usefulness scale achieved a reliability of .95 and .92, and a
four-item ease of use scale obtained reliability coefﬁcients of .91 and .92, respectively.
We also adopted the multiple-item measure for the sake of the structural equation
model (SEM) testing. In SEM, all concepts correspond to latent variables, and the
observed variables indicate their latent variables. Latent variables are unobserved
variables implied by the covariances among two or more indicators (Hoyle, 1995).
Although observed, or manifest, variables of a latent variable may contain random or
systematic measurement errors, it is granted that such multiple-item scales are most
suitable for indirect measurement of abstract latent variables, such as intelligence, social

class or expectations (Bollen, 1989).

Data Collection Instruments

The questionnaire items were carefully chosen from the TAM, SIM and TTF
literature. Some wordings of these items have been slightly modiﬁed for the purpose of
m-commerce study. A total of eighty-six (86) items were formulated to measure the
perceived ease of usability, perceived usefulness, perceived social inﬂuence, enjoyment,
purchase intention, trainer’s communication styles, demographic data, prior experience,
and self-efﬁcacy (see Appendix 1). Questions concerning the trainer’s communication
styles were excerpted from the discipline of interpersonal communication. Especially, the
dominant Communication Styles Measures (Rubin et al., 1994) were taken into

consideration to statistically control, if necessary, for the variance among leaders’

47

subjective characteristics. In addition to the self-reported questionnaire, success of trials
was measured by the search time, number of retraced Web pages it took to ﬁnd the
targeted ﬁles, and the number of correct ﬁles found and downloaded. The following
paragraphs speciﬁcally address the items put on the questionnaires and the trial data for
the measurement of the dependent variables. The actual statements of the questionnaire

and their corresponding construct are illustrated later in Table l.

Perceived Task- Technology Fit

Task-technology ﬁt is deﬁned as “the degree to which a technology assists in
individual in performing his or her portfolio of tasks. .. [it] is the correspondence between
task requirements, individual abilities, and the functionality of the technology” (Goodhue
and Thompson, 1995). As previously described, Task-technology ﬁt was arbitrarily
manipulated by dividing the treatments into two experimental groups to which subjects
were randomly assigned prior to participation. However, since we did not maneuver the
actual task completion sequence for each participant, the extent to which a person feels
how well the technology supports the task would be different in many cases. Therefore it
is obvious that a measure that reﬂects their evaluation of the ﬁt between the task and the
technology is required. Such appliance of user evaluations as surrogates for task-
technology ﬁt has been commonly supported in the TTF literature (Goodhue, 1995).
According to Goodhue (1998), various perceptive constructs are cited to reﬂect the
concept of TTF. Among the listed ones, accuracy, compatibility, rightness of data,
tightness of level of detail, locatability, accessibility, ﬂexibility, meaning, assistance, ease

of use, system reliability, currency, training, authorization, presentation, and confusion

48

were deemed as indicators of task-technology ﬁt. Seemingly, the level of technology ﬁt
can be measured in multiple aspects through a collection of diverse sub-constructs. In this
study, this concept is measured in terms of rightness of features, support, and general ﬁt
between the functions and the speciﬁc tasks. Each item was measured on a seven-
point scale, in which a score of one (1) would indicate a strong disagreement, a four (4)
signifying a neutral state, and a seven (7) suggesting a strong agreement to the
corresponding statement. While it is arguable that some other constructs reﬂecting TTF
could also be considered, these were excluded ﬁom the list of questionnaire items
because the current study speciﬁcally examines ﬁle searching functions of a mobile
device’s Web interface. Moreover, since the constructs of ease of use and usefulness of
the system are the key determinants of the Task Acceptance Model, items for these
constructs would be more appropriate to be treated as indicators of perceptive measures

rather than the perception of TTF.

Perceived Ease of Use

One of the main variables in the TAM model is the perceived ease of use
(PEOU) of the system. The perceived ease of usecan be deﬁned as “the degree to which
a person believes that using a particular system would be free of effort” (Davis, 1989).
Following numerous previous studies in TAM literature (Davis et al., 1989; Malhotra and
Galletta, 1999; Venkatesh et al., 2002; Klopping and McKinney, 2004; Saade and Bahli,
2005), for our research the perceived ease of use was measured with items indicating the
subjects’ sense of time consumption in accomplishing the task, learning of the device,

and overall easiness. As with the case of perceived TTF construct, each of the above

49

items was measured on a seven-point scale, with one (1) implying a strong disagreement,
a four (4) suggesting a neutral state, and a seven (7) indicating a strong agreement to the

corresponding statement.

Perceived Usefulness

Another key variable in the TAM literature is perceived usefulness (PU) of the
technology. Perceived usefulness is, by deﬁnition, “the degree to which a person believes
that using a particular system would enhance his or her job performance” (Davis, 1989).
In our study, this construct was assessed with questionnaire items suggesting the
usefulness in searching and locating ﬁles, appropriateness of use, and general usefulness
of the interface. Again, every statement was set on a seven-point scale measure: a strong
disagreement for a score of one (I), a neutral state for a score of four (4), and a strong

agreement for a score of seven (7).

Use Intent

According to TAM, intention to use, or use intent (U1), of a new technology is
determined by PEOU and PU, and also leads to the actual behavior (usage). Therefore,
this variable has been vastly treated as a mediating variable rather than a dependent one.
However, one stream of technology acceptance research focuses on individual acceptance
of technology by using intention as a dependent variable (Compeau and Higgins, 1995).
Though the role of intention as a predictor of usage is crucial and has been well-founded
(Sheppard et al., 1988; Ajzen, 1991; Taylor and Todd, 1995), our research implements an

introduction of a new system and service that are yet hardly practiced in today’s

50

information technology market. The fact that there is limitation of collectable data in our
study, and that the path between intention and usage had been conﬁrmed by numerous
extant studies (Venkatesh et al., 2003) yielded us to drop the measurement of actual
usage, resulting in conﬁning the research model to use intent as a dependent variable. The
following statements represent ﬁve questionnaire items that were included in the surveys.
By the same token as the perceptive measures, each item was assessed on a seven-point
scale, in which a score of one (1) would indicate a very low intention, whereas a seven

(7) a very high likeliness to use the service.

Success of Trials

Success of trials is an initial performance measure, which reﬂects how easily and
well a user utilizes the technology (Goodhue and Thompson, 1995). However,
performance is associated with a user’s overall ability to accomplish certain given tasks,
and encompass an idea of continuity beyond the effectiveness of an initial usage. Since
our study introduces new systems and services not frequently used by the participants, it
is believed that success of initial trial of the new technology would better describe our
variable than the word “performance.”

Success is an objective measure assessed with various tools of measurement
evaluating task accomplishment. Since one can apply various criteria to analyze the
degree and quality of success in task-completion, a multiple-item approach, rather than
single-item, would be appropriate. Unlike the self-reported perceptive constructs, which
were interpreted with survey question items, measures of successful use were collected

by employing an automated computer data collection program that monitored the

51

sequence of activities and stored it on a Web-based database ﬁle. When a subject began
searching and navigating the Web site, the database ﬁle started recording the time it took,
the number of Web pages the person went through, and the number of correct ﬁles found
until all the downloads were conﬁrmed. The subject was not aware of the monitoring
process, nor was she told how well she did on the tasks. Thus three indicators, namely,
time, number of pages, and number of correct ﬁles, were measured to determine success

of trials.

Reliability Test

Prior to analyzing the collected data to examine the results, a reliability test was
run to verify the authenticity of the measures to their respective constructs. Reliability,
which is often estimated by Cronbach’s alpha values, indicates the extent to which
measurements are repeatable (N unnally and Durham, 1975), or have a relatively high
component of true score and relatively low component of random error (Carmines and
Zeller, 1979). Since the number of items for the key constructs of the study range from
three to eight, problems concerning their internal consistency might be critical if some
dimensions fail to correctly indicate the corresponding variable.

Results show that at Trial 1, which is immediately after the completion of the
ﬁrst set of tasks, the reliability test of all ﬁve items measuring Perceived TTF yielded a
Cronbach’s alpha value of .949. Also, the reliability estimate of the eight items for
Perceived EOU at Trial 1 proved to be .930, while the value of ﬁve measures for
Perceived Usefulness was .948. As for the intention to use, also indicated by ﬁve items,

we found a value of .948, and for the dimension of success of trials, which had three

52

measures, the Cronbach’s alpha showed .680. Quite similar ﬁndings were obtained for
the measures at Trial 2, i.e. after the second set of tasks. At Trial 2, Perceived TTF
measures indicated an alpha value of .973, Perceived EOU Showed a value of .964,
Perceived Usefulness revealed .937, intention to use exhibited .949, and success of trials

generated a Cronbach’s alpha of .755.

Exploratory Factor Analysis (EFA)

Although the above reliability indices conﬁrm that the measures for each

 

variable are internally consistent, a contention suggesting that some indicators would Ir"
denote equally high coefﬁcient values toward other unintended constructs may arise. To

verify whether there exist cross loadings among the perceptive items, we executed an

exploratory factor analysis (EF A) for the indicators of the cognitive variables, i.e.

Perceived TTF, Perceived EOU, and Perceived Usefulness. Varimax rotation (Kaiser,

1958) of the three factors yielded EFA results suggesting that, indeed, some items

proposed for perception measures were highly related to items planned for evaluative

measures, and vice-versa (see Table 2 & 3). By eliminating the potentially problematic

measures, cleaner groupings emerged as items originally projected to reﬂect their

relevant variables were less associated to others (Table 4 & 5). Statistical analysis

software, SPSS version 11.0, was implemented to compute the estimates.

53

Table I: Questionnaire Items

 

Construct & Item

Questionnaire Statement

 

 

 

 

 

 

 

Features The interface has the exact features for me to carry out the tasks
C‘-
.§ Task Fit The interface ﬁts well with the tasks I was required to perform
S
3 Support The interface fully supports the purpose of the tasks
u. It was difficult to complete the tasks because some features were not
E: Lack Features .
avarlable
General Fi t In general, the Interface had functIonalItIes that corresponded to the
tasks
Task Difﬁcult The tasks I performed were difficult
Interaction Clear My interaction with the interface was clear and understandable
8
a Skillful I quickly became skillful at using this interface
0
5': Learning Easy Learning how to fully use the interface was easy for me
«I
to
‘3 Interface Find It was easy to get the interface search and ﬁnd what I was looking for
.2.
o . .
8 Too Much Time It took too much tIme to ﬁnd the ﬁle I was seekrng through the
3: Interface
. The interface was so complicated, it was difﬁcult to understand what
Complicated
to do
General EOU In general, the interface was easy to use
m Useful Search The interface would be useful for any type of ﬁle search
(h
d)
g Improve Ability This Web site would improve my ability to locate other music ﬁles
0
m
; Useless I think the interface is useless for locating music ﬁles
>
'§ Appropriate Use This was an appropriate use of a Web site interface on the PDA
«‘3
9" Generally Useful Generally, I ﬁnd the interface useful
Public Place If you were in an airport with a mobile device, what would be the
probability that you consider buying songs through such services?
Home If you were at home with the device, what would be the probability
"‘9’ that you consider buying songs through such services?
If; Outdoors If you were relaxing on the beach with the device, what would be the
E. probability that you consider buying songs through such services?
0
‘5 Present If you had to decide whether or not to buy the songs using this tech-
nology at this moment, how likely is it for you to make the purchase?
Future In the future, if you had access to similar services on mobile devices,

how likely is it for you to buy music?

 

54

 

 

 

Table 2: Rotated Component Matrix of Perceptive Items at Trial 1

 

 

 

 

 

Loadings

Item Factor 1 Factor 2 Factor 3
Features (TTFI) .595
Task ﬁt (TTF2) .603
Support (TTF3) .576 .549 1.1
Lack features (TTF4) .506 .527 1
General ﬁt (TTFS) .618
Task difficult (PEUl) .581 L
Interaction clear (PEU2) .861
Skillful (PEU3) .892
Learning easy (PEU4) .901
Interface ﬁnd (PEU5) .599 .558
Too much time (PEU6) .879
Complicated (PEU7) .784
General EOU (PEU8) .771
Useful ﬁle search (PUI) .780
Improve ability (PU2) .705
Useless (PU3) .809
Appropriate use (PU4) .514 .706
General useful (PU5) .648 .609

 

 

 

 

 

 

(Rotation Method: Varimax with Kaiser Normalization)
(Values under .50 omitted)

55

7
I
q.
-
"

Table 3: Rotated Component Matrix of Perceptive Items at Trial 2

 

 

 

 

 

 

Loadings
Item Factor 1 Factor 2 Factor 3

Features (TTFl) .646

Task ﬁt (TTF 2) .607 .661

Suvport (TTF3) .695

Lack features (TTF4) .628 .622
General ﬁt (TTFS) .647 .641 .
Task difﬁcult (PEUI) .825 L}
Interaction clear (PEU2) .799 ’
Skillful (PEU3) .836

Learning easy (PEU4) .841

Interface ﬁnd (PEU5) .621

Too much time (PEU6) .681 ‘

Complicated (PEU7) .640

General EOU (PEU8) .767

Useful ﬁle search (PUI) .808

Improve ability (PU2) .773

Useless (PU3) .627
Appropriate use (PU4) .722
General useful (PUS) .672 .611

 

 

 

 

 

 

(Rotation Method: Varimax with Kaiser Normalization)
(Values under .60 omitted)

56

Table 4: Rotated Component Matrix at Trial 1 (after elimination of problematic items)

 

 

 

 

 

 

 

Loadings
Item Factor 1 Factor 2 Factor 3

Features (TTFl) .775

Task Fit (TTF 2) .782

Support (TTF 3) .750

Interaction Clear (PEU2) .839

Skillful (PEU3) .867

Learning Easy (PEU4) .889

Useful File Search (PUl) .771
Improve Ability (PU2) .838
Appropriate Use (PU4) .682

 

 

 

 

Rotation Method: Varimax with Kaiser Normalization)
(Values under .60 omitted)

Table 5: Rotated Component Matrix at Trial 2 (after elimination of problematic items)

 

Loadings

 

Item

Factor 1

Factor 2

Factor 3

 

 

Features (TTF 1)

Task Fit (TTF2)

Support (TTF3)
Interaction Clear (PEU2)
Skillful (PEU3)
Learning Easy (PEU4)
Useful File Search (PUl)
Improve Ability (PU2)
Appropriate Use (PU4)

 

.782
.853
.848

 

.71 1
.757
.806

 

.728
.683
.633

 

 

(Rotation Method: Varimax with Kaiser Normalization)
(Values under .60 omitted)

57

Conﬂrmatory Factor Analysis (CFA)

Since measures of all three cognitive variables, that is Perceived TTF, Perceived
EOU, and Perceived Usefulness, were theoretically and empirically supported to Show
construct validity (Davis, 1989; Goodhue, 1998; Venkatesh and Davis, 2000; Klopping
and McKinney, 2004), a conﬁrmatory factor analysis (CFA) was performed to

corroborate the literature ﬁndings concerning these constructs. Structural equation

 

modeling (SEM) was run on EQS version 6.1 to perform the computation for the CFA of

the model of perceptive measurements. The CFA tests the construct validity of the factor

 

..
loadings to the latent variables. Figure 10 and 11 illustrate the CFA models with all factor L"
loadings indicating unidimensionality

Results show that both CFA models proved to be well-ﬁtting. As for the Trial 1
CFA model (Figure 10)goodness of ﬁt measures indicated that Bentler and Bonett’s

(1980) Norrned Fit Index (N F I) was .973, Bentler and Bonett’s Nonnorrned Fit Index

 

(BFI) showed .972, and the Comparative Fit Index (CPI) (Bentler, 1988) was .981, while
the Goodness-of-F it Index (GFI) measured .922, and the Adjusted Goodness-of—Fit Index
(AGFI) (Joreskog and Sorbom, 1984) estimated a value of .854. Since 0.90 is generally
considered to be the threshold value for these ﬁt indices, such highly signiﬁcant values
clearly indicate that the analysis proved support for the model. Also, although the Chi-
square ()8) value was 75.829 based on 24 degrees of freedom (df), lending a probability
value of less than 0.00] (meaning that the model is signiﬁcantly different from the
predicted population), the Root Mean Square of Approximation (RMSEA) index (Steiger,

1989; Browne and Mels, 1990) at .102 supported a reasonably good ﬁt as well.

58

F tyre 10: CFA Model for Perceptive Measures at Trial 1

 

 

 

TTF Perceived Perceived
Evaluation Ease of Use Usefulness

 

 

 

 

 

 

 

 

PU4

TTF 2
TTF3

 

 

 

 

 

 

 

 

 

 

 

 

PEU4
PU 1

 

 

 

 

 

 

 

 

 

 

 

ill 1

.38 .34 .36

——+ .

.21 .27 .42
(* based on test statistics, all factor loadings signiﬁcant at p<.001)

g _, PEU2
’43, _,w PU2

\0—>

F zgure 11: CFA Model for Perceptive Measures at Trial 2

 

 

 

 

TTF Perceived Perceived
Evaluation Ease of Use Usefulness

 

 

 

 

 

 

 

TTF2
TTF3

 

 

 

 

 

 

 

PEU2

 

 

PEU3
PEU4
PUl

 

 

 

 

 

 

 

 

 

 

 

 

 

e
l

——>
—>

ll ll

.41 .26 .28 .34 .24 .23 .40

5—, PU2
g_, PU4

(* based on test statistics, all factor loadings signiﬁcant at p<.001)

59

The Trial 2 CF A model (Figure l 1) equally conﬁrmed the construct validity of
the variables by generating an NF I value of .979, a BFI of 982, a CF I of .988, a GP I
of .943, an AGFI of .892, and an RMSEA of .081. The Chi-square estimate marked
56.878 on 24 degrees of freedom. The reason for the drop in index value of the AGF I can
be explained by the nature of this measure. Because the AGFI adjusts for the degrees of
freedom of a model relative to the number of variables, it rewards simpler models with
fewer parameters (Bollen, 1989). Shown below, Table 6 arranges all ﬁt indices for both

CFA models.

Table 6: Fit Indices for CF A Models of Cognitive Measures at Trial 1 and Trial 2

 

 

 

 

Indices

Fit Measure Trial 1 CFA Trial 2 CFA
NF I .973 .979
BFI .972 .982
CF I .981 .988
GFI .922 .943
AGFI .854 .892
RMSEA .102 .081
x’ (dt) 75.829 (24) 56.878 (24)

 

 

 

 

Measures selected, we repeated the reliability tests of the variables without the
items dropped. Measures of success of trials underwent revision as well, and an item
showing weaker overall correlation within the construct was eliminated. Speciﬁcally, the
number of correct ﬁles found during the task completion was removed. Since the
requirement of the tasks involving ﬁnding and downloading particular music ﬁles was

not tortuous, most participants somehow managed to successfully complete the tasks

60

regardless of the number of pages or time it took to locate them. Only 27 out of 105
subjects in the poor-ﬁt groups had one or more incorrect downloads at Trial 1, and
merely 5 out of 105 chose wrong ﬁles at Trial 2 (in the good-ﬁt groups, all subjects but
one person downloaded incorrect titles). Also, among the 27 underachieved subjects, just
5 showed error rates of 50% or more. At Trial 1, the Pearson’s r correlation between
number of correct ﬁles and search time was .073 (r2 = .0054) and for the number of pages,
the r value was .011 (r2 = .0001). Therefore, it is assumed that, in our study, the number
of correct music ﬁles fails to adequately indicate the degree of a trial’s success.

After the reliability tests were rerun, the initial values were compared with the
ﬁxed values. Cronbach’s alphas showed minimal changes for all cognitive and readiness
dimensions, namely TTF evaluation, perceived ease of use and usefulness, and use
intention, while in the case of success of trials, the alpha value increased by almost 0.3 to
show .955 at Trial 1. The initial and the reassessed reliability estimates are presented on

the following Table 7.

Table 7: A Comparison of Cronbach’s Alphas between the Initial and Selected Items

 

 

 

 

 

 

Reliability Estimates
Trial 1 Trial 2 Trial 1 Trial 2
Construct Without Without After Dropping After Dropping
Dropping Items Dropping items items items
TTF Evaluation .949 .973 .953 .961
Perceived EOU .930 .964 .977 .973
Perceived U .948 .937 .948 .912
Use Intent
(no item dropped) .948 .949 .948 .949
Success of Trials .680 .755 .955 .746

 

 

 

 

 

61

 

Other Measures

Other measures include: demographic data, prior experience, enjoyment,
trainer’s communication style, self-efﬁcacy, and familiarity with the task object.
Although these variables were not reﬂected on the research questions and hypotheses as
being critical antecedents of technology acceptance, they are taken into consideration to

control for potential intervening factors.

Demographic Data

Since our research involved experimental studies with mostly undergraduates,
the data collected were restricted within the college level students. Although no
signiﬁcant results were assumed from the demographics of the participants, items were
included in the questionnaires to collect the subjects’ information on their gender, age,

major, and year in college.

Prior Experience

Measures of prior experience with electronic media were also gathered.
Speciﬁcally, subjects were asked to indicate the amount of time they have been spending
daily on the Internet, and whether or not they had ever used any peer-to-peer (P2P) ﬁle
sharing programs, cell phones, personal data assistants (PDAs), and/or MP3 players.
Those who gave positive responses regarding their media usage experience were also

asked to provide the length of time they had used them in terms of months and years.

62

 

 

 

Enjoyment

Enjoyment is an abstract term that can be explained by the state of having
pleasure or keen satisfaction from an act or object. Previous electronic commerce and
TAM studies cite enjoyment as one of important factors that leads to satisfaction and
usage of online shopping systems (Li et al., 2001; Goldsmith and Goldsmith, 2002;
Katerattanakul, 2002; Van Dolen and De Ruyter, 2002). Contradictions to such claims
state that goal-oriented motives are more common among online Shoppers than are
experiential motives, and therefore, shopping is planned ahead hardly with any impulse
purchases, such that freedom of control is considered to be more important than fun
factors (Wolﬁnbarger and Gilly, 2001). Since our study involves experimental sessions,
which are mandatory rather than voluntary to the participants, we concur with this
argument, and omitted enjoyment from the model’s key variables. However, a series of

ﬁve items were included in the questionnaire to test the presence of this factor.

Trainer ’s Communication Style

Social inﬂuence through the words-of-mouth of the trainer is one of the crucial
determinants in our model. However, since a subject’s impression on the trainer’s
communication style is likely vary from person to person, it is deemed that not all
participants would be heavily inﬂuenced by the trainer’s messages. For this matter, the
dominant Communication Styles Measures (Rubin et al., 1994) were recorded on a six-

item dimension.

63

 

Self-Efﬁcag/

Self-efﬁcacy is another factor that could affect a user’s perception on using a
new technology. Based on the social cognitive theory (Bandura, 1982), self-efﬁcacy is
deﬁned as the belief that one has the capability to perform a particular behavior. To date,
empirical studies on TAM and self-efficacy have been abundant yet contradicting. For
instance, while Igbaria and Livari (1995) found strong support for direct and indirect
effects of self-efﬁcacy on users’ perceptions in using computers, Chau (2001) contended
that self-efﬁcacy shows negative or minor effects on perceptive variables. Although we
followed Chau’s ﬁndings and assumed no critical effects of self—efﬁcacy on users’
perception and use intent of a new technology, measures concerning users’ self-
awareness and evaluation of personal capabilities have been put into consideration to

verify potential effects of this factor.

Familiarity with the Tauk Object

Familiarity with the task object can play an important role in users’ success in
achieving certain tasks. Because of the mandatory nature of the tasks in our experiment,
which requested all participants to ﬁnd and download speciﬁc music titles from the
designated Web sites, it is believed that those who coincidentally had previous
knowledge and profound interests in the musicians or songs we speciﬁed would
demonstrate high level of success in locating the required music titles. Although
familiarity might also concern perception and use intent of new media, this variable only
relate to individual recognition of the artists, albums, and songs in question by measuring

familiarity on a seven-point scale ranging from “don ’t know at all” to “know very well. ”

64

 

Manipulation Check

A t-test was performed to examine whether the participants in the social inﬂuence
treatment groups differed from each other in terms of levels of trainer credibility. The
indicators of the trainer’s communication style included measures of the trainer’s
persuasiveness, trustworthiness, and perceived trainer knowledge. However, items
indicating the degree of positively or negatively perceived messages were not taken. We
believe that the two types of experimental treatments, termed as positive and negative,
delivered enough favorable or unfavorable (according to the SI group the subject was
assigned) WOM messages just as the research pursued by Galletta et al. (1995) proved to
be effective when training the end-users on a new computer software. In their study, a
confederate acted to perform three unsanctioned outbursts during the training sessions to
express his appreciation or disapproval of the software to socially inﬂuence the other
participants.

As expected, the t-test indicated that there were no signiﬁcant differences between
the two groups in terms of persuasiveness (t = 0.385, p > .05), trustworthiness (t = 1.827,
p > .05), and perceived trainer knowledge (t = 1.894, p > .05). The results, as illustrated
in Table 8, Show that both groups were above the 5-point (for persuasiveness) and 6-point
(for trustworthiness and perceived trainer knowledge) in the scale, indicating high

credibility.

65

 

Table 8: Manipulation Check of Trainer Credibility

 

Measure Group (N) Mean St. Dev. t (If p

. . . Neg. (72) 5.22 1.746
Trarnerrs Persuasrve 0.385 207 .701

P08. (74) 5.34 1.682

. . Neg. (72) 6.11 1.069
Trarner Is Trustworthy 1.827 207 .069
P05. (74) 6.39 0.737

. . Neg. (72) 6.28 0.938
Trarnerrs Knowledgeable 1.894 207 .060

P08. (74) 6.53 0.602

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Chapter Summary

In this chapter we described the process of data collection methods, and provided
the rationale for choosing certain measures for testing the hypotheses. Issues and
procedures applied in gathering and measuring the key factors were explained and

construed. The following chapter discusses the results of data analyses from the 210

 

subjects who participated in our experiment. Accordingly, the proposed model of SIM

and TTF in TAM is revised and adjusted for an overall better ﬁt.

66

CHAPTER 4
RESULTS

The current chapter illustrates the quantitative data obtained from the
experiments and reports the results analyzed with statistical methods. As for the data 1
analysis methods, analysis of variance (ANOVA) multivariate analysis of variance

(MANOVA), and structural equation modeling (SEM) had been employed. Since data

 

integrity checks were provided in the previous chapter, we ﬁrst begin with an outline of
the descriptive statistics by examining the composition of the participants, and then
describe the outcomes from hypotheses testing. The model is reassessed and revised to

create a ﬁnal model.

Descriptive Statistics

As explained in the preceding chapter, a total of 210 subjects participated in the
experiments. As suggested by Wirnmer and Dominick (1997), every subject was
randomly assigned to one of the six treatment groups to allocate 36 participants in poor-
ﬁt/negative condition, another 36 in good-ﬁt/positive condition, 32 in poor-ﬁt/neutral
condition, 32 in good-ﬁt/neutral condition, 37 in poor-ﬁt/positive condition, and 37 in
good-ﬁt/positive condition. Among them, 144 were male (68.6%) and 66 were female
(31.4%) college students, and the average age of the subject was 21.21 years old with a
median of 21. Also, a majority was majoring Telecommunication (151 students, 71.9%),

followed by Advertising (22 students, 10.5%), Computer Science (6 students, 2.9%),

67

Communication (4 students, 1.9%), and other majors (27 students, 12.8%). Students in

their freshman year counted only 3 (1.4%), while sophomores were 36 (17.2%), juniors

92 (43.8%), seniors 76 (36.2%), and 3 graduate students (1.4%).

The mean value of daily lntemet use time was 251.32 minutes (4 hours, 11

minutes and 20 seconds) per day, with a median of 210 minutes (3.5 hours). When asked

whether or not they ever had or currently use a mobile device, 205 subjects (97.6%) gave
a positive response about cell phones, while 81 (38.6%) agreed with ever owning an MP3
player, and 58 (27.6%) responded positively about having a PDA. Regarding the use of a

P2P ﬁle-sharing program, a large number (190 respondents, 90.5%) reported having ever

used it. Table 9 partially illustrates these basic descriptive statistics in an organized

manner.

Table 9: Study Participants’ Descriptive Statistics Summary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Male =l44 (68.6%) Female = 66 (31.4%)
Yes No Yes No
Cell Phone 139 (96.5%) 5 (3.5%) 66 (100%) 0 (0%) Agﬁ‘;
MP3 Player 59 (41.0%) 85 (59.0%) 22 (33.3%) 44 (66.7%) ziugeﬁi
PDA 47 (32.6%) 97 (67.4%) I I (16.7%) 55 (83.3%) Daily
P2P File—sharing 133 (92.4%) 11 (7.6%) 9 (13.6%) 57 (86.4%) average
lntemet
Freshmen = 3 (1.4%) 3 (100%) 0 (0%) use 11::
Sophomores = 36 (I 7.2%) 19 (52.8%) 17 (47.2%) 1 1 mi"-
20 sec.
Juniors = 92 (43.8%) 70 (76.1%) 22 (23.9%)
Seniors = 76 (36.2%) 51 (67.1%) 25 (32.9%)
Graduates = 3 (1.4%) 1 (33.3%) 2 (66.7%)

 

68

 

Hypothesis Testing

The purpose of testing the hypotheses is to verify the relationships among the
variables and observe whether or not the effects were present as predicted earlier in the
dissertation. The ultimate goal, however, is to make inferences about the beliefs and
behavior of the subjects who have not been tested in the experiment (Keppel, 1991). Thus,
data analyses must show statistical power, which indicates the probability that a statistical
test of a null hypothesis will result in the conclusion that the phenomenon under study
actually exists (Cohen, 1988). For this matter, we employed a series of statistical tests,
including ANOVA, MANOVA, and SEM to corroborate the expected results from the

pursued study.

Analysis of Variance (AND VA)

The research design of the current study implied that the acceptance of a new
technology is inﬂuenced mainly by task-technology ﬁt and social inﬂuence. The
anticipated relationships between the independent variables and dependent variables state
that there will be a main effect of task-technology ﬁt (TTF) on users’ success of trials,
perceptions and use intention of the new technology (Hypothesis 4), and a main effect of
social inﬂuence (SI) on users’ perceptions and use intention of the new technology
(Hypothesis 5). Corresponding to these hypotheses, a three-by-two (3x2) two-factor
between subjects factorial design was adopted, and data gathered were ﬁrst analyzed with

the method of one-way multi-factor analysis of variance (AN OVA) using SPSS version

11.0.

69

Main Effect of TTF on Success Qf Trials

As explained in the previous chapter, success of trials was determined by the
number of Web pages and time it took for a user to ﬁnd all required ﬁles for the given
tasks. Data analysis results clearly indicate that there is a signiﬁcant treatment effect of
TTF on the success of trials at Trial 1, i.e. after the completion of the ﬁrst set of tasks.
The number of Web pages browsed averaged (M) 72.46 for the case of the good-ﬁt
groups, whereas the poor-ﬁt groups’ subjects went through (_111) 180.26 pages (E=756.632,
p<.001). Similar ﬁndings were observed during the second set of tasks, or at Trial 2,
when the good-ﬁt groups searched through (M) 16.87 Web pages, the poor-ﬁt groups had
to navigate through (M) 29.08 pages (E=69.086, p<.001).

Time was also an indicator that showed Signiﬁcant differences between the two
TTF groups. At Trial 1, each subject of the good-ﬁt groups took an average of (M)
419.32 seconds (almost 7 minutes) to complete the ﬁrst set of tasks, while the poor-ﬁt
groups’ users spent (M) 1034.52 seconds, or approximately 17 minutes and 15 seconds
(E=535.913, p<.001). Also, at Trial 2, the good-ﬁt groups showed a mean score of 132.36
seconds (2 minutes and 12 seconds), and the poor-ﬁt groups spent (1l_4) 182.34 seconds, or
3 minutes and 2 seconds (E=21.494, p<.001). Therefore, Hypothesis 4a (there will be a

main effect of TTF on users’ success of trials) was successfully supported.

Main Eﬂect of TTF on Perceived TTF
For the sake of the AN OVA method, we took the average score of the three
perceptive items for “features,” “task-ﬁt,” and “support.” Our ﬁndings suggest that

there is a signiﬁcant treatment effect of TTF on users’ evaluation with a mean score of

70

 

(M) 3.254 for the poor-ﬁt group, and (M) 6.165 for the good-ﬁt group (£=333.232,
p<.001). Similar outcomes were generated when we compared the evaluated TTF at Trial
2, with a mean score of (M) 4.118 for the poor-ﬁt group, and (M) 6.200 for the good-ﬁt
group (E=156.869, p<.001). Thus, Hypothesis 4b (there will be a main effect of TTF on

users’ TTF perception) was supported.

Main Effect of IT F on Perceived Ease of Use

By the same token as the perceived TTF, we took an average score of the three
items measuring perceived ease of use, namely “clear interaction,” “become skillful,” and
“easy learning,” and checked the signiﬁcance of score difference. The ANOVA results
indicate a signiﬁcant main effect on Perceived EOU at both Trial 1 and Trial 2. At Trial 1,
the mean score of perceived ease of use of the poor-ﬁt group was M) 4.400, while that of
the good-ﬁt group was (M) 6.610 (E=194.186, p<.001). At Trial 2, the same measures
scored (M) 4.622 and (M) 6.489 for the poor-ﬁt and good-ﬁt groups, respectively
(E=110.011, p<.001). We concluded that Hypothesis 4c (There will be a main effect of

TTF on perceived EOU) was indeed supported.

Main Eﬂect of 77F on Perceived Usefulness

The ANOVA test also supported the main effect of TTF on perceived usefulness.
Again, by taking the average of the three perceived usefulness measures, poor-ﬁt group
scores were compared with the good-ﬁt group. The mean score of the poor-ﬁt group at
Trial 1 Showed (M) 3.362, and that of the good-ﬁt group Showed (M) 5.746 (E=272.511,

p<.001). Also at Trial 2, the poor—ﬁt group’s mean score was (M) 3.911, and the good-ﬁt

71

group marked (M) 6.016 (£=166.698, p<.001). Therefore, Hypothesis 4d (There will be a

main effect of TTF on perceived usefulness) was supported as well.

Main Eﬂect of TTF on Use Intention
With the single-itemized use intent measures, another ANOVA test was

conducted, resulting in signiﬁcant difference between the poor-ﬁt and good-ﬁt groups.

 

The scores at Trial 1 indicated (M) 3.149 and 4.921 for the poor-ﬁt and good—ﬁt groups,

respectively (1_7=126.002, p_<.001), while at Trial 2, they showed (M 3.212 and (M) 5.044

 

(E=IZ6.420, p<.001). The assumption of null hypothesis for Hypothesis 4e (There will be LI

a main effect of TTF on use intention) was once again rejected.

Main Effect of Social Inﬂuence on Perceived TTF

Since social inﬂuence was another crucial independent variable in our research
model, the treatment effect of this factor on users’ perception of TTF was also analyzed.
The ANOVA results show that at Trial 1, the negative group scored (M 4.194, the
control group (M 4.542, and the positive group (M 5.356 (E=19.405, p<.001), while at
Trial 2, the scores were (Ms) 4.607, 4.953, and 5.874 for each of the three respective
groups (5:21.533, p<.001).

Although the observed F values suggest signiﬁcant difference among the three
socially inﬂuenced groups, there is no indication whether the differences are Signiﬁcant
between two speciﬁc groups (e. g. negative and neutral groups). We therefore ran contrast
tests involving only two of the three groups, and obtained evidence that, indeed, there

existed no statistical differences for the case between the negative versus neutral groups

72

(Triall: E=3.059, p>.05; Tria12: E= 2.806, p>.05), whereas the difference between the
neutral and positive groups was still signiﬁcant (Triall: _E=17.040, p<.001; Tria12:
5:20.044, p<.001). Thus, Hypothesis 5a (there will be a main effect of social inﬂuence
on users’ perceived TTF) was partially supported for the case of positively inﬂuenced

groups.

Main Effect of Social Inﬂuence on Perceived Ea_se of Use

The main effect of Social Inﬂuence on perceived EOU was also tested with the
mean values of the three groups. At Trial 1, the each of the three groups scored (Ms)
5.222, 4.938, and 6.270 for the negative, neutral, and positive conditions, respectively
(1_7=20.406, p<.001). As for Trial 2, these scores showed (Ms) 5.259, 5.130, and 6.212
(E=l4.962, p<.001). However, when contrast tests were conducted, only the differences
between the positive and neutral groups (Triall: 5:46.186, p<.001; Tria12: E=24.138,
p<.001), and positive and negative groups (Triall: E=30.371, p<.001; Tria12: 5:19.909,
p<.001) showed statistical signiﬁcance. The neutral and negative groups failed to provide
statistically signiﬁcant differences regarding the perceived ease of use (Triall: E=2.082,
p>.05; Tria12: _E=0.340, p>.05). Hypothesis 5b (There will be a main effect of social
inﬂuence on perceived EOU) was again partially supported within the constraint of

positively inﬂuenced groups.

Main Effect of Social Irnluence on Perceived Ugﬁylness

As for the treatment effect of social inﬂuence on perceived usefulness, data

analysis results evidenced statistical differences among all three groups. The negative

73

group showed a mean score of (M 3.699, the neutral group (M) 4.589, and the positive
group (M 5.356 at Trial 1 (5:45.788, p<.001), while at Trial 2, the three groups
averaged (Ms) 4.421, 4.823, and 5.613, respectively (F_‘=19.213, p<.001). When
compared by contrast tests, the usefulness-perceptive measures of the negatively
inﬂuenced groups were signiﬁcantly lower than those of the neutrally inﬂuenced groups
(Triall: E=24.68l, p<.001; Tria12: _=3.916, p<.05), and the scores of the positive groups
were Signiﬁcantly higher than the control groups (Triall: E=18.456, p<.001; Tria12:
5:15.343, p<.001). Hypothesis So (there will be a main effect of social inﬂuence on

perceived usefulness) was fully supported.

Main Effect of Social Inﬂuence on Use Intent

The mean score differences among the groups showed statistical signiﬁcance in
terms of users’ behavioral intention. The mean scores of the negative group indicated
(Ms) 2.838 and 2.801 , those of the control group showed (Ms) 3.912 and 4.073, and those
of the positive group were (MS) 5.306 and 5.469 at Trial 1 (£=85.565, p_<.001) and Trial
2 (5:93.284, p<.001), respectively. Contrast tests also proved that these mean values
were improbably similar even when only negative and neutral conditions were compare
(Triall: _E=29.866, p<.001; Tria12: _E=39.338, p<.001), and when neutral and positive
conditions were contrasted (Triall: E=51.065, p<.001; Tria12: _E=47.969, p<.001)
Social inﬂuence’s main effect on new technology use intention, as predicted by
Hypothesis 5d, was thus statistically fully supported. Figures 12 -— 23 represent the mean
values of the perceptive and intention measures at Trial 1 and Trial 2 of the experiment.

Also, the ANOVA test results are summarized in Table 10 below.

74

Table 10: ANOVA Summary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PTTF PEU pU UINT
T1 T2 T1 T2 T1 T2 T1 T2
@2303}; 3.25 4.12 4.40 4.62 3.36 3.91 3.15 3.21
p (5 (1.63) (1.73) (1.91) (1.97) (1.49) (1.60) (1.53) (1.62)
63331033; 6.17 6.20 6.61 6.49 5.75 6.02 4.92 5.04
p (g) (0.92) (0.94) (0.66) (0.72) (1.10) (1.00) (1.56) (1.63)
mg“? 15:3332 £=156.9 5:1942 5:1100 5:2725 E=166.7 E=126.0 E=IZ6.4
TTF 15001 94101 P<001 p<001 p<001 p<(X)1 p<(X)1 p<(X)1
Giigii’ﬁ 4.19 4.61 5.22 5.26 3.70 4.42 2.84 2.80
p (5 (1.93) (1.75) (1.92) (1.89) (1.52) (1.67) (1.51) (1.62)
Neutral
Groups, M 4.54 4.95 4.94 5.13 4.59 4.82 3.91 4.07
(g; (227) (2.04) (2.18) (2.05) (2.00) (1.90) (1.47) (1.52)
6:38:13: 5.36 5.87 6.27 6.21 5.36 5.61 5.31 5.47
p (g; (1.52) (1.11) (0.82) (0.97) (1.37) (1.31) (1.37) (1.34)
Main _ __ _ _ _ _ _ _
Effectof 5-194 _E—21.5 5—204 15—150 15—458 5—192 15—856 E933
31 $001 95001 D<001 P<001 p<001 p<001 p<001 p<001
Negative 3:30; :20? 5:26; a)“; _=24.7 _=3.9 5:299 5:393
VS. Neutral n.s. n.s. n.s. n.s. p<m1 9<05 D<m1 p<m1
Neutral vs. 5:170 5:200 5:462 5:241 _E=18.5 5:153 15:51.1 E=48.0
Positive p<(X)1 941)] “m1 D<m1 p<.(X)1 p<ml p<ml 124111

 

75

 

 

 

Figure 12: Number of Web Pages Browsed at Trial 1

200
180 1— + A
160
140
120

 

 

 

 

 

 

—O— Good-F it
+ Poor-Fit

100

 

 

 

 

80
60
40
20

0 1 1
Negative Neutral Positive

 

 

 

 

 

 

 

 

Figure 13: Number of Web Pages Browsed at Trial 2

35

 

 

30

 

 

25

 

20 -

/\o + Goal-Fit
+ Poor-Fit

15 .7

 

 

 

 

 

10

 

 

 

 

 

Negative Neutral Positive

76

Figure 14: Time Spent at Trial 1

I200

 

1000 ‘/‘\A

 

 

800

600

 

 

 

 

+ Good-Fit
+ Poor-Fit

 

 

 

400 4

 

200

 

 

 

Negative Neutral Positive

Figure 15: Time Spent at Trial 2

200

 

160

 

 

/ #0
120

 

 

 

—O— Good-Fit
+ Poor-F it

 

 

 

 

 

 

 

80
40
O l J
Negative Neutral Positive
77

 

Figure 16: Perceived TTF at Trial 1

 

 

 

7
A

6 ./

5

4 /‘

 

 

 

 

—O— Good-Fit
+ Poor-F it

 

 

 

 

 

 

Negative Neutral

Figure 17: Perceived TTF at Trial 2

Positive

 

 

 

 

 

 

 

+ Good-Fit
+ Poor-F it

 

 

 

 

 

 

Negative Neutral

78

Positive

 

 

Figure 18: Perceived EOU at Trial 1

 

 

 

 

 

+ Good-Fit
3 + Poor-Fit

 

 

 

 

 

 

 

 

 

Negative Neutral Positive

Figure 19: Perceived EOU at Trial 2

 

 

 

 

 

\/ _._ GOOd-FII

3 + Poor-Fit

 

 

 

 

 

 

 

 

 

Negative Neutral Positive

79

F igre 20: Perceived Usefulness at Trial 1

 

 

 

 

7
6 f A
5 //

 

 

 

 

 

+ Good-Fit
+ Poor-Fit

 

 

 

 

 

 

3 | f
k
2
1
O L l
Negative Neutral Positive

Figure 21: Perceived Usefulness at Trial 2
7

 

/.—\-O

 

6 /

 

.5. /‘

 

 

.\/

 

 

+ Good-Fit
+ Poor-Fit

 

 

 

 

 

 

Negative Neutral Positive

80

 

 

 

F igre 22: Use Intention at Trial 1

 

 

 

 

 

./ / + GOOd-FII
+ Poor-Fit

 

 

 

 

 

 

 

 

 

Negative Neutral Positive

Figure 23: Use Intention at Trial 2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

./ + Good-Fit
3 + Poor-Fit
2
1
O l 1
Negative Neutral Positive

81

Multivariate Analysis of Variance (MANO VA)

Along with the main effects of the independent variables, we anticipated an
interaction effect between TTF and social inﬂuence, the two key determinants of our
study. An interaction is observed when the simple effects of one independent variable are
not equal at all levels of another independent variable, thus making the effects of one of
the independent variables to be conditionally related to the levels of the second
independent variable (Keppel, 1991). The ANOVA results demonstrated the presence of
the main effects of TTF and social inﬂuence. However, evidence was not provided
regarding the difference of amount of variation in users’ perception and reaction along
the TTF levels in regard to the social inﬂuence levels. In other words, we suspected that
the strength of the main effect of social inﬂuence would be different according to the
TTF levels, and vice-versa. Such interaction effects on users’ perceptions, and use
intention were predicted in Hypothesis 5. A multivariate analysis of variance
(MANOVA) was conducted on SPSS 11.0 to test the hypothesis and verify if the data

were meaningful.

Interaction Effect of 77F and Social Iuﬂuence on Perceived TTF

To verify whether there is an overall interaction effect between the treatments, an
omnibus ANOVA was initially performed. The outcome showed that there exists
signiﬁcant interaction effect between TTF and social inﬂuence factors at both Trial 1
(_E=15.55, p<.001) and Trial 2 (_E=15.30, p<.001). However, since this result fails to
indicate any particular differences between any two speciﬁc social inﬂuence treatment

groups, an interaction contrast test was then pursued to observe and analyze each of the

82

relations more closely. Through a MANOVA test, ﬁrst we performed an interaction
effect analysis with the negative and neutral groups in contrast. The observed F values at
both Trial 1 (E=7.672, p<.01) and Trial 2 (E=8.667, p<.01) indicate a clear interaction
effect between TTF and social inﬂuence in the case of negative and neutral treatments.
Next, the interaction effect was analyzed for the neutral and positive groups. The
signiﬁcance tests once again suggested that both at Trial 1 (_E=31.036, p<.001) and Trial
2 (5:30.592, p<.001), the average TTF perception score difference between the good-ﬁt
group and poor-ﬁt group is not the same when users are socially inﬂuenced in different
ways. Hypothesis 6a (there will be an interaction effect between social inﬂuence and

task-technology ﬁt on users’ perception of TTF) was therefore not rejected.

Interaction Effect of TTF and Social Influence on Perceived EOU

AS with the analyses of interaction effect on perceived TTF, the average scores
to be measured were replaced with users’ perceived ease of use and a MANOVA for the
two treatments was performed. Since the omnibus F test provided evidence for a
signiﬁcant interaction effect (Trial 1: 5:32.62, p<.001; Trial 2: _E=20.36, p_<.001), we
proceeded with the interaction contrast effect test to isolate and compared two of the
three social inﬂuence treatments. Again, an interaction contrast effect between the poor-
ﬁt and neutral groups was conﬁrmed (Trial 1: 5:11.290, p<.005; Trial 2: E=6.703,
p<.05), as well as the interaction contrast effect between neutral groups and positive
groups (Trial 1: 5:64.096, p<.001; Trial 2: E=39.879, p<.001). Hypothesis 6b (there will
be an interaction effect between social inﬂuence and task-technology ﬁt on perceived

ease of use) was also retained.

83

Interaction Effect of TTF and Social Inﬂuence on Perceived Usefulness

By the same token as the above methods, the omnibus F test for perceived
usefulness provided a signiﬁcant overall interaction effect of TTF and social inﬂuence
(Trial 1: 5:14.83, p<.001; Trial 2: 5:11.91, p<.001). The interaction contrast tests also
generated statistically signiﬁcant differences of perceived usefulness variation for both
negative versus neutral group case (Trial 1: E=17.539, p<.001; Trial 2: E=9.878, p<.005),
and neutral versus positive group case (Trial 1: E=26.801, p<.001; Trial 2: 5:23.377,
p<.001). The results, therefore, successfully supported Hypothesis 6c, which predicted an
interaction effect between social inﬂuence and task-technology ﬁt on perceived

usefulness.

Interaction Effect of TTF and Social Inﬂuence on Use Intention

MANOVA tests also analyzed the interaction effect of the treatments on new
technology use intention. However, similar outcomes as of the perceptive measures were
not found for the case of this variable. Omnibus F tests failed to endorse the presence of
the interaction effect at both Trial 1 (E=l .31, p>.05) and Trial 2 (E=1.53, p>.05).
Although it is needless to illustrate further interaction contrast tests, the results reafﬁrmed
the absence of treatment interactions when it comes to the subjects’ intention to use, as
change in the negative group was not signiﬁcantly different from that of the positive
group (Trial 1: E=0.01, p>.05; Trial 2: 1_7=0.176, p>.05), and the average scores of
negative group and neutral group showed no statistical difference in change Of intention
measures with the positive group for different TTF conditions (Trial 1: 1_7 =0.564, p>.05;

Trial 2: i=1 .538, p>.05). Thus, Hypothesis 6d (there will be an interaction effect between

84

social inﬂuence and task-technology ﬁt on use intent) failed to reject null hypothesis. The

following Table l 1 summarizes the MANOVA results of the interaction effect tests.

Table 11: Summary of Interaction Effect Tests from MANOVA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Trial 1 Trial 2
15' value Sig.(2-tailed) 1;" value Sig.(2-tailed)
E Omnibus F 15.55 p<.001 15.30 p<.001
1: 76‘
:3 g Cont. “—"vs. “a" 7.67 n<.01 8.67 p<.01
'3: v
1?. Cont. “a"vs. “+ " 31.04 p<.001 30.59 p_<.001
Omnibus F 32.62 p<.001 20.36 p<.001
D 3
g g Cont. “—"vs. "a" 11.29 p<.005 6.70 p<.05
Du v
Cont. “o"vs. “+ " 64.10 p_<.001 39.88 n<-001
Omnibus F 14.83 p<.001 1 1.91 p<.001
E3 g3 Cont. “—"vs. "a" 17.54 p<.001 9.88 n<.005
Cont. "e"vs. "+ " 26.80 p<.001 23.38 p<.001
Omnibus F 1.31 p>.05; n.s. 1.53 p>.05; n.s.
1" ‘6‘
:5, g Cont. “—"vs. 11+" 0.01 p>.05; n.s. 0.176 9205; n.s.
“—"& “0"vs. “+ " 0.564 12>.05; n.s. 1.538 p>.05; n.s.

 

85

 

Repeated Measures

Since the study incorporated two trials that were performed successively by the
participants, a repeated measure analysis of the outcomes is needed to verify whether or
not there exists a type of ‘comforting’ effect during continual testing. A repeated measure
MANOVA test was performed to examine the main effect of successive trial sessions on
each dimension of perception and use intention. Also, the test explored the interaction
effects between trials and TTF, as well as trials and SI. Results from repeated measure
MANOVA indicate that for perceived TTF there is a signiﬁcant comforting effect after
the ﬁrst trial (E=35.66, p<.001). It has been also revealed that there is a signiﬁcant
interaction effect of trials and TTF on perceived TTF (E=30.42, p<001). Looking at the
mean score changes, it is noticeable that the poor-ﬁt group subjects expressed higher TTF
perception scores at the second trial, compared to the ﬁrst trial (See Figure 24). However,
no signiﬁcant trials by SI effect was observed (5:023, p>.05).

A similar result was seen for the case of the interaction effect between trials and
TTF on perceived EOU (1_7=6.27, p<.05). Yet, there was no signiﬁcant main effect of
trials on perceive EOU (5:066, p>.05), and no signiﬁcant interaction effect either of
trials and SI (5:1 .06 p>.05), as depicted in Figure 25. When the repeated trials’ effect on
perceived usefulness was tested, a signiﬁcant main effect (E=35.05, p<.001) as well as a
signiﬁcant trials by TTF interaction effect (E=4.19, p<.05), and trials by S1 interaction
effect (5:550, p<.01) emerged. Subjects, it seems, tend to perceive the mobile music
download service more useful over a repeated trial sessions; and the increase was even

higher in the case of the poor-ﬁt and negatively inﬂuenced conditions (Figures 26 and 27).

86

As for the case of use intent, a signiﬁcant main effect of trials was detected
(E=6.46, p<.05), as well as an interaction effect between trials and SI (E=3.17, p<.05).
However, no signiﬁcance was found for the interaction effect between trials and TTF
(_E=0.69, p_>.05). Figure 28 shows the repeated measure effect of trials in accordance to
different social inﬂuences. These ﬁndings are summarized in Table 12 below.

Although most of the repeated trial sessions’ effects proved to be signiﬁcant,
with results indicating that there exists a “comforting” factor after the subjects initially
experience the new system, it is still arguable that such effects might have been driven by
the difference of the load of the tasks between the two trials. Since the ﬁrst set of tasks
required ﬁnding and downloading a total often songs, the amount of time and number of
Web pages it took to ﬁnish the ﬁrst trial were highly different from those of the second
trial, in which subjects were asked to ﬁnd and download no more than two music titles
(See difference between Figures 12 and 13, and Figures 14 and 15). Therefore, it is
questionable that an intervening factor, namely task load, could have affected the change
in perceptive measures among the participants who went through both trials continually.
Moreover, by eyeballing the histograms in Figures 24, 25, 26, 27 and 28, it is obvious
that the shifts are still minimal and no perceptive or intentional shifts exceed more than
one point in the overall seven-point measure scale. An in-depth discussion and

interpretation of the repeated measure results will be provided in Chapter 5.

87

Figure 24: Repeated Measure Effect of Trials and TTF on Perceived TTF

NWAMO‘N

 

 

 

 

 

 

 

 

 

 

 

T1

T2

 

+ Poor-Fit
1 + Good-F it

 

F igre 25: Repeated Measure Effect of Trials and TTF on Perceived EOU

NWAUIO\\I

 

0

 

 

 

 

 

 

 

 

 

T1

T2

f_—'—

+ Poor-Fit

 

 

+ Good-Fit

 

Figure 26: Repeated Measure Effect of Trials and TTF on Perceived Usefulness

Nw-RUIO\\I

 

 

 

-...-l

 

 

 

 

 

 

 

 

 

TI

88

T2

 

 

+ Poor-Fit
+ Good-Fit

 

 

 

Figure 27: Repeated Measure Effect of Trials and SI on Perceived Usefulness

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7 ,. ..L. _ .2. _.__, ....--“ __L ..L
6 *m——_ —— —_.__-.
5 e —w ~- —— k .
4 —-~**‘74:/‘ —— +Neg'
-l— Cont.
3
—O— Pos.
2 e —— n-_ ——————————— —— ——————-————————4
1 - _ _._ .4
0
T1 T2
Figure 28: Repeated Measure Effect of Trials and SI on Use Intent
7
6
5 ~— ——e ~— ‘ﬁ -
4 k _i’__#_'; i~____ 4 +Neg.
-l— Cont.
3 s I
r 1' -+- Pos.
2
I __
0
T1 T2
Table 12: MANOVA Results for Repeated Measures Effects of Trials
. Trials Trials x TTF Trials x SI
Variable
F Sig. F Sig. F Sig.
PTTF 35.66 12<-001 30.42 p<.001 0.23 p>.05 (n.s.)
PEOU 0.66 12>.05 (n.s.) 6.27 n<.05 1.06 p>.05 (n.s.)
PU 35.05 p<.001 4.19 p<.05 5.50 p<.01
UINT 6.46 n<.05 0.69 p>.05 (n.s.) 3.17 p<.05

 

 

 

 

 

 

 

 

89

 

Structural Equation Model Testing

To test the hypothesized process model of this study, which was illustrated in
Chapter 2, Figure 7, we applied SEM using EQS version 6.1 and investigated the effects
expected by the hypotheses. It is believed that the structural equation modeling (SEM)
approach has several advantages over traditional analyses even when it comes to
experimental studies. Bagozzi and Yi (1989) state that the SEM procedures do not
involve the restrictive assumption of homogeneity in variances of the dependent variables
across groups, and therefore can handle cases in which homogeneity assumptions are
violated. This approach also provides a way to correct for measurement error in the
measures of variables, reducing the likelihood of making type II errors. Another
advantage of SEM is that it allows for a more complete modeling of theoretical relations,
whereas traditional analyses are limited to associations among individual items.

Whit this in mind, a Multiple-group SEM (MSEM) analysis approach was
performed. Figure 29 represents the results of a three-group between-subjects model test
for positively, neutrally, and negatively inﬂuenced groups at Trial 1 of the experiment,
while Figure 30 shows the outcomes of the same model tested at Trial 2. Since each of
the path coefﬁcients designate a hypothesis, the signiﬁcance levels of the parameter
estimates indicate either support or rejection of the corresponding predictions. In general,
both models proved to be good ﬁts as, for the Trial 1 model, Bentler and Bonett’s (1980)
Normed Fit Index (NF 1) showed .913, Bentler and Bonett’s Nonnormed Fit Index (BFI)
showed .950, and the Comparative Fit Index (CPI) (Bentler, 1988) indicated .962.
However, the Goodness-of—F it Index (GFI) was a little below the suggested value

with .818, and the Adjusted Goodness-of-Fit Index (AGFI) (Joreskog and Sorbom, 1984)

90

was also moderate at .727. Yet, the Root Mean Square Of Approximation (RMSEA)
index (Steiger, 1989; Browne and Mels, 1990) was satisfactory at .058, overall
supporting a reasonably good ﬁt of the proposed model. As for the Trial 2 model, the NFI
denoted a ﬁt value of .903, the BF I a .946, the CFI a .958, the GFI a .823, the AGFI

a .734, and the RMSEA a .056, reporting a reasonably good ﬁt as well. The models’

goodness-of-ﬁt indices are summarized in Table 13. l’

 

The Effects of the Moderating Variable

MSEM analysis also enables the statistical examination of the moderating

 

variable through Lagrangian Multiplier (LM) and Likelihood Ratio (LR) tests (Bollen,
1989; Scott-Lennox and Lennox, 1995) to examine whether the parameter estimate
between each perception measure and success of trials differs across the three different
social inﬂuences. These techniques are advantageous, since they allow for controlling the
measurement errors, while estimating the path coefﬁcients with less bias (Matsuno and
Mentzer, 2000). Hypotheses 12, 13, and 14 projected a moderating effect of social
inﬂuence on the relationship between the success of trials and the perceptive variables.
Both LM and LR tests were performed to statistically verify whether there exist

signiﬁcant differences between these path coefﬁcients.

Lagrangian Multiplier (LM) Test
The LM test compares the ﬁt of restrictive to less restrictive models. When a
structural model fails to adequately ﬁt the sample data, it may suggest that certain ﬁxed

parameters or constraints need to be released to produce better ﬁt. The LM test evaluates

91

the statistical necessity of the restrictions, based on the calculations that can be obtained
on the restricted model alone (Bentler, 1995). Though the method was developed in a
single-group context, where model ﬁt improvement is suggested through relieving
speciﬁc restricted paths, it has been also adopted to verify the viability of the equally
constrained paths across groups in multi-sample analysis. Therefore, if the paths to be
tested are equally constrained in a MSEM, the LM test will try to ﬁt the constraints into F
the model and indicate if any of the constraints are invalid. If a constraint is invalid, a

signiﬁcant difference of chi-square values is shown on that path between groups.

built ”I'LL - P. {1‘ .lil' .‘-I .

EQS 6.1 was used to run the LM tests. First, we assumed equality of the

 

parameters to be estimated (i.e. a null case) by constraining the paths between success of
trials and perceived EOU measures. More Speciﬁcally, we set the Trial-Perceived EOU
estimate value of the positively inﬂuenced group to be equal to that of the control group.
We then set equality to the path in the case of the control versus negatively inﬂuenced
group, and also to that for the positive versus negative groups. LM test results showed
that, at Trial 1, two constraints, those including positive conditions, needed to be released
with chi-square values showing 16.825 (p<.001), 2.170 (p>.05), and 17.746 (p<.001) for
the positive-control, control-negative, and positive-negative respectively, which suggests
that there exist some signiﬁcant cross-group differences for the path coefﬁcients between
Trial and Perceived EOU. Since the standardized coefﬁcients of this relationship were .34
for the positive group, .09 for the control group, and -.004 for the negative group,
Hypothesis 13 was partially supported. However, an examination of LM tests at Trial 2
showed that all three groups were equal when it comes to the correlation between Trial

and Perceived EOU (chi-squares of 2.582, p>.05 for positive-control; 2.117, p>.05 for

92

control-negative; and 2.792 p>.05 for positive-control). Thus, Hypothesis 13 was only
partially supported.

By the same token, the constrained paths between Trial and Perceived TTF were
examined. At Trial 1, the chi-square tests yielded a value of 32.170 (p<.001), 8.516
(p_<.01), and 10.060 (p<.01) for positive-control, control-negative, and positive-negative
comparisons, respectively. However, with the control group’s path showing the highest F
estimate, Hypothesis 12 was rejected. Similar results came out at Trial 2, as the chi- 1

squares indicated 1.939 (p>.05) for positive-control, 0.533 (p>.05) for control-negative,

 

and 13.015 (p<.001) for positive-negative comparisons, with a stronger correlation for the L
case of the negatively inﬂuenced group.
The LM test also failed to reject the constrained paths between Trial and
Perceived Usefulness for the case of positive-control comparison both at Trial 1 (0.025,
p>.05 for positive-control; 0.176, p>.05 for control-negative; and 0.197, p>.05 for
positive-negative comparisons) and Trial 2 (0.369, p>.05 for positive-control; 0.106,
p>.05 for control-negative; and 0.68, p>.05 for positive-negative comparisons), resulting

in a null case for Hypothesis 14.

Likelihood Ratio (LR) Test

Also known as chi-square difference test, the LR test uses a statistic which has a
limiting chi-square distribution when the restrictive model is valid. The degree of
freedom (df) is referred to the difference in the df for the two models, and the usual chi-
square estimators for the restricted and unrestricted models are compared. Since the

MSEM produces one set of ﬁt indices with a chi-square value, when all the constraints

93

are held for the parameters to be estimated, the outcome would indicate a “null” case,
because this assumes the equality of those paths. As the constraints are released, the
indices will improve, indicating which paths are different, and enabling the examination
of the moderator’s effects.

For our study, the LR test was conducted for a pairwise comparison of the
restricted and free path coefﬁcients between Trial and perceptive measures across the
three social inﬂuences. First, a pairwise comparison was done by observing the chi-
square differences between two models, in which one model constrained all three paths to
be equal, while the other model was left with free parameters. The resulting chi-squares
were compared to observe whether the equality constraint model produced a better ﬁt
than the free model or vice-versa. For the Trial-Perceived EOU estimate, when
unconstrained, the model at Trial 1 produced a chi-square value of 353.802 with a df of
210, whereas when all three conditions were equally constrained, the chi-square
increased by 10.872 (364.674) with a difference of dfof 2 (212). Since this chi-square
value is statistically signiﬁcant (p<.01), it is highly unlikely that the restrictive model is
correct. Therefore we continued to compare the chi-square differences of the constrained
versus unconstrained models across two speciﬁc groups. The positive versus control
conditions were signiﬁcantly different (p<.01), as the chi-square difference between the
models with unconstrained and constrained paths showed 10.777 (364.579) at 1 df.
Positive and negative groups were not equal either as the chi-square difference was
12.347 (366.149) at 1 df(p<.001). However, the control versus negative condition
comparison suggested that the path was not signiﬁcantly different across these two

groups (p>.05) with a chi-square of 354.801 (difference of 0.999) at 21 1 df (difference of

94

l). Null hypothesis was retained for the case of control-negative group comparison. The
unconstrained and constrained Trial-Perceived EOU estimates comparison at Trial 2
delivered a chi-square value difference of 2.468 (350.501 — 348.033) at 2 df(212 — 210),
which had no statistical signiﬁcance (p>.05). Thus the Trial-Perceived EOU path in the
Trial 2 model failed to signiﬁcantly differ across the three social inﬂuence conditions.
Hypothesis 13 was therefore only partially supported for the case of the positive
inﬂuence.

By the same token, we tested the moderating effect of social inﬂuence on the
Trial-Perceived TTF path coefﬁcient. Once again, LR tests were conducted and for the
Trial 1 model, we found a signiﬁcant difference between the constrained and free paths
across all three groups (chi-square of 387.787 at 210 df, difference of 33.985 at 2 df,
p<.001). Each comparison for a pair of groups proved signiﬁcant difference as well
(positive = control groups’ chi-square 387.634 at 211 df, difference of 33.832 at 1 df,
p<.001; control = negative groups’ chi-square 361.355 at 211 df, difference of 7.553 at 1
df, p<.01; and positive = negative groups’ chi-square 361.032 at 211 d]; difference of
7.23 at 1 df, p<.01). However, these ﬁndings indicate that the control group’s coefﬁcient
is the strongest, while that of the positive group is the weakest, contradicting our
Hypothesis 12, which predicted a higher term of correlation for the positive group.
Similar results were found in the Trial 2 model, where the negative groups’ path showed
a signiﬁcantly stronger coefﬁcient that that of the positive group and control group. This
time, the chi-square difference tests generated a 363.163 (212 df) when the Trial-
Perceived TTF parameter was constrained to be equal across all three conditions, yielding

a chi-square difference of 15.13 at dfof 2 (p<.001) compared to the free model’s chi-

95

square value of 348.033 at 210 df While the positive and control groups’ path
coefficients did not signiﬁcantly differ from each other (350.144, chi-square difference of
2.111 at 1 df, p>.05), and those of control and negative were equal as well (349.409, chi-
square difference of 1.376 at 1 df, p_>.05), only the equality assumption between positive
and negative groups was statistically refuted (362.453, chi-square increase of 14.42 at 1
df, p<.001). Yet, as the negative group’s Trial-Perceived TTF estimate showed the
strongest correlation at Trial 2, Hypothesis 12 was once more not supported by our data.
Since Hypothesis 14 predicted a moderating effect of social inﬂuence on the
estimator of Trial-Perceived Usefulness relationship, the chi-square differences were also
tested though LR tests. Results showed that for both paths in Trial 1 and Trial 2 models,
equally constrained (null) cases generated better ﬁts, with a chi-square difference of
0.418 at 2 df(p>.05, with the constrained model’s chi-square of 354.220, 212 dj) when
we assumed equality across all three groups at Trial 1, and a chi-square difference of
0.013 at 2 df(p>.05, with the constrained model’s chi-square of 348.046, 212 df) when
the three groups’ paths were constrained in Trial 2 model. No ﬁrrther pairwise analyses
were needed in this case, since the generic null hypothesis was supported in both models.
Although it seems redundant to perform two different data analyses to test the
moderating effect of social inﬂuence, discovering that both LM and LR tests produced
identical results provided us with higher conﬁdence in presenting and discussing the
outcomes and implications. The LM and LR ﬁndings are summarized in Table 15 at the

end of this Chapter.

96

Chapter Summary

The data analysis results conﬁrm that Hypotheses 3, 4, 5, 6 (a, b, and c), 7, 10
and 11 are retained, while Hypotheses 8 and 13 are partially supported, and Hypotheses 1,
2, 6d, 9, 12 and 14 are rejected. According to the results, it seems that people’s successful
trial in using a new technology positively inﬂuences their perception of TTF (H7). This
perceived TTF subsequently affects their perception of the device’s ease-of-use (H10)
and usefulness (H11), which leads to the intention to use the technology. The positive
correlation between success of trials and perceived ease of use (H8) was only present
when subjects were positively inﬂuenced by the trainer (H12). However, when the
subjects were negatively or neutrally inﬂuenced prior to participate in the study, their
trials’ success level did not affect their perception of ease of use. Also, the measure of
success of trials showed no relation with the level of perceived usefulness (H9). However,
since perceived TTF was highly associated with perceived ease of use and perceived
usefulness, the relation between success of trials and perception of a technology’s
easiness and usefulness was mediated by perceived TTF. An interesting ﬁnding from the
model analysis is the absence of the link between perceived ease of use and perceived
usefulness (H1). While most of the previous TAM studies contend that perceived ease of
use positively inﬂuences perceived useﬁrlness, our data failed to reject the null
hypothesis for the relation between these two factors. It appears that the notion of
easiness or convenience does not always concede the feeling of usefulness to the users of
a new technology. The perception of usefulness had a positive effect on the willingness to
use a new technology (H3), whereas in most cases, the level of perceived ease of use

failed to directly inﬂuence the use intent. Interestingly, the relation between use intent

97

and perceived ease of use appeared to be partially moderated by social inﬂuence, since
the effect was only Signiﬁcant when subjects were positively introduced to a new
technology. The detailed discussion on the implications of the ﬁndings is presented in the

following Chapter 5.

98

 

 

 

 

 

 

 

 

           
      

      

   

      

       

 
 

 

  
  

 

 

  

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

99

 

 

 

 

 

 

 

 

.32 8ch 3 “56553
amigo. Lough =< o ~> N. ~> 0 ~ >
—oo.Vm..:.:.. 8. H2 m. HZ mm. HZ
we. ”0 5. H0 3. HO
Eva: 8. a 3. H. 8. a
3v?
”$56503 Sum 805383
wozoeom
13$. HZ p45 ma. “Z
...—r. "U L A615 5:. "U
eeeNv. "Am «Scab. uZ Add-v v—. "A—
tenab. “U
Steam. “A—
o p.95 =. “Z
G £686: Adév on. "U ME teach. "Z mECH.
83 3.5 _N. "m «o cornea?” :38. "U .«o 382%
see—b. "a—
«each. HZ
5...: 3.- Hz 6.33. "U
6.... .u . . . . .1"
. . A rem". m ......S ... Jake”... ”w N6. “2 3. Hz
3 . . . ca. HU 3. ”U
S. no :3 ... be. a we. a
on. ” 83 mo ommm
vo>mooeom
aw. HZ
: 3:2 36 ~> _>
:5 a. ..z a. o a. a
3. ”z 3. ”2 eh. m 8 a
mm. ”D mo. HO
3. i am. ”A— m> v>
\ e>
:> o_> o>
om mum“:

_ a5. a 8:80 Beebe—E 2V bozawoz as Q :13:on 32:68 Le 2332 2mm: ..

 

 

  

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.32 Sch a 2.85ch
@533 .585 =< Nm> ow> av>

 

 

 

 

 

 

 

               
        

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

qumi: 5. ”2 hm. NZ
a. NO ma. ”0
avg... a. H . a
nova...
”358580 5.3 305333
3&3qu
..z ass...
:2 .U . . . .
.. .. 2 5 2 .u
.252. .m .2.ch .Z .5". "a—
3395. "U
.3165. "n—
33: 5. NZ
Gomuﬂowcu 9...: 3.- "U m: 353. " Z $th
3.5 :. ".— mo somaoeom A .3. no Mo 3326
3. .z .3. a m
2.5 2.- 2 .mm... a 1
3.5 m—.. uU .. nu. E 3.5 8. "z . . . .
ma. H 3.5 8.- E .. .. 3.5 3. "U mm. ..z 3. ..z
. . .n . we .0 mm .0
can 353 .m aﬁnm 3Jm
S. ”a . H 3.3 m0 8mm
boﬁoﬁom
ow. HZ
. . ma. ”z 3. no mm> Vm>
vo> 8 .z 8. no Q. ”n.
. . «a. HD 3. ”m
8 .z 8. ”A.
ma. uU hm>
g. ”m
wm>
om>
35 mv> Nv>

 

 

 

 

 

 

 

m E: a 385 88.35 25 33.332 3... Q €582 .85 boats; .8.“ 33.3 2%: slag 2 E

Table I3: Goodness—of-ﬁt Summary for the Three-Group Structural Equation Models

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Indices
Fit Measure Trial 1 SEM Trial 2 SEM

NFI .913 .903

BFI .950 .946

CFI .962 .958

GF I .818 .823

AGFI .727 .734

RMSEA .058 .056

x2 (dt) 353.802 (210) 348.033 (210)

Table 14: Summary of Path Coefficients in Trial 1 and Trial 2 Models
Path Coefﬁcient (73 and Bs) Trial 1 Trial 2
31:12:: 5321;: Pos. Cont. Neg. Pos. Cont. Neg.

Success of Trials Perceived TTF .71‘" .88"* .76*“ .28“ .40“ 63*"
Success of Trials Perceived EOU .34" .09 -.004 -.15 .04 .02
Success of Trials Perceived Useful. .14 .07 .05 .20“ .12 .09
Perceived TTF Perceived EOU 89*" 86*” .76”* .73"* 95*" .89“*
Perceived TTF Perceived Useful. 59*" .78""' .79*" .70*" 94“" 86*"
Perceived EOU Perceived Useful. .2] .16 .11 .ll -.01 .01
Perceived EOU Use Intention .43*" .08 .02 -.02 -.15 -.08
Perceived Useful. Use Intention .42*** .71“ .68*" .72*" .92" .50"
*Q<.05 ; **p<.01 ; **Q<.001 ; all other parameters non signiﬁcant at p>.05

 

10]

 

Table 15: LM and LR Tests’ Results Summary

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hypo. Cont. Model LM Test Statistics LR Test - )8 Difference H Support
P-c—N T1 N/A 11’: 33.985, 2 df(9<.001) H12 rejected in
a _ _
g E T2 N/A 11’: 15.13, dﬁ 2 (9<.001) T1 model:
14...
2 E T1 11’: 32.170 (9<.001) x2: 33.832,df: 1 (9<.001) P(-71)<N(-76)
1.1.1 5 T2 112: 1.939(9205) x2: 2.lll,dﬁ 1 (9205)
on
1: I . .
"g g T1 112: 8.516 (9<.01) x22 7.553, dﬁ 1 (9<.01) H12 rejected m
o o C=N T2 model:
g (:35 T2 112: 0.533 (9205) 12: 1.376, df. 1(9>.05)
.. ._ P(.28) = C(.40)
g E F_N T1 112: 10.060 (p<.01) 11’: 7.23, df. 1(9<.01) géggrmgg
__ . < .
T2 11’: 13.015 (9<.001) 12: 14.42, df 1 (9<.001)
T1 N/A 112: 10.872, df: 2 (9<.01) H13 partially
8 :3 P=C=N supported in T1
71 8 T2 N/A 112: 2.468, df. 2 (9205) model.
C.—
§ E T1 112: 16.825 (9<.001) 12: 10.777, df: 1 (9<.01) N(.004) = C(99)
'as P=C
E g T2 112: 2.582, (9205) x2: 2.463, df. 1 (9205) < ”'34)
DD
.5 l 2, .
g g c= T1 11 . 2.170(p>.05) 212. 0.999, df. 1 (9205) ”Oremined for
o .
g 5); T2 12: 2.117,(9>.05) 112: 0.209. df. I(12>-05) “'13?
(«1 .76
.... c : . . : . 4 ,1 <. 01
:1: E_ P=N T1 11’ 17 746(9<001) x2 123 7 df(9 0 ) P(-.15)=C(.O4)
T2 112: 2.792 (9205) 11’: 0.051, 1 (M9205) = N902)
9, T1 N/A 112: 0.418.dﬁ 205205) HO retained for
5 _g P=C=N 2 H14 in T1
5 “3 T2 N/A X I 0.013, df. 2 (p>.05) model:
‘0— m
.3. g P-c T1 112: 0.025, (9205) 182 0.089, df. 1 (9205) 9914) = C(07)
0 - =N .05
33 .73 T2 112: 0.369, (12>.05) 112: 0.01 1, df. 1 (9205) ( )
DD 1...
: O
.5 2, .
g8; °l- (2: TI 11 . 0.176, (9205) )8. 0.102, df. 1 (9205) H0 retained for
523 g. T2 ,3; 0.106, (9205) )6: 0.001, dﬁ 102205) ”'1'ng
o .
.. =3
<- U) 2, 2.
E .‘é [EN T1 x-0-197,(12>-05) x . 0.415, df. 1 (9205) P020) = C(.12)
I" T2 112: 0.68, (9205) 112: 0.002, dﬁ 1 (9205) = NW9)

 

102

 

CHAPTER 5

DISCUSSION

The main goal of our research was to investigate the key factors of technology
acceptance within the context of m-commerce. Most of the ﬁndings from the analyzed
data agree with previous work that examined the adoption of new information systems,
with a few exceptions. The current chapter discusses the implications of the results by
addressing each particular hypothesis, explaining the situational circumstances of our

experiments, and providing insight into the very meaning of the outcomes.

Experimental Study

A major advantage of implementing experimental designs is that researchers
have control over the subjects’ values on factors that can inﬂuence the variable of interest
in the study (Agresti and Finlay, 1997). Thus, through experiments, effects can be
maximized and causalities are better observed. To exploit and compare the effects of the
degree of ﬁt between task and technology, and social inﬂuence through the WOM of a
trainer, we recruited over 200 college students, mostly undergraduate, and gave each of
them a PDA wirelessly connected to the Internet. Since we randomized the process of
subject assignment to one of the six preset conditions, it was not possible to control for
their gender, college grade, or prior experience, and therefore, these variables were not
considered as the main factors to test the hypotheses. One-third of the sample constituted

of male student, and a majority of the participants were newly introduced to a PDA at the

103

time of their participation (see Table 9). During the experiment, we assigned each subject
two sets of tasks to be completed at the location of the trial, and thus collected two data
sets (Trial 1 and Trial 2). The objective for employing a repeated data collection method
was mainly to obtain a robust explanation of the main factors driving technology
acceptance. Also, since we created two different Web sites, the appearance or
navigational characteristics of a particular site had to be controlled to prevent interference
with a subject’s impression or personal preference of a Speciﬁc mobile lntemet search

system.

Repeated Measures

The research design incorporating successive trials necessitates a test of repeated
measures. Although the repeated measure results indicate that there is a high probability
of learning, or “comforting” effect of the device between Trial 1 and Trial 2, we believe
that such an effect might have been caused by the difference of the tasks’ load between
the two trials. The ﬁrst set of tasks involved searching, ﬁnding, and downloading a total
of ten songs that required more than 17 minutes and 180 Web pages in average to
complete for the poor-ﬁt groups’ subjects, while the second set of tasks, which needed
just two music ﬁles to ﬁnd, took only about 3 minutes and 29 Web pages to browse.
Moreover, since the Web sites’ navigational interface was different for each trial (if the
ﬁrst trial was completed on a horizontal-depth navigation, the second trial switched to an
interactive interface), learning effects had a small chance to occur. Also, not all
perceptive measures increased after Trial 1. For instance, perceived EOU Slightly (but not

signiﬁcantly) decreased among the good-ﬁt group subjects (Trial 1’s M=6.6l vs. Trial 2’8

104

_=6.49), without any signiﬁcant main effect of trials. The outcomes were not consistent
either, as the interaction effects between trials and TTF were observed on perceived TTF,
perceived EOU and perceived usefulness, while the interaction effects between trials and
81 were only veriﬁed on perceived usefulness and use intent. Therefore, although the
repeated measure tests indicated the existence of a “comforting” effect, we contend that
such an effect could have been absent if the second trial carried the same amount of task

load as the ﬁrst trial.

Task-Technology Fit in Mobile Commerce

The mobile access technologies through which wireless users connect to a
network are the foundation for high quality m-commerce services (Maitland, 2004).
Since m-commerce rely on advances in middleware technologies, systems, and services
(Lee and Ke, 2001), an important issue regarding its adoption involves the
correspondence between the systems designed by the service providers and the contents
of the mobile services. A well-matched system to the characteristics of the needs is,

without any doubt, likely to evoke successful trial and positive reactions from the users.

Effect of T T F on Success of Trials

One of the main m-commerce technology acceptance drivers we tested in this
study is the ﬁt between task and technology. TTF is a function of task requirements,
individual abilities, and the functionality and interface of the technology (Goodhue,
1997). Therefore, if a speciﬁc technology ﬁts the task requirements, its users are expected

to Show higher trial success than those using a system that fails to match the speciﬁc job

105

characteristics. Although success is a multifaceted concept that implies multiple
dimensions, such as quality and perfection of the accomplishment, or quantity and size of
the outcome, our study considered two objective measures quantiﬁable in numbers: time
and Web pages it takes to achieve the tasks. AS evidenced by numerous previous TTF
studies (e.g. Keller, 1994; Goodhue, 1995; D'Ambra and Wilson, 2004; Staples and
Seddon, 2004), the ANOVA results support with high signiﬁcance the fact that TTF is
highly correlated with success of trials both in terms of time and number of Web pages
(H4a). Therefore, a well-matched TTF signiﬁcantly differ from a poor-matched TTF in
number of procedures through which a user undergoes when using an m-commerce
service. These results also indicate that success of trials measured in terms of time and
Web pages offers effective manipulation checks on the control of TTF in our
experimental study. Thus, success of trials represents an objective assessment of TTF that

was widely deemed as measurable only through users’ TTF evaluation.

Effect of T T F on Perception and Intention to Use

With Hypothesis 4a strongly supported by the data, it has become obvious that
success of trials reﬂects the extent of task-technology ﬁt adequacy. However, one of the
main questions of our study entailed the relation between TTF and technology acceptance.
The TAM literature presents external variables, such as system usability, to have an
impact on perceptions (Venkatesh and Davis, 1996). Yet, little has been empirically
shown as to represent the antecedents of perceived ease of use and perceived useﬁilness.
We denote that TTF can represent an objective construct for these exogenous variables

causing variance of perception in technology acceptance. Since the TAM theory suggests

106

that positive perception of EOU and usefulness leads to high intention to use, our
predictions stated that a well-ﬁt TTF would associate with positively perceived TTF
(H4b), perceived ease-of-use (H4c), perceived usefulness (H4d), and high willingness to
use the new technology (H4e). The results supported our proposed statements, thus
conﬁrming the strong correlations both between TTF and perception, and between TTF
and use intention. From these results, it can be inferred that TTF provides a solid ground
for controlling and measuring the external factors of system usability. The major
implication of such ﬁndings is that TTF not only increases the level of success of a new
system’s trial, but also positively enhances the users’ perceptions on TTF, EOU,
usefulness, and their intention to use. Since the latest small mobile devices increasingly
need to simultaneously fulﬁll two features, namely simplicity and functionality, the
importance of a user interface that meets the functional expectation by maintaining the

conciseness of the system can never be overemphasized.

Technology Acceptance Factors: Implications of the Process Model Test Results

As the main effects of TTF were supported, a question remains on whether or not
there are causal chains among the variables. Again, the TAM literature speciﬁes two key
factors of technology acceptance as perceived usefulness and perceived ease of use, and
since success of trials was an objective reﬂection of TTF, we examined whether the
effect of success of trials directly links to each individual dimension of perception. The
paths to be tested were: (1) the association between success of trials and perceived TTF,
(2) the association between success of trials and perceived ease of use, and (3) the

association between success of trials and perceived usefulness. Additionally, the

107

relationship between perceived ease of use and perceived usefulness was examined, and

the links to the use intent variable were analyzed.

Implications of the MSEM Results

According to the MSEM analyses, the direct and positive relationship between
success of trials and perceived TTF (H7) was supported. However, ceteris paribus, the
direct links from success of trials to perceived ease of use (H8) and perceived usefulness
(H9) were not observed. Instead, these relations seemed to be mediated by perceived TTF,
which in turn, showed signiﬁcant positive correlations with both perceived ease of use
(H10) and perceived usefulness (H11). The implications of our ﬁndings state that system
usability can be represented in terms of TTF level, which comprises trials’ success, and
consequently causes perception on the goodness of TTF. Therefore, based on the MSEM
results, it can be inferred that the links of a trial’s success on perceived ease of use and
usefulness are not straight, but rather indirect. People tend to perceive new mobile
devices and services easy to use and useful only when they consent to the fact that the
new technology highly supports the expected functions. Thus, from a practical point of
view, it is suggested that mobile system developers are constantly aware of the needs of

the potential users when implementing a new service in the market.

Effect of Perceived Ease of Use on Perceived Usefulness
Equally, if not more, interesting ﬁnding in our model is that the relation between
perceived ease of use and perceived usefulness (H1) was absent. Although it is not

common, similar results are reported by numerous TAM studies (Lee et al., 2003). For

108

example, Bjorn et al. were compelled to abandon the correlation between perceived ease
of use and perceived usefulness after pursuing an in-depth qualitative research on
technology acceptance of Groupware in virtual learning teams. Also, in a study of
implementation of an IT training program to inform mid-career employees, Neville and
Fitzgerald (2002) found that perceived ease of use is not necessarily a precursor of
perceived usefulness, but instead, high perception of usefulness can supercede the
importance of perceived ease of use. Our data analysis results concur with those ﬁndings,

implying the disparity of the feeling of easiness from the perception of usefulness.

Effect of Perception on Use Intent

Another interesting ﬁnding indicates that the behavioral intention is only
inﬂuenced by perceived usefulness (H3), and not by perceived ease of use (H2). Also
claimed in many extant studies is the argument that users’ readiness to accept a new
technology does not require the attribute of perceived convenience (Lee et al., 2003). An
exemplary study is reported by Szajna (1996) in which she performed a longitudinal
research on the procedure of the adoption of an electronic mail system. In her paper, she
contends that only perceived usefulness leads to use intention, while ease of use is
identiﬁed as a dependent measure of experience. Koufaris (2002) also found that
perceived ease of use of a Web store failed to correlate with the customers’ intention to
return to the same site. It is therefore arguable that the extent to which users ﬁnd a new

system to be easy does not fully guarantee their willingness to continue using it.

109

Social Inﬂuence in Technology Acceptance
Main Effects of Social Inﬂuence

Social inﬂuence is an independent variable in the variance model and a
moderating variable in the process model. In the two-way ANOVA and MAN OVA
results, social inﬂuence proved to be a critical independent factor that generated a
signiﬁcant treatment effect on users’ perception and use intention (H5). However, the
effect was partially present when it came to two perceptive variables: perceived TTF
(H5a) and perceived EOU (H5b). According to the ﬁndings, the effect of social inﬂuence
was not observed when the mean values of these two perceptive measures were compared
between the negatively inﬂuenced group and the control group. Only the positively
inﬂuenced subjects Showed Signiﬁcantly higher perceptive and intentional scores than the
negative and neutral groups. These ﬁndings suggests that positive social inﬂuence is
indispensable to positively affect perceived TTF and ease of use, regardless of the actual
degree of ﬁt between task and technology.

When it comes to perceived usefulness (H50) and use intent (H5d), all three
social inﬂuences were signiﬁcant determinants, implying that perceiving a device or a
system to be useful highly depends on the opinion of signiﬁcant others. This certainly can
be explained by the nature of the process by which an individual agrees with his(her)
peers on the usefulness of a new device. The opinion of signiﬁcant others about a new
device’s usefulness is very likely to be acquired through social exchange. By the same
token, the willingness to use a speciﬁc system or service is also highly dependent on
signiﬁcant others’ opinions, especially regarding a new and unknown technology, and

more speciﬁcally in the case of young college students, who tend to freely exchange and

110

share personal interests more so than the older generation. Therefore, our experimental
setting, which was established in a collegiate environment, turned out to be an
advantageous location, if not ideal, for testing the effect of social inﬂuence on technology

acceptance.

Interaction Effects between Social Inﬂuence and T ask- Technology Fit

Interaction effects between social inﬂuence and TTF were observed in all three
perceptive factors of technology acceptance. In other words, the difference of perception
shift between poor TTF and good TTF among the neutrally inﬂuenced individuals was
greater than those of the positively inﬂuenced and negatively inﬂuenced people. Since
the subjects in neutral inﬂuence condition formed a control group without any affective
messages addressed to them, a difference in the goodness of TTF was highly likely to
orient them to a more extreme direction, be it positive or negative, than those who were
previously exposed to the opinion of another social member. Thus, the survey results
showed clear and signiﬁcant interaction effects of social inﬂuence and TTF on perceived
TTF (H6a), perceived EOU (H6b), and perceived usefulness (H6c). Social inﬂuence
proved to be an effective method to form a less dispersed cognitive responses among the
users of a newly-introduced technology.

Quite interestingly, however, this interaction effect was absent in the case of use
intent measures (H6d). Put it differently, the difference of use intent levels between good-
ﬁt and poor-ﬁt of the technology to the task was practically the same in all three social
inﬂuence conditions. This result indicates that when combined, each of the objective

(TTF) and subjective (social inﬂuence) factors can add up to each other’s effect and

111

mount to the highest degree of willingness. Yet, no factor causes the increase or decrease
of intensity of another factor’s effect, and the changes appear parallel on a scatter plot
(see Figures 22 and 23). A key suggestion to the m-commerce practitioners according to
theses ﬁndings underscore the use of a social inﬂuence strategy when they seek to diffuse
the medium in the potential market. Because each individual hardly has complete
information on any new medium, media choice oﬁen occur through cognitions of other
people’s opinions rather than through personal learning of media attributes. Indeed,

“[m]edia perceptions are determined to some degree by objective

features. . . However, they are also determined to a substantial degree by

the attitudes, statements, and behaviors of coworkers.” (Fulk et al., 1990)

“Social interaction in the workplace shapes the creation of shared

meanings and these shared deﬁnitions provide an important basis for

shared patterns of media selection.” (Schmitz and Fulk, 1991)

Moderating Effect of Social Inﬂuence

From our MSEM analyses that tested the process model, social inﬂuence showed
little Signiﬁcance as a moderating factor. Its predicted effect on the relation between
success of trials and perceived TTF (H12) was not consistent at Trial 1 and Trial 2, and
practically absent in the relation between success of trials and perceived usefulness (H14).
The only prediction that was observed from the model testing was its moderating effect
on the relation between success of trials and perceived ease of use (H13) at Trial 1. The
positive effect of successful trials on perceived ease of use was signiﬁcant only for the

positively inﬂuenced group, whereas the control and negative groups showed no effect of

112

trials’ success whatsoever on ease of use. These results suggest that users’ perception on
a new technology’s traits is little dependent on social inﬂuence, but rather strongly
associated with its goodness of ﬁt between task and technology.

A surprising result, though, is found in the relation between usefulness
perception and use intent. At Trial], the negative and neutral groups generated a
standardized path coefficient of .70 and .69, respectively, while the positive group
showed a weaker effect of .40. Also at Trial 2, the parameter estimates of negative and
neutral groups were .71 and .86, whereas the positive group’s coefﬁcient was lower at .49.
It can be inferred, according to these numbers, that the moderating effect of social
inﬂuence is affecting the relation between perceived usefulness and use intention of a
user such that the relationship is stronger for negatively and neutrally inﬂuenced users,
and weaker for positively inﬂuenced users. The implication of such ﬁnding is that
negatively and neutrally inﬂuenced individuals would rely on their perception of the
system’s usefulness when deciding whether or not they want to use the new technology,
while positively inﬂuenced individuals tend to have an overall high intent to use the
system, and therefore their perception of usefulness would have a lesser impact on their
intention to use. This result once again emphasizes the importance of positive social
inﬂuence from peers or Signiﬁcant others in forming the willingness to use a certain new

technology.

Chapter Summary

The current chapter inspected the data analysis results and explained the

implications of the ﬁndings. The ANOVA and MANOVA results supported the main

113

effects of TTF and social inﬂuence on technology acceptance, and conﬁrmed the
interaction effect of these two independent variables on users’ perception. However, no
interaction effect was found on the users’ willingness to further use our tested mobile
system. Instead, TTF and social inﬂuence proved to be both independently important in
leveraging use intention. The process model’s MSEM analysis results showed signiﬁcant
support for an indirect effect of success of trials on perceived ease of use and usefulness,
while only perceived usefulness was statistically proven to impact use intent. Meanwhile,
only a positive social inﬂuence was found to partially moderate the relation between
success of trials and perceived ease of use. Overall, regardless of the goodness of ﬁt
between task and technology, we identiﬁed that positive social inﬂuence is a crucial
factor in increasing potential users’ willingness to adopt a particular technology.

In the next chapter we provide a summary of this dissertation, present the
managerial implications of ﬁndings for the m-commerce practitioners, exploit the
limitations of the current study, discuss the issues concerning future studies, and conclude

the dissertation with closing remarks.

114

CHAPTER 6

SUMMARY AND CONCLUSIONS

The current study entails theoretical improvement of the literature of the TAM,
TTF, and SIM within the discipline of telecommunication, as well as practical
contribution in the domain of business-to-consumer m-commerce. The empirical results
obtained from the experimental data provide a solid foundation for an insightful
understanding of the factors driving technology acceptance, and an oriented suggestion
for practitioners and marketers of mobile lntemet services. We believe that, however, this
dissertation has its limitations which should not be disregarded but rather addressed with
humble acknowledgement. This chapter concludes our dissertation by providing a
summary of the study, presenting the limitations and directions for future studies, and

discussing the managerial implications of the current study.

Summary of the Study

During the last few years, high-speed 3G mobile technologies have been
proliferating rapidly, signaling prospects for the mobile lntemet (Racanelli, 2005).
Although m-commerce is still in its infancy, some foreign markets, including Japan and
South Korea, are already experiencing soaring increase of mobile lntemet subscribers
(Minges, 2005), with high demands of multimedia services over the mobile devices (Kim,
2005). However, such great expectation is only an anticipation of the marketers in mobile

telecommunication industry, and while the projected numbers are not fully guaranteed to

115

be attained, there still exist controversies regarding the future of mobile lntemet and m-
commerce. After an extensive review of the m-commerce literature, we found that little
has been studied on the factors that drive the users’ acceptance of this new technology.
This dissertation speciﬁcally examines the main factors that drive technology acceptance
by testing the effects of TTF and social inﬂuence on the TAM model’s variables in the
context of m-commerce.

We started by construing today’s situation of mobile commerce and identifying
crucial acceptance factors to raise theoretical questions in the discipline of
telecommunication. Thus we generated the following research questions:

What are the main factors inﬂuencing the individual acceptance of mobile
Internet services in mobile commerce dealing with entertainment applications such as
music ﬁle downloads?

Three theoretical paradigms that were selected to formulate our research model
are as follows: The Task-Technology Fit (TTF) Theory, Social Inﬂuence Model (SIM),
and Technology Acceptance Model (TAM). By adopting variables from the TTF and
SIM to the TAM, our hypotheses predicted the following relations:

H1: Perceived ease of use will be positively related to perceived usefulness.

H2: Perceived ease of use will have a positive eﬂect on use intention.

H3 .' Perceived usefulness will have a positive effect on use intention.

H4: There will be a main eﬂect of task-technology ﬁt on the success of triasl, users’
perceptions and use intention of the new technology.

H5: There will be a main effect of social inﬂuence on users 'perceptions and use

intention of the new technology.

116

H6: There will be an interaction effect between social inﬂuence and task-technology ﬁt
on users ’ perceptions and intention.

H7: The degree of a trial ’s success in achieving a task with a new technology will be
positively associated with the users ’ perception of TTF.

H8: Success of trials will be positively related to perceived ease-of-use.

H9: Success of trials will be positively related to perceived usefulness.

H10: Users ’ perception of TTF will be positively associated with perceived ease-of-use.

H l I : Users ’ perception of TTF will be positively associated with perceived usefulness.

H12: The eﬂect of the trials’ success on perceived TTF will be moderated by social
inﬂuence, such that the relationship will be stronger for positively inﬂuenced users,
moderate for neutrally inﬂuenced users, and weaker for negatively inﬂuenced users.

H13: The eﬂect of the trials’ success on perceived ease-of-use will be moderated by
social inﬂuence, such that the relationship will be stronger for positively inﬂuenced
users, moderate for neutrally inﬂuenced users, and weaker for negatively
inﬂuenced users.

H14: The eﬂect of the trials’ success on perceived usefulness will be moderated by social
inﬂuence, such that the relationship will be stronger for positively inﬂuenced users,
moderate for neutrally inﬂuenced users, and weaker for negatively inﬂuenced users.

An experimental setting was conceived to artiﬁcially create a 3x2

(positive/neutral/negative social inﬂuence by good/poor task-technology ﬁt) between-

subjects factorial design, in which subjects were randomly assigned to one of six

conditions. After the participants were introduced and instructed on the method of using a

PDA, data were collected over two experimental sessions (Trial 1 and Trial 2), each of

117

which requested a set of tasks to complete (to ﬁnd and download the required MP3 music
titles) using a speciﬁc type of mobile lntemet navigation interface system. Two different
Web sites were created for this purpose. The time and number of Web pages it took for a
user to complete a task were monitored to measure the level of success of the trials. Also,
two separate questionnaires were handed out to measure the subjects’ perceptive
responses. When data collection was ﬁnished, a total of 21 0 college students participated
in the study.

Data analyses were performed by employing two-way ANOVA and MAN OVA,
and Multiple group Structural Equation Modeling (MSEM) tests. The results showed full
support for H3, H4, H5, H7, H10 and H11, partial support for H6, H8 and H13, and no
signiﬁcant support for H1, H2, H9, H12 and H14. Based on these ﬁndings the followings
could be inferred: First, positive social inﬂuence was crucial in forming positive
perception on a new technology and high intention to use it, and when combined with
appropriate goodness of ﬁt between task and technology, the effect on use intent would
maximize. Second, although task-technology ﬁt impacts the users’ success of trials in
ﬁnding music titles, it does not directly affect their perceived ease of use and perceived
usefulness, but rather has an indirect effect through perceived TTF. Third, perceived ease
of use has almost no effect on perceived usefulness and use intent, but instead, perceived
usefulness was found to Signiﬁcantly associate with users’ willingness to use a new
technology. These ﬁndings suggest that as much as TTF plays an important role in
shaping and changing users’ perception on new mobile technologies, social inﬂuence is
equally, if not more, inﬂuential when it comes to the decisions of these new users toward

an intention to actually use them.

118

Managerial Implications

This dissertation ﬁnds strong support for the extended model of TAM with TTF
and SI factors for the use of m-commerce. In other words, we found that the external
variable that was presented as a rather abstract construct in TAM can be efﬁciently
measured by TTF and SI. This yields to several suggestions for the m-commerce
practitioners and marketers. First, the importance of a navigational interface system
appropriate to the purpose of a service should be emphasized. Since mobile devices have
limited screen capacity to display all functions and contents of a Web store, businesses
should actively seek ways to improve the ﬁt of their systems by constantly revising and
researching the functionality of navigational interfaces adequate to the needs of the users.
Second, in doing so, the developers Should mainly consider improving the level of
usefulness of their services, and by a lesser degree, enhance the easiness of browsing.
Third, marketers must recognize that consumers of new telecommunication technologies,
especially those from the young generation, are prone to continually exchange
information with each other and listen to the opinions of early adopters. Therefore, to
approach this potentially huge market and gain success, the opinion leaders Should be the

ﬁrst to be identiﬁed and persuaded.

Limitations and Directions for Future Studies

As it is the case with all studies, this dissertation should be interpreted with
regards to the limitations inherent in the study. The ﬁrst limitation relates to the
methodological practice of using college students as subjects in behavioral experiments,

which has been widely criticized in the ﬁeld of social sciences. For more than half a

119

century, numerous scholars expressed concerns over the validity of results generated in
this manner, including McNemar (1946), who argued that researchers tend to over-rely
on ﬁndings derived from the study of the “behavior of sophomores,” Enis et al. (1972),
who reported that students were studied in more than one-half of the consumer behavior
articles, Cunningham et al. (1974), who suggested that students are not equal to the
household consumers, Morgan (Morgan, 1979), who questioned the value of students as
role-players because they could not anticipate the reactions of the typical consumers, and
more recently, Winer (1999), who expressed skepticism about experimental studies’
external validity. Although we agree that a selection of student samples is problematic to
represent a speciﬁc set of “rea ” people, we argue that the purpose of experimental
designs is to isolate and emphasize the treatment effects to observe the reactions of the
individual participants, and therefore, students should be no less real than regular people.
As claimed by Lynch (1999), just as in the laboratory, the real world varies in
background facets of subject characteristics, setting, context, relevant “history,” and time.
Moreover, since young generation markets constitute the majority and target consumers
of wireless and mobile lntemet (Okazaki, 2004), the ﬁndings from our experimental
sample of young undergraduate college students has the external validity of
generalizability across (Cook and Campbell, 1979) this speciﬁc population. However,
concerns might arise for the small number of subjects (N=210) to perform MSEM
analyses, since our study implemented a multiple-group approach, allowing only about
70 subjects in each social inﬂuence group. Some insigniﬁcance of path coefﬁcients might
have been caused by the small sample size, and therefore, we suggest a larger sample size

for future studies employing a MSEM analyses with experimental designs.

120

Another method-related issue in our experimental study is the lack of actual
manipulation checks of the social inﬂuence. Although the measures for the trainer’s
communication styles allowed the subjects to evaluate the credibility of the trainer, we
omitted to verify whether the positively inﬂuenced group signiﬁcantly differed from the
negatively inﬂuenced group in terms of perceived favorability of the WOM messages.
Also, some students with an already strongly established opinion toward m-commerce
and mobile devices might have felt uneasiness listening to the partiality of the
expressions cued by the trainer. Such manipulation check measures should not be
overlooked in other studies that test the effect of WOM social inﬂuences

The second limitation is the use of a trainer as confederate to practice social
inﬂuence on the subjects through word-of-mouth messages. Most social inﬂuence studies,
including those associated with TAM, measure social inﬂuence in terms of social or
subjective norms. Subjective norms can be described as the degree to which an individual
believes that people who are important to him think he should perform the behavior in
question (Ajzen and Fishbein, 1975). Such socially-established norms are found to have
both short- and long-term effects on people’s acceptance of a new technology (Taylor
and Todd, 1995; Venkatesh and Davis, 2000; Venkatesh and Morris, 2000). Although our
method introduced a seemingly expert trainer to the subjects, it is not exactly known
whether they value his messages with the same importance as those of their peers or
signiﬁcant others. Therefore, the effect of such word-of-mouth messages expressed by
an unknown person should be considered to have a short-term impact rather than a long-

term effect on their eventual perception and decision on using the new system.

121

The third limitation of this study involves the lack of actual usage measurement.
Use intent is a mere indication of readiness or willingness to use, which does not
necessarily trigger real-life usage. In fact more than 80% of TAM-related studies omitted
to verify the link between behavioral intention and actual behavior (Lee et al., 2003).
Since our study tested a system that is currently not offered by m—commerce service
providers, the unavailability of measures of usage was a major hindrance we could not
overcome in our experimental design. Perhaps a follow-up survey on each participant’s
m-commerce usage aﬁer the launch of similar services in the mobile market would serve
as an examination of the relation between use intent and actual use.

Fourth, the context of m-commerce can not be applied to all new technological
devices. Especially, social inﬂuence, which is one of the key independent variables, may
as well be a less important factor when it comes to a totally different technology. Thus,
the ﬁndings of this study should be addressed with careful consideration of the products’
characteristics if one wished to apply our extended model to explain a new system
irrelevant to the mobile lntemet.

In light of the above limitations, we recommend that future studies of m-
commerce acceptance consider conducting a more robust research with “real” people, by
assessing highly interpretable and controllable social inﬂuence, checked for the
manipulation effects, with a complete data on actual usage of the new technology, and a
larger sample size if MSEM analyses are to be performed. Also, a suggestion for another
interesting research idea relating to social inﬂuence beyond the WOM messages is a
study of the effect of Web vendors’ list of recommendation systems. Since most of

today’s online stores, such as iTunes and Amazon.com, sort popular items by product

122

categories through data mining techniques and show a list of recommended titles to the
online customers by indicating such statements as: “Customers who bought this product
also bought: ...”, these mediated social inﬂuences could signiﬁcantly affect the
consumers’ perception of usefulness, intention to purchase, and actual buying behavior of
the listed products. If the results were to generate signiﬁcant social effects of these
messages, interesting ﬁndings of the impact of a non-face-to-face and asynchronous
computer-mediated interaction might as well suggest for exciting theoretical and practical

implications in the discipline of information systems studies.

Conclusion

This dissertation attempts to identify and examine the main factors driving the
acceptance of new technology of m-commerce. For this purpose, the TTF and SIM
variables have been integrated into the TAM theory and successﬁrlly tested through
experimental processes. While a number of factors have been considered and analyzed to
prove our predicted causal relations, the fact that the current study did not contemplate all
possible factors should not be neglected. We deem, however, that the selection of
variables was theoretically adequate, and the results valid in many aspects.

M-commerce represents the convergence of the lntemet and the mobile
technology, which, through mobile devices, has added a dimension to e-commerce. M-
commerce is forecasted to continue to grow and show a signiﬁcant impact on the global
business environment. This study provides theoretical and practical implications for both
scholars and practitioners of m-commerce. As the technological advances in mobile

communication and data transmission is likely to stimulate new types of services, further

123

investigation on the drivers of diffusion and adoption of the new technology would be

needed to improve the model of technology acceptance.

124

APPENDIX 1: QUESTIONNAIRES

Questionnaire 1 (completed by the participants after the first set of tasks)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly Strongly
Disagree Neutral Agree
1 The tasks I performed were difﬁcult. 1 2 4 6 7
2 The interface has the exact features for me 1 2 4 6 7
to carry out the tasks.
3 The interface ﬁts well with the tasks I was 1 2 4 6 7
required to perform.
4 The interface fully supports the purpose of 1 2 4 6 7
the tasks.
5 It was difficult to complete the tasks 1 2 4 6 7
because some features were not available.
6 In general, the interface had functionalities 1 2 4 6 7
that corresponded to the tasks.
7 My interaction with the interface was clear 1 2 4 6 7
and understandable.
8 1 quickly became skillful at using this 1 2 4 6 7
interface.
9 Learning how to fully use the interface was
1 2 4 6 7
easy for me.
10 It was easy to get the interface search and 1 2 4 6 7
ﬁnd what I was looking for.
11 It took too much time to ﬁnd the ﬁle I was 1 2 4 6 7
seeking through the interface.
12 The interface was so complicated, it was 1 2 4 6 7
difﬁcult to understand what to do.
13 In general, the interface was easy to use. 1 2 4 6 7
14 The interface would be useful for any type
1 2 4 6 7
of ﬁle search.
15 This Web site would improve my ability to 1 2 4 6 7
locate other music ﬁles.
16 I think the interface is useless for locating
. 1 2 4 6 7
musuc ﬁles.
17 This was an appropriate use of a Web site 1 2 4 6 7
interface on the PDA.
18 Generally, I ﬁnd the interface useful. 1 2 4 6 7

 

 

125

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly Strongly
Disagree Neutral Agree
19 Using the interface was enjoyable. 1 2 4 6 7
20 The actual process of searching with the 1 2 4 6 7
interface was pleasant.
21 I had fun using the interface on a PDA. 1 2 4 6 7
22 I ﬁnd the interface exciting. 1 2 4 6 7
23 The interface was so boring, I felt annoyed 1 2 4 6 7
while using it.
24 This type of service will become popular in 1 2 4 6 7
meMMm.
Very Very
Low Neutral High
25 If you were in an airport with a mobile
device, what would be the probability that
. . 1 2 4 6 7
you consnder buying songs through such
services?
26 If you were at home with the device, what
would be the probability that you consider 1 2 4 6 7
buying songs through such services?
27 If you were relaxing on the beach with the
device, what would be the probability that 1 2 4 6 7
you consider buying songs through such
services?
28 If you had to decide whether or not to buy
the songs using this technology at this 1 2 4 6 7
moment, how likely is it for you to make the
purchase?
29 In the future, if you had access to similar
services on mobile devices, how likely is it 1 2 4 6 7

for you to buy music?

 

 

126

 

 

 

 

 

 

 

 

 

 

Don't know Know
at all very well
30 How familiar are you with the artist of the
first task whose song you had to ﬁnd? 1 2 6 7
(Counting Crows)
31 How familiar are you with the song of the
first task? 1 2 6 7
( “This Desert Life”)
32 How familiar are you with the artist of the
second task whose song you had to ﬁnd? 1 2 6 7
(Urban Knights)
33 How familiar are you with the song of the
second task? 1 2 6 7
(“Urban Knights V”)
Survey 1 Completed.

Please return this form to the experiment administrator.

127

 

 

Questionnaire 2 (completed after the second set of tasks)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly Strongly
Disagree Neutral Agree

34 The tasks I performed were difﬁcult. 1 2 4 6 7

35 The interface has the exact features for me 1 2 4 6 7
to carry out the tasks.

36 The interface ﬁts well with the tasks I was 1 2 4 6 7
required to perform.

37 The interface fully supports the purpose of 1 2 4 6 7
the tasks.

38 It was difﬁcult to complete the tasks 1 2 4 6 7
because some features were not available.

39 In general, the interface had functionalities 1 2 4 6 7
that corresponded to the tasks.

40 My interaction with the interface was clear 1 2 4 6 7
and understandable.

41 I quickly became skillful at using this
interface. 1 2 4 6 7

42 Learning how to fully use the interface was 1 2 4 6 7
easy for me.

43 It was easy to get the interface search and 1 2 4 6 7
ﬁnd what I was looking for.

44 It took too much time to ﬁnd the ﬁle I was 1 2 4 6 7
seeking through the interface.

45 The interface was so complicated, it was 1 2 4 6 7
difﬁcult to understand what to do.

46 In general, the interface was easy to use. 1 2 4 6 7

47 The interface would be useful for any type 1 2 4 6 7
of ﬁle search.

48 This Web site would improve my ability to 1 2 4 6 7
locate other music ﬁles.

49 I think the interface is useless for locating 1 2 4 6 7
music ﬁles.

50 This was an appropriate use of a Web site 1 2 4 6 7
interface on the PDA.

51 Generally, I ﬁnd the interface useful. 1 2 4 6 7

 

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Strongly Strongly
Disagree Neutral Agree
52 Usrng the interface was enjoyable. 1 2 4 6 7
53 The actual process of searching with the
. 1 2 4 6 7
Interface was pleasant.
54 I had fun usrng the interface on a PDA. 1 2 4 6 7
55 I ﬁnd the Interface exerting. 1 2 4 6 7
56 The interface was so boring, I felt annoyed
. . . 1 2 4 6 7
while usrng It.
57 This type of service will become popular in 1 2 4 6 7
the future.
Very Very
Low Neutral High
58 If you were in an airport with a mobile
device, what would be the probability that 1 2 4 6 7
you consider buying songs through such
services?
59 If you were at home with the device, what
would be the probability that you consider 1 2 4 6 7
buying songs through such services?
60 If you were relaxing on the beach with the
device, what would be the probability that 1 2 4 6 7
you consider buying songs through such
services?
61 If you had to decide whether or not to buy
the songs using this technology at this 1 2 4 6 7
moment, how likely is it for you to make the
purchase?
62 In the future, if you had access to similar
services on mobile devices, how likely is it 1 2 4 6 7

for you to buy music?

 

 

129

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Don't know Know
at all very well
63 How familiar are you with the artist of the
first task whose song you had to ﬁnd? 1 2 4 5 7
(Gerald Levert)
64 How familiar are you with the song of the
first task? 1 2 4 6 7
(“Awesome”)
65 How familiar are you with the artist of the
second task whose song you had to ﬁnd? 1 2 4 5 7
(Nickelback)
66 How familiar are you with the song of the
second task? 1 2 4 6 7
(“Someday”)
Strongly Strongly
Disagree Neutral Agree
67 The instruction I received during the training 1 2 4 6 7
session was sufﬁcient for the tasks.
68 The trainer was frrendly. 1 2 4 6 7
69 The trainer was persuasive.
1 2 4 6 7
70 The trainer was trustworthy. 1 2 4 6 7
71 The trainer was knowledgeable. 1 2 4 6 7
1 . . . .
72 n general, the trainer did a good 10b 1 2 4 6 7
73. Gender: (1) Male (2) Female
74. Age:
75. Major:
76. What is your year in college?
(1) Freshman (2) Sophomore (3) Junior (4) Senior

(5) Other (please specify)

 

130

 

 

77. Please indicate the amount of time you spend on an averagg gr day using the lntemet:
hours, minutes a day.

 

78. Are you using, or did you ever use, any peer-to-peer (P2P) ﬁle sharing application?
(for example, Napster, Kazaa, Morpheus, iMesh, Soribada etc...)

(1)Yes (2)No
If yes, how long have you used a P2P program? years, months

79. Do you have, or did you ever have, a cell phone? (1)Yes (2)No
If yes, for how long have you used a cell phone? years, months

80. Do you have, or ever had, a PDA? (1)Yes (2)No
If yes, for how long have you used a PDA? years, months

81. Do you have, or ever had, a portable MP3 player? (1)Yes (2)No

If yes, for how long have you used an MP3 player? years, months

 

 

Strongly

Neutral Agree

 

82 How do you rate your skill in manipulating
electronic devices?

 

83 How experienced are you in using the
lntemet?

 

84 How do you rate your computing ability?

 

85 How do you rate your interests in new
technologies?

 

86 How skilled are you in using a PDA?

 

 

 

 

End of Survey 2.
Please return this form to the experiment administrator.

Thank you for participating!

131

APPENDIX 2: TRAINER’S SCRIPT
Hello, and welcome to the PDA study.

I’m , I will show you how to use the PDA and which Web sites you will browse
during the experiment.

First, I’d like you to read and sign this consent form. (Give the subject a consent form to
sign) It basically tells you that the experiment is completely safe, that your participation
will be anonymous and your answers will not affect your grade in class. So everything
that comes out from this study will be kept conﬁdential.

This experiment is about using a PDA to download MP3 ﬁles from certain commercial
Web sites. Don’t worry! We are not going to ask you to do anything weird.

To turn on the PDA, you simply need to operate this power button. (Show the subject how
to turn on)

The stylus is located on the side of the device; just pull it down and take it out. (Show the
subject how to pick up the stylus)

To choose a program, just tap on the icon on the screen, like this. (Show how to tap by
using a stylus)

If you want to go back to the menu page, just tap on the “Home” icon on the lower left
corner.

For this experiment, you will need to use a browser, NetFront 3.0. Tap on it, and here’s a
Web page that has been already downloaded on this PDA.
To go to another site, just tap on the upper left comer and choose “Enter Address.”

Tap on the “Past Input” option, and choose among the Web sites that have been pre-
indicated.

F or the purpose of this study, we want you to remain on these two Web sites that we
provide. (Show the subjects both “Music World. Com ” and “MP3 TopiaCom ”)

To download a song from MP3Kingdom.Com:

(1) choose the genre of the song you wish to search,
(2) tap on first letter of the artist’s name,

(3) choose the artist from the list,

(4) choose a speciﬁc album,

(5) and ﬁnally, select a song in that album.

So, for example, to download the song “Hotel Paper” in the “Hotel Paper” album by

“Michelle Branch,” go under “Rock,” choose “M,” choose “Michelle Branch”, tap on
“Hotel Paper” album, and select “Hotel Paper.” Here’s the last page that shows the

132

song’s information: running time, album, artist, genre etc... tap on the “Download”
button, and just wait until the ﬁle is stored in the PDA.

The other site provides a different interface.
To download a song from MP3Topia.Com:

(I) begin typing the ﬁrst letter of the artist’s name,
(2) select a name in the list below,

(3) then begin typing the song title,

(4) and again, select a song from the list.

So, for instance, to download “Island In The Sun” by “Weeper,” start typing, like this
(show how to type by using the virtual keyboard, Graﬂiti, or the keypad), and whenever
you want to select, just tap on the name.

Then, type again the song title, “Island. ...” and see? Select the song, and the last page
looks similar to that from MP3Topia.Com. Tap on “Download” to get the song.

You can listen to these songs by using “Audio Player” program.
Go to the menu, choose “Audio Player” icon, and tap on the play button.
(Show them how to play the songs)

Positive Influence

Isn’t this cool? Of course, you might be able to download songs for free by using Kazaa
or Morpheus, but that is now considered illegal, right? And you don’t actually need to
synchronize with your computer to put your music on your PDA. I wonder if they are
planning to commercialize this type of MP3 service, because if they do, and if I have the
device, I’ll deﬁnitely use it. I heard there exist similar services abroad, like in Japan,
Australia, and Europe. They can actually download songs on cell phones and use them as
MP3 players! You may know that using P2P ﬁle sharing programs can actually
penalize the individual users since it’s now legally punishable by law.

Negative Influence

I’m doing the training sessions because I was asked to, but personally, I don’t think this
will work commercially. You can buy MP3 players for a fraction of the price of a PDA,
and you can download songs for free through any P2P ﬁle sharing program, such as
Kazaa or Morpheus. I heard that there exist such services in overseas, like in Japan and
Europe, but it seems that they are not quite popular. Downloading is fast here because
we’re using a wireless LAN, but if we were using a cellular technology, the process will
take so much more time. I guess this will end up being useless.

I33

APPENDIX 3: HANDOUTS son TASK Descmrrlon

First Set of Tasks
1” Subtask: Find/Download 5 songs by Counting Crows, in the album “This Desert
Life” (music genre: Rock)
2"d Subtask: Find/Download 5 songs by Urban Knights, in the album “Urban

Knights V” (music genre: Jazz)

> Please see the experiment administrator right after you have completed each subtask

Second Set of Tasks

1St Subtask: Find/Download the song “Awesome ” by Gerald Levert (music genre:
R&B)

2"d Subtask: Find/Download the song “Someday” by Nickelback (music genre: Rock)

> Please see the experiment administrator right after you have completed each subtask

134

ﬂ

M

' O\

\I

00

\0

APPENDIX 4: EXPERIMENT PROTOCOL AND DEBRIEFING STATEMENT

. Welcome the participant, tell them to read and sign the consent form

. Introduction to the tasks
The purpose of this study is to record your evaluation of using commercial Web
sites on PDAs. You will be asked to perform a few tasks that involve searching,
finding and downloading MP3 files on the PDA. There will be a total of four
tasks, and you will eventually complete two questionnaires. Each of them will
take approximately 5-10 minutes. Once you are done with a task, you will be
able to freely use the PDA.

The tasks are written on folded papers. Please, unfold the paper only when you
are told to do so.

. Instructions on using PDA
This is how you use a PDA. (Provide instruction about activating, tapping and
writing)

The PDA is connected online wirelessly. You will be able to browse through any

Web site with the device. However, I’d like you to remain on the Web sites that
are provided for the tasks.

This is how you use a Web browser. (Provide instruction about Web browser use)
Let me show you how you can ﬁnd a song on the Web site. To download a song
from a Web site you simply need to tap on the “download” button...
(This is when you express your positive/negative opinion)

. Ask: Do you have any questions?

. Give the subject the task descriptions. Tell them not to read until they are told to.

. Begin lst task
Unfold the ﬁrst paper and perform the task as it is written on it. Tell me when

it’s over.

. Begin 2nd task
Now perform the second task. Tell when you are done.

. Hand out ﬁrst survey

. Begin 3rd task
Unfold the paper and perform the third task. Notify me when you are through.

135

10. Begin 4th task
- Begin the fourth task. Let me know when you have finished.

1 1. Hand out second survey

12. When surveys are completed, debrief subjects

- The true purpose of the study is to examine how social inﬂuence and interface
design affects your perception and intention in using mobile devices for
purchasing digital goods. The system is not real; it was made up by the
researcher for the purpose of the study.

- Please do not inform any of your classmates of the purpose of the study because
other students will also be participating throughout the week.

As mentioned in the consent form, if you want to know about the result or have any

questions about the study you can contact the researcher. Thank you again for
participating in the study

136

APPENDIX 5: CONSENT FORM

Experimental Mobile Commerce Study

To be conducted by
Zoo-Hyun Chae
Department of Telecommunication and Department of Communication
Michigan State University, East Lansing, Michigan

Please ask the study investigator tO explain any words or information that you do not clearly
understand.

INTRODUCTION

As a student of Communication Arts and Sciences, you are being invited to participate in a study
that investigates the motivational factors for using mobile commerce. Researchers at Michigan
State University will conduct this research.

PURPOSE

The purpose Of this study is to determine the factors that inﬂuence perceptions, intentions and
performance on the wireless lntemet accessed via mobile devices. You will evaluate a mobile
system that provides online music ﬁles.

PROCEDURE

This survey is anonymous and your name will not be associated with your answers. You will be
provided with a personal digital assistant (PDA) and given a short instruction on how to use it.
Afterward, you will be asked sit anywhere in the room and perform four tasks by browsing the
wireless lntemet through the PDA. Your tasks involve ﬁnding a number of music titles and
downloading them on your device. After each set of two tasks, you will be given a short survey
in order to record your perceptions and purchase intention. We estimate that your total
participation for this project will require about twenty minutes. Your completion Of the
questionnaires is completely voluntary and will have no effect on your evaluation in any class.

RISKS

Although it is highly unlikely, there is a chance that you might feel uncomfortable with the
mobile device and/or questionnaire. If for any reason you ﬁnd a survey question inappropriate,
you have the right to decline answering that question, or just stop participating. The study can be
postponed and/or terminated in order to make the experience a positive one.

BENEFITS
We cannot and do not guarantee you will directly beneﬁt if you take part in this study. Through
your comments we can Obtain valuable information about the factors that inﬂuence the use of

mobile devices for electronic commerce. Your participation in this study is completely optional,
and you may earn extra credit in your class by doing an alternate assignment.

137

PAYMENT OF SUBJECTS

As a participant in this research study you will receive classroom extra credit. If you choose not
to participate, then you may do an alternate assignment for extra credit.

ALTERNATIVES

The alternative to participating in this research study is to decline. Participation is purely
voluntary, and for research purposes only. Those who decline participation in any aspect Of this
study will not be penalized for any reason.

CONFIDENTIALITY/PRIVACY

The investigators will keep conﬁdential all research related to records and information in this
study. Throughout the study, all information from the survey will be kept conﬁdential. Results of
the study will only be published in aggregate. The raw data will only be accessible to the
researcher, ZOO Chae. All data will be recorded on the paper survey forms, which will be kept in
a secure location, and then be entered into a computer for storage. The computer, protected by
password authentication, is only accessible to the researcher. Your privacy will be protected to
the maximum extent allowable by law.

QUESTIONS
If you have any questions after reading this form, please contact ZOO-Hyun Chae at the following

address: 442 Communication Arts Building, Michigan State University, East Lansing, Michigan
48824-1212, telephone: (517) 432-9855, email: chaezoo@msu.edu.

Questions concerning your rights as a participant in a research study may be directed to:

Dr. Peter Vasilenko at the University Committee on Research Involving Human Subjects,
Michigan State University, 202 Olds Hall, East Lansing, Michigan 48824-2180, or call at (517)
355-2180.

CONSENT

If you have no additional questions at this time, please sign below to indicate your voluntary

participation in this study. A copy Of this form will be provided for your records.

Signature Date

 

 

Please print your name

 

138

APPENDIX 6: PICTURE SAMPLES OF THE NAVIGATIONAL SYSTEMS

MP3Kingdom: Horizontal-Depth

SQN‘Y

 

 

 

139

MP3Topia: Interactive

 

 

Search Songs for artist
Gerald Levert:

 

 

 

140

REFERENCES

Adams, D. A., Nelson, R. R. and Todd, P. A. (1992). "Perceived usefulness, ease Of use
and usage Of information technology: A replication." MIS Quarteg 16: 227-247.

 

Agresti, A. and Finlay, B. (1997). Statistical Methods for the Social Sciences. Upper
Saddle River, NJ, Prentice-Hall.

Ajzen, I. (1985). From Intentions to Actions: A Theory of Planned Behavior. Action
Control: From @gnition to Behavior. J. Kuhl and Beckmann, J. New York, Springer
Verlag: 11-39.

Ajzen, I. (1989). Attitude Structure and Behavior. Attitude Structure and Function. A. R.
Pratkanis, Breckler, S. J. and Greenwald, A. G. Hillsdale, NJ, Lawrence Erlbaum
Associates: 241 -274.

Ajzen, I. (1991). "The Theory of Planned Behavior." Organizational Behavior and
Human Decision Processes 50: 179-211.

Ajzen, I. and F ishbein, M. (1975). Theories of Attitude. Belief, Attitude, Intention an_c_l_
Behavior: An Introduction to Theory and Research. Reading, MA, Addison-Wesley: 21-
52.

Ajzen, I. and Fishbein, M. (1980). UnderstandingAttitudeL and Predictingﬁocial
Behavior. Englewood Cliffs, NJ, Prentice-Hall.

Alanen, J. and Autio, E. (2003). Mobile Business Services: A Strategic Perspective.
Mobile Commerce: Technology, Theoryand Applications. B. E. Mennecke and Strader,
T. J. Hershey, PA, Idea Group Publishing: 162-184.

Al-Gahtani, S. S. and King, M. (1999). "Attitudes, satisfaction and usage: Factors
contributing tO each in the acceptance Of information technology." Behaviour and
lnforrnation Technology 18(4): 277-297.

Allport, F. H. (1940). "Polls and the Science Of Public Opinion." Public Opinion
Quarterly 4(2): 249-257.

 

141

Anckar, B. and D'Incau, D. (2002). "Value Creation in Mobile Commerce: Findings from

a Consumer Survey." Journal of Information Technology Theory and Application 4(1):
43-64.

Babbie, E. (1998). Survey Research Methods. Belmont, CA, Wadsworth Publishing.

Bagozzi, R. P. and Yi, Y. (1989). "On the Use Of Structural Equation Models in
Experimental Designs." Journaiof Marketing Reseath 26(August): 271-284.

Balasubramanian, 8., Peterson, R. A. and Jarvenpaa, S. L. (2002). "Exploring the
Implications of M-commerce for Markets and Marketing." Acaderm of MaLketigg
ScienceLJournal 30(4): 348-361.

Bandura, A. (1982). "Self-Efﬁcacy Mechanism in Human Agency." American
Psychologist 37(2): 122-147.

Barnes, 8. J. (2003). The Wireless Application Protocol: Strategic Implications for
Wireless lntemet Services. Mobile Commerce: TecMOlongeoqand Application_s_. B.
E. Mennecke and Strader, T. J. Hershey, PA, Idea Group Publishing: 145-161.

Benoit, G. (2002). "Data Discretization for Novel Relationship DiscOvery in Information
Retrieval." Journal Of American Society for Information and Technology 53(9): 736-746.

Bentler, P. M. (1988). "Comparative Fit Indexes in Structural Models." Psychological
Bulletin 107: 238-246.

 

Bentler, P. M. (1995). EQS Structural Equationa Program Manual. Encino, CA, Seventh
Printing.

Bentler, P. M. and Bonett, D. G. (1980). "Signiﬁcant tests and goodness-of-ﬁt in the
analysis of covariance structures." Psychological Bulletin 88: 588-606.

Betts, M. (2001). Optimistic Forecasts Fuel Wireless Hype. Computerworld. 35: 20.

 

Blau, P. M. (1987). Microprocess and Macrostructure. Social Exchange Theory. K. S.
Cook. Newbury Park, CA, Sage Publications: 83-100.

142

Bollen, K. A. (1989). Structural Eguations with Latent Variables. New York, Wiley.

Brodsky, I. (2002). Games: Wireless Killer App? NeMorLWorld. 19: 3S.

Browne, M. W. and Mels, G. (1990). RAMONA User's Guide. Columbus, OH,
Department Of Psychology, Ohio State University.

Buellingen, F. and Woerter, M. (2004). "Development Perspectives, Firm Strategies and
Applications in Mobile Commerce." Journal of Business Research 57(12): 1402-1408.

Carmines, E. G. and Zeller, R. A. (1979). Reliability and Validity Assessment. Beverly
Hill, CA, Sage Publications.

Chae, M. and Kim, J. (2003). An Empirical Study on the Breadth and Depth Tradeoffs in
Very Small Screens: Focusirg on Mobile lntemet Phones. Proceedings of the 9th
Americas Conference on Information Systems (AMCIS), Tampa, FL.

Chau, P. Y. K. (2001). "Inﬂuence of Computer Attitude and Self-efﬁcacy on IT Usage
Behavior." Journal of End User Computing 13(1): 26—33.

Chevillat, P. R. and Schott, W. (2003). "Broadband Radio LAN 5 and the Evolution of
Wireless Beyond 3G." IBM Journal of Research and Development 47(2/3): 327-336.

Chittaro, L. and Cin, P. D. (2001). Evaluating Interface Design Choices on WAP Phones:
giggle-choice List Selection and Navigation amorgCards. Proceedings of Mobile HCI
2001: Third lntemational Workshop on Human-Computer Interaction with Mobile
Devices, Lille, France.

Choi, S.-Y., Stahl, D. O. and Whinston, A. B. (1997). The Economics Of Electronic
Commerce. Indianapolis, Macmilland Technical Publishing.

Clarke III, I. (2001). "Emerging Value Propositions for M-commerce." Journal of
Business Strategi_e_s 18(2): 133-148.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ,
Lawrence Erlbaum Associates.

143

Cohen, W. M. and Levinthal, D. A. (1990). "Absorptive Capacity: A New Perspective on
Learning and Innovation." Administrative Science Quarterly 35: 128-152.

Compeau, D. R. and Higgins, C. A. (1995). "Application of Social Cognitive Theory to
Training for Computer Skills." Information Systems Research 6(2): 118-143.

Cook, T. and Campbell, D. T. (1979). Quasi-Experimentation: Design and Analysis
Issues in a Field Setting. Chicago, Rand McNally.

Coursaris, C., Hassanein, K. and Head, M. (2003). "M-commerce in Canada: An
interaction Framework for Wireless Privacy." Canadian Journal Of Administrative
Sciences 20(1): 54-73.

Crockett, R. O. and Reinhardt, A. (2003). America Zoorns In On Camera Phones
Suddenly, the Market is Hot - and Asian companies are grabbing it ﬁrst. Business Week:
44.

 

Cunningham, W. H., Anderson, T. W. J. and Murphy, J. H. (1974). "Are Students Real
People?" Journal of Business 47: 399-409.

Daﬁ, R. L. and Lengel, R. H. (1984). Information Richness: A New Approach to
Managerial Behavior and Organization Design. Research In Organizational Behavior. B.
M. Staw and Cummings, L. L. Greenwich, CT, JAI Press. 6: 191-233.

Dahlberg, T., Mallat, N. and Oomi, A. (2003). Trust Enhanced Technology Acceptance
Model - Consumer Acceptance of Mobile Payment Solutiona. The Stockholm Mobility
Roundtable 2003, Stockholm, Sweden.

D'Ambra, J. and Wilson, C. S. (2004). "Use of the World Wide Web for International
Travel: Integrating the Construct Of Uncertainty in Information Seeking and the Task-
Technology Fit (TTF) Model." Journal Of the American Society for Informzaion Science
and Technology 55(8): 731-742.

Davis, F. D. (1989). "Perceived Usefulness, Perceived Ease of Use, and User Acceptance
of Information Technology." MIS Quarterly 13: 319-340.

144

Davis, F. D., Bagozzi, R. P. and Warshaw, P. R. (1989). "User Acceptance Of Computer
Technology: A Comparison of Two Theoretical Models." Management Science 35(8):
982-1003.

 

Davis, F. D., Bagozzi, R. P. and Warshaw, P. R. (1992). "Extrinsinc and Intrinsic

Motivation to Use Computers in the Workplace." Journal Of Applied Social Psychology
22(14): 1111-1132.

Dishaw, M. T. and Strong, D. M. (1999). "Extending the Technology Acceptance Model
with Task-Technology Fit Constructs." Information; and Management 36(1): 9-21.

Dishaw, M. T., Strong, D. M. and Bandy, D. B. (2002). Extending the Task-Technology
Fit Model with Self-Efﬁcacy Constructs. Proceedings of the Eighth American
Conference on Information Systems (AMCIS), Dallas, TX.

Drucker, J. (2003). Parts Shortage Crimps Motorola; Demand for Camera Phones Won't
Be Fully Satisﬁed; 'Triplets' Line Affected. The Wall Street Journal. New York, NY: B.5.

Dunn, D. (2002). DSP Group to Partner on MPEG-4 - Will work with Meicom on
licensable solution for 2.50, BG cell phones. EBN: 22.

Durkheim, E., Ed. (1958). Professional Ethics and Civic Moraﬁ. New York, The Free
Press.

Ekeh, P. P. (1974). Social Exchange Theory: The Two Traditions. London, UK,
Heinemann.

Emmett, P. (1989). Attitude Measurement. Measurement of Communication Behavior. P.
Emmer and Barker, L. L. New York, Longman: 134-153.

Emmert, P. (1989). Philosophy of Measurement. Measurement of Communication
Behavior. P. Emmer and Barker, L. L. New York, Longman: 87-116.

Enis, B. M., Cox, K. K. and Stafford, J. E. (1972). "Students as Subjects in Consumer
Behavior Experiments." Journal of Marketing Researah 9: 72-74.

Fairlie, R. (2001). A Peek at the Future Of Digital Domiciles. Fortune. 143: 102-109.

 

145

Ferratt, T. W. and Vlahos, G. E. (1998). "An Investigation Of Task-Technology Fit for
Managers in Greece and the US." European Journal of Information Systems 7(2): 123-
136.

 

Figge, S. (2004). "Situation-dependent Services - A Challenge for Mobile Network
Operators." Journal of Business Reseaﬂh 57(12): 1416-1422.

Fishbein, M. and Ajzen, I. (1975). Belief, Attitude, Intention and Behavior: An
Introduction to Theory and Research. Reading, MA, Addison-Wesley.

Fransman, M. (2002). Explainiag the Success Of NTT Docomo's I-Mode Wireless
lntemet Service. lntemational Telecommunication Society (ITS) 14th Biennial
Conference, Seoul, Korea.

Frick, O. (2000). "WAP - Designing for Small User Interfaces." SAP Design Guild,
Edition 7.

Frolick, M. N. (2004). "Assessing M-commerce Opportunities." Infogation Systems
Managgment 21(2): 53-61.

Fulk, J ., Schmitz, J. and Steinﬁeld, C. W. (1990). A Social Inﬂuence Model of
Technology Use. Organizations and Communication Technology. J. Fulk and Steinﬁeld,
C. Newbury Park, Sage Publications.

Galletta, D. F ., Ahuja, M. K., Hartman, A., Teo, T. and Peace, A. G. (1995). "Social
Inﬂuence and End-user Training." Association for Computing Machinery,
Communications Of the ACM 38(7): 70-79.

Gefen, D., Karahanna, E. and Straub, D. W. (2003). "Trust and TAM in Online
Shopping: An Integrated Model." MIS Quarterly 27(1): 51-90.

Gerstheimer, O. and Lupp, C. (2004). "Needs versus Technology - The Challenge to
Design Third-generation Mobile Applications." Journal of Business Research 57(12):
1409-1415.

146

Giaglis, G. M. (2003). Towards a Classiﬁcation Framework for Mobile Location Services.
Mobile Commerce: Techonolgy, Theory and Applications. B. E. Mennecke and Strader,
T. J. Hershey, PA, Idea Group Publishing: 67-85.

Goldsmith, R. E. and Goldsmith, E. B. (2002). "Buying Apparel over the Internet." The
Journal of Product and Brand Management 11(2/3): 89-100.

Goodhue, D. L. (1995). "Understanding User Evaluations of Information Systems."
Management Science 41(12): 1827-1844.

Goodhue, D. L. (1997). "The Model Underlying the Measurement Of the Impacts of the
IIC on the End-Users." Journal of the American Society for Information Science 48(5):
449-453.

Goodhue, D. L. (1998). "Development and Measurement Validity of a Task-Technology

Fit Instrument for User Evaluations of Information Systems." Decision Sciences 29(1):
105-138.

 

Goodhue, D. L. and Thompson, R. L. (1995). "Task-Technology Fit and Individual
Performance." MIS Quarterly 19(2): 213-236.

Hammond, K. (2001). "B2C E-commerce 2000-2010: What Experts Predict." Business
Strategy Review 12(1): 43-50.

Haque, A. (2003). "Mobile Commerce: Customer Perception and Its Prospect on
Business Operation in Malaysia." Journal of American Academy Of Business, Cambridga
4(1/2): 257-262.

Heise, D. R. (1970). The Semantic Differential and Attitude Research. Attitude
Measurement. G. F. Summers. Chicago, Rand McNally & CO.: 235-253.

Henneman, R. L. and Rouse, W. B. (1984). "Human Performance in Monitoring and
Controlling Hierarchical Large-Scale Systems." IEEE Traasactions on System, Man and
Cybernetics (SMC) 14: 184-191.

Homans, G. C. (1958). "Social Behavior as Exchange." American Sociological Review
12(Feb): 13-26.

147

Hoyle, R. H. (1995). The Structural Equation Modeling Approach: Basic Concepts and
Fundamental Issues. Structural Equation Modeliag: Concggta, Issues, and Applications.
R. H. Hoyle. London, Sage Publications.

Hsu, C.-L. and Lu, H.-P. (2004). "Why DO People Play On-Line Games? An Extended
TAM with Social Inﬂuences and Flow Experience." Information and Managament 41(7):
853-868.

Igbaria, M. and Livari, J. (1995). "The Effects of Self-efﬁcacy on Computer Usage."
OMEGA 23(6): 587-605.

Igbaria, M., Zinatelli, N., Cragg, P. and Cavaye, A. L. M. (1997). "Personal Computing
Acceptance Factors in Small Firms: A Structural Equation Model." MIS Quarterly 21(3):
279-302.

 

Jasperson, J., Sambamurthy, V. and Zmud, R. W. (1999). Social Inﬂuence and Individg
IT Use: Unravelirg the Pathways Of Apprgiation Moves. Proceedings of the 20th
lntemational Conference on Information Systems (ICIS), Charlotte, NC.

Jones, M., Marsden, G., Mohd-Nasir, N., Boone, K. and Buchanan, G. (1999).
"Improving Web Interaction on Small Displays." Computer Networks 31: 11291137.

Jones, T. (2004). "Put It On My Mobile." Telecommunicationa lntemational 38(11): 37-
38.

Joreskog, K. G. and Sorbom, D. (1984). Lisrel 6: User's guide. Mooresville, IN,
Scientiﬁc Software.

Kaiser, H. F. (1958). "The Varimax Criterion for Analytic Rotation in Factor Analysis."
Psychometrika 23: 187-200.

Katerattanakul, P. (2002). "Framework Of Effective Web Site Design for Business-to-
Consumer lntemet Commerce." INFOR 40(1): 57-70.

Keller, R. T. (1994). "Technology-Information Processing Fit and the Performance of
R&D Project Groups: A Test of Contingency Theory." Academy Of Management Journal
37(1): 167-179.

148

Keppel, G. (1991). Design and Analysis: A Researcher's Handbook. Upper Saddle River,
NJ, Prentice Hall.

Kim, Y. (2005). "Estimation of Consumer Preferences on New Telecommunications
Services: [MT-2000 Service in Korea." Information Economics and Policy 17(1): 73-84.

Kleijnen, M., De Ruyter, K. and Wetzels, M. G. M. (2003). Factors Inﬂuencing the
Adoption Of Mobile Gaming Services. Mobile Commerce: Technology, Theory, and
Applications. B. E. Mennecke and Strader, T. J. Hershey, PA, Idea Group Publishing.

Klopping, I. M. and McKinney, E. (2004). "Extending the Technology Acceptance
Model and the Task-Technology F it Model to Consumer E-Commerce." Information
Technolgy, Learning, and Performance Journal 22(1): 35-48.

 

Koh, J. and Kim, Y. M. (2000). Mobile lntemet Applications Primer, UBS Warburg,
Global Equity Research Report.

Koufaris, M. (2002). "Applying the Technology Acceptance Model and Flow Theory to
Online Consumer Behavior." Information Systems Research 13(2): 205-223.

Kraut, R., Steinﬁeld, C., Chan, A., Butler, B. and Hoag, A. (1998). "Coordination and
Virtualization: The Role of Electronic Networks and Personal Relationships." Journal of
Computer Mediated Communication 3(4).

Kuk, G. and Yeung, F. T. (2002). "Interactivity in E-commerce." Quarterly Journal Of
Electronic Commerce 3(3): 223-234.

 

Landler, M. (2002). In Europe, a Promise of Mobile Multimedia. The New York Times.
New York, NY: C.3.

 

Lee, C. and Ke, C.-H. (2001). "A Prediction-Based Query Processing Strategy in Mobile
Commerce Systems." Journal Of Database Management 12(3): 14-26.

Lee, Y., Kim, J ., Lee, I. and Kim, H. (2002). "A Cross-cultural Study Of the Value
Structure Of Mobile lntemet Usage: Comparison Between Korea and Japan." Journal of
Electronic Commerce Research 3(4): 227-239.

149

Lee, Y., Kozar, K. A. and Larsen, K. R. T. (2003). "The Technology Acceptance Model:
Past, Present, and Future." Communications of the Association for Information Systems
12(50): 752-780.

Lee, Y., Lee, J. and Lee, Z. (2001). The Effect of Self Identity and Social Identity on
Technology Acceptance. Proceedings of the 22nd lntemational Conference on
Information Systems (ICIS), New Orleans, LA.

Leibenstein, H. (1950). "Bandwagon, Snob, and Veblen Effects in the Theory Of
Consumer Demand." The Quarterly Journal of Economics(August): 183-207.

Leung, K. and Antypas, J. (2001). "Improving Returns on M-commerce Investments."
The Journal of Business Strategy 22(5): 12-13.

Levi-Strauss, C. (1969). The Elementary Structures of Kinship. Boston, Beacon Press.
Lewis, P. (2003). Winners and Losers of 2003. Fortune. 148: 187.

Li, H., Daugherty, T. and Biocca, F. (2001). "Characteristics of Virtual Experience in

Electronic Commerce: A Protocol Analysis." Journal Of Interactive Marketing 15(3): 13-
30.

Lim, K. H. and Benbasat, I. (2000). "The Effect of Multimedia on Perceived Equivocality
and Perceived Usefulness of Information Systems." MIS Quarterly 24(3): 449-471.

Lin, X., White, H. D. and Buzydlowski, J. (2003). "Real-time Author Co-citation
Mapping for Online Searching." Information Processirgand Management 39(5): 689.

LOO, R. (2002). "A Caveat on Using Single-item Versus Multiple-item Scales." Journal Of
Managerial Psychology 17(1/2): 68-75.

Lynch, J. G. J. (1999). "Theory and External Validity." Journal of the Academy of
Marketig Science 27(3): 367-376.

 

Magura, B. (2003). "What Hooks M-commerce Customers?" MIT Sloan Management
Review 44(3): 9.

150

Maitland, C. F. (2004). Mobile Commerce Technologies and Services: A SME User's
Perspective. E-Life After the Dot Com Bust. B. Preissl, Bouwman, H. and Steinﬁeld, C.
Heidelberg, Physica-Verlag.

Malhotra, Y. and Galletta, D. F. (1999). Extending the Technology Accgptance Model to
Account for Social Inﬂuence Theoretical Bases and Empirical Validation. Proceedings of
the 32nd Annual Hawaii lntemational Conference on System Sciences (HICSS), Maui,
Hawaii.

Massey, A. P., Montoya-Weiss, M., Hung, C. and Rarnesh, V. (2001). "Cultural
Perceptions of Task-Technology F it. " Association for ComputingMachinery.
Communications of the ACM 44(12): 83.

Mathieson, K. (1991). "Predicting User Intentions: Comparing the Technology
Acceptance Model with the Theory of Planned Behavior." Information Systems Research
2(3): 173-191.

Mathieson, K. and Keil, M. (1998). "Beyond the Interface: Ease of Use and
Task/Technology F it. " Inergration and Management 34(4): 221-230.

Mathieson, K., Peacock, E. and Chin, W. W. (2001). "Extending the Technology
Acceptance Model: The Inﬂuence of Perceived User Resources." The DATA BASE for
Advances in Information Systems 32(3): 86-112.

Matsuno, K. and Mentzer, J. T. (2000). "The Effects of Strategy Type on the Market
Orientation-Performance Relationship." Journal of Marketing 64(4): 1-16.

May, P. (2001). Mobile Commerce: Opmrtunities, Applications, and Technologias of
Wireless Business. Cambridge, UK, Cambridge University Press.

McClelland, S. (2004). "Japan: Lessons in Growth." Telecommunications International
38(9): S3-SS.

McKenzie, R. B. (2003). Digital Economics: How Information Technology has
Transformed Business Thinking. Westport, CT, Praeger Publishers.

McNemar, Q. (1946). "Opinion-Attitude Methodology." Psychological Bulletin 43: 289-
374.

151

Meuter, M. L., Ostrom, A. L., Bitner, M. J. and Roundtree, R. (2003). "The Inﬂuence of
Technology Anxiety on Consumer Use and Experiences with Self-Service
Technologies." Journal Of Business Research 56(11): 899.

Minges, M. (2005). "Is the lntemet Mobile? Measurements from the Asia-Paciﬁc
Region." Telecommunications Policy 29(2,3): 113-125.

Morgan, F. W. J. (1979). "Students in Marketing Research: Surrogates vs. Role-Players."
Academy Of Marketing Science, Joumaﬂ 7: 255-264.

Murthy, U. S. and Kerr, D. S. (2000). Task/Technology Fit and the Effectiveness Of
GroupSupport Systems: Evidence in the Context of Tasks Requirig Domain Speciﬁc
Knowledge. Proceedings Of the 33rd Hawaii lntemational Conference on System
Sciences.

Nadel, B. (2002). Waiting For the Wireless Revolution; 3G cell phones promise tons of
features. Are Dick Tracy video phones in you future? PC Magazine. 21: 84.

Neville, K. and Fitzgerald, B. (2002). "An Innovative Training Model for an
Organization Embracing Technology." Journal of Information Technology Education
1(3): 193-200.

Nunnally, J. and Durham, R. (1975). Validity, Reliability, and Special Problems of
Measurement in Evaluation Research. Handbook of Evaluation Research. E. Struening
and Guttentag, M. Beverly Hills, CA, Sage Publications. 1: 289-352.

Okazaki, S. (2004). "How do Japanese Consumers Perceive Wireless Ads? A
Multivariate Analysis." Intematiopal Journal Of Advertising 23(4): 429-454.

Osgood, C. E., Suci, G. J. and Tannenbaum, P. H. (1970). Attitude Measurement.
Attitude Measurement. G. F. Summers. Chicago, Rand McNally & CO.: 227-234.

Paap, K. and Cooke, N. (1997). Design Of Menus. Handbook of Human-Computer
Interaction. M. Helander and Landauer, T. K. Amsterdam, Netherlands, Elsevier Science
B. V.

152

Pavlou, P. A. (2001). Consumer Intention_s_ to Adopt Electronic Commerce -
Incorporatig Trust and Risk in the TechologpAcceptance Model. DIGIT Workshop,
New Orleans, LA.

Peterson, C. (1997). Psychology: A Biopsychosocial Approach. New York, Longman.

Racanelli, V. J. (2005). "The Latest Gee-Whiz Forecasts for 3G." Barron's 85(9): MW9.

Raisinghani, M. (2002). "Mobile Commerce: Transforming the Vision into Reality."
Information Resources Management Journal 15(2): 3-4.

Ralph, D. and Aghvami, H. (2001). "Wireless Application Protocol Overview." Wireless
Communications and Mobile Computing 1(2): 125-140.

Riesman, D. (1952). The Lonely Crowd. New Haven, Yale University Press.

Rogers, E. M. (1983). Diffusion of Innovations. New York, The Free Press.

Rosson, M. B. and Carroll, J. M. (2002). Usability Engineering Scenario-Based
Development of Human-Computer Interaction. San Francisco, CA, Academic Press.

Rowello, R. (2001). The lntemet Unplugged: Consumer Electronics in a 3G World,
Pittiglio Rabin Todd & McGrath (PRTM).

Royer, S. and Van Der Velden, R. (2003). "Economics, E-commerce and Strategy
Development: Resources and Rent Creation for Digital Goods Providers on the lntemet."
lntemational Journal of Management & Decision Making 4(2/3): 116-132.

Rubin, R. B., Palmgreen, P. and Sypher, H. E. (1994). Communication Research
Measures: A Sourcebook. New York, The Guilford Press.

 

Saade, R. and Bahli, B. (2005). "The Impact of Cognitive Absorption on Perceived
Usefulness and Perceived Ease of Use in On-line Learning: An Extension Of the
Technology Acceptance Model." Information and Management 42(2): 317-327.

153

Schiff, M. R. (1970). Some Theoretical Asgcts of Attitudes and Perception. Toronto,
Natural Hazard Research.

Schmitz, J. and Fulk, J. (1991). "Organizational Colleagues, Media Richness, and
Electronic Mail: A Test Of the Social Inﬂuence Model of Technology Use."
Communication Research 18(4): 487-523.

Schone, S. (2004). "M-Commerce: The New "Anywhere, Anytime" Sales Channel."
Computer Technology Review 24(10): 1, 38.

Scott-Lennox, J. A. and Lennox, R. D. (1995). Sex-Race Differences in Social Support
and Depression in Older Low-Income Adults. Structural Eauation Modeling: Concepts,
Issues, and Applications. R. H. Hoyle. Thousand Oaks, Sage Publications.

Sheppard, B. H., Hartwick, J. and Warshaw, P. R. (1988). "The Theory Of Reasoned
Action: A Meta-Analysis of Past Research with Recommendations for Modiﬁcations and
Future Research." Journal Of Consumer Research 15(3): 325-343.

Shirani, A. 1., Tafti, M. H. A. and Afﬁsco, J. F. (1999). "Task and Technology Fit: A
Comparison of Two Technologies for Synchronous and Asynchronous Group
Communication." Information and Management 36(3): 139-150.

Siau, K., Lim, E.-P. and Shen, Z. (2001). "Mobile Commerce: Promises, Challenges, and
Research Agenda." Journal of Database Management 12(3): 4-13.

Staples, S. D. and Seddon, P. (2004). "Testinf the Technology-to-Performance Chain
Model." Journal of Oaganizational and End User Computing 16(4): 17-36.

Steﬂ-Mabry, J. (2003). "A Social Judgment Analysis of Information Source Preference
Proﬁles: An Exploratory Study to Empirically Represent Media Selection Patterns."
Journal Of the American Society for Information Science ml Technology 54(9): 879.

Steiger, J. H. (1989). Causal Modeling: A Supplementary Module for SYSTAT and
SYGRAPH. Evanston, IL, SYSTAT.

Straub, D., Limayem, M. and Karahanna-Evaristo, E. (1995). "Measuring System Usage:
Implications for IS Theory Testing." Management Science 41(8): 1328-1342.

154

Szajna, B. (1996). "Empirical Evaluation Of the Revised Technology Acceptance Model."
Management Science 42(1): 85-92.

Takahashi, D. (2002). The Next MPEG Step. Electronic Business. 28: 80-83.

Tarasewich, P. (2003). Wireless Devices for Mobile Commerce: User Interface Design
and Usability. Mobile Commerce: Technology, Theory, and Application. B. E. Mennecke
and Strader, T. J. Hershey, PA, Idea Group Publishing: 26-50.

Taylor, S. and Todd, P. A. (1995). "Assessing IT Usage: The Role of Prior Experience."
MIS Quarterly 19(4): 561-570.

Taylor, S. and Todd, P. A. (1995). "Understanding Information Technology Usage: A
Test Of Competing Models." Information Systems Research 6(2): 144-176.

Thompson, R. L., Higgins, C. A. and Howell, J. M. (1991). "Personal Computing:
Toward a Conceptual Model Of Utilization." MIS Quarterly 15(1): 124-143.

Tushman, M. L. and Nadler, D. A. (1978). "Information Processing as an Integrating
Concept in Organization Design." Academy of Management Review 3: 613-624.

Van de Ven, A. H. and Drazin, R. (1985). The Concept Of Fit in Contingency Theory.

Research in Organizational Behavior. L. L. Cummings and Staw, B. M. Greenwich, CT.,
JAI Press. 7: 333-365.

Van Dolen, W. M. and De Ruyter, K. (2002). "Moderated Group Chat: An Empirical
Assessment of a New E—service Encounter." lntemational Journal of Service Industry
Management 13(5): 496-51 1.

Venkatesh, V. and Davis, F. D. (1996). "A Model of the Antecedents Of Perceived Ease
of Use: Development and Test." Decision Sciences 27(3): 451-481.

Venkatesh, V. and Davis, F. D. (2000). "A Theoretical Extension of the Technology
Acceptance Model: Four Longitudinal Field Studies." Management Science 46(2): 186-
204.

155

Venkatesh, V. and 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. and Davis, F. D. (2003). "User Acceptance Of
Information Technology: Toward a Uniﬁed View." MIS Quarterly 27(3): 425-478.

Venkatesh, V., Speier, C. and Morris, M. G. (2002). "User Acceptance Enablers in
Individual Decision Making About Technology: Toward an Integrated Model." Decision
Sciences 33(2): 297-316.

Wimmer, R. D. and Dominick, J. R. (1997). Mass Media Research: An Introduction.
Belmont, CA, Wadsworth Publishing.

Winer, R. S. (1999). "Experimentation in the let Century: The Importance of External
Validity." Journal of the Academy of Marketiag Science 27(3): 349-358.

Wolﬁnbarger, M. and Gilly, M. C. (2001). "Shopping Online for Freedom, Control, and
Fun." California ManaLement Review 43(2): 34-55.

Young, J. (2002). Will PDAs Ever Be All-In-One? Office Solution_s. 19: 46-48.

 

Zigurs, I. and Buckland, B. K. (1998). "A Theory Of Task/Technology Fit and Group
Support Systems Effectiveness." MIS Quarterly 22(3): 313-334.

156

AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA

lllﬂlllllllllllIlllllllllllljlllljjljlljjlll

II
1293 027