Factors influencing the post-adoption consequences of online

Southern Cross UniversityePublications@SCU

Theses

2009

Factors influencing the post-adoptionconsequences of online securities trading inSingapore’s retail investorsAnthony YeongSouthern Cross University

ePublications@SCU is an electronic repository administered by Southern Cross University Library. Its goal is to capture and preserve the intellectualoutput of Southern Cross University authors and researchers, and to increase visibility and impact through open access to researchers around theworld. For further information please contact [email protected].

Publication detailsYeong, A 2009, 'Factors influencing the post-adoption consequences of online securities trading in Singapore’s retail investors', DBAthesis, Southern Cross University, Lismore, NSW.Copyright A Yeong 2009

http://epubs.scu.edu.au

http://epubs.scu.edu.au/theses

mailto:[email protected]

GRADUATE COLLEGE OF MANAGEMENT

Factors Influencing the Post-Adoption Consequences of

Online Securities Trading

in Singapore’s Retail Investors

Anthony Yeong

Higher Diploma (Staffordshire University, UK, 1991) MBA (Southern Cross University, Australia, 1998)

THESIS

Submitted to The Graduate College of Management, Southern Cross University Australia

In partial fulfilment of the requirement for the degree of DOCTOR OF BUSINESS ADMINISTRATION

August 2009

- i -

I certify that the substance of the research thesis has not been submitted for any

degree and is not currently being submitted to any other degree.

I also certify that to the best of my knowledge any help received in preparing this

thesis, and all sources used have been acknowledged in this thesis.

Signed:

(Anthony Yeong)

Date: August 2009

STATEMENT OF ORIGINAL AUTHORSHIP

- ii -

The completion of this Doctor of Business Administration dissertation would not

have been possible without the support, advice and encouragement from a number of

people.

First and foremost, I would like to express my deepest appreciation to my supervisor,

Dr Chris McDowell, the Graduate College of Management, Southern Cross

University. His prompt advice and assistance has encouraged me to carry through to

the completion of this research study. He has always shown a keen interest and

patience in my research study.

Secondly, my heartfelt thanks also go to Dr Margo Poole who provided me with

excellent guidance and advice in SPSS and data analysis.

Thirdly, I am thankful to Adjunct Professor Dr. C.S. Teo who has introduced me to

this doctoral program. He has been a great mentor for the past many years and has

provided much advice and support especially during the early stage of this research

study. In addition, I am especially grateful to Sue White and Susan Riordan for their

administrative support. Their genuine support and patience were greatly appreciated.

Last, but not least, I would like to offer a special thanks to my beloved wife,

Jacqueline and daughter, Amanda. They have been supportive and understanding

during my long journey in completion of this program.

Anthony Yeong

ACKNOWLEDGEMENTS

- iii -

This research aims to establish the important dimension of pre-adoption factors’

influence on the consequences, or post-adoption usage behaviour, of online securities

trading by the retail investors in Singapore.

While several studies have been conducted by researchers on the factors that lead

people to adopt new innovations, few have actually explored the consequences of the

innovations or post-adoption usage behaviour. Therefore, this research fills the gap in

the diffusion and adoption studies. The researcher aimed to address the following

three research issues in this study:

RI1) What pre-adoption variables in the Diffusion of Innovations model affect the

post-adoption usage behaviour of retail investors using online securities trading?

RI2) Does Perceived usefulness affect the post-adoption usage behaviour of the

retail investors using online securities trading?

RI3) Will Consumer loyalty affect the post-adoption usage behaviour of retail

investors using online securities trading?

The parent discipline of this research is Consumer Behaviour which has been

elaborated on in the literature review chapter. The immediate disciplines are

Consequences of innovations from the Diffusion model; Perceived usefulness from

the Technology Acceptance Model and Consumer loyalty. They have been discussed

in the literature review chapter as well. Seven propositions have been derived from

ABSTRACT

- iv -

the literature review and have subsequently formed the hypotheses of the research

study.

A positive paradigm has been selected for this research, and the data analysis

undertaken used a quantitative method and an online survey questionnaire to gather

research data from the retail investors using online securities trading in Singapore.

There were 232 data elements collected from the online survey. The data were then

further analysed by: using factor analysis by the Varimax rotation extraction method,

conducting reliability testing using Cronbach Alpha testing, and testing of the

theoretical model and hypotheses using multiple linear regression analysis.

The findings concluded that not all of the variables in the pre-adoption factors of

Roger’s Diffusion model are influencing the post-adoption usage behaviour of the

retail investors trading stock online. The pre-adoption variables tested to have

significant influence on the post-adoption usage behaviour are: Compatibility,

Complexity, Trailability and Observability. The Optional decision variable was found

to have an influence on the post-adoption usage behaviour but not the variables

Authority decision and Collective decision.

Nature of social system and Change agent’s promotion efforts are shown to have

significant influence on post-adoption usage behaviour of the retail investors.

Perceived usefulness and Consumer loyalty variables have also been tested and

concluded to have an influence on post-adoption usage behaviour of the retail

investors trading stock online.

Finally, contributions to the knowledge, research limitations and areas for further

research, especially in the Consequences of Innovations discipline, were discussed.

- v -

From the findings of this research study, a framework has been set for future

researchers to investigate further the pre-adoption factors’ influence on the post-

adoption usage behaviour or consequences of innovations in other products and

services. In addition, this study provides some useful findings and implications for

the retail users, researchers, practitioners and brokerage firms in the area of online

securities trading usage.

- vi -

Page Statement of Original Authorship i Acknowledgements ii Abstract iii Table of Contents vi List of Figures xii List of Tables xiv List of Charts xvi

Contents

Chapter 1 Introduction..............................................................................................17

1.1 Introduction to Chapter One ........................................................................ 17

1.2 Background to the research.......................................................................... 19

1.3 Research problem, research issues and contributions.................................. 22

1.3.1 Research problem.....................................................................................22

1.3.2 Research issues ........................................................................................23

1.3.3 Research contributions.............................................................................25

1.4 Justification for the research ........................................................................ 25

1.4.1 Limited research in consequences of innovations ...................................25

1.4.2 Importance of online securities trading....................................................26

1.4.3 Online survey method ..............................................................................26

1.4.4 Potential usage of research findings ........................................................27

TABLE OF CONTENTS

- vii -

1.5 Methodology................................................................................................ 27

1.6 Definitions.................................................................................................... 28

1.7 Delimitations of scope and key assumptions ............................................... 33

1.8 Outline of the thesis ..................................................................................... 34

1.9 Summary of the chapter ............................................................................... 37

Chapter 2 Literature Review.....................................................................................38

2.1 Introduction to Chapter Two........................................................................ 38

2.2 Stock Exchange and Online Securities Trading........................................... 40

2.2.1 Retail Stock Investors ..............................................................................46

2.3 Consumer Behaviour ................................................................................... 47

2.3.1 History of Consumer Behaviour ..............................................................47

2.3.2 Process of Consumer Behaviour..............................................................53

2.3.3 Consumer Behaviour and Decision Making ............................................54

2.3.4 Consumer Behaviour and Adoption.........................................................54

2.4 Review of Diffusion Model Literature ........................................................ 55

2.4.1 History of the Diffusion Model................................................................55

2.4.2 Adoption Theory......................................................................................58

2.4.3 Diffusion Theory......................................................................................63

2.4.4 Acceptance and Consequences of Innovations ........................................78

2.4.5 Consumer loyalty and Consequences of Innovations ..............................86

2.5 Theoretical Model........................................................................................ 88

2.5.1 The Variables ...........................................................................................88

2.5.2 Proposed Independent Variables..............................................................89

2.5.3 Dependent Variable .................................................................................94

2.5.4 Linkages amongst the Variables ..............................................................95

- viii -

2.5.5 Research Model .......................................................................................96

2.5.6 Research Propositions Formulation .........................................................98

2.6 Summary of the chapter ............................................................................. 103

Chapter 3 Methodology..........................................................................................105

3.1 Introduction to Chapter Three.................................................................... 105

3.2 Research Paradigms ................................................................................... 107

3.2.1 Positivism...............................................................................................109

3.2.2 Critical Theory .......................................................................................110

3.2.3 Constructivism .......................................................................................111

3.2.4 Realism ..................................................................................................112

3.2.5 Justification of the selected research paradigm .....................................113

3.3 Research Methods...................................................................................... 114

3.3.1 Qualitative Research ..............................................................................115

3.3.2 Quantitative Research ............................................................................116

3.3.3 Justification of the selected research method.........................................116

3.3.4 Limitation of the selected research method ...........................................117

3.4 Survey Objects ........................................................................................... 119

3.4.1 Theoretical Population...........................................................................119

3.4.2 Accessible Population............................................................................122

3.5 Sampling .................................................................................................... 123

3.5.1 Sampling Design....................................................................................123

3.5.2 Sampling Size ........................................................................................126

3.5.3 Validity and Reliability..........................................................................126

3.6 Questionnaire Design................................................................................. 127

3.6.1 Questionnaire Objective.........................................................................127

- ix -

3.6.2 Questionnaire Types ..............................................................................133

3.6.3 Independent Variables ...........................................................................136

3.6.4 Dependent Variable and Components ...................................................136

3.7 Mode of Survey.......................................................................................... 137

3.7.1 Web Survey / Email ...............................................................................137

3.8 Data Processing Procedures....................................................................... 140

3.8.1 Descriptive statistics - Cross Tabulation................................................140

3.8.2 Factor Analysis ......................................................................................142

3.8.3 Regression Analysis...............................................................................146

3.8.4 Multiple Regression...............................................................................147

3.8.5 Data Processing Tools............................................................................150

3.8.6 Ethical Considerations ...........................................................................150


Chapter 4 Data Analysis .........................................................................................153

4.1 Introduction to Chapter Four ..................................................................... 153

4.2 Data Profile Examination........................................................................... 155

4.2.1 Data Summary .......................................................................................155

4.2.2 Demographic Profiles ............................................................................155

4.3 Development of Constructs........................................................................ 162

4.3.1 Reliability Analysis................................................................................163

4.4 Factor Analysis .......................................................................................... 164

4.4.1 Factor Analysis of Independent Variables.............................................165

4.4.2 Reliability Testing of Factors (Independent Variables).........................181

4.4.3 Factor Analysis for Dependent Variables ..............................................191

4.4.4 Reliability Testing of Factor (Dependent Variable) ..............................196

- x -

4.5 Multiple Regression Analysis .................................................................... 197

4.5.1 Multiple Regression Model....................................................................197

4.5.2 Model Summary – R Square..................................................................200

4.5.3 ANOVA Table .......................................................................................201

4.5.4 Model Parameters ..................................................................................202

4.6 Hypotheses Testing.................................................................................... 204

4.6.1 Test of Hypothesis 1a.............................................................................204

4.6.2 Test of Hypothesis 1b ............................................................................204

4.6.3 Test of Hypothesis 2a.............................................................................205

4.6.4 Test of Hypothesis 2b ............................................................................206

4.6.5 Test of Hypothesis Three .......................................................................206

4.6.6 Test of Hypothesis Four.........................................................................207

4.6.7 Test of Hypothesis Five .........................................................................207

4.6.8 Test of Hypothesis Six ...........................................................................208

4.6.9 Test of Hypothesis Seven.......................................................................208

4.6.10 Summary of Hypotheses Testing .......................................................209


Chapter 5 Conclusions and implications ................................................................215

5.1 Introduction to Chapter Five...................................................................... 215

5.2 Restatement of the research problem and hypotheses ............................... 217

5.3 Conclusion about the hypotheses and research problem ........................... 219

5.3.1 Conclusion for hypothesis one...............................................................220

5.3.2 Conclusion for hypothesis two...............................................................221

5.3.3 Conclusion for hypothesis three.............................................................223

5.3.4 Conclusion for hypothesis four..............................................................223

- xi -

5.3.5 Conclusion for hypothesis five ..............................................................224

5.3.6 Conclusion for hypothesis six................................................................225

5.3.7 Conclusion for hypothesis seven ...........................................................226

5.4 Contributions to the body of knowledge.................................................... 227

5.4.1 Consequences of Innovations ................................................................228

5.4.2 Technology Acceptance Model .............................................................228

5.4.3 Consumer loyalty ...................................................................................229

5.5 Managerial Implications ............................................................................ 230

5.6 Limitations of the research......................................................................... 232

5.6.1 Scope limitation .....................................................................................232

5.6.2 Geographical limitation .........................................................................232

5.6.3 Online questionnaire ..............................................................................233

5.7 Recommendations for future research ....................................................... 233

5.7.1 Other products........................................................................................233

5.7.2 Other geographical regions ....................................................................234

5.7.3 Other factors affecting post-adoption consequences .............................234

5.7.4 Other post-adoption consequences ........................................................234

5.8 Research conclusion................................................................................... 235

Bibliography and References 1..................................................................................237

Appendix A Letter of ethic approval form ..........................................................245

Appendix B Letter of Introduction ..........................................................................247

Appendix C Questionnaire.......................................................................................248

Appendix D SPSS Outputs ......................................................................................253

- xii -

Figure 1-1 Overview of Chapter One ........................................................................18

Figure 1-2 Online securities trading process .............................................................33

Figure 1-3 Outline of this thesis.................................................................................36

Figure 2-1 Overview of Chapter Two........................................................................39

Figure 2-2 Traditional trading process flowchart ......................................................42

Figure 2-3 Online securities trading process flowchart .............................................43

Figure 2-4 Bass's Model ............................................................................................57

Figure 2-5 AIDA Model ............................................................................................59

Figure 2-6 Hierarchy-of-Effects Model .................................................................61

Figure 2-7 Innovation-Decision Model .....................................................................63

Figure 2-8 The diffusion process ...............................................................................66

Figure 2-9 POEMS demo ..........................................................................................69

Figure 2-10 Variables determining the rate of adoption.............................................71

Figure 2-11 Adopter categorisation on the basis of innovativeness ..........................73

Figure 2-12 Technology Acceptance Model (TAM)...................................................79

Figure 2-13 Framework of online consumer behaviour ............................................83

Figure 2-14 Customer's intention to use and usage of electronic ..............................85

Figure 2-15 Consumer loyalty and purchase behaviour ..............................................87

Figure 2-16 Independent variables and post-adoption usage behaviour....................96

Figure 2-17 Conceptual theoretical model of post-adoption usage behaviour of

online securities trading ...............................................................................................97

LIST OF FIGURES

- xiii -

Figure 3-1 Overview of Chapter Three....................................................................106

Figure 3-2 Theoretical model developed for this research ......................................131

Figure 4-1 Overview of Chapter Four .....................................................................154

Figure 4-2 Revised theoretical model ......................................................................212

Figure 5-1 Overview of Chapter Five......................................................................216

- xiv -

Table 1-1 Number of trades online ............................................................................20

Table 1-2 The seven research propositions................................................................24

Table 2-1 Online stock trading in Asia ......................................................................44

Table 2-2 Proposed independent variables for the research ......................................90

Table 2-3 Seven research propositions ......................................................................99

Table 3-1 Basic systems of alternative enquiry paradigms .....................................108

Table 3-2 Characteristics of quantitative and qualitative paradigms.......................114

Table 3-3 Internet penetration rate and percentage of online trade .........................120

Table 3-4 Securities trading members in Singapore ................................................120

Table 3-5 Securities trading members with online trading service..........................122

Table 3-6 Literature based constructs, hypotheses and survey questions................132

Table 3-7 Demographic factors and related questions.............................................136

Table 3-8 Post-adoption behaviour and related questions .......................................137

Table 3-9 Internet survey versus mail survey ..........................................................139

Table 3-10 Cross tabulation of age versus occupation ............................................141

Table 4-1 Correlation matrix for independent variables..........................................167

Table 4-2 KMO and Bartlett’s Test for dependent variables...................................172

Table 4-3 Communalities for independent variables ...............................................173

Table 4-4 Total variance explained for independent variables................................174

Table 4-5 Rotated component matrix for independent variables.............................178

Table 4-6 Reliability test of FS1-H7........................................................................182

LIST OF TABLES

- xv -



Table 4-9 Reliability test of FS4-H1a......................................................................184

Table 4-10 Reliability test of FS5-H3......................................................................185

Table 4-11 Reliability test of FS6-H1b....................................................................186

Table 4-12 Reliability test of FS7-H5......................................................................187

Table 4-13 Reliability test of FS8-H2a.....................................................................188

Table 4-14 Reliability test of FS9-H2b....................................................................189

Table 4-15 Summary of extracted components .......................................................190

Table 4-16 Correlation matrix for dependent variables...........................................192

Table 4-17 KMO and Bartlett’s Test for dependent variables.................................193

Table 4-18 Communalities for dependent variables ................................................193

Table 4-19 Total variance explained based on Eigenvalues....................................194

Table 4-20 Initial component matrix .......................................................................196

Table 4-21 Reliability statistics ...............................................................................196

Table 4-22 Model summary.....................................................................................200

Table 4-23 ANOVA table........................................................................................201

Table 4-24 Coefficients table...................................................................................203

Table 4-25 Coefficients table (without constant).......................................................203

Table 4-26 Summary of the results of hypotheses testing .......................................210

Table 5-1 Summary of results from testing of hypotheses ......................................219

- xvi -

Chart 4-1 Age profiles of respondents .....................................................................156

Chart 4-2 Education profiles of respondents ...........................................................157

Chart 4-3 Occupation profiles of respondents .........................................................158

Chart 4-4 Income profiles of respondents................................................................159

Chart 4-5 Martial status of respondents ...................................................................160

Chart 4-6 Gender profiles of respondents................................................................161

Chart 4-7 Scree Plot for independent variables .......................................................176

Chart 4-8 Scree Plot for dependent variables ..........................................................195

LIST OF CHARTS

Chapter One Introduction

- 17 -

Chapter 1 Introduction

1.1 Introduction to Chapter One

The focus of this thesis is on exploring what underlying factors influence the post-

adoption usage behaviour of retail investors in Singapore after adopting online

securities trading. The aim of the research is to contribute to the research and

understanding of the consequences of innovations. The research is based on Rogers’

model of Diffusion of Innovations. There seems to be a lack of previous research

conducted in the consequences of innovations or post-adoption usage behaviour

specifically in the context of online securities trading usage.

As illustrated in Figure 1.1, this first section of Chapter One provides an overview of

the structure of the chapter. Section 2 covers the background to the research. Section

3 discusses the research problem and research issues in the consequences of

innovations and online securities trading. Section 4 justifies this research study with

the underlying potential contributions, and Section 5 summarises the research

methodology. Section 6 contains a definition of terms. Section 7 covers the

limitations and assumptions in this research. A summary of the overall structure of

the thesis is to be found in Section 8, followed by a summary of the chapter in the

final section.


- 18 -

Figure 1-1 Overview of Chapter One

Source: developed for this research


- 19 -

1.2 Background to the research

Grounded in Rogers’ Diffusion of Innovation Theory (Rogers 2003), and Davis’

Technology Acceptance Model (Davis, F. D. 1989) together with a consumer loyalty

framework, this research develops an integrative model to study the pre-adoption

influences on post-adoption usage behaviour of retail investors using online securities

trading in Singapore.

Electronic commerce and online technologies are considered to be the breakthrough

innovations here, not only in information technology but also in the business world.

Stockbroking firms have openly adopted information and communication technology,

especially in online securities trading, to improve their competiveness and

responsiveness to market conditions (Gharavi, Love & Cheng 2004). This

competition and consequent challenges for the financial sector are a feature of the last

decade.

High-speed communication networks and Internet technology have created an

opportunity for a new online trading platform that transforms the century-old

brokerage industry. This new platform could be considered as one of the most

successful innovative tools in the financial industry (Gharavi, Love & Cheng 2004)

(DeForge 2001). Online securities trading has the highest usage in terms of online

tools in comparison to other online products like online banking. In the simplest

terms, online securities trading is the buying and selling of financial products like

equities or mutual funds via the Internet. Online brokerage is provided by the security

firm who provides the trading services using the Internet and web browser


- 20 -

technologies as the transaction tools (DeForge 2001) . According to a survey done by

the Thailand Securities Institute, the percentage of online trading value by the retail

investors is significant in most Asian countries (see Table 1.1).

Table 1-1 Number of trades online

Korea (2005)

Taiwan (2006)

Singapore (2004)

Hong Kong (2005)

Thailand (2006)

% Online Trading Value of Retail Trading

60.9% N/A N/A 11.5% 11.97%

% Online Trading Value of Total Market

47.3% 31.7% 10% 3.7% 6.56%

% Online Investors N/A 26.27% N/A 31.8% 23.96% # Total Investors 3,537,000 38,079,336 2,500,000 1,631,000 478,585

Source: Adapted from Thailand Securities Institute (TSI 2006b)

According to Parthasarathy and Bhattacherjee (1998), it is important to investigate

the post-adoption behaviour of the users, especially in the context of online services

(Parthasarathy & Bhattacherjee 1998). The factors influencing the post-adoption

usage behaviours of the adopters will determine whether they continue to use the

product or service (Parthasarathy & Bhattacherjee 1998). According to an

investigation done by Glaser, retail investors tend to trade frequently using online

securities trading as compared to retail investors who have not adopted the new

innovations (Glaser 2003a).

The above literature triggered the interest of this research, that is, to explore the pre-

adoption factors influencing the post-adoption usage behaviour of online securities

trading by the retail investors in Singapore. The theoretical foundation of the research

is based on the Diffusion of Innovations model pioneered by Everett M Rogers


- 21 -

(Rogers 2003) and the Technology Acceptance Model established by Davis (Davis, F.

D. 1989). Consumer loyalty factors have also been introduced in the theoretical

model in this research.

Singapore Stock Exchange and Online Securities Trading

This section describes the mechanism of the stock exchange and online securities

trading that are the objects of study for this research. This is to facilitate

understanding of the financial terminologies described in this research.

Stock Exchange is defined as an organised financial market for buying and selling

financial instruments, including stocks, options, and futures (MAS 2001). Most stock

exchanges have specific locations where commissioned, or paid, intermediaries

called brokers conduct trading - that is, buying and selling. Stocks are not always

traded on a stock exchange. Some are traded ‘over the counter’, without a specific

central trading location.

Stocks are shares of ownership in companies. People who buy a company’s stock are

entitled to dividends, or shares of any profits. A company can list its stock on a stock

exchange for trading. Stock brokers are the persons or firms that are registered with

the stock exchange in which they are allowed to trade. Stock brokers earn

commission from the customers who utilise their services to trade stock.

Singapore Exchange is Asia-Pacific's first de-mutualised and integrated securities

and derivatives exchange and was inaugurated on 1 December 1999, following the

merger of two established and well-respected financial institutions - the Stock

Exchange of Singapore (SES) and the Singapore International Monetary Exchange


- 22 -

(SIMEX).

On 23 November 2000, SGX became the first exchange in the Asia-Pacific to be

listed via a public offer and a private placement. The Straits Times Index is a

component of benchmark indices of Singapore stock listed in the exchange. There are

about thirty securities trading brokerage firms in Singapore and their primary role is

to provide a mechanism for stock investors to purchase or sell stocks at the stock

exchange. The brokers will charge a certain commission to the investors for the

transaction service (Teo, Tan & Peck 2004). However, with advances in Internet

technology, securities firms have gradually adopted online securities trading to cut

down manpower resources and operation costs. Almost all the securities firms are

now equipped with online securities trading capability. The traditional brokers have

shifted their job function as purely a middleman to new roles like financial planner or

financial advisor (Smart-Investor 2000).

1.3 Research problem, research issues and contributions

1.3.1 Research problem

The main objective of this research entitled “Factors influencing the post-adoption

consequences of online securities trading in Singapore’s retail investors”, is to study

the underlying pre-adoption factors, perceived usefulness and consumer loyalty

affecting the post-adoption usage behaviours of the online retail investors.

This research will, therefore, explore the individual variables in the model and

identify how a clear understanding of the interactions among these variables can

contribute towards the knowledge of diffusion research, especially in the area of


- 23 -

consequences of innovations. In this research context, the consequences of

innovations are the post-adoption usage behaviours of the retail investors in

Singapore. The research focuses on the study of factors influencing the post-adoption

usage behaviours of the retails investors after adopting online securities trading. The

theoretical model will be tested to answer the following research problem:

1.3.2 Research issues

To address the research problem, three groups of research issues are developed based

on the literature review of past research studies.



RI2) Does Perceived usefulness affect the post-adoption usage behaviour of retail




The research issues can be further elaborated in the form of seven research

propositions as represented in Table 1.2. These were developed to identify the

relationship of Rogers’ Diffusion attributes; Davis’ Perceived usefulness, and

Consumer loyalty with the dependent variable; the post-adoption usage behaviour of

the retail investors using online securities trading.

Research propositions one to five cover the components of the Diffusion Model

proposed by Rogers while proposition six addresses the relationship between


- 24 -

Perceived usefulness and post-adoption usage behaviour. In addition, research

proposition seven, based on Consumer loyalty, has been included in the research

model.

Table 1-2 The seven research propositions

RP Research Propositions

RP1 A relationship exists between the perceived attributes of

innovations and post-adoption usage behaviour.

RP2 A relationship exists between the type of innovation decision and

post-adoption usage behaviour.

RP3 A relationship exists between the communication channels and


RP4 A relationship exists between the nature of the social system and


RP5 A relationship exists between the change agent’s promotion efforts

and post-adoption usage behaviour.

RP6 A relationship exists between perceived usefulness and post-

adoption usage behaviour.

RP7 A relationship exists between consumer loyalty and post-adoption

usage behaviour.


These research propositions will be further justified and elaborated in Chapter Two,

the literature review of this research study.


- 25 -

1.3.3 Research contributions

There are four potential research contributions identified in this research study.

Firstly, the research adds to the literature on consequences of innovations. As stated

by Rogers, there is limited study by researchers in this area (Rogers 2003).

Secondly, the research strengthens confirmation of the correlation between Perceived

usefulness and post-adoption usage. Several researchers have conducted research

exploring the influence of Perceived usefulness, a factor of the Technology

Acceptance Model (Davis, F. D. 1989), on post-adoption usage behaviour (Naidoo &

Leonard 2007) (Parthasarathy & Bhattacherjee 1998) (Kurnia & Chien 2003).

Thirdly, the research examines the influence of Consumer loyalty factors (loyalty and

confidence) and aims to confirm whether these affect the post-adoption usage

behaviour and thus contribute knowledge to the Consumer loyalty study.

A fourth contribution the research provides is a better understanding of the post-

adoption usage behaviour of retail investors who use online trading technology. This

understanding is potentially useful to service providers and may have implications for

other similar technology innovations.

1.4 Justification for the research

1.4.1 Limited research in consequences of innovations

This research will contribute to the body of knowledge in the area of Diffusion of

Innovations and online securities trading. A diffusion model is the theoretical

framework, and the study aims to identify the interrelationships between the pre-


- 26 -

adoption factors, or the diffusion attributes of innovations, and post-adoption usage

behaviour of those who take up the innovation.

From the literature review, it seems that there are limited diffusion researches in the

field of online securities trading especially the consequences or post-adoption usage

behaviours of the retail investors trading stock online. There are very few academic

studies of online securities trading.

1.4.2 Importance of online securities trading

It is important to investigate the post-adoption usage behaviour of online securities

trading used by retail investors as it will contribute to an understanding by the

financial industry as they introduce more advanced technologies in the future. The

securities firms might make use of this framework to predict the post-adoption usage

behaviour of retail investors. Thus, the brokerage firm would be able to formulate

new a product development strategy and provide a chance to enhance the usage of

online securities trading.

1.4.3 Online survey method

This research makes use of an online questionnaire to collect information from the

retail investors in Singapore. This approach of survey using an online questionnaire is

limited in post-graduate research. The more common survey methods are mail

questionnaire, face-to-face interview and case study. This research may be a

reference to other researchers who would like to adopt an online survey method as an

alternative survey method.


- 27 -

1.4.4 Potential usage of research findings

The research findings describe the pre-adoption factors affecting the post-adoption

usage behaviour of online securities trading by the retail investors. These are of

practical use to the financial industry, especially the brokerage firms, in

understanding the usage of online securities trading. In addition, the research findings

can be a basis for other researchers who are keen to investigate the pre-adoption

factors’ influence on post-adoption usage behaviour in other products and services.

1.5 Methodology

A questionnaire was developed based on a theoretical model. This questionnaire was

created and posted on a web server of a third party web hosting vendor. Retail

investors trading stock online from the selected securities firms were invited to

respond to the online survey. Participants were requested to fill in their answers

online and the data was stored on the hosting server of the online questionnaire. The

data was subsequently downloaded for further investigation and analysis.

The online survey method was selected based on the assumption that this group of

retail investors should be using Internet surfing frequently and would be comfortable

in responding to the questions online since they are already trading stock online.

Cross tabulation is used to examine and describe the profiles of the data collected.

Data is further investigated and analysed using factor analysis and to filter out the

valid and reliable construct variables. These are described in further detail in Chapter

Four. Finally, the research used multiple linear regression analysis to examine the


- 28 -

data against the hypotheses proposed by the theoretical model. All statistical analysis

and output presentation is conducted using SPSS statistical package.

1.6 Definitions

Key words and terminologies used in this research are defined below.

Diffusion of Innovations

Diffusion research focuses upon how new technologies and concepts spread

throughout social systems (Rogers 2003). From its beginnings in sociology research

in the early 1900s, Diffusion of Innovation studies have been conducted in disciplines

as diverse as anthropology, education, and geography. Communication has become

the second largest area of investigation (Rogers 2003). The Diffusion of Innovation

Model has most recently become an important and frequency used analytical

framework for studies that focus on information technologies, especially on computer

and communication impacts on the society.

Relative Advantage

This attribute is the degree to which an innovation is perceived as better than the idea

it supersedes. The relative advantage of an innovation, as perceived by members of a

social system, is positively related to its rate of adoption (Rogers 2003).

Compatibility

This attribute is the degree to which an innovation is perceived as consistent with the

existing values, past experiences, and needs of potential adopters. The perceived


- 29 -

compatibility of an innovation is positively related to its rate of adoption (Rogers

2003).

Complexity

Complexity is the degree to which an innovation is perceived as relatively difficult to

understand and to use. The perceived complexity of an innovation is negatively

related to its rate of adoption. One of the obstacles of promoting online trading is the

complexity in terms of technical implementation of the E-Commerce platform. It

involves bandwidth issues, security issues, and software design issues etc. that hinder

the adoption of this innovation.

Trialability

This is the degree to which an innovation may be experimented with on a limited

basis. The perceived trialability of an innovation is positively related to its rate of

adoption (Rogers 2003).

Observability

This attribute is the degree to which the results of an innovation are visible to others.

The perceived observability of an innovation is positively related to its rate of

adoption (Rogers 2003).

Type of Innovation-Decisions

Rogers (2003) stated that the innovation-decision process is the process through

which an individual passes from gaining initial knowledge of an innovation, to


- 30 -

forming an attitude toward the innovation, to making a decision to adopt or reject, to

implementation of the new idea, and to confirmation of this decision (Rogers 2003,

p.168).

Rogers (2003) classifies three main types of innovation-decisions. These represent

the various ways decisions are accepted by the individual to adopt an innovation

(Rogers 2003).

Optional innovation decisions, the choices made by individuals are independent of

others;

Collective innovation decisions, the choices are decided by consensus and then

adopted by all in that consensus;

Authority innovation decisions, the choices are made by a few with power and are

then adopted by the whole (usually by some form of directive).

Communication Channels

Rogers (2003) defined communication as “a process in which participants create and

share information with one another in order to reach a mutual understanding”(Rogers

2003, p.5). This communication occurs through channels between sources. Rogers

(2003) states that “a source is an individual or an institution that originates a message.

A channel is the means by which a message gets from the source to the receiver”

(Rogers 2003, p.204). Rogers (2003) states that diffusion is a specific kind of


- 31 -

communication and includes these communication elements: an innovation, two

individuals or other units of adoption, and a communication channel. Mass media and

interpersonal communication are two communication channels.

Nature of Social Systems

Rogers (2003) defined the social system as “a set of interrelated units engaged in

joint problem solving to accomplish a common goal” (Rogers 2003, p.23). Rogers

(2003) claimed that since diffusion of innovations takes place in the social system, it

is influenced by the social structure of the social system.

Change Agent’s Promotion Effects

As stated by Rogers (2003), a change agent is an individual who influences clients’

innovation-decisions in a direction deemed desirable by a change agency. The rate of

adoption of innovations is influenced by the extent of change agents’ efforts in

diffusing the innovation (Rogers 2003, p.400).

Perceived Usefulness

According to Davis (1989), perceived usefulness is defined as the probability of the

user’s belief that the adoption of the technology will enhance his performance in the

organisational context (Davis, F. D. 1989).

Consumer Loyalty

Oliver (1999) defined consumer loyalty as “a deeply held commitment to rebuy or

repatronise a preferred product or service consistently in the future, causing


- 32 -

repetitive same brand of same brand-set purchasing, despite situational influences

and marketing efforts” (Oliver 1999, p.34).

Consequences of Innovations

Rogers (2003) defined consequences of innovations as “the changes that occur to an

individual or to a social system as a result of the adoption or rejection of an

innovation” (Rogers 2003,p.470). In this research context, it is referring to the post-

adoption usage behaviour of the retail investors trading stock online.

Online Securities Trading

DeForge (2001) defined online securities trading as the technology of buying and

selling stocks or securities over the Internet (DeForge 2001). It is also known as

Internet trading, web trading or electronic trading. Online securities trading has

significantly changed the financial service industry as the business operating model

and commission of transactions have all been modified. Online securities trading has

more than changed the mode of trading from personal communication with a broker

to Internet based transactions. It has changed the behaviour of the retail investors in

trading stocks and their decision processes relating to investment. The process of

online securities trading is summarised in Figure 1.2 below.


- 33 -

Figure 1-2 Online securities trading process

Source: adapted from (DeForge 2001)

1.7 Delimitations of scope and key assumptions

The scope of the research study is restricted to online securities trading and will not

cover other online financial services like online banking or securities trading via

mobile phone. The geographical scope is limited to Singapore. The environment does

not cover online securities trading in other countries. The sampling population is the

retail investors in Singapore and not other investors from different countries.

The methodology limitation is that the survey was conducted online and there is

difficulty to prove the validity of the participants. The validity of the responses was

controlled by distributing the online questionnaire website via selected securities

firms in Singapore and a selected newsgroup related to financial investment in

Singapore.

The major assumption of the research is that the participants who responded to the

online questionnaire were willing to provide insight into their opinion of online

Retail Investor

Online Securities Trading Platform

Brokerage firm Stock Exchange


- 34 -

securities trading. It was assumed that there would be sufficient participants for valid

data analysis from the online questionnaire. It was assumed the retail investors have

the requisite knowledge to go online to the survey website to participate in answering

the questionnaire.

1.8 Outline of the thesis

The outline of this thesis is based on Perry’s structured approach to presenting theses

(Perry 2002). It is structured into five chapters (see Figure 1.3).

The study consists of five chapters and is completed with appendices and references.

A brief summary of the chapters’ contents follows:

Chapter One: Introduction

This chapter is an introduction and overview of the research, covering the research

problem, issues, contributions and justification. It briefly outlines the methodology,

and states delimitations of scope and key assumptions of the study.

Chapter Two: Literature Review

Here the researcher provides an in-depth study of the background to the Diffusion of

Innovations Model and online securities trading which are related to this study. It

explores the research studies conducted within the domain of study. Hypotheses are

justified and derived based on the research problem and issues identified in Chapter

One and further explained in Chapter Two.


- 35 -

Chapter Three: Methodology

Chapter Three explains the research methodology adopted. It includes the research

approach, and the methods of data collection based on the research issues and

hypotheses identified. The method to test the data validity and reliability is discussed.

Finally, this chapter explains the processes used to test hypotheses.

Chapter Four: Data Analysis

In this chapter, details of data analysis are provided with results of statistical analysis.

The statistical processes are explained and the statistical output is interpreted. The

chapter concludes with a summary of findings and a revised theoretical model of the

variables studied in the research.

Chapter Five: Conclusion and Implications

The last chapter summarises the research and its findings and provides a direction for

future research. The limitation and implications of the research are presented.

Appendices and References

The ethics approval form, survey questionnaire, introduction letter to the securities

brokerage firms and selected data analysis outputs from SPSS are attached.

The bibliography and references for the research are listed in this section.


- 36 -

Figure 1-3 Outline of this thesis



- 37 -

1.9 Summary of the chapter

Chapter One outlined the foundation and objectives for this research. It introduced

the research problem, research issues, methodology and potential contributions to the

body of knowledge. This research was justified and the significance of the study was

pointed out. Overall thesis content was outlined and definitions were presented. The

scope and limitation of the study were highlighted. Chapter One provides a holistic

picture of the research before moving on to the literature review chapter that

elaborates further on theoretical models and the formulation of research propositions

which will form the hypotheses of this research study.


- 38 -

Chapter 2 Literature Review

2.1 Introduction to Chapter Two

In Chapter One, the scope of the research, questions to be explored and outline of the

thesis were discussed. The research problem and research issues were introduced

with potential contributions of the research and its significance.

In this chapter, contemporary literature on consumer behaviour, innovation adoption

and diffusion processes are reviewed. The research focuses on the adoption and

diffusion process and consequential behaviours of individual investor users of online

securities trading. Section 2.2 provides an introduction to online securities trading.

This is followed by a discussion of consumer behaviour. There are many different

areas of research within the broad consumer behaviour discipline. This study focuses

on one aspect of consumer behaviour, namely, ‘Diffusion of Innovations’. Diffusion

of Innovations is one of the major chapters found in many consumer behaviour books

(Neal, Del & Haskins 2004; Wells & Presky 1996) (Schiffman et al. 2005) (Solomon

1999, p.281). The history of diffusion research and models of diffusion of

innovations among consumers are reviewed in section 2.4. ‘Consequences of

Innovations’, the final and arguably most important stage of Diffusion of Innovation,

which is also the main focus of this study, is reviewed in this section.

Section 2.5 is devoted to the development of a theoretical model for use in this study.

The last section summarises the chapter.


- 39 -

Figure 2-1 Overview of Chapter Two



- 40 -

2.2 Stock Exchange and Online Securities Trading

As outlined in Chapter One, this research is focused on online securities trading as an

innovation taken up by retail investors, who are envisaged as consumers of the

technology and services, and their consequential trading behaviours. This section

describes the terminologies in the securities industry, and in particular the financial

tools of online securities trading.

a. Securities

According to the Securities and Futures Act issued by the Monetary Authority of

Singapore, the central bank, securities means stock issued or proposed by a

government or a commercial corporation (MAS 2001). There are many other

financial instruments like derivatives, options, unit trusts and contracts for difference

that could be referred to as securities in general. However, most commonly, securities

traded by retail investors are in the form of stock, which is a certificate representing

ownership of one or more shares of a corporation’s equity. (MAS 2001)

b. Stock Exchange

Stock exchange means an approved exchange in respect of the operation of its

securities market (MAS 2001). The term stock exchange and securities exchange are

used interchangeably.

The stock exchange is a market which brings together people who want to buy or sell

shares in a company.


- 41 -

The laws of supply and demand determine the prices buyers and sellers settle on.

The companies that use the service provided by the stock exchange to sell their

shares are referred to as listed companies. The stock exchange provides a common

platform for retail investors to buy or sell their stock of a listed company via a

brokerage firm.

The stock exchange also provides a market for a growing number of secondary

products derived from shares like stock indexes, options and contracts for difference.

c. The roles of the Stock Exchange

The stock exchange is classified as an organised institution for the sale and purchase

of shares where prices are determined by the forces of supply and demand in the

market (Teweles & Bradley 1998). The market works more efficiently when an

exchange allows for ready communication between many buyers and sellers of stocks

offered by the companies.

d. Stock broker

A broker is an individual or institution that, acting as an agent, brings buyers and

sellers together enabling them to enter into contracts to which the broker is not a

principal. Brokers charge a commission for this service.


- 42 -

e. Securities Trading Process

The securities trading process involves four steps: (i) the trading order by the

investor; (ii) the routing of this order for execution; (iii) the establishment of the

price; and finally, (iv) the execution of the order. (Sharma, M. K. & Bingi 2000)

In the traditional model of securities trading, the retail investor calls a broker of a

brokerage firm by telephone. The broker receives the order and transfers it to the

brokerage society. The society then routes the order again to the agent who will try to

find the best offer for the execution of the order. This is elaborated in Figure 2.2

(Sharma, M. K. & Bingi 2000).

Figure 2-2 Traditional trading process flowchart

f. Online Securities Trading

In conventional broking, retail investors usually make a phone call to their broker in a

securities brokerage firm to place an order to buy or sell shares.

Source: (Sharma, M. K. & Bingi 2000)


- 43 -

In contrast, online securities trading is considered as a new innovation in the financial

industry which allows retail investors to buy and sell securities via the Internet, using

a browser to navigate the brokerage firm’s website (DeForge 2001). The retail

investor who owns a user account with the online securities trading provider will log

in to the system via the Internet and place a trading order directly through his or her

computer. The system will route the order to the stock exchange for matching,

execution and settlement (Sharma, M. K. & Bingi 2000). The online securities

trading flow is presented in Figure 2.3.

The brokerage staff are not involved in the online securities trading process but in

some cases they provide research and advisory services to the retail investors for

additional fees.

Figure 2-3 Online securities trading process flowchart

Online securities trading is leading a revolution of the financial industry, utilising

online technology to provide a more efficient and effective trading tool. Buying and

selling stocks on the Internet is one huge success story in the stock-broking industry

(Shankar 2002). The success of online securities trading can be supported by the

survey conducted by Thailand Securities Institutes (see Table 2.1) on the online

Source: (Sharma, M. K. & Bingi 2000)


- 44 -

securities trading in Asia (TSI 2006a). The percentage of online trading value out of

retail trading in Korea is 60.9 percent. In Taiwan, the percentage of online trading

value of the total market is 31.7 percent. The total value of stock trading done via

online securities trading in Asia is substantial.

Table 2-1 Online stock trading in Asia

Korea (2005)

Taiwan (2006)

Singapore (2004)

Hong Kong (2005)

Thailand (2006)

% Online Trading Value of Retail Trading

60.9% N/A N/A 11.5% 11.97%

% Online Trading Value of Total Market

47.3% 31.7% 10% 3.7% 6.56%

% Online Investors N/A 26.27% N/A 31.8% 23.96% # Total Investors 3,537,000 38,079,336 2,500,000 1,631,000 478,585

Following are the advantages and disadvantages of using online securities trading as

compared to traditional broking:

Cost Advantage:

It is generally far cheaper for investors to trade stock online as compared to using a

traditional service from a broker. For conventional broking, the brokerage firm needs

to pay the brokers’ wages and have a physical location for the broker to operate from.

When the same trade is done online, the transaction is handled by a computer system.

Most brokerage firms reduce the price they charge for execution of trades done by

investors online. Using the commission charges from Phillip Securities in Singapore

as a reference, an investor pays 0.50 percent of the trade if the transaction is done by

phone through a broker. However, the commission charge is only 0.28 percent of the

trade if the investor trades online (POEMS 2008).

Source: (TSI 2006a)


- 45 -

Faster Execution

If a retail investor has an online stock brokerage account, trading and execution is

almost immediate. The retail investor can log on to the online securities trading

website to check on the price of the shares he has decided to buy or sell. Most of the

securities trading websites provide real time prices of all the shares listed in the stock

exchange - the investor places the order for the share immediately if the price

matches the value at which he will like to buy or sell.

There is no need to phone a broker to check for the prices of the shares one at a time

and place the order for the share the investor has decided upon via the broker. The

time taken for the investor to decide on placing an order will be shorter through

online securities trading. In addition, online broking accounts enable the user to

access their transaction status immediately online.

All shares purchased are due for delivery on transaction plus 3 market days. In the

terminology of the securities industry, it is known as (T+3). For example, if the

investor buys a share on Monday, then the due date for the contract is on Thursday.

In traditional broking via the broker, you have to wait for your phone call to get

through to your broker in order for him to place the order for you using the brokerage

firm’s system. The settlement would still fall within T+3.

Dependency on Technology

In order to trade online, the investor must be equipped with a computer and have a

connection to the Internet. The investor should also possess certain competencies in


- 46 -

information technology such as how to get online and how to use an Internet browser

to connect to the online securities trading site to buy and sell shares. There is also a

security risk of leaking confidential information if the investor’s computer is not well

protected by appropriate software to avoid attack by a computer virus or malicious

computer hackers.

Most online brokerage firms will still have brokers taking orders by phone from

retail investors who are not comfortable in going online, and also to act as a back-up

in case the online securities trading services provided by the brokerage firm are down.

Professional Research and Advisory Services

Without professional advice from a broker, the investor has to depend on his own

resources to conduct research on investment strategy. The investor may not have

accurate or updated financial industry information as compared to the broker and

does not have the benefit of a ‘last minute’ conversation with brokerage staff. Many

traditional brokers have changed their roles to emphasise advisory services to provide

financial research and advice to the clients in addition to the online securities trading

service that the retail investors are using. In some brokerage firms, the brokers will

conduct regular training courses for retail investors as is the case for Phillip

Securities in Singapore.

2.2.1 Retail Stock Investors

An investor is a party who invests in assets in order to produce income or capital

appreciation, or both (MAS 2001).

In the context of this research, the investors referred to are the retail investors who


- 47 -

are utilising online securities trading as a trading tool to buy or sell securities via the

Internet.

According to Rogers (2003), future investigations need to focus on the effects of

adopting innovations. In the past diffusion research, focus has mainly been on

innovativeness and the research has often stopped with analysis of the decision to

adopt new ideas (Rogers 2003). It is the interest of this research to focus on the

consequences for retail stock investors of adopting online securities trading. For

example, how will online securities trading affect the trading habits and frequency of

trading by the retail investors?

In a similar context, Rigopoulos has conducted research on the consequential usage

of online electronic payments after being adopted by the retail customers (Rigopoulos

& Askounis 2007). McKechnie argues that there is strong support for the case that

those who have already purchased goods and services over the Internet are also more

likely to purchase a larger variety of financial services online (McKechnie,

Winklhofer & Ennew 2006).

2.3 Consumer Behaviour

2.3.1 History of Consumer Behaviour

The primary discipline of this thesis rests on Consumer Behaviour. Selling and

buying of products and services are common occurrences in our daily life style.

Marketers find ways to identify products and services, set prices, promote the

products or services and place them so that they are available to the consumers. This

is the ‘Four P’ concept of the marketing mix (Kotler 2000); Product, Promotion,

Price and Place. In this research, the product is the online trading service provided by


- 48 -

a broker to facilitate online securities trading and the consumers are the retail stock

investors. As a consumer, it is common behaviour to decide to adopt certain new

products or services. There are changes in the usage behaviour of the consumers after

they adopt new products or services. Cheung and Chan (2005) conducted an

empirical research on online consumers and reported there are many factors like

Perceived usefulness, Perceived ease of use and demographics that affect consumers’

adoption intent and continuance of the products (Cheung, Chan & Limayem 2005).

The main objective is to study the correlation between the pre-adoption factors,

perceived usefulness, and consumer loyalty and the consequences of the consumers’

behaviours after the retail investors have adopted the online securities trading.

Antonides and Raaij (1998) stated that we are all consumers of certain products or

services. The attitudes and usage patterns of behaviour that we exhibit as we adopt

products will then generally be known as consumer behaviour, a field of study which

covers the mental and physical acts of the consumers, including the motives to buy

and the consequences of buying (Antonides & Raaij 1998). In this research,

consumer behaviour refers to the behaviours of the retail investors respond to when

choosing to adopt online securities trading, as well as the post-adoption usage

behaviour.

Consumer Behaviour is concerned about the motives for, and causes of consumers’

actions. To understand motivation is to understand why consumers do what they do.

Consumer Behaviour studies what motivates or causes consumers to purchase certain

goods or services (Solomon 1999).


- 49 -

Park (1998) argues that consumer behaviour research into post-purchase behaviour

was more concerned with psychological aspects (such as consumer satisfaction and

dissatisfaction) and has paid less attention to behavioural aspects (such as level of use

or quality of use) of usage behaviour (Park 1998).

Schiffman and Bednall (2005) stated that consumer behaviour is not just about

making the decision or the act of purchasing, it also includes a full range of

experiences associated with using or consuming the product or service (Schiffman et

al. 2005).

Antonides and Raaij (1998) mentioned that the ultimate goal of acquiring or using

goods and services is consumer satisfaction and well-being. Consumers must feel

happy in adopting the goods(Antonides & Raaij 1998).

It is the objective of this research to contribute to the knowledge by studying

consequential behaviours of retail investors after adopting online securities trading,

such as trading volume and trading frequency.

Antonides and Raaij (1998) indicated that Consumer Behaviour is not only a study of

the individual consumer but the group of consumers as well. A household or a

community may have a similar buying behaviour (Antonides & Raaij 1998).

In the case of online securities trading, the retail investors who trade online might

have a different behaviour compared to those investors who still prefer traditional

broking via the brokers. Glaser (2003) conducted a study of over 3,000 online retail

investors and found that this group of investors trade frequently compared to retail

investors using traditional broking. Glaser found that online retail investors are


- 50 -

mostly younger, below 40, and the vast majority of the investors are male, as

compared to traditional retail investors (Glaser 2003b) .

According to Rogers (2003), consequences of innovations are difficult to measure.

The consumers of new products or services are often not fully aware of all of the

consequences of the adoption (Rogers 2003). Due to the time constraint of this

research, it is not possible to study the differences in the trading behaviours of retail

investors before and after they adopt online securities trading. The research studies

the correlations of pre-adoption attributes and the post-adoption trading behaviours of

retail investors using online securities trading.

In the following section, the history of literature on Consumer Behaviour is briefly

summarised.

The Pre-Scientific Approach

Early studies of consumer behaviour were based on observation and were studied

from the point of philosophical and socio-critical views. According to Antonides and

Raaij (1998), one of the representatives of this view was Thorsein Velben in his

study of “The Theory of the Leisure Class” in 1899 (Antonides & Raaij 1998).

In Velben’s work, consumption is an expression of power and status. During this

period, another important person who influenced consumer behaviour theory was

Gabriel Tarde, a French philosopher. He is considered as the father of economic

psychology and has written several books such as the “The Law of Imitation” and

“On Communication and Social Influence” which explain the role of imitation and

consumer behaviour, especially in the upper classes at that time (Antonides & Raaij

1998).


- 51 -

The Motivation Approach

Further advances in Consumer Behaviour include motivation research in the period

of 1940 to 1964. According to Antonides and Raaij, Ernst Dichter in 1964 wrote a

book named “Handbook of Consumer Motivations” based on Freud’s psychoanalytic

theory. Dichter used in-depth interviews with consumers to reveal their deeper, often

unconscious motives in purchasing and using goods and services (Antonides & Raaij

1998).

The Single-Concept Approach

Antonides and Raaij (1998) described the Single-Concept approach, which emerged

in the 1960s, and has included the study of personality, perceived risk and cognitive

dissonance in consumer behaviour. One example of the study of personality is to

investigate the differences between the owners of BMWs, Citroens and Fiats.

The concept of perceived risk is relevant in consumer decisions associated with high

costs, physical danger or criticism from other people. Cognitive dissonance is the

study of knowledge that is inconsistent with behaviour. One of the interesting

examples is smoking while knowing that it is harmful to one’s health (Antonides &

Raaij 1998).

The Grand Theories

According to Antonides and Raaij (1998), in the period of 1966 to 1972, consumer

behaviour researchers attempted to integrate all existing knowledge into a larger

picture that is known as the ‘Grand Theories’. Antonides and Raaij mentioned that

during that period, Andreasen developed a general model of consumer choice


- 52 -

behaviour. Antionides and Raaij (1998) stated that Andreasen included three major

factors: opinions, affect and attitudes in consumer behaviour. According to

Antionides and Raaij, another researcher, Nicosia, in 1966 derived another model and

included four factors: mass communication, search behaviour, choice and

consumption, including feedback to preceding areas. Around this time frame, a

number of textbooks on consumer behaviour were published. Most of them are

concerned with the consumer’s preference and decision processes (Antonides &

Raaij 1998).

The Information-Processing Approach

Antonides and Raaij (1998) considered the 1970s as the years of consumer

information processing research. There are many experiments investigating how

much information consumers use and how they use it to arrive at a decision.

Antionides and Raaij indicated that the key researcher is Bettman, who in 1979

published a research work on this aspect of information processing and consumer

choice (Antonides & Raaij 1998).

The Affective Approach

Antionides and Raaij (1998) reported that in the 1980s, consumer behaviour

researchers re-focused their studies on affect or emotion. That consumers decide on

certain products or services, could be based purely on emotional factors (Antonides

& Raaij 1998) .

The theory of reasoned action written by Fishbein and Ajzen (1975) is the typical

study on this aspect. They assert that one’s intention has an influence over the

person’s behaviour (Fishbein & Ajzen 1975).


- 53 -

The Experiential Approach

Antionides and Raaji (1998) assert that in recent developments, consumer behaviour

study is focusing on the experiences and emotions during the course of consumption

rather than the process of buying the products. Two major concepts have derived:

categorisation and behavioural economics.

Categorisation is closely related to mental schemata of products, acts and consumers

themselves. A mental schema is the association that a consumer has with the

purchased product or service. Behavioural economics concerns the way consumers

perceive, evaluate and process information (Antonides & Raaij 1998).

After going through the several studies of consumer behaviour literatures, it appears

that experiences and perception of the consumers could have influenced the

continued usage or re-purchase of the products or services. Thus, the pre-adoption

factors, perceived usefulness and consumer loyalty could have significant influence

to the post-adoption behaviour of the retail investors and determine the continued

usage of the product. In this research context, the consumers are the retail investors

in Singapore using online securities trading.

2.3.2 Process of Consumer Behaviour

Generally, people are involved daily in activities of deciding, obtaining, consuming,

adopting and disposing of products and services. The above activities could be

classified as consumer behaviour. The study of consumer behaviour focuses on how

individuals make decisions to spend their available resources (time, money, effort) on

consumption-related items. That includes what they buy, why they buy it, when they


- 54 -

buy it, where they buy it, how often they buy it, and how often they use it (Schiffman

et al. 2005).

2.3.3 Consumer Behaviour and Decision Making

The decision of a consumer to buy or not to buy a product or service is a complex

process that involves an economic view, passive view, cognitive view and emotional

view (Schiffman et al. 2005).

In this research context, we are looking into the ways that investors decide to use

online securities trading and the ways they use the product when compared to a

traditional brokerage service.

2.3.4 Consumer Behaviour and Adoption

According to Schiffman (2005), the framework for exploring consumer acceptance of

new products is based on the model of ‘Diffusion of Innovations’. This consists of

two closely related but different processes: the adoption process and the diffusion

process (Schiffman et al. 2005). Diffusion is concerned with how extensively and

with how much depth the innovation spreads to the community. It could be

considered a macro view of the process of spreading the innovation from the source

to the consuming public. For example, the online securities trading service offered by

a brokerage firm being promoted to the retail investors.

Adoption, in contrast, is a micro view which focuses on the stages which an

individual consumer goes through to accept or reject the new product or service. For

example, how and why the individual retail investor adopted the online securities

trading as his trading method.


- 55 -

The research focus of this thesis is on what are the consequences or change of

behaviours of the retail investors after they adopted online securities trading.

2.4 Review of Diffusion Model Literature

2.4.1 History of the Diffusion Model

Everett Rogers’s book on “Diffusion of Innovations” (2003) is one of the most cited

works within the field of diffusion theory (Dearing & Singhal 2006).

The following section, mainly based on Rogers’s work, addresses the origin of

diffusion theory and how it grew to its present recognition by commercial and

academic researchers. According to Rogers (2003), the earliest diffusion study can be

traced back to the1900s in the area of social science (Rogers 2003). A French lawyer,

Gabriel Tarde (1903), who is also a pioneer in sociology and social psychology,

initiated the study of Diffusion of Innovations. The objective of his research was “to

learn why, given one hundred different innovations conceived at the same time –

innovations in the form of words, in mythological ideas, in industrial processes, etc. –

ten will spread abroad while ninety will be forgotten”(Rogers 2003).

Rogers (2003) stated another root of diffusion research was a group of

anthropologists that emerged in England and in Germany-Austria soon after Gabriel

Tarde. They believed that the social change in one society was the result of the

introduction of innovations from another society (Rogers 2003).


- 56 -

Rogers (2003) identified an influential 1943 diffusion study undertaken by Ryan and

Gross (Rogers 2003). They studied the diffusion of hybrid corn amongst farmers in

Iowa. The study covered the four main elements of diffusion: innovation,

communication channels, time and social system. It was the study of how an Iowa

farmer’s social relationships with his neighbours influenced the individual’s decision

to adopt hybrid corn. Based on Ryan and Gross’ research and experience, many

studies and exploration on diffusion in various fields have since started (Rogers

2003). According to Rogers (2003), the diffusion perspective was introduced in

consumer behaviour literature in the mid 1960s (Rogers 2003).

The Bass model, developed by Frank Bass, is another major contribution to diffusion

theory (Bass 1969). Bass tried to associate diffusion theory with a quantitative

formula, using the derived model to forecast the rate of diffusion. He assumed

potential adopters are influenced by two means of communication, mass media or

word-of-mouth; adopters are categorised into two groups, innovators and imitators.

Bass implied that the initial purchase of a new product by the consumers has

exponential growth to a peak and then exponential decay (see Figure 2.4). The

model is widely used in product forecasting and technology forecasting (Sohn & Ahn

2003).


- 57 -

Figure 2-4 Bass's Model

During the early 1980s, expanding from the earlier works on diffusion research,

Rogers (2003) proposed six elements that formed the fundamental framework of

studying the consumer diffusion process (Rogers 2003). The six elements are:

a. the innovation

b. its diffusion process over time

c. the personal influence and opinion leadership process

d. the adoption process

e. the roles of the innovator and other adopter categories

f. the social system or market segment within which diffusion occurs.

The other diffusion researcher in recent history who contributed knowledge to the

diffusion school is T. Robertson. He introduced a “Propositional inventory for new

diffusion research” by adding the role of marketing action (change agents) and the

Source: (Bass 1969)


- 58 -

role of competitive actions. Together with H. Gatignon, they have contributed

several journal publications about the diffusion discipline. Robertson and Gatignon

argued that diffusion concepts have proven useful to the change agents in promoting

new products. They have also asserted that competitive pressure from the industry

has shown to have an influence on a company’s decision to adopt new innovations

(Gatignon & Robertson 1985; Gatignon & Robertson 1989; Robertson, Thomas S &

Gatignon 1986; Robertson, Thomas S. & Gatignon 1991).

2.4.2 Adoption Theory

There are two major processes in diffusion of innovations models. Adoption process

is one of them and diffusion being the next. The focus of the adoption process is the

stages through which an individual consumer passes while arriving at a decision to

continue or discontinue the new product (Schiffman et al. 2005).

There are four major components of adoption theory and they are discussed below.

AIDA Model

Garber and Dotson (2002) stated that the AIDA model, shown in Figure 2.5, is said

to have been originated by E.K. Strong, an American advertisement guru. AIDA is

the acronym made up of Attention, Interest, Desire and Action. The concept was first

published in 1925 in “The Psychology of Selling” (Strong 1925). The concept is still

widely acceptable in the consumer adoption area (Hanekom 2006). For example,

when the online securities trading product POEMS, by Philip Securities in Singapore,

was introduced to the financial market, it attracted the attention of the retail investors

through heavy marketing promotions (Kong 1999). The brokerage firm


- 59 -

demonstrated the features and advantages of the online securities trading services and

raised the interest of the retail investors.

The online securities trading providers convinced the retail investors that it was the

trading tool they needed. The retail investors were convinced and started to subscribe

to online securities trading offered by the brokerage firms.

Figure 2-5 AIDA Model

Attention

Desire

Interest

Action

Attention

Desire

AIDA Model

Source: (Garber & Dotson 2002)


- 60 -

Hierarchy-of-Effects Model

The Hierarchy-of-Effects model, shown in Figure 2.6, was published by Robert J.

Lavidge and Gary A. Steiner (1961) in their article “A Model for Predictive

Measurements of Advertisement Effectiveness” (Lavidge & Steiner 1961) . Lavidge

and Steiner stated that advertising may be thought of as a force which must move

consumers up a series of steps. According to this model, to get a consumer to adopt a

product, the marketer of the product has to create awareness first by advertising the

product. The consumer might have product awareness but not know much. The

marketer should educate the consumer about the product. The consumer might know

the product but may not like it. So, the marketer must create a means to persuade the

consumer to select the product. The consumer may now start to like the product but it

may not be the preferred product if there are many alternatives. The marketer has to

promote the product’s quality, value, performance and other features in order to make

the consumer feel it is the preferred product. The consumer might prefer the product

but may not develop a conviction to buy it. The marketer’s task is to get the

consumer into action to buy the product. There are many means to convince the

consumer to execute the final decision, for example, offering a premium price or

allowing a trial of the product. Finally, the consumer may purchase the product and

adopt it.


- 61 -

Figure 2-6 Hierarchy-of-Effects Model

Innovation-Decision Model

The Innovation-Decision model in Figure 2.7 was presented by Everett M Rogers in

his book “Diffusion of Innovations” (Rogers 2003).

According to Rogers (2003), it is believed that there are five stages consumers go

through when deciding on adopting a new innovation. The first stage being: the

innovation is exposed to an individual and he gains knowledge about this new

product or service. For example, the online securities trading service offered by

Phillip Securities being advertised on TV and newspaper in 1996 (Smart-Investor

Awareness

Hierarchy-of-Effects Model

Knowledge

Liking

Preference

Conviction

Purchase

Source: adapted from (Lavidge & Steiner 1961) (Kotler 2000)


- 62 -

2000) (Kong 1999) . The retail investors were exposed to the advertisement and

gained more knowledge about the online securities trading service.

Next, persuasion occurs to form a favourable or unfavourable attitude toward the

innovation by the targeted consumer. If the retail investors are interested in the online

securities trading services, they will call up the brokerage firms to ask for more

product information.

The third stage is decision, where the targeted consumer engages in the activities that

lead to adoption or rejection of the innovation. Before the decision, the retail

investors will usually have a chance to sign up for a trial account, to try out online

securities trading like POEMS offered by Phillip Securities and to decide whether to

take on the service.

The implementation phase is where the consumer puts the innovation into use after

the decision of adopting it. Once the retail investors have decided to subscribe to the

online securities trading service, they will start to trade their first stock via online

securities trading all by themselves instead of going through the trade with the broker

by phone.

Finally, in the confirmation stage, consumers show reinforcement of their decision to

adopt the innovation. They will confirm by continuing to use the adopted innovation

or reject it and look for other alternatives. When the retail investors have gained

confidence in using online securities trading, they will continue to trade shares via

this method instead of through traditional broking.


- 63 -

These are the five stages which the consumer goes through to adopt an innovation

according to Rogers (Rogers 2003). This model provides a more complete view of

how the retail investors adopt online securities trading as a new product in

comparison to the previous two models. Rogers’ model provides the pre-adoption

and post-adoption stages while the AIDA model and Hierarchy-of-Effect model stop

at the adoption decision stage.

Figure 2-7 Innovation-Decision Model

2.4.3 Diffusion Theory

The following describes the context of diffusion theory based on Rogers’ Diffusion

of Innovations Model (Rogers 2003). It covers definitions of the major terminologies

used in the diffusion model proposed by Rogers (2003).

Knowledge

Persuasion

Decision

Implementation

Confirmation

Innovation-Decision Model

Source: adapted from (Rogers 2003)


- 64 -

Diffusion –

Rogers (2003) defined diffusion as "the process by which an innovation is

communicated through certain channels over time among the members of a social

system" (Rogers 2003 , p. 5).

Innovation –

Rogers (2003) states that an innovation is an idea, practice, or object that is perceived

as new by an individual or other unit of adoption (Rogers 2003). In the finance

industry, new financial services like online securities trading could be classified as

innovations as compared to the traditional broking method via the broker. For the

purposes of this research, the innovation upon which attention is focused is limited to

the computer / Internet technology of an online securities trading platform and the

brokerage service that is provided to accompany it.

Rogers (2003) also noted that innovations may be accepted or rejected by the system

as a whole, either as a collective decision or influenced by some authorities. The

decisions of adoption are broadly categorised into three types:

Optional innovation decisions are where the individual could make the decision

independently of the other members in the society.

Collective innovation decisions are options to adopt or reject the new ideas, product

or services that are influenced by the members of the society in a particular system.

Authority innovation decisions are options to adopt or reject an innovation that are

decided by a powerful authority. The individual has little or no option to adopt or

reject the innovation.


- 65 -

Communication Channels –

The term Communication channels "are the means through which knowledge about

innovations are conveyed" and may influence the rate of adoption (Rogers 2003).

Rogers (2003) asserted that mass media tended to be more effective at the knowledge

stage of diffusion, while opinion leaders, via interpersonal contact, were more

effective in persuading people to adopt. Accordingly, leaving out one of the channels

or reversing their placement in the stages of diffusion usually slowed the rate of

diffusion, while appropriate placement of both channels increased it. For the

purposes of this research, communication channels were limited to public media like

newspapers and TV advertisements, investors’ interpersonal communications and the

Internet.

Time –

Time refers to the rate or speed of adoption by potential users. According to Rogers

(2003) it was represented numerically as "the steepness of the curve" (Rogers 2003).

The time-line begins after the development stage of an innovation. A curve such as in

Figure 2.8, depicts the cumulative number of adoptions at designated time-units. The

curve typically shown in most diffusion studies reviewed by Rogers (2003) was the

S-shaped or sigmoidal-distribution curve (Rogers 2003), which shows total

cumulative adoptions at each stage of a time line.


- 66 -

Figure 2-8 The diffusion process

Due to the limited research time frame, it is not possible to conduct an accurate

survey across a long time period. It is not the focus of this study to investigate the

decision factors of adopting online securities trading. The research is based on the

retail investors who have already adopted online securities trading, and research is

focused on the consequences of the retail investors adopting the online securities

trading as a new innovation.

Social System –

Rogers (2003) classifies Social System as a set of interrelated units that are engaged

in joint problem-solving to accomplish a goal (Rogers 2003). The units may consist

of individuals or groups organised according to accepted structures and norms. The

Source: adapted from (Rogers 2003)


- 67 -

two components that are included in this variable are norms and degree of network

interconnectedness.

Rogers (2003) noted that "the social or communication structure of a system

facilitates [or impedes] the diffusion of innovations in the system"(Rogers 2003,

p.25). For the purposes of this research, social system scope is confined to the retail

investors of Singapore. It can be categorised as all retail investors who owned an

account with the brokerage firms that offered online securities trading to trade shares

for Singaporeans. There were about 100,000 retail investors in Singapore who used

online securities trading at the time of survey. It would not be impossible to conduct

a survey of the whole population; hence a sampling of the total population was used

for the investigation.

Perceived Attributes of Innovations

Besides the major terms stated in the definition of Diffusion of Innovations, the

innovation attributes comprise other important terms. There are five attributes as

defined by Rogers (2003): Relative advantage, Compatibility, Complexity,

Trialability and Observability.

Relative advantage is "the degree to which an innovation is perceived as being better

than" a competing or preceding idea (Rogers 2003, p.229). It can be a comparison of

the relative advantage of online securities trading adopted by the retail investors

versus the traditional broking method via the broker.

Compatibility is the "degree to which an innovation is perceived as consistent with

existing values that are in line with the social norm, past experiences, and needs of


- 68 -

the potential adopters" (Rogers 2003, p.241). Consumers cannot deal with an

innovation except on the basis of being familiar with the previous practice.

Another indication of the compatibility of an innovation is the degree to which it

meets the needs of the consumers. The promoters of the product need to determine

the needs of their clients and then recommend a suitable product that fulfils these

needs. In the case of online securities trading, the questions arising are how will the

retail investors feel about this new tool in terms of their past experience of trading

shares via the broker through the phone, and how effectively will this new tool meet

their needs?

Complexity is defined as the "degree to which an innovation is perceived as

relatively difficult to understand and use". Innovations can be ranked on a

numerically ascending scale that designates the lowest number as the simplest and

the highest number as the most complex (Rogers 2003, p.257). Accordingly, Rogers

(2003) accrued evidence to support his logic relative to this attribute - namely, that

the more complex the innovation, as perceived by users, the less likely it was to be

adopted. A learning curve may be required for the retail investors to use online

securities trading as they need to get online to the Internet and, using the Internet

browser, connect to the online securities trading website of the brokerage firm. For

those consumers who are not familiar with computer technologies, the online

securities trading might seem to be complex.

Further experience, more trading, and more time spent will take the user further

along this learning curve.


- 69 -

Trialability is "the degree to which an innovation may be experimented with on a

limited basis". Rogers’ studies found that "The trialability of an innovation, as

perceived by members of a social system, is positively related to its rate of adoption.

Early adopters are models for later adopters" (Rogers 2003, p.258).

For most online securities trading providers, consumers are able to view a

demonstration of how to use online securities trading via the website. In Figure 2.9, it

shows the demo site of POEMS, the online securities trading tool provided by Phillip

Securities in Singapore. The demo site has a step by step guide to show the features

and functions of the tool for the potential clients.

Figure 2-9 POEMS demo

Source: access from POEMS Internet website: www.poems.com.sg


- 70 -

Phillip Securities has also conducted a POEMS stock challenge game every year

since 2003. The participants are given $80,000 virtual cash to trade any stock listed

in the stock exchange using POEMS. The simulated game allows participants to

experience online securities trading without the risk of using actual money. The

participant who gains the most money from the simulated trading game is awarded

with $1,000 in real cash (POEMS 2008).

Observability is an attribute which describes the extent to which an innovation can be

"seen" in the process of being used or tried out by others. Observing others using the

innovation not only increases "its probability of being adopted, but it also strengthens

the perceived ability to judge whether an innovation has a relative advantage over

another, whether it is compatible with existing [or previous similar] innovations, and

whether or not it is sufficiently simple to understand and implement" (Rogers 2003,

p.258).

Before many of the retail investors decided to adopt online securities trading, they

may have had opportunities to obtain experience by seeing early adopters who had

started online securities trading earlier.

Figure 2.10 summarises the independent variables affecting the rate of adoption of

innovations as stated by Rogers (2003).


- 71 -

Figure 2-10 Variables determining the rate of adoption

of Innovations

Source: (Rogers 2003, p. 222, p.222)

Adopter Categories

Another critical concept of diffusion theory is the adopter categories. The adopter

categories indicate where a consumer stands in relation to other consumers in terms

of the time taken by the consumer in adopting a new product (Schiffman et al. 2005).

Five categories are observed in the adopters.

Innovators (Venturesome)

As noted by Rogers (2003), this category of consumer is venturesome. They have a

keen interest in new ideas that differentiate them from other consumers. They play an

important role in launching a new innovative initiative. The innovators are known to


- 72 -

have the ability to understand complex technical information. They are willing to

take risks and to cope with uncertainty about innovations at the time of adoption.

They are able to accept setbacks when an innovation is not successful.

According to Rogers (2003), based on a normal frequency distribution of the five

categories of adopters, the innovators tend to make up around 2.5 percent of the total

consumer population of a general product or service as shown in figure 2.9 (Rogers

2003).

Early Adopters (Respect)

Rogers (2003) considers the early adopters to be more integrated into the local social

system; the people to check with before adopting a new idea. This category contains

a greater number of opinion leaders; they are the role models. They make up

around13.5 percent of the total consumer population (Rogers 2003).

Early Majority (Deliberate)

Rogers (2003) found that these consumers adopt new ideas just prior to the average

time; seldom hold leadership positions; and deliberate for some time before adopting.

They make up 34 percent of the total consumer population (Rogers 2003).

Late Majority (Skeptical)

Rogers (2003) indicates that these consumers adopt new ideas just after average time;

adopting may be both an economic necessity and a reaction to peer pressures; and


- 73 -

innovations are approached cautiously. They make up also another 34 percent of the

total consumer population (Rogers 2003).

Laggers (Traditional)

According to Rogers (2003), these are the last people to adopt an innovation; most

are “localite” in outlook; oriented to the past; and suspicious of the new ideas. This

category is around 16 percent of the consumer population (Rogers 2003).

The following diagram (Figure 2.11) shows a clear picture of the adopter categories

and their respective percentages in the overall category.

Figure 2-11 Adopter categorisation on the basis of innovativeness

The Change Agent

It is noted by Rogers (2003) that innovations decisions and adoption can be

influenced or, intervened in, by a change agent. A change agent is an individual who

influences the users’ innovation decisions in a direction deemed desirable by the

Source: (Rogers 2003 , p.281)


- 74 -

change agency. Rogers (2003) classifies seven roles of the change agent (Rogers

2003):

- to develop a need for change on the part of the users

- to establish an information exchange relationship

- to diagnose a problem

- to create an intent to change in the user

- to translate intentions into action

- to stabilise adoption and prevent discontinuance

- to achieve a terminal relationship with users.

In the case of online securities trading, the change agents are either the sales

representatives or the customer relations executives of the brokerage firms. They

have the responsibility to encourage users to migrate from traditional securities

trading using telephone contact with brokers to the online securities trading. In most

brokerage firms in Singapore, there are free public demonstration sessions of online

securities trading conducted by the customer relations section.

The following website shows the locations and schedules of the free demonstration

sessions conducted by Philip Securities:

http://www.poems.com.sg/Seminar/seminar.asp (POEMS-Seminars 2008) .

Change agents play an important role in influencing the adoption of online securities

trading by retail investors. The online trading service offered by Phillip Securities


- 75 -

was awarded the Broker of the Year consecutively in 1999 and 2000. The success can

be attributed to the customer relations of Phillip Securities in promoting the product

to the retail investors through the demonstration sessions (Tan, A. 2000).

Consequences of Innovations

Rogers (2003) indicated that consequences are the changes that occur for an

individual or a social system as a result of the adoption or rejection of an innovation

(Rogers 2003). According to Rogers (2003), in spite of the importance of the

consequences of innovations, they have received relatively little attention by

diffusion researchers. Rogers (2003) states that most of the data about consequences

are collected through case studies, which are rather “soft” in nature, making it

difficult to generalise about consequences of certain innovations (Rogers 2003). It

may be difficult for researchers to predict when and how the consequences will

happen. Change agents may give little attention to consequences and they assume

that the adoption of a certain innovation will have beneficial results for the adopters

(Rogers 2003). Most diffusion research stopped with an analysis of the decision by

the consumers to adopt new products or services, ignoring how this choice is

implemented and with what consequences. According to Rogers (2003), there are

three main reasons why there are few studies of consequences (Rogers 2003):

“Change agencies, which often sponsor diffusion research, overemphasise adoption

per se, tacitly assuming that the consequences of innovation-decisions will be

positive” (Rogers 2003 , p.440) .

It seems obvious that most providers of online securities trading will be keen to know

how many retail investors adopted their products and the factors influencing them to


- 76 -

do so. However, the commission charge for online trading transactions by retail

investors is much lower than the traditional brokerage commission. In the case of

Phillip Securities, the commission charge for the first S$50,000 stock transaction is

0.28 percent as compared to the 0.50 percent commission charge for a phone

transaction via a broker (POEMS 2008). To remain profitable in providing the online

securities trading service to retail investors, the provider will be interested in

understanding the consequences for retail investors after adoption of online trading.

Providers may wish to understand what factors affect retail investors’ trading volume

and frequency contra to the lower commission charges.

More trades or a higher volume of trades through online trading by retail investors

means greater revenue to the provider. As such, understanding the consequences of

adopting online securities trading by the retail investors is critical to the provider.

Consequences are difficult to measure. Individuals using an innovation are often not

fully aware of all of the consequences of their adoption. For the purposes of this

research, it is assumed that there are changes in trading habits after adopting online

securities trading and this will be the main research focus of the thesis.

Rogers (2003) describes consequences as being classified into three categories: (1)

desirable versus undesirable, (2) direct versus indirect, and (3) anticipated versus

unanticipated (Rogers 2003 , p.442). Desirable consequences are the functional

effects of an innovation for an individual or for a social system. In contrast,

undesirable consequences are the dysfunctional effects of the innovation. Many

innovations cause both positive and negative consequences for the adopters.


- 77 -

Sharma and Maleyeff (2003) found that education via Internet technology has

positive and negative consequences for students. The students could obtain access to

websites to find information and make better decisions, which is a desirable or

positive consequence of Internet education. However, the use of the Internet as an

education medium can lead to certain communication behaviours.

The students are less aware of how other people perceive them while being more

aware of themselves, which is a negative consequence according to the findings by

Sharma and Maleyeff (Sharma, P. & Maleyeff 2003).

Direct consequences are the changes to an individual or a system that occur in

immediate response to an innovation. Indirect consequences are the result of the

direct consequences of the innovation. For example, Sharma and Maleyeff (2003)

found a direct consequence of Internet education is that students can access class

notes on a website from anywhere in the world. However, indirect consequences are

the legal issues of accessing the website, and changes in social customs or behaviour

patterns when using Internet education (Sharma, P. & Maleyeff 2003).

According to Rogers (2003), anticipated consequences are changes due to an

innovation that are anticipated by the members of the system. Unanticipated

consequences are changes due to an innovation that are neither intended nor

recognised by the members of a system (Rogers 2003). It could be anticipated that

the retail investors would trade more frequently via online trading as the commission

charge is cheaper and they are able to access real time stock information and monitor

the whole stock transaction process via the web.


- 78 -

Sharma and Maleyeff discovered that an unanticipated consequences of Internet

education was the heavy Internet usage by the students (Sharma, P. & Maleyeff

2003).

The online trading system by Phillip Securities being disabled due to a power failure,

as the company did not have a backup power generator, resulted in many online retail

investors not being able to trade. This is an unanticipated consequence of using

online trading (Chua 2008).

2.4.4 Acceptance and Consequences of Innovations

Davis (1989) introduced a Technology Acceptance Model (TAM) as a model for

researching information technology adoption (Davis, F. D. 1989). In the broadest

sense, information technology refers to both the hardware and software that are used

to store, retrieve, and manipulate information. As shown in figure 2.12, Davis (1989)

proposed that the behavioural intention of adoption of a computer technology by the

users is determined by two factors: Perceived Usefulness (PU) and Perceived Ease of

Use (PEOU). Perceived usefulness is defined as the probability of the user’s belief

that the adoption of the technology will enhance his performance in the

organisational context. It is the task oriented outcome that looks at how the adoption

of a certain IT technology helps a user in completing tasks. It is similar to the

Relative Advantage factor in Roger’s diffusion model. Perceived ease of use is the

degree to which the user expects the target system to be easily useable by themselves.

It is similar to the Complexity factor in Roger’s diffusion model. Davis argues that

the perceived ease of use has a direct impact on perceived usefulness, so one factor is

dependent, to some extent, on the other as indicated in Figure 2.12 (Davis, F. D.

1989). Both factors have influences on the user’s intention of adoption to the IT


- 79 -

technology. Davis (1989) indicates that the intention to use the technology ultimately

has an influence on the actual usage of the system.

Figure 2-12 Technology Acceptance Model (TAM)

There are some academic criticisms about Davis’s Technology Acceptance Model

and in particular the work by Dillon and Morris (1996). They indicated the major

theoretical limitation of Technology Acceptance Model is the “exclusion of the

possibility of influence from institutional, social, and personal control factors’

(Dillon & Morris 1996). This criticism could be applicable to the study of the

influence of perceived usefulness to post-adoption usage behaviour. However, the

theoretical model of this research has been modified to include other social factors

(eg. nature of the social system) and external factors (Type of Innovation Decision)

to modulate the limitation in Technology Acceptance Model.

Source: (Davis, F., Bagozzi & Warshaw 1989 , p.985)


- 80 -

There are many competing theoretical models co-exist that are relevant to Davis’

Technology Acceptance Model. They are about the study of innovation acceptance

and adoption but with different focus in their models. Some of the well known

theoretical models are Theory of reasoned action (TRA), Theory of planned

behaviour (TPB), Decomposed theory of planned behaviour and Unified Theory of

Acceptance and Use of Technology (UTAUT).

Theory of reasoned action (TRA).Based on social psychology research, Fishbein &

Ajzen (1975) and Ajzen & Fishbein (1980) developed the Theory of Reasoned

Action (TRA) to predict the individual behaviour in a social context (Ajzen &

Fishbein 1980; Fishbein & Ajzen 1975). TRA is a well-known general theory to

explain behaviour beyond adoption of technology. The model asserts that specific

beliefs influence behavioural perceptions and actual behaviour. According to TRA

model, the behaviour intention is the immediate antecedent of an individual’s

execution of a piece of behaviour. Ajzen & Fishbein (1975) (1980) indicates that the

behavioural intention is a function of two determinants: a personal factor termed

‘attitude toward the behaviour’ and a person’s perception of social pressures or

‘subjective norm’ (Ajzen & Fishbein 1980; Fishbein & Ajzen 1975). Technology

Acceptance Model by Davis (1989) can be considered as an adaptation of the TRA

model and was used to explain individual user’s intentions of using a system.

Theory of planned behaviour (TPB). Theory of planned behaviour or TPB is an

extension of the theory of reasoned action (TRA) and was proposed by Ajzen (Ajzen

1985). TPB model added a third construct, “perceived behavioural control” compared

to TRA. Ajzen (1985) extended the TRA model to cover willing behaviours for


- 81 -

predicting behavioural intention and actual behaviour. TPB specifies the nature of

relationships between beliefs and attitudes. According to Ajzen (1985), people’s

evaluations of behaviour are determined by their accessible beliefs about the

behaviour. TPB model links belief, attitude and expectation of an individual. The

behavioural intention is influenced by the strength of the belief according to Ajzen

(Ajzen 1985).

Decomposed Theory of planned behaviour (DTPB). Taylor and Todd (1995) have

included additional constructs such as relative advantage, compatibility, influence of

significant others, and risk, which are derived from Rogers’ (2003) diffusion model,

and decomposing the perceptions in TPB model into a series of specific belief

dimensions. The DTPB model is considered to be more complete as it focuses on

specific factors influencing adoption and usage (Talyor & Todd 1995).

Unified Theory of Acceptance and Use of Technology (UTAUT). The UTAUT

model is a Technology Acceptance Model unified by Venkatesh and Others (2003).

The UTAUT model aims to explain the user intentions to use an information system

and the subsequent usage behaviour (Venkatesh et al. 2003). The model has four

major constructs: performance expectancy, effort expectancy, social influence, and

facilitating conditions) are direct determinants of usage intention and behaviour.

There are also other moderating variables like gender, age, experience, and

voluntariness of use that might have impact to usage intention and behaviour.


- 82 -

In recent years, there has been an increase in studies of the post-adoption behaviour

of consumers or the consequences of innovations. Parthasarathy and Bhattacherjee

(1998) examined post-adoption behaviours in relation to continued adoption versus

discontinuance within the context of online service use. The results of their study

indicate that potential discontinuers can be discriminated from continued adopters of

online service based on their sources of influence, perceived usefulness, and service

utilisation during the time of initial adoption (Parhasarathy & Bhattacherjee 1998).

Karahanna et al (1999) studied the pre-adoption and post-adoption views and

attitudes of information systems users. They argued that the behaviour of the

information systems users varies across different phases of the innovation process

and the beliefs of usefulness affect the post-adoption behaviour (Karahanna, Straub

& Chervany 1999).

Based on Davis’ Technology Acceptance Model, Kurnia and Chien (2003) studied

the acceptance of online grocery shopping and identified factors that may foster or

hinder the acceptance. Kurnia and Chien (2003) argue that the perceived usefulness

of online grocery shopping is influenced directly by its perceived ease of use. They

discovered that these factors affect the attitude of consumers towards using online

grocery shopping and in turn influence the behavioural intention and the actual usage

of online grocery shopping.

Going beyond the adoption and non-adoption research, Zhu and Kraemer (2005)

studied post-adoption usage in electronic business (e-business) by organisations in

the retail industry (Zhu & Kraemer 2005). Their findings indicate that technology


- 83 -

competence is the strongest factor that has a significant influence on the extent of e-

business use in an organisation.

Zhu and Kraemer (2006) also add an international dimension to the innovation

diffusion literature, showing that careful attention must be paid to the economic and

regulatory environment in different countries, as that will affect the acceptance of

technology. To further the investigation into the post-adoption stages in innovation

diffusion of e-business, Zhu and others (2006) studied the determinants of post-

adoption usages using relative advantage and compatibility. They conclude both

relative advantage and compatibility factors have a positive influence over e-business

usage (Zhu et al. 2006) .

Cheung, Chan and Limayem (2005) have conducted a thorough analysis of the

literature in the area of online consumer behaviour and their finding, shown in Figure

2.13, indicates that intention influences adoption of online product, which in turn

influences continuance or repurchase of an online product. However, compared to

intention and adoption, studies on continuance is an under-researched area (Cheung,

Chan & Limayem 2005).

Figure 2-13 Framework of online consumer behaviour

Intention

Adoption

Continuance

Source: adapted from (Cheung, Chan & Limayem 2005)


- 84 -

Jasperson, Carter and Zmud (2005) believe that the features of information

technology-enabled work systems in most organisations have been under-utilised by

the users (Jasperson, Carter & Zmud 2005). They study the post-adoptive behaviours

of users and analyse the factors affecting the continuance of using the functionality of

the IT-enabled work systems. They advocate that organisations should strongly

consider capturing users’ post-adoptive behaviours as well as the outcomes

associated with these behaviours over time, so as to identify factors which influence

them to continue to use the IT-enabled work systems.

McKechnie, Winklhofer and Ennew (2006) conducted a similar research based on

Davis’ Technology Acceptance Model in the content of online retailing of financial

services. They argue that Perceived usefulness is positively related to the attitude

towards using the Internet as a distribution channel for financial services. They also

find that Perceived ease of use is positively related to attitudes towards using the

Internet as a distribution channel for financial services. McKechnie and others

summarised that there is strong support for the case that consumers who have

already purchased other goods or services over the Internet are also more likely to

purchase a larger variety of financial services online (McKechnie, Winklhofer &

Ennew 2006).

Chea and Luo (2007) believe that the customer satisfaction and post-adoption

behaviours of customers using an electronic service are important factors in

determining whether customers continue to use the electronic service and recommend

use to other customers (Chea & Luo 2007). Using Davis’ Technology Acceptance

Model, they confirm Perceived usefulness has a significant influence over customer


- 85 -

satisfaction, which in turn affects the continuance of electronic service use by

consumers.

In another similar study, Rigopoulos and Askounis (2007) state that both Perceived

usefulness and Perceived ease of use have a strong positive relationship to

behavioural intention to use online electronic payments by consumers, as shown in

Figure 2.14.

They indicated behavioural intention has a strong positive influence to system usage

of electronic payments by the consumers. Finally, Rigopoulos and Askounis confirm

that both Perceived usefulness and Perceived ease of use have a strong positive

relationship to actual usage of electronic payments by consumers (Rigopoulos &

Askounis 2007).

Figure 2-14 Customer's intention to use and usage of electronic

payment service

PerceivedUsefulness

(PU)

PerceivedEase of Use

(PEOU)

Actual System UseBehavioral Intention to Use

Source: adapted from (Rigopoulos & Askounis 2007)


- 86 -

2.4.5 Consumer loyalty and Consequences of Innovations

McMullan and Gilmore (2008) recognise the importance of consumer loyalty that is

affecting many organisations in competitive industries. They have done an empirical

study on consumer loyalty literature and state that the most widely accepted

definition of loyalty is by Jacoby and Kyner (1973). They describe loyalty as the

“biased (i.e. non-random), behavioural response (i.e. purchase), expressed over time,

by some decision making unit, with respect to one or more alternative brands out of a

set of such brands, and is a function of psychological (i.e. decision making,

evaluation) processes” (McMullan & Gilmore 2008) (Jacoby & Kyner 1973).

Another well cited definition of consumer loyalty is by Oliver (1999). He defines

consumer loyalty as “a deeply held commitment to rebuy or repatronise a preferred

product or service consistently in the future, causing repetitive same brand or same

brand-set purchasing, despite situational influences and marketing efforts (Oliver

1999, p.34).

As illustrated in Figure 2.15, Donio, Massari et al. have stated trust (confidence) has

been considered as an outcome of customer satisfaction and as an antecedent of

customer commitment and customer loyalty (Donio, Massari & Passiante 2006).

They have studied and confirmed the relationship between customer satisfaction;

customer trust; customer commitment and the influence to purchase behaviour of

consumers.


- 87 -

Figure 2-15 Consumer loyalty and purchase behaviour

Customer Satisfaction

Customer Trust

Customer Commitment

Purchase Behaviour

Source: adapted from (Donio, Massari & Passiante 2006, p. 453)

Floh and Treiblmaier (2006) have conducted research on loyalty of customers using

online products and have concluded that there is a strong significant association

between trust and loyalty (Floh & Treiblmaier 2006).

In this research, the confidence and loyalty of retail investors in online securities

trading are proposed to be factors that might affect the post-adoption usage behaviour.

Thus, confidence and loyalty factors are suggested to be significant constructs that

might affect the consequences of adoption of online securities trading. Both factors

are to be grouped as the consumer loyalty construct in this research.


- 88 -

2.5 Theoretical Model

Grounded in the Diffusion of Innovation theory of Rogers (Rogers 2003) and the

Technology Acceptance Model of Davis (Davis, F. D. 1989), this research goes

beyond the adoption and non-adoption boundary to study the factors of perceived

attributes of Innovation, Perceived usefulness, Consumer loyalty and the

consequences of adopting online securities trading by retail investors. The research

focuses on pre-adoption factors that influence the post-adoption consequences or

behaviours of the retail investors such as trading patterns, trading frequency and

trading volume. The research objective and thus the proposed theoretical model is to

address the research issues identified in Chapter One:







A theoretical model is proposed with a dependent variable and a number of

independent variables.

2.5.1 The Variables

In the previous sections, literature on Consumer Behaviour, Diffusion of Innovation

and Technology Acceptance Model have been reviewed. Based on these studies, it is

anticipated that several attributes, or independent variables, might influence the


- 89 -

consequences of innovations. There are six groups of independent variables and one

dependent variable identified by the researcher, to be the theoretical framework of

this study. Each of these variables is described in the following sections.

2.5.2 Proposed Independent Variables

This section introduces the framework for the empirical research. It provides an

overview of the proposed relationship among independent variables and the

dependent variables which are the post-adoption behaviours of the retail investors

using online securities trading. The research model is based on Roger’s model on

Diffusion of Innovations (Rogers 2003) and Davis’ Technology Acceptance Model

(Davis, F. D. 1989).

The researcher proposes that selected independent pre-adoption variables from

Rogers’ Diffusion model and selected independent variables from Davis’ Technology

Acceptance Model have a positive influence on the post-adoption consequences of


Table 2.2 lists all the independent variables that are possible items for further

investigation.


- 90 -

Table 2-2 Proposed independent variables for the research

Independent

Variables

Attributes Description

Perceived

Attributes of

Innovations

(Rogers 2003)

Relative

advantage

The degree to which an

innovation is perceived as

better than what it supersedes.

Compatibility The degree to which an

innovation is perceived as

consistent with the existing

values, past experiences, and

needs of potential adopters.

Complexity The degree of difficulty

perceived to understand or use

an innovation. In this research

context, it is referring to the

lack of complexity of similar to

Perceive Ease of Use.

Trialability The degree to which an

innovation may be

experimented with on a limited

basis.


- 91 -

Observability The degree to which the results

of an innovation are visible to

others. The perceived

observability of an innovation

is positively related to its rate

of adoption.

Type of Innovation

Decision

(Rogers 2003)

Optional

The choices to adopt or reject

an innovation that are made by

an individual independent of

the decisions by other members

of a system.

Collective The choices to adopt or reject


consensus among the members

of a system.

Authority The choices to adopt or reject


a relatively few individuals in a

system who posses power,

status, or technical expertise.

Communication

Channels

(Rogers 2003)

Mass media The communication channel

that spreads information to the

general public or large number

of people.


- 92 -

Inter personal It involves a face-to-face

exchange between two or more

individuals.

Nature of Social

System

(Rogers 2003)

Norms This is about the nature of the

system and has an impact on

the rate of adoption.

Degree of

network

The complexity of the

communication network. If the

society is highly

interconnected, the

communication will be more

effective and hence affect the

rate of adoption.

Extent of Change

Agent’s Promotion

Efforts

(Rogers 2003)

A change agent is an individual

who influences a client’s

innovation-decisions. The

change agent’s effort has an

impact on the client’s adoption

intent and rate of adoption.

Perceived

usefulness

(Davis, F. D. 1989)

The term is defined as a

person’s opinion that using a

certain application system will

increase his or her work

performance.


- 93 -

Perceived Ease of

Use

(Davis, F. D. 1989)

The term is defined as the

degree to which a person

believes that using a particular

system would be free from

effort. It is similar to the level

of complexity and is to be

covered under Perceived

attributes of innovations.

Consumer Loyalty

(Donio, Massari &

Passiante 2006)

Confidence The consumers’ trust in the

product is considered as a

significant factor affecting the

purchase behaviour.

Loyalty

This study focuses on the consequences of adopting an online securities trading

service by the retail investors and not the pre-adoption decision to use the service.

It is anticipated that the independent variables may have a direct influence on the

consequences of adoption. For example, it is proposed that the perceived attributes of

online securities trading, like relative advantages of flexibility and mobility, have an

influence on the usage of trading shares online by the retail investors.

It is proposed that the perceived usefulness of online securities trading by the retail

investors will have an influence on the trading behaviours like trading frequency.

Source: adapted from (Rogers 2003) and (Davis, F. D. 1989)


- 94 -

2.5.3 Dependent Variable

The dependent variable is referring to the consequences of adopting the online

securities trading as an innovative product. It is also referring to the post-adoption

behaviours of the retail investors using online securities trading. It is proposed that

there will be significant changes in buying or selling behaviour in securities trading

using the online securities trading tool, as compared to a more traditional trading

approach used previously. The changes in the post adoption behaviour of the retail

investors could be in the following areas:

Type of shares – retail investors may trade other categories of shares, as they are

able to access more information from the online trading website about other shares

that they were not familiar with previously;

Location – retail investors may tend to trade from locations other than home, such as

a work place, Internet café or overseas while travelling, as they are able to access

shares information and an electronic trading facility wherever there is Internet

available;

Volume – retail investors may trade a different quantity (size of share trades), as it is

more convenient to access information and to trade online;

Frequency – retail investors may trade shares more frequently online, as they are no

longer dependent upon calling the broker by phone to trade.

Investment tips exchange – retail investors may increase the opportunities to share

opinions and comments with other investors using online securities trading through

email, investment newsgroups or online chat.


- 95 -

Investment information frequency – retail investors may tend to increase the

frequency of searching for financial investment information after adopting online

securities trading.

This study will contribute to knowledge in the area of Diffusion of Innovations

research, Technology Acceptance research as well as the Online Securities Trading

commercial interests.

By better understanding the factors affecting the consequences of the retail investors’

use of online securities trading, the commercial providers of online securities trading

would able to enhance their product and the process associated with its adoption and

use, to improve value for retail investors who continue to use the product.

As indicated by Rogers (2003) there is limited research on the consequences of

innovations (Rogers 2003). It is the purpose of this research to contribute to

knowledge in this area.

2.5.4 Linkages amongst the Variables

As shown in Figure 2.16, this research proposes that all of the seven groups of

independent variables have a direct or indirect influence on post-adoption usage

behaviour or on the consequences of adopting online securities trading by the retail

investors. The propositions are based on the Rogers’ diffusion model (2003), Davis’

Technology Acceptance Model (1998) and consumer factors by Donio et al (2006).

The literatures discussed in previous sections have indicated that pre-adoption factors,

Perceived usefulness and consumer factors have significant influence on post-

adoption behaviours. This forms the basis of a number of more specific propositions

of the research, which are discussed in the following section.


- 96 -

Figure 2-16 Independent variables and post-adoption usage behaviour

2.5.5 Research Model

The model has adapted the Diffusion of Innovations model by Rogers (2003), and the

Technology Acceptance model by Davis (1998). Additional consumer loyalty factors

have been included as independent variables adapted from Donio et al (2006). Figure

2.17 illustrates the structure of the relationships proposed, showing dependent and

independent variables selected for closer investigation by testing a series of

hypotheses that are converted from the propositions. The hypotheses are to be tested

empirically and elaborated in Chapter Three. Zikmund (2000) states a hypothesis is a

proposition that can be tested using empirical evidence. Thus, hypothesis is an

empirical statement concerned with the relationship among the variables to be studied

(Zikmund 2000).

Independent

Variables

Post-Adoption

Usage Behaviour



- 97 -

Figure 2-17 Conceptual theoretical model of post-adoption usage behaviour of online securities

trading

Hypothesis 7

Communication Channels- Mass Media- Interpersonal

Perceived Attributes of Innovations

- Relative Advantage- Compatibility- Complexity- Trailability- Observability

Extent of Change Agent’s Promotion Efforts

Technology Acceptance Model- Perceived Usefulness

Consumer Loyalty

- Loyalty- Confidence

Independent variables Dependent variables

Hypothesis 1

Hypothesis 2

Hypothesis 3

Hypothesis 4

Hypothesis 5

Nature of the Social System- Norms- Degree of network interconnectedness

Type of Innovation- Decision- Optional- Collective- Authority

Hypothesis 6

Post-Adoption Usage Behaviour of Online Securities Trading

- Type of shares- Location- Volume- Frequency- Investment tips exchange

- Investment information frequency

Source: Adapted from (Rogers 2003), (Davis, F. D. 1989) and (Donio, Massari &

Passiante 2006)


- 98 -

2.5.6 Research Propositions Formulation

As shown in Table 2.3, seven research propositions are outlined based on the

theoretical model discussed in the previous section. The research propositions are to

identify the research issues addressed in Chapter One:







Based on the several literatures discussed in earlier sections especially in consequence

of innovations studies, there are strong indications that pre-adoption variables have

influence to post-adoption usage behaviour. The Perceived usefulness variable in

Technology Acceptance Model by Davis (1999) provided evidence that the variable

might influence the post-adoption usage behaviour of the retail investors trading stock

online. Literatures and in particular Oliver (1999) indicated that Consumer loyalty

influences the post-adoption usage behaviour.


- 99 -

Table 2-3 Seven research propositions

RP Research Propositions

RP1 A relationship exists between the perceived attributes of

innovations and post-adoption usage behaviour.

RP2 A relationship exists between the type of innovation decision and


RP3 A relationship exists between the communication channels and


RP4 A relationship exists between the nature of the social system and


RP5 A relationship exists between the change agent’s promotion efforts

and post-adoption usage behaviour.

RP6 A relationship exists between perceived usefulness and post-


RP7 A relationship exists between consumer loyalty and post-adoption

usage behaviour.

Research Proposition 1

RP1: A relationship exists between the perceived attributes of the innovations and

post-adoption behaviour.

It is proposed that there is correlation between the attributes of innovations (Relative

advantage, Compatibility, Level of complexity, Trialability and Observability) and


- 100 -

the post-adoption behaviour (increase in trading volume; increase in trading

frequency; increase in trading locations and increase in variety of shares traded). For

example, Relative advantage may have an effect of increasing the frequency of shares

trading by the retail investors using online securities trading.


RP2: A relationship exists between the type of innovation decision and post-adoption

usage behaviour.

Here the researcher proposes there may be correlation between the type of innovation

decision (Optional decision, Collective decision and Authority decision) and post-

adoption usage behaviour. For example, the optional decision made by retail

investors to trade online or to call the broker will affect the usage of online securities

trading by the retail investors.


RP3: A relationship exists between the Communication channels and post-adoption

usage behaviour.

This proposition states that Communication channels (mass media or personal) will

influence post-adoption behaviour. For example, there may be influences from using

mass media or a personal communication approach that will affect the usage of

online securities trading by the retail investors.


- 101 -


RP4: A relationship exists between the Nature of the social system and post-adoption

usage behaviour.

It is proposed that there is correlation between the Nature of the social system (norms

and degree of network interconnectedness) and post-adoption usage behaviour. For

example, actions of peers and perceptions of normal behaviour may affect the usage

of online securities trading by the retail investors.


RP5: A relationship exists between the change agent’s promotion efforts and post-


It might be expected that there is correlation between the Change agent’s promotion

efforts and post-adoption behaviour. For example, there may be direct influences of

the Change agent’s promotion efforts on the retail investor’s usage of online

securities trading.


RP6: A relationship exists between Perceived usefulness and post-adoption usage

behaviour.

RP6 proposes that there is correlation between Perceived usefulness and post-


- 102 -

adoption behaviour. For example, the retail investor’s opinion on the usefulness of

online securities trading will affect his trading frequency.


RP7: A relationship exists between Consumer loyalty and post-adoption usage

behaviour.

Some correlation between Consumer loyalty (loyalty and confidence) and post-

adoption usage behaviour would be expected. For example, the trust in online

securities trading would affect the usage behaviours of retail users.


- 103 -


In this chapter, a theoretical framework was built based on Rogers’ Diffusion of

Innovation Model, Davis’ Technology Acceptance Model and a consumer loyalty

framework. The focus of this research study is on pre-adoption factors that might

influence the post-adoption usage behaviour of retail investors using online securities

trading. In this research context, the post-adoption usage behaviour of the retail

investors using online securities trading is the consequences of adopting the new

innovations. Rogers’ Diffusion of Innovations Model has been elaborated in the

chapter and Davis’ Technology Acceptance Model has also been described. These

models have been adapted for this study. The Consumer loyalty framework including

the loyalty and confidence factors has also been described. The five different

categories of independent variables based on Roger’s Diffusion Model: Perceived

attributes of innovations, Type of innovation decision, Communication channels, the

Nature of the social system and Extent of the change agent’s promotion effects have

been discussed. The Perceived usefulness factors derived by Davis have also been

adapted as independent variables to study their correlations with post-adoption usage

behaviours of retail investors using online securities trading. Another independent

variable group, Consumer loyalty factors, has been selected for research investigation

and has been discussed and elaborated in the chapter. The dependent variable group,

post-adoption usage behaviour, or consequences of adopting the innovations, the

main focus of study, was also elaborated. A theoretical model has been derived

showing proposed interrelationships between independent variables and dependent

variables. On the basis of the research model, seven general research propositions

have been developed that form the basis for future empirical research investigation.


- 104 -

In the following chapter, the research method is explained. A method for

investigating the research propositions is outlined. The methods for collecting data

and analysing the data are described. A number of testable hypotheses based on the

research propositions in this chapter are specified, and a method of testing these

hypotheses is explained.


- 105 -

Chapter 3 Methodology

3.1 Introduction to Chapter Three

The theoretical background of this research that is based on Roger’s Diffusion of

Innovations Model (Rogers 2003) and Davis’ Technology Acceptance Model (Davis,

F. D. 1989), together with the research model and the hypotheses built for this

research were detailed in Chapter Two. Chapter Three of the thesis focuses on

research methodology.

There are nine sections in this chapter as indicated in Figure 3.1. The first section is

the introduction of the chapter. The second section discusses several common

research paradigms and justifies the selected research paradigm for this research.

Section three identifies the qualitative research and quantitative research. It is then

followed with a justification of the selected research method and states it limitations.

Section four highlights the theoretical and accessible population. This is followed by

section five which outlines the several approaches in samplings as well as validity

and reliability of the data. Section six describes the detailed questionnaire used to test

the research model and hypotheses. In section seven, different modes of survey such

as web surveys and interviews are discussed. Section eight describes the data

processing procedures and the methods used to analyse the data. The ethical

considerations are also discussed in section eight. The last section summarises the

topics covered in this chapter.


- 106 -

Figure 3-1 Overview of Chapter Three

3.1 Introduction to Chapter Three

3.4 Survey objects 3.4.1 Theoretical population 3.4.2 Accessible population

3.5 Sampling3.5.1 Sampling design3.5.2 Sampling size3.5.3 Validity and reliability

3.6 Questionnaire design3.6.1 Questionnaire objective3.6.2 Questionnaire types3.6.3 Independent variables

3.6.4 Dependent variables and components

3.7 Mode of survey3.7.1 Web survey / email

3.8 Data processing procedures 3.8.1 Descriptive statistics – cross tabulation 3.8.2 Factor analysis 3.8.3 Regression analysis 3.8.4 Multiple regression 3.8.5 Data processing tools 3.8.6 Ethical considerations


3.2 Research Paradigms 3.2.1 Positivism 3.2.2 Critical Theory 3.2.3 Constructivism 3.2.4 Realism 3.2.5 Justification of the selected research paradigm

3.3 Research Methods 3.3.1 Qualitative Research 3.3.2 Quantitative Research 3.3.3 Justification of the selected research method 3.3.4 Limitation of the selected research method



- 107 -

3.2 Research Paradigms

As suggested by Perry (2002), it is important to understand the body of knowledge of

methodology and to choose the appropriate research paradigm which is most suitable

for the research undertaking (Perry 2002). A research paradigm is a systemic

conceptual framework which guides researchers in how to conduct appropriate

research (Guba & Lincoln 1994). There are four research paradigms: positivism,

critical theory, constructivism and realism (Perry, Riege & Brown 1999) (Guba &

Lincoln 1994). These research paradigms are reviewed in the following sections.

Guba and Lincoln (1994), define a paradigm as a set of basic beliefs that deal with

ultimates or first principles. It is a basic belief system based on ontological,

epistemological, and methodological assumptions (Guba & Lincoln 1994). Ontology

can be viewed as the reality that researchers investigate, epistemology is the

relationship between that reality and the researcher, and methodology is the

technique used by the researcher to investigate the reality (Healy & Perry 2000).

Krauss (2005) defines ontology as the philosophy of reality, epistemology addresses

how we come to know that reality, while methodology identifies the methods of

attaining the knowledge of the reality (Krauss 2005). A paradigm can also be viewed

as an overall conceptual framework within which a researcher may work (Perry,

Riege & Brown 1999). Easterby-Smith et al. have classified paradigms as positivist

paradigms and phenomenological paradigms. The basic belief of a positivist

paradigm is that the world is external and objective and the observer is independent.

The phenomenological paradigm is to believe that the world is socially constructed


- 108 -

and subjective, and that the observer is part of what is observed (Easterby-Smith,

Thorpe & Lowe 1991).

The following table based on Perry, adapted from Guba and Lincoln (1994)

summarises the four major research paradigms.

Table 3-1 Basic systems of alternative enquiry paradigms

Paradigm Item POSTIVISM CRITICAL

THEORY CONSTRUCTIVISM REALISM

ONTOLOGY naïve realism: reality is real and apprehensible

historical realism: ‘virtual’ reality shaped by social, economic, ethnic, political, cultural, and gender values, crystallised over time

critical relativism: multiple local and specific ‘constructed’ realities

critical realism: reality is ‘real’ but only imperfectly and probabilistically apprehensible and so triangulation from many sources is required to try to know it

EPISTEMOLOGY objectivist: findings true

subjectivist: value mediated findings

Subjectivist: created findings

modified objectivist: findings probably true

METHODOLOGY experiments / surveys: verification of hypotheses: chiefly quantitative methods

dialogic/dialectical: researcher is a ‘transformative intellectual’ who changes the social world within which participants live

hermeneutical/ dialectical: researcher is a ‘passionate participant’ within the world being investigated

case studies / convergent interviewing: triangulation, interpretation of research issues by qualitative and quantitative methods such as structural equation modelling

Note: Essentially, ontology is ‘reality’, epistemology is the relationship between the reality and the researcher and methodology is the technique used by the researcher to discover the reality.

Source: (Perry, Riege & Brown 1999) based on (Guba & Lincoln 1994)


- 109 -

3.2.1 Positivism

According to Easterby-Smith et al. (1991), the positivist view is that the social world

exists externally and that the facts of this social world can be discovered by a set of

scientific methods. The researcher using this world view is usually using a

quantitative method where the technique can provide wider coverage of the range of

situations, and is fast and economical (Easterby-Smith, Thorpe & Lowe 1991). The

investigator and the investigated ‘object’ are assumed to be independent entities. The

researcher is able to study the object without influencing it or being influenced by it

(Guba & Lincoln 1994). The objectives of the research based on a positivism

paradigm include the measurement and analysis of causal relationships between

variables that are consistent across time and context (Perry, Riege & Brown 1999).

The ontology perception of positivism is that the researcher discovers a single

apprehensible reality concerning a research problem based on independent

observation, and the resulting knowledge is considered to be trustworthy (Guba &

Lincoln 1994).

The epistemology of the positivist paradigm indicates that the inquiry takes place as

if through a one-way mirror according to Guba and Lincoln (1994). The researcher

prevents any influence on the outcomes by his own values and biases. The findings

should be repeatable and true in nature. The investigator and the investigated ‘object’

are independent from each other (Guba & Lincoln 1994).

The methodology commonly used in the positivism paradigm includes hypotheses in

the form of propositions which are subjected to an empirical test to verify them.


- 110 -

There are controlled conditions to prevent research outcomes being improperly

influenced (Guba & Lincoln 1994).

3.2.2 Critical Theory

The second paradigm, critical theory, is where reality is held to be based on historical

structures, and the researcher aims at criticising and transforming social, political,

cultural, economic, ethnic and gender values. Marxists, feminists and action research

all fall under this category of research paradigm (Perry, Riege & Brown 1999).

The ontology of critical theory is considered as historical realism, where the reality

was shaped over time (Guba & Lincoln 1994).

The epistemology of critical theory assumes that the investigator and the investigated

object are interactively linked, and that the values of the investigator influence the

inquiry (Guba & Lincoln 1994). Perry et al. state that this paradigm is not appropriate

for marketing research unless the researcher aims to be part of the investigation

(Perry, Riege & Brown 1999).

The methodology of critical theory requires the researcher to have dialogue with the

subjects of the inquiry, the dialogue must be dialectical in nature, and truth should be

reasoned out from the dialogue (Guba & Lincoln 1994).


- 111 -

3.2.3 Constructivism

The third paradigm, constructivism, assumes truth is a construction which refers to a

particular belief system held in a particular context. Meaning has more value than the

measurement, and the perception is the most important reality. It enquires about the

ideologies and values underlying the research findings (Perry, Riege & Brown 1999).

The ontology of constructivism argues that humans construct knowledge and

meaning from their experiences (Guba & Lincoln 1994). The realities in the

constructivism paradigm appear as multiple realities which are socially and

experientially based, and are the intangible mental constructions of individual

persons (Perry, Riege & Brown 1999). Perry et al. argue that the constructivist

approach is rarely appropriate for business research because the approach excludes

concerns about the economic and technological dimensions of business (Perry, Riege

& Brown 1999).

Guba and Lincoln (1994) describe the methodology of constructivism as

hermeneutical and dialectical. It is suggested that the variables of the research can be

elicited and refined through interaction between and among the investigator and the

respondents. The researcher is a ‘passionate participant’ within the world being

investigated (Guba & Lincoln 1994).


- 112 -

3.2.4 Realism

The last research paradigm to be discussed is the realism paradigm. It assumes that

reality exists but is imperfect because of the flaw in human intellectual mechanisms

and the fundamentally intractable nature of phenomena (Guba & Lincoln 1994). It

has the elements of both positivism and constructivism (Perry, Riege & Brown 1999).

The realism paradigm is also known as critical realism or post-positivism (Guba &

Lincoln 1994).

The ontology of the realism paradigm assumes that reality is ‘real’ but only imperfect

and probabilistically apprehensible and so triangulation from many sources is

required to try to understand the reality (Guba & Lincoln 1994; Perry, Riege &

Brown 1999).

The epistemology of the realism paradigm is considered as modified objectivist and

assumes that the findings of the researcher are probably true but always subject to

falsification. The findings of the reality have to fit the pre-existing knowledge (Guba

& Lincoln 1994).

The methodology of the realism paradigm includes case studies and interviewing.

The interpretation of the findings can be by quantitative methods and / or qualitative

methods (Perry, Riege & Brown 1999).


- 113 -

3.2.5 Justification of the selected research paradigm

The above section has reviewed four different types of research paradigms. A

positivism paradigm has been selected for this research for the following reasons:

a. From the ontology perspective, the post-adoption usage behaviour of online

securities trading is an independent reality. A positivist approach is suited to the

measurement and analysis of causal relationships between the pre-adoption variables

and post-adoption variables.

b. The pre-adoption and post-adoption variables related to online securities trading

are quantifiable where personal perceptions are important; they can be quantified by

asking participants to ‘score’ perceptions on a scale.

c. From an epistemological perspective, the researcher is independent of the subjects

in this research. It is free of personal bias or values. The researcher has no influence

on the data collected from the retail investors. So researcher bias or interaction with

subjects will have no influence on the research findings.

d. The researcher does not have any relationship with the survey participants and is

also not part of the object of this particular research.

e. This research is about observable phenomena in a market. It is appropriate

therefore, from a methodology perspective, that the hypotheses of this research are

stated in propositional form and subjected to empirical tests to verify them.

In summary, the research findings are in a quantifiable form to verify the causal

relationships between the variables. The researcher is independent and not


- 114 -

influencing or influenced by this research. Thus, a positivism paradigm has been

selected.

3.3 Research Methods

There are two main types of research methods: quantitative research that can be

associated with a positivist paradigm, and qualitative research that is associated with

a realism paradigm. The strength of quantitative research methods is that they can

provide wide coverage of the range of situations and a survey is fast and economical.

The strength of qualitative research is the ability to look at change processes over

time, to understand people’s meanings, to adjust to new issues and ideas as they

emerge, and the findings can contribute to a new theory (Easterby-Smith, Thorpe &

Lowe 1991).

The following table illustrates the characteristics of quantitative paradigm and

qualitative paradigm.

Table 3-2 Characteristics of quantitative and qualitative paradigms

Qualitative Paradigm Quantitative Paradigm 1. Qualitative method preferred. 1. Quantitative methods preferred. 2. Concerned with understanding human behaviour from the actor’s frame of reference.

2. Seeks the facts or causes of social phenomena without advocating subjective interpretation.

3. Phenomenological approach. 3. Logical positivistic approach. 4. Uncontrolled, naturalistic observational measurement.

4. Obtrusive controlled measurement.

5. Subjective; ‘Insider’s’ perspective; close to the data.

5. Objective; ‘outsiders’ perspective; distanced from the data.

6. Grounded, discovery-oriented, exploratory, expansionist, descriptive, inductive

6. Ungrounded, verification-oriented, confirmatory, reductionist, inferential, hypothetical, deductive.

7. Process-oriented. 7. Outcome-oriented. 8. Validity is critical; ‘real’, ‘rich’, and ‘deep’ data.

8. Reliability is critical; ‘hard’ and replicable data.

9. Holistic – attempts to synthesise. 9. Particularistic – attempts to analyse.

Source: (Deshpande 1983) who adapted from Reichardt and Cook (1979)


- 115 -

3.3.1 Qualitative Research

According to Bryman (1984), there are other terms used interchangeably by

qualitative research: ‘naturalistic’ field research, ‘ethnographic’, ‘interpretivist’, and

‘constructivist’. The philosophical underpinnings of quality research are typically

attributed to phenomenology and symbolic interactionism (Bryman 1984).

The qualitative research approach requires the researcher to be involved closely in

the investigation and findings, and views the social world from the point of the

researcher (Bryman 1984). Similarly, Zikmund (2000) describes qualitative research

as subjective in nature, and much of the measurement process is at the discretion of

the researcher (Zikmund 2000). Qualitative research can be in the form of

exploratory research to diagnose a situation, to screen alternatives, or to discover new

ideas (Zikmund 2000). Qualitative research tends to be appropriate where the aims of

the research involve building theory rather than testing pre-existing theory in the

social environment.

According to Zikmund (2000), a qualitative approach usually involves gathering

information about a small group of people or organisations, and the data is generally

not presented in a numerical form for statistical study (Zikmund 2000).

As this research is based on an existing theoretical model and large number of data

are available from the retail investors using online securities trading, qualitative

research is deemed to be inappropriate.


- 116 -

3.3.2 Quantitative Research

Quantitative research is a common method in conducting social research and can be

attributed to a positivist research paradigm. This research method is objective,

replicable and is concerned about causality. Quantitative research is frequently

described as being positivist or empiricist (Bryman 1984).

According to Zikmund (2000), quantitative research primarily involves statistical

analysis of the data gathered in the research. Using the numerical evidence of the data

gathered, the researcher objectively draws conclusions, or may test hypotheses built

from the existing theory (Zikmund 2000).

Johnson and Onwuegbuzie (2004) state the characteristics of quantitative research to

be; a focus on deduction, confirmation, hypothesis testing, explanation, prediction,

standardised data collection, and statistical analysis (Johnson & Onwuegbuzie 2004).

Quantitative research methods seem appropriate for this research based on the above

characteristics described.

3.3.3 Justification of the selected research method

Quantitative research is selected for this research for several justifiable reasons.

Firstly, quantitative research is attributed to the positivism research paradigm that has

also been selected as the research paradigm for this study.

Secondly, the research aims to gather a large number of data from retail investors

who use online securities trading in Singapore, which indicates that a quantitative

research method seems more appropriate over qualitative research. The research is to

investigate relationships between the variables stated in the questionnaire that have


- 117 -

potential influence over the post-adoption usage behaviour of the retail investors. A

qualitative research approach was deemed to be too time consuming and costly for

this research. It would be unnecessary to conduct in-depth interviews for each and

every individual retail investor. The researcher is not required to participate in the

research observation but to gather the data objectively and independently.

Furthermore, this research is based on the existing theoretical models of Rogers’

Diffusion Model (Rogers 2003) and Davis’ Technology Acceptance Model (Davis, F.

D. 1989); it is not trying to build a new theory from the research findings.

The subjects of this research are users of online securities trading facilities; well

suited for the use of an online survey questionnaire to gather data from a large

number of retail investors. Objective data can then be analysed using statistical

methods, with the aim of measuring the relationships among variables. Thus

quantitative research methods are much more suitable for this study.

A quantitative research method allows the researcher to examine the effects of

independent variables on an outcome of a dependent variable that can be expressed

numerically. Causal relationships among the variables can be deduced to explain

observed market behaviours.

3.3.4 Limitation of the selected research method

The first limitation of the quantitative research approach selected for this research is

that a large amount of data is required to be gathered for analysis, and thus it could be

resource intensive and time consuming. This limitation is overcome by using an

online questionnaire survey posted to the selected retail investors. Data are gathered


- 118 -

by the online survey application. The online survey facilitates the researcher in

gathering a large amount of research data using limited resources, and also reduces

the time required to collect data for analysis.

The standardisation of the questions in the questionnaire tends to limit the testing and

findings towards the pre-determined hypotheses of the research. More in-depth

research observation and findings are limited by the quantitative method. The

objective of this research is not to build a new theory through participative

observation and interaction with survey objects, but to determine the causal

relationship among the variables using existing theories.

Johnson and Onwuebuzie (2004) argued that the knowledge produced by quantitative

research may be too general for direct application to specific local situations, contexts,

and individuals (Johnson & Onwuegbuzie 2004). However, this research does not

have such a research objective. The limitation of generalisation of knowledge is

accepted, as the research domain is aimed at explaining post-adoption usage

behaviour by a broad class of consumers.


- 119 -

3.4 Survey Objects

3.4.1 Theoretical Population

The survey objects in this research are the adopters of online securities trading in

Singapore. These adopters are users who buy or sell equity stocks in Singapore

through an online securities trading tool. They are referred to as retail investors, as

they trade stocks in a relatively small volume on their own behalf. To do so, the retail

investors have to be above 21 and possess a Central Depository Account (CDP) from

Singapore Exchange Ltd (SGX) (CDP 2008). In addition, these users must also open

valid trading accounts with an authorised securities broker in Singapore. These

authorised securities brokers are registered with the Singapore Exchange Ltd (SGX

2008).

According to the CDP as of 30 Dec 2006, there are around 1.29 million account

holders registered with CDP (SGX 2006). Out of the 1.29 million accounts, there are

278,793 active accounts who have conducted trading in 2006 (Liew 2006). However,

as shown in Table 3.1, of the total trading value, only ten percent is traded online

(Liew 2006).


- 120 -

Table 3-3 Internet penetration rate and percentage of online trade

Country Internet Penetration

Rate

% of online trades

(Trading Value)

Singapore 67.2% 10%

Korea 67% 47%

Taiwan 60% 17%

Australia 68.4% 50%

Japan 67.2% 29%

According to the Singapore Exchange, there are 27 Securities trading members, or

brokerage firms in Singapore. Out of the 27 brokerage firms, only 11 offer an online

securities trading service to the retail investors (SGX 2008).

Table 3-4 Securities trading members in Singapore

S/No Securities Trading Members Online Securities Trading Service

1 ABN AMRO Asia Securities (Singapore) Pte Limited

No

2 AmFraser Securities Pte Ltd Yes 3 BNP Paribas Securities (Singapore) Pte. Ltd. No 4 CIMB - GK Securities Pte. Ltd Yes 5 Citigroup Global Markets Singapore Securities Pte Ltd

No

6 CLSA Singapore Pte Ltd

No

7 Credit Suisse Securities (Singapore) Pte Ltd

No

8 Daiwa Securities SMBC Investment Services Pte Ltd

No

9 DBS Vickers Securities Online (S) Pte Ltd Yes 10 DBS Vickers Securities (Singapore) Pte Ltd

No

11 DMG & Partners Securities Pte Ltd Yes 12 Fortis Clearing Singapore Pte Ltd Yes

Source: (Liew 2006; SGX 2008)


- 121 -

13 Instinet Singapore Services Private Limited No 14 J.P. Morgan Securities Singapore Pte Ltd

No

15 Kim Eng Securities Pte. Ltd Yes 16 Lehman Brothers Private Limited

No

17 Lim & Tan Securities Pte Ltd Yes 18 Macquarie Securities (Singapore) Pte Limited

No

19 Merrill Lynch (Singapore) Pte Ltd

No

20 Morgan Stanley Asia (Singapore) Securities Pte Ltd

No

21 Nomura Securities Singapore Pte Ltd

No

22 OCBC Securities Pte Ltd Yes 23 Phillip Securities Pte Ltd Yes 24 SBI E2-Capital Asia Securities Pte Ltd

Yes

25 UBS Securities Pte. Ltd No 26 UOB Kay Hian Private Limited Yes 27 Westcomb Securities Pte Ltd

No

The proportion of stock trading volume online in Singapore is relatively low as

compared to other Asian countries like Korea and Japan. In one of the Singapore

Exchange conferences, the Head of Retail Investing of SGX has highlighted that only

10 percent of total stock turnover on the SGX is done via online trading (Liew 2006).

Based on the research figures provided by SGX, it is estimated that there were

170,000 active online traders in Singapore for the year of 2006 (Liew 2006).

Source: (SGX 2008)


- 122 -

3.4.2 Accessible Population

As mentioned in the earlier section, the targeted survey objects are the adopters of

Online Securities Trading from the Online Brokerage firms in Singapore. There are

six major Online Securities Trading service providers in Singapore as listed below

(SGX 2006):

Table 3-5 Securities trading members with online trading service

S/No Securities Trading Members with online

trading service

Website

1 POEMS by Phillip Securities www.poems.com.sg

2 Fraser Securities by AmFraser Securities

Pte Ltd

www.amfraser.com.sg

3 Lim & Tan Securities www.limtan.com.sg

4 DBS Vickers Securities Online (S) Pte

Ltd

www.dbsvonline.com.sg

5 DMG Online by DMG & Partners

Securities

www.dmg.com.sg

6 iOCBC by OCBC Securities Pte Ltd www.iocbc.com

(SGX 2006)


- 123 -

The six major brokerage firms account for about 50 percent of the online securities

trading users - around 85,000 subscribers. If permission was granted from all six of

the major brokerage firms to access these subscribers, this would be the estimated

total of users accessible through the brokers who provide an online securities trading

service.

Using this accessible population, a sample could be obtained for further analysis to

test the hypotheses in this research.

The details of selecting the sample and sample size will be discussed in the next

section.

3.5 Sampling

3.5.1 Sampling Design

According to Zikmund (2000), there are 4 common sampling designs used by

researchers (Zikmund 2000):

Simple Random Sampling

A simple random sample requires that every unit in the population has a known and

equal chance of being selected. An example is the drawing of participants’ names

from a hat to get the lucky gift in a party.


- 124 -

Systematic Random Sampling

Systematic random sampling requires the units in the population to be ordered in

some way or another. For example, names that are ordered alphabetically, or

customers who walk into an entrance or follow one another (Zikmund 2000).

Cluster Sampling

A cluster sample is a sample where the units making up the population and sample

are divided into clusters. They are also known as first-stage units and primary

sampling units. Each cluster has more basic units called second-stage units or

secondary sampling units. The cluster is usually divided by geographic area and used

by census surveys (Zikmund 2000).

The geographic region in this research is restricted to Singapore and the sample is

drawn from the retail investors and no further classification is necessary for the

survey.

Stratified Sampling

For stratified sampling, the researcher divides the population into groups and

randomly selects subsamples from each group (Zikmund 2000).

Stratified sampling allows the large population to be divided into several subgroups

or strata to draw a random type sample from each stratum. The online retail investors

can be divided into six stratums represented by the major brokerage firms in

Singapore.


- 125 -

However, the most prominent brokerage firms are Phillip Securities and Fraser

Securities, and they could represent the typical retail investors who trade online in

Singapore.

The online securities trading service by Phillip Securities is called Phillip Online

Electronic Securities Mart System (POEMS). The POEMS customers are the

adopters to be surveyed in this research. POEMS has been selected as it is the earliest

online securities trading tool adopted in Singapore, as well as having the largest

number of subscribers compared with other brokerage firms (Smart Investor

Magazine 2000). POEMS has also been awarded as the Broker of the Year 2000 by

Finance Magazine and Channel News Asia in Singapore (Tan, A. 2000). POEMS

has been awarded the Hitwise Singapore Online Performance Award for 2005 and

2006 as the No.1 ranking website most visited by Singapore Internet users (POEMS

2008).

In addition to POEMS, FraserDirect Online Securities Trading service by Fraser

Securities has been selected, as it has one of the most vibrant chat rooms providing

cyber-advisors for the online investors. (Financial Planner Magazine, July 1999)

FraserDirect is also the most advertised brokerage firm in Singapore. Fraser

Securities is the nation’s oldest brokerage firm and has been established since 1873

(AMInvestment 2007). Both the users of POEMS and FraserDirect Online have been

selected as the survey sample.


- 126 -

3.5.2 Sampling Size

The researcher is not able to get the number of retail investors subscribed to online

securities trading offered by Phillip Securities and Fraser Securities, as both

companies have declined to disclose the numbers. It is estimated that there were

170,000 active online traders in Singapore for the year of 2006 (Liew 2006). As the

samples are gathered from only the retail investors using POEMS and FraserDirect,

the available population will be lower than 85,000.

3.5.3 Validity and Reliability

Sampling Validity

Sampling validity refers to the idea that the sample must allow the research process

to measure what the researcher intend to measure (Zikmund 2000). Pelham and

Blanton (2003) state that validity refers to the relative accuracy or correctness of a

research statement (Pelham & Blanton 2003). This will only be achieved if the

sample is taken from a population which is valid in relation to the research statement.

It is established by the degree to which the measure confirms a network of related

hypotheses generated from the theoretical model. The sampling targets are selected

from the adopter group and so are closely related to the Diffusion model and the

hypotheses derived from it. The hypotheses are about adopter behaviour, and the

post-adoption usage of online securities trading.


- 127 -

Construct Validity

Zikmund (2000) states that construct validity is about the ability of a measure from

the research to confirm the network of variables related to the hypotheses generated

from the theoretical model. To achieve construct validity, the researcher must

establish convergent and discriminant validity. A theoretical model has convergent

validity when it is highly correlated with different measures of similar constructs

(Zikmund 2000). For example, there are repeated questions to be asked which relate

to a single variable in a questionnaire.

The research uses Cronbach’s Alpha coefficient to test scale reliability for the

variables constructed. Shorter-Judson uses Cronbach’s Alpha for reliability testing in

a similar research (Shorter-Judson 2000).

Sampling Reliability

Zikmund (2000) defines reliability as the ability to provide consistent results in

repeated uses of the measuring instrument (Zikmund 2000). It is about the degree of

measures that are free from error. There are two underlining dimensions of reliability:

repeatability and internal consistency.

3.6 Questionnaire Design

3.6.1 Questionnaire Objective

In Chapter Two, the literature on Consumer behaviour, Diffusion of Innovation

Model, Technology Acceptance Model and online securities trading in Singapore

were reviewed. The focus of the research is to investigate the correlation between


- 128 -

pre-adoption factors which influence adoption of online securities trading, and the

consequences, or behaviour, of adopting the online securities trading services by the

retail investors in Singapore. This is to address the research issues identified in

Chapter One:







The literature is built on Roger’s Diffusion of Innovations Model (Rogers 2003) and

Perceived usefulness derived from Davis’ Technology Acceptance Model (Davis, F.

D. 1989). In addition, Consumer loyalty factors have also been taken into

consideration for the study (Oliver 1999).

Generally, it is proposed in Chapter Two that there may be a relationship between

factors preceding adoption and patterns of usage which follow adoption. To test this

relationship, a number of pre-adoption factors are envisaged, derived from the

models of Diffusion of Innovations and Technology Acceptance. These were

explained in Chapter Two. In addition to these, a measure of post-adoption behaviour,

or usage, of the technology is envisaged. Thus, from the general proposition that

post-adoption behaviour might be influenced by specific pre-adoption factors, a


- 129 -

number of specific propositions were developed in Chapter Two. Each proposition

relates to a specific pre-adoption factor. So the form of each proposition is that post-

adoption usage of the technology, or post-adoption behaviour, will be influenced by

each of the specific pre-adoption factors.

From these propositions in Chapter Two, seven hypotheses were developed in this

chapter. Zikmund (2000) states a hypothesis is a proposition that can be tested using

empirical evidence. Thus, hypothesis is an empirical statement concerned with the

relationship among the variables to be studied (Zikmund 2000).

Deriving from the propositions, the hypotheses state that there will be a significant

relationship between each pre-adoption factor and post-adoption behaviour. To test

these hypotheses, it was necessary to find a measurable, or quantifiable construct for

each pre-adoption factor and for post-adoption behaviour.

In order to specify the factor, a questionnaire was devised which asked a large

number of specific questions about what influenced the decision to adopt the

technology of online trading of securities. A large number of questions were devised

so that quantifiable answers to each question could become a variable, where a

number of variables would form a factor as derived from the models in the literature.

As well, for post-adoption behaviour, a number of aspects of behaviour were

incorporated into questions about the extent or degree of post-adoption usage of the

technology. The questionnaire was divided into parts, where each part comprised

questions, as variables, related to the pre-adoption factors outlined in Chapter Two.


- 130 -

There are seven hypotheses established for the research as stated in Figure 3.2. For

the questions 1a, 1b, 1c, 1d, 1e, 1f, 1g , 1h , 1i an 1j, they are addressing the element

related to Perceived attributes of innovation which form hypothesis 1. To answer

questions related to type of innovation-decision associated with hypothesis 2,

questions are found in 2a, 2b, 2c, 2d, 2e and 2f. In regard to Communication channels

related to hypothesis 3, questions are asked in 3a, 3b, 3c and 3d. The Nature of social

system variables are found in hypothesis 4 and the questions are addressed in 4a, 4b,

4c and 4d. For hypothesis 5 that is regarding the Extent of the change agent’s

promotion efforts, questions are found in 5a, 5b and 5c. Hypothesis 1, hypothesis 2,

hypothesis 3, hypothesis 4 and hypothesis 5 are founded on the literature review of

Everett Rogers. Using the Technology Acceptance Model, hypothesis 6 has been

established, and questions 7a, 7b and 7c are to address the correlation between

Perceived usefulness and post-adoption behaviours of the retail investors. In

hypothesis 7, which is to address the question related to Consumer loyalty and post-

adoption behaviours, the questions are found in 8a, 8b, 8c, 9a, 9b and 9c. Table 3.6

summarises the hypotheses in this research and the respective questions indicated in

the questionnaire.


- 131 -

Figure 3-2 Theoretical model developed for this research



- 132 -

Table 3-6 Literature based constructs, hypotheses and survey questions

HYPOTHESIS Variables and Related literature

Questions

H1: A relationship exists between the perceived attribute factors and post-adoption behaviour.

Relative advantage (Rogers 2003)

Q1(a) - Quicker to trade online Q1(b) - Cheaper to trade online

Compatibility (Rogers 2003)

Q1c -Process is not much different from calling the broker Q1d - Trading information is not much different from calling the broker

Complexity (Rogers 2003)

Q1e- Easier to access investment information online Q1f – Easier to trade using Online Securities Trading

Trialability (Rogers 2003)

Q1g -Able to do a trial trade which is not possible via the broker Q1h - Easier to obtain an online demonstration and explanation

Observability (Rogers 2003)

Q1i - More investors signed up for Online Securities Trading system Q1j - More investors started to trade using Online Securities Trading

H2: A relationship exists between the type of innovation-decision and post-adoption behaviour.

Optional (Rogers 2003)

Q2a – Have considered other options like automatic voice trading or WAP trading via phone? Q2b – Have considered online trading as an additional method of trading

Collective (Rogers 2003)

Q2c – Have consulted other investors using online securities trading Q2d – Have consulted my friends using online securities trading

Authority (Rogers 2003)

Q2e – Have been advised by investment experts to sign up online securities trading Q2f – Stock Exchange has liberalised brokerages’ commission rate for online securities trading

H3: A relationship exists between the communication channels and post-adoption behaviour.

Mass media (Rogers 2003)

Q3a - Online advertisements like Internet or email Q3b - Mass media like TV or newspaper advertisements

Interpersonal (Rogers 2003)

Q3c - Broker's explanation and demonstration Q3d - Friends' and other investors' advice

H4: A relationship exists between the nature of the social system and post-adoption behaviour.

Norms (Rogers 2003)

Q4a – Know more friends who used online securities trading Q4b – I feel left out if I do not sign up online securities trading

Degree of network interconnectedness (Rogers 2003)

Q4c – Consult or discuss with friends or other investors when I trade online Q4d – Exchange information with friends or other investors using online securities trading

H5: A relationship exists between the extent of change agent’s promotion efforts and

The change agent’s promotion efforts (Rogers 2003)

Q5a - Constantly received Online Securities Trading information Q5b - Satisfaction with the broker's


- 133 -

post-adoption behaviour. promotional efforts on Online Securities Trading Q5c - Satisfaction with the brokerage firm's promotional efforts on Online Securities Trading

H6: A relationship exists between the Technology Acceptance Model and post-adoption behaviour.

Perceived usefulness (Davis, F. D. 1989)

7a - Trading Online increases my trading profit 7b - The system facilitates diversification of my portfolio 7d - I can react to the stock market quicker

H7: A relationship exists between consumer loyalty and post-adoption behaviour.

Loyalty (Oliver 1999) (Hennig-Thurau, Gwinner & Gremler 2002) (Werner & Murphy 2007)

8a – Online securities trading will be my major method to trade 8b – I will introduce online securities trading to non-users 8c – I will not consider other new method of trading in near future

Confidence (Hennig-Thurau, Gwinner & Gremler 2002) (Iyengar 2004) (Gefen, Karahanna & Straub 2003)

9a – It is a highly secured system 9b – It is easily accessible 9c – It is very reliable

3.6.2 Questionnaire Types

Questionnaire Sections

The questionnaire is designed with three Sections to cover questions to be answered

in the seven hypotheses. Section A covers independent, or pre-adoption, variables

related to perceived attributes of innovations (Relative advantage, Compatibility,

Complexity, Trialability and Observability). Section A has seven parts, as explained

above, relating to seven factors influencing adoption. It covers independent variables

on Type of innovation-decision, Communication channels, the Nature of the social

system and Change agent’s promotion effort. The other independent variables include

the Perceived usefulness that is a component of the Technology Acceptance Model

and the Consumer loyalty (Loyalty and Confidence).



- 134 -

Section B covers the dependent, or post-adoption, variables about trading behaviours:

frequency of trading, volume of shares trading, types of stock trading, location of

trading shares, exchange of investment tips and frequency of checking investment

information. Lastly, Section C consists of questions that are related to demographic

characteristics like age range, education level, current occupation, income level,

marital status and gender of the participants of the survey. (see Appendix C)

Type of Questions and Options of Answers

A questionnaire was created which took the form of 42 questions in the form of a

statement about pre-adoption influence or post-adoption behaviour. Respondents

were asked to answer these questions stated in the questionnaire

The questionnaire for this research is designed around a seven point Likert scale with

a range from “Strongly Disagree” to “Strongly Agree”. The range is associated with

the value from 1 to 7. The value of each question is entered into SPSS for data

analysis in a later phase. In most of the researches related to adoption intent, post-

adoption and the Technology Acceptance Model, a seven point Likert scale is

adopted (EL-Gayar & Moran 2007; Lau, Yen & Chau 2001; Luo, Remus & Chea

2006; Parthasarathy & Bhattacherjee 1998; Rigopoulos & Askounis 2007).

For all of the questions linked to the variables to be answered in the questionnaire,

the scale was arranged as follows:


- 135 -

Strongly Disagree

Disagree Slightly Disagree

Neutral Slightly Agree

Agree Strongly Agree

Easier to trade using Online Securities Trading

There are also other types of questions with a range of different selections mainly

catering for demographic factors. For example, the different range of figures for the

average annual income question. The data gathered will be for reference purposes

and not to be verified against the hypotheses. It is to show the demographic profiles

of the retail investors who responded to the survey.

What is your average annual income range in SGD?

20,000 & below 21,000 - 35,000 35,001 - 50,000

50,001 - 65,000 65,001 - 80,000 80,001 - 100,000 100,001 & above


- 136 -

3.6.3 Independent Variables

As shown in Table 3.6, there are seven groups of independent variables and one

group of dependent variables that is, the post-adoption consequences. In total, there

are 42 variables including both independent variables and dependent variables.

Beside the independent variables constructed from Rogers’ Diffusion of Innovation

Model (2003), Davis’ Acceptance of Technology Model (1998) and the Customer

loyalty, demographic characteristics of the retail investors have been gathered to

illustrate the respondents’ profiles (see Table 3.7).

Table 3-7 Demographic factors and related questions

Demographic Factors Questions Age Part C Q13 Education Part C Q14 Occupation Part C Q15 Income Part C Q16 Marital Status Part C Q17 Gender Part C Q18

3.6.4 Dependent Variable and Components

In this dissertation, the dependent variable is mainly the change of behaviour or the

consequences in the retail stock investors after adopting Online Securities Trading.

The variable has several components that will be described.

The following table, Table 3.8 lists the dependent variable and components versus

the questions to be tested in the questionnaire.

Source: designed for this research


- 137 -

Table 3-8 Post-adoption behaviour and related questions

Dependent Variable Components Questions

Post-Adoption Behaviour Frequency Part B Q11a I trade more frequently now

Volume Part B Q11b I trade in smaller lot sizes now

Trading Type Part B Q11c I buy certain categories of stock (eg. High Tech Stocks)frequently

Location Part C Q11d I trade from more locations like office, Cybercafé and overseas in addition to home

Investment tips exchange Part C Q11e I exchange investment tips with other online investors easily

Investment information Part C Q11f I check investment information frequently

3.7 Mode of Survey

3.7.1 Web Survey / Email

The primary mode of data gathering was to undertake a survey via the Internet. The

questionnaire was firstly designed in a manual form and then converted to HTML

(Hyper Text Mark-up Language) to be hosted on a web server. The web page was

then established with an address that could be accessed via the Internet. For this

research, the web design and hosting for the questionnaire used FormSite, a third

Source: designed for this research


- 138 -

party online survey tool. FormSite is owned by Vronam Systems Inc and has been a

tool to build forms easily on the web since 1998 (Formsite 2008).

The researcher used Formsite to build the questionnaire (appendix C) and publish it

on the Internet. The website of the online questionnaire was :

http://fs3.formsite.com/onlinetrading/index.html

The website of the online questionnaire was sent to Phillip Securities and Fraser

Securities respectively to request them to make it known to their online securities

trading subscribers in Singapore. The researcher also posted the website of the online

questionnaire to the online discussion forum of POEMS provided by Phillip

Securities and that of FraserDirect, provided by Fraser Securities, which were

accessible by their online securities trading subscribers.

Once respondents completed the questionnaire form on the website, the FormSite

server was arranged to send an email to alert the researcher. The result also resided

on the FormSite web server so that it could be downloaded by the researcher.

The Online Survey was preferred in this dissertation as it has been proven successful

in other similar surveys conducted in the areas of adoption related research (Chea &

Luo 2007; Tan, M. & Teo 2000; Vitartas et al. 2007). The samples are drawn from

retail investors using online securities trading and it was convenient for this group of

online users to respond to the survey through an online questionnaire when they saw

the survey website through the online forums.


- 139 -

The following table, Table 3.9 is a comparison between an Internet survey and a mail

survey conducted by the National University of Singapore (Tan, M. & Teo 2000).

Table 3-9 Internet survey versus mail survey

Characteristics Mail Survey Internet Survey

Manpower Insert survey into envelopes, paste stamps or frank envelopes.

Design Web page and write Javascript. For this research, the Web page used FormSite that could automate the web design.

Cost Envelopes, stamps, photocopying of questionnaires.

Rental/maintenance of server space to host Web page. The Online questionnaire was rented on FormSite Web Server.

Sampling frame Restricted to sample that received the questionnaire.

Restricted to people with access to the Internet who chose to respond.

Response rate Can be computed. Percentage of respondents is dependent on follow-up mailings.

Cannot be computed. Response dependent on publicity of the survey as well as follow-up reminders via emails to potential respondents.

Time frame Usually takes about a month for surveys to be returned.

Responses can usually be collected within two weeks.

Quality of data Dependent on whether targeted respondents respond to questionnaire. Systematic bias is reduced with the use of random sample and also by obtaining a high response rate.

Adequate if the targeted respondent is the general Internet user population. Potential for systematic bias if only people with certain characteristics respond.

Generalisability of results Results generalisable to target population if response rate is adequate.

Difficult to determine since there may be systematic bias in terms of who actually responds to the questionnaire.

Suitability Must be able to identify potential respondents. Can reach out to the general public regardless of computer access.

Survey of people with Internet access. Very suitable as only people with Internet access would be part of target population of online traders of securities.


- 140 -

Problems Costly and slow.

Unable to control who responds. Data must be screened for unsuitable respondents.

Pilot Testing

The researcher conducted some pilot testing for the online web survey by sending it

to DBA students as well as MBA students from Southern Cross University in

Singapore. The pilot survey population size was about 100 and 13 responded to the

survey. Based on the feedback, the respondents were able to understand the online

questionnaire easily and were able to complete all the questions without any

difficulty.

3.8 Data Processing Procedures

3.8.1 Descriptive statistics - Cross Tabulation

This questionnaire contained a section to collect the statistics of the demographic

factors of the retail investors.

Zikmund (2000) defines descriptive analysis as the process of transforming raw data

into a format that is easy to understand and interpret (Zikmund 2000).

Zikmund (2000) defines a categorical variable as any variable that has a limited

number of distinct values like gender. A continuous variable, according to Zikmund

(2000), is any variable that has an infinite number of values, like time taken to

complete a test (Zikmund 2000).

Source: (Tan, M. & Teo 2000)


- 141 -

For this study, descriptive statistics about the population are collected with questions

that provide categorical responses.

A cross tabulation method is used to describe the demographic data gathered before

getting into further detailed study of other data.

Cross tabulation can be defined as the tables displaying the number of cases falling

into each combination of the categories of two or more categorical variables. The

table may display the counts, percentages, expected values, and residuals.

Following is an example of cross tabulation (see Table 3.10) comparing the age of

the retail investors and their respective occupations. For example, in the respondents

of between the age ranges of 21 to 30, 29 out of 100 are professional.

Table 3-10 Cross tabulation of age versus occupation

Age * Occupation Crosstabulation

Count

19 29 15 27 7 3 10030 22 18 13 11 1 955 10 4 2 6 2 291 1 2 0 2 2 8

55 62 39 42 26 8 232

21-3031-4546-5556&up

Age

Total

Executive andManagerial Professional

Technical,Production& related

Administration, Sales& Services

Self-employed Others

Occupation

Total

Source: developed for this research based on survey data


- 142 -

After the descriptive analysis is completed, factor analysis is used to reduce the

number of variables to a smaller set of constructs for further analysis. The variables

deemed to be significant are then to be analysed by regression analysis to test against

the hypotheses established for the study.

3.8.2 Factor Analysis

A similar factor analysis approach has also been used in other researches conducted

in Singapore. Kendall and others have conducted research on “Electronic Commerce

Adoption by SMEs in Singapore” and used factor analysis to reduce the large number

of variables into groups of constructs for further regression analysis. (Kendall et al.

2001)

Other researchers used the factor analysis technique, taken from National University

of Singapore, in a similar research on “Factors Influencing the Adoption of Internet

Banking”. (Teo, Tan & Peck 2004)

Zikmund (2000) stated that the purpose of factor analysis is to summarise the

information contained in a large number of variables into a smaller number of factors

(Zikmund 2000). Factor analysis is used as a data reduction technique using principal

component analysis (PCA) as it takes a large set of variables and looks for a way the

data may be reduced into a smaller number of components (Pallant 2007). Rummel

(1967) indicated that factor analysis is useful in reducing a mass number of data into

economical description (Rummel 1967). The researcher may collect a large variety of

data, however, the large number of variables should be reduced to certain underlying


- 143 -

constructs that will summarise the information contained in the variables. In general,

the objective of factor analysis is to reduce a large number of variables to as few

constructs as possible (Zikmund 2000). Rummel (1967) mentioned that factor

analysis can be used to reduce the independent variables into a group of meaningful

and descriptive categories (Rummel 1967).

Similarly, Hair (1998) stated that factor analysis is a useful tool that is part of the

multivariate statistical technique to extract information from a large database and

identify the interrelated data (Hair et al. 1998).

Factor analysis can be used to study the correlation among the variables by analysing

the correlation matrix. A correlation matrix is simply a rectangular array of numbers

which gives the correlation coefficients between a single variable and every other

variable in the investigation. It identifies the relationships among the variables

(Rummel 1967).

As a data reduction technique, factor analysis is useful in reducing a large number of

variables to a smaller number. This allows for data to be used in subsequent analysis

such as multiple regression. When there is correlation among a set of independent

variables in a multiple regression, this is known as multicollinearity and it exists when

the variables are highly correlated (r=0.9) and above(Pallant 2007). Independent

variables that exhibit multicollinearity are not suitable for multiple regression and do

not contribute to good regression model (Pallant 2007).

According to Zikmund (2000), independent regression variables derived from factor

scores taken from factor analysis will reduce the problem of multicollinearity in

multiple regression. Factor score is the number that represent each observations’


- 144 -

calculated value on each factor in a factor analysis (Zikmund 2000). It is also

mentioned in SPSS manual that factor scores computed from factor loadings on each

orthogonal component help to minimise issues of multicollinearity (SPSS 2000).

Rummel (1967) stated the factor loadings through orthogonal varimax rotation returns

factors that are uncorrelated (Rummel 1967). Thus, independent variables to be used

in multiple regression that are created from factor scores will be uncorrelated.

42 questions contained in the questionnaire became 42 categorical variables with a

range of scores from 1 to 7. These forty-two questions were arranged so that a

number of questions related to each of seven independent, or pre-adoption, attributes,

or factors, related to online trading facilities as well as dependent, or post-adoption

usage.

Factor analysis was used to reduce these variables to a smaller number of constructs.

If the result of the factor analysis was that just seven independent factors emerged,

then this would tend to confirm the applicability of the model derived in Chapter

Two, to this sample and population. The results of the factor analysis are discussed in

detail in Chapter Four. Detailed below is a summary of the factor analysis method

and process used.

The first step was to conduct a data reduction, factor analysis by selecting all of the

42 variables using SPSS V15.0. The KMO and Barlett’s test of sphericity was

selected as the correlation matrix. This was used to determine the appropriateness of

factor analysis. The Principal Component extraction method was used and

Eigenvalues greater than 1 applied as the criteria for choosing the number of factors.


- 145 -

Pallant (2007) suggested that Principal component analysis is the most commonly

used method among researchers (Pallant 2007).

If a number of factors is extracted such that one or more of the factors has an

Eigenvalue less than 1, then that factor is not explaining any more variation than the

average single variable (Pallant 2007).

Scree plots were printed with a view to see whether a different number of factors was

indicated. The rotation method used is Varimax with maximum iterations for

convergence of 25 set as the default. Pallant (2007) stated that the Varimax is the

most commonly used factor rotation method (Pallant 2007).

When printing and displaying a rotated factor matrix, it is helpful in interpreting

factors to suppress, factors landings that are low are not shown in the matrix. In this

way, most variables will show a loading, or score, on only one factor, and so make

interpretation of factors easier. A factor can be interpreted, or described, in terms of

the few variables that load onto it.

Pallant (2007) indicates that variables with greater loading scores than 0.3 may be

considered as statistically significant (Pallant 2007). Field (2005) recommends the

suppression of scores below an absolute value showed depend on sample size or a

value reflecting the expected of a significant loading. For example, with a sample

size of 100, the loading should be greater than 0.512 to be a significant result (Field

2005). In this study, the number of respondents is 234 and the sample size is less than

85,000. It is recommended to use the loading as 0.5 as suggested by Field (2005).


- 146 -

From the data reduction result, a number of significant factors are derived. Each

factor will be related to a group of significant components.

The factors identified through this data reduction method are used as construct

variables, used in further analysis to test hypotheses. All steps in conducting factor

analysis and the respective results are presented in the data analysis chapter.

3.8.3 Regression Analysis

According to Zikmund (2000), linear regression analysis is a technique for estimating

a relation between a dependent variable and one or more independent variables

(Zikmund 2000).

The objective of regression analysis is to determine the values of parameters for a

function that causes the function to best fit a set of data observations. In linear

regression, the function is a straight-line equation, hence the name of linear. Manson

and others (2000) state that a regression analysis technique can be used to develop an

equation for the straight line fit to data, and make predictions of the value of the

dependent variable (Y) based on the independent variable (X) (Mason, Lind &

Marchal 2000).

Single regression analysis is used as the analysis of bivariate correlation while

multiple regression is a method for analysing the relationship between two or more

independent variables against the single outcome dependent variable.


- 147 -

3.8.4 Multiple Regression

Rummel (1967) mentioned that factor analysis can be used to transform the data to

meet the application of multiple regression analysis by means of deriving the factor

scores from the reduced variables (Rummel 1967).

The objective of multiple regression analysis is to determine the changes in the

dependent variable in response to changes in the independent variables (Hair et al.

1998). Multiple regression analysis is also used to explore the predictive ability of a set

of independent variables on a dependent variable (Pallant 2007).

Multiple linear regression technique estimates an equation expressing the relationship

between a dependent (Y) and a number of independent variables X1 …Xi. The

equation is specified in the general form:

Y = a + b1X1 …biXi

Where Y is the dependent variable

A is an intercept term, which provides a value for Y when all X i = 0

b1 is a coefficient of the independent variable Xi

X1 …Xi are independent variables.

When a linear regression equation is estimated, it is normal practice to test the

coefficient bi for significance. If the equation is not being used for predictive

purposes, then finding that a value bi is significant is a way of confirming a

relationship between any independent variable Xi and the dependent variable Y. If bi


- 148 -

is positive, this indicates a positive relationship between Xi and Y that is, as Xi

increases, Y will increase. Alternatively, bi can be interpreted as the rate of change of

Y as Xi changes, all else being constant.

Multiple regression analysis is deemed appropriate for this research as there are

several independent variables but one dependent variable (post adoption behaviours)

that is made up of a number of components (Zikmund 2000).

Specifically, in this research, the expression of the regression model is an equation to

test all of the hypotheses derived in Chapter Two. Regression is used to determine the

relationship between the various independent variables, which are pre-adoption

factors or attributes of the innovation of online securities trading. The post-adoption

behaviour is the dependent variable. The multiple regression equation is:

Y = α + β1X1 + β 2X2 + β 3X3 + β 4X4 + β 5X5 + β 6X6 + β 7X7 + ε

Where;

Y = Dependent variable (post-adoption behaviours)

X1 = Independent variable (perceived attributes of Innovations)

X2 = Independent variable (type of innovation-decision)

X3 = Independent variable (communication channels)

X4 = Independent variable (nature of the social system)

X5 = Independent variable (extent of change agent’s promotion effects)


- 149 -

X6 = Independent variable (TAM: perceived usefulness)

X7 = Independent variable (customer loyalty)

α = Intercept

β1, β 2 , β 3, β 4, β 5, β 6, β 7 = Regression coefficient (slope) for each independent

variable

ε = Random error

The β coefficients show the relationship between Xi and Y. Each of the coefficients

β1, β2, β3, β4, β5, β6 and β7 is tested for significance using the ‘t’ statistics. An

examination of the t-values indicates the contribution of the independent variables to

the dependent variable (Coakes, Steed & Price 2008). Using the significant t-value of

the β coefficients, it is able to draw a conclusion about the significance of any

attribute. Thus, a test of significance can be used as a test of a hypothesis that a

significant relationship exists between Xi and Y. The statistical analysis software

package SPSS is used here to conduct regression analysis.

SPSS provides the R Square value in multiple regression analysis that indicates how

much of the variance in the dependent variable is explained by the regression model

which made up of the independent variables if the regression equation is to be used

for predictive purpose (Pallant 2007). In the first run of the multiple regressions, if a

certain coefficient of the Xi is not significant, it should be removed and the regression

testing re-run. In the final model, only the coefficients that are statistically significant

are measured to determine the variance between the dependent variable and


- 150 -

independent variables. In this research, the objective is not prediction, but to test the

significance of relationships.

The steps and results of the multiple regression analysis are presented in Chapter

Four.

3.8.5 Data Processing Tools

The data processing tool used is SPSS version 15.0 which is a common statistical

analysis package. The descriptive statistics use cross tabulation to illustrate the

demographic profiles of the respondents. In this research, the factor analysis and

regression analysis of the data gathered is conducted using SPSS.

3.8.6 Ethical Considerations

Ethical considerations are important when dealing with research involving human

beings. The research methodology and questionnaire has been reviewed and

approved by the Southern Cross University’s Human Research Ethics Committee.

The research was conducted in accordance with the guidelines and policies provided

by the committee. The committee’s approval number for this research is ECN-03-20

(see Appendix A).

The respondents to the online survey are not forced into answering the

questionnaires; they are given the free choice to participate in the survey.

At any point of time during the survey, respondents are free to terminate the survey at

their will. Another consideration is the privacy of the respondents; the researcher will


- 151 -

not publish or reuse the respondent’s information anywhere else besides this research

without the consent of the respondents.

Lastly, the researcher will not mislead the respondents in any form but will follow

closely the guidelines and policies provided by the committee.

The researcher will not utilise the results of the research in other sources without

prior consent from the respondents. The researcher is to be honest and keep the

integrity of the survey’s results and not distort or manipulate the results dishonestly.

The researcher will avoid disseminating a conclusion of the research that is

inconsistent with the survey’s results.


- 152 -


After reviewing the literature in Chapter Two, a unique theoretical model for this

research has been established using Roger’s Diffusion of Innovations Model and the

Perceived usefulness factor derived from Davis’ Technology Acceptance Model,

with the additional independent variable of Consumer loyalty factors.

The model was adapted and used as a basis to test for a proposed relationship

between the attributes of technology, online securities trading, and the degree of post-

adoption usage of the technology.

In this chapter, various research paradigms have been reviewed, and from those a

positive approach was selected. Following a review of alternative methodological

approaches, the research deemed most appropriate was a quantitative approach to test

the model. Data was collected using an online survey of users of the new technology.

The chapter then follows with an outline of the specific way in which questions in the

survey were specified, and how responses were analysed using factor analysis and

regression analysis.

In Chapter Four, the results of data analysis are explained and interpreted.


- 153 -

Chapter 4 Data Analysis

4.1 Introduction to Chapter Four

The methods of data collection and data analysis have been discussed in previous

chapters. Proceeding from Chapter Three, this chapter focuses on the analysis of data

with the intention of testing hypotheses formulated from a proposed theoretical

model.

There are seven sections in this chapter as indicated in Figure 4.1. The first section is

the introduction. Section two presents the profile of the sample from which data have

been collected. A cross tabulation technique is used to present the information.

Section three describes the development of constructs in this research and examines

the reliability of the data through Cronbach’s alpha technique.

In section four, factor analysis is used to reduce a large number of variables into

significant factors of online trading behaviour. From these factors, a revised set of

hypotheses is proposed. The factor scores are retained to form construct variables that

are used for multiple linear regression analysis. Regression analysis is used to test the

revised hypotheses. These regression statistics are represented in section five. The

results of hypothesis tests are explained in section six.

The last section summarises the topics covered in this chapter.


- 154 -

Figure 4-1 Overview of Chapter Four

4.1 Introduction to Chapter Four

4.2 Data Profile Examination 4.2.1 Data summary 4.2.2 Demographic profiles

4.4 Factor analysis4.4.1 Factor analysis for independent variables4.4.2 Reliability testing of factors (independent variables)4.4.3 Factor analysis for independent variables4.4.4 Reliability testing of factors (dependent variable)

4.5 Multiple regression analysis4.5.1 Multiple regression model4.5.2 Model summary – R Square

4.5.3 ANOVA Table 4.5.4 Model Parameters

4.6 Hypothesis testing 4.6.1 Test of hypothesis 1a 4.6.2 Test of hypothesis 1b 4.6.3 Test of hypothesis 2a 4.6.4 Test of hypothesis 2b

4.6.5 Test of hypothesis three 4.6.6 Test of hypothesis four 4.6.7 Test of hypothesis five 4.6.8 Test of hypothesis six 4.6.9 Test of hypothesis seven 4.6.10 Summary of hypothesis testing


4.3 Development of Constructs 4.3.1 Reliability Analysis

Source: Developed for this research


- 155 -

4.2 Data Profile Examination

4.2.1 Data Summary

There were 246 responses collected via the online survey questionnaires as explained

in Chapter Three. Out of the 246 data elements, 9 records are incomplete and 5

records are inaccurate. The inconsistent records are removed leaving 232 records as

the raw data for analysis.

4.2.2 Demographic Profiles

Data was collected on demographic variables. Presented below is a summary of the

demographic profile of the sample based on age, education, occupation, income,

marital status and gender.

Age

Of the total population (N = 232), 43.1 percent of valid respondents are of the age

range from 21 to 30, while 40.9 percent are aged 31 to 45. Just 12.9 percent of the

sample is 46 to 55 years old, and only 3.0 percent are 56 years old and above. In

summary, around 83 percent of the sample is aged 21 to 45 years. Chart 4.1

illustrates this age distribution.


- 156 -

Chart 4-1 Age profiles of respondents

Age

Age

56&up46-5531-4521-30

Freq

uenc

y

120

100

80

60

40

20

0

Source: Survey data

Education

Of the 232 respondents, only 5.2 percent reported having Secondary and below

education. 12.9 percent or 30 respondents had Junior College and equivalent level.

The second largest group of respondents is those with a Diploma and equivalent

which make up 28.0 percent. The largest group of respondents is those with a

Bachelor degree and equivalent which is 38.4 percent of the sample. There are 15.5

percent of respondents with a Master degree and above. Chart 4.2 illustrates the

education levels of the sample respondents. When Chart 4.1 and 4.2 are considered


- 157 -

together, it is apparent that the sample is generally below middle age and quite well

educated.

Chart 4-2 Education profiles of respondents

Education

Education

Master degree & abovBachelor degree & eq

Diploma & equivalentJunior College & equ

Secondary & below

Freq

uenc

y

100

80

60

40

20

0

Source: Survey data

Occupation

From the records gathered, 23.7 percent of people who responded are holding

executive and managerial positions. 62 professionals responded which is about 26.7

percent of the total population and it is the largest group of people. People who are

holding administration, sales and services related positions make up 18.1 percent,

while 11.2 percent are self –employed. Only 3.4 percent or 8 cases out of the total


- 158 -

population reported an occupation identified as ‘others’. Chart 4.3 illustrates the

occupation categories of the sample respondents.

Chart 4-3 Occupation profiles of respondents

Source: Survey data

Income

There were 11.6 percent of the respondents who reported an income of $20,000 and

below. The next income range, from $21,000 to $35,000, has 19.0 percent of

respondents. 30.6 percent of respondents had an income level of $35,000 to $50,000.

There are 19.0 percent of respondents with an income level range from $50,0001 to

$65,000. The next income range, $65,001 to $80,000 has 10.8 percent of respondents.

There are 4.3 percent of respondents belonging to an income range of $80,001 to


- 159 -

$100,000. There are only 4.7 percent of respondents with an income level more than

$100,001. Chart 4.4 illustrates the income distribution of the sample respondents with

a model income range of $30,000 to $50,000; the remainder of the sample is

reasonably even, though quite widely distributed around that.

Chart 4-4 Income profiles of respondents

Income

Income

100,001 & above80,001 - 100,000

65,001 - 80,00050,001 - 65,000

35,000 - 50,00021,000 - 35,000

20,000 & below

Freq

uenc

y

80

60

40

20

0

Source: Survey data

Marital Status

In the marital status factor, the single respondents are the largest group which

is 51.7 percent of the total sample population. Respondents who are married make up

47 percent. Only 1.3 percent respondents indicated ‘others’ in the marital status.

(refer to Chart 4.5)


- 160 -

Chart 4-5 Martial status of respondents

marital status

marital status

OthersMarriedSingle

Freq

uenc

y

140

120

100

80

60

40

20

0

Source: Survey data

Gender

From the data of 232 respondents, 72 percent are male while only 28 percent are

female. Chart 4.6 shows that more than twice as many males have responded to the

survey as compared to females.


- 161 -

Chart 4-6 Gender profiles of respondents

genderFemaleMale

Cou

nt

200

150

100

50

0

Source: Survey data

Based on the above demographic profiles, most of the respondents belong to the age

group of 21 to 30, with a Bachelor degree and equivalent, working as professionals

with an average income level of $35,000 to $50,000, and are single males.

The average age range of the respondents is between the ages of 31 to 45. The

general education level of the respondents is a Diploma and equivalent. The

respondents are from a broad spectrum of occupational categories.


- 162 -

4.3 Development of Constructs

The objective of this study is to test the relationship between the pre-adoption

variables and post-adoption behaviour of retail investors trading stocks online. The

data gathered are to study the research issues addressed in Chapter One:







As the research is based on Roger’s Diffusion Theory and Davis’s Technology

Acceptance Model as well as the inclusion of a Consumer loyalty factor, there are a

large number of variables built into the theoretical model as illustrated in Chapter

Two. The large number of variables, both dependent and independent, was reduced to

meaningful constructs through factor analysis. Cronbach’s alpha tests were used to

test the reliability of the constructs. A small number of construct variables reduces

the large number of dependent variables to a small set of variables that can be used

for further analysis. By ensuring that that the construct variables were reliable, the

researcher was able to have greater confidence in subsequent analysis. The new

construct variables were further analysed using regression analysis to test the

theoretical model.


- 163 -

4.3.1 Reliability Analysis

There are several types of reliability testing but the most commonly used is

Cronbach’s alpha coefficient (Coakes, Steed & Price 2008). The Cronbach’s alpha

coefficient is to test the internal consistency of the selected variables and is the

degree to which the items that make up the scale are all measuring the same

underlying attribute (Pallant 2007).

According to the definition from UCLA, Cronbach’s alpha illustrates how well a set

of variables measure a single unidimensional latent construct. Cronbach’s alpha is

usually low when the research data have a multidimensional structure. Cronbach’s

alpha is considered a coefficient of reliability and not really a statistical test (UCLA

2009a). The following formula shows the standardised Cronbach’s alpha where it can

be illustrated as a function of the number of test items and the average inter-

correlation among the items.

Where;

N = the number of items

C-bar = the average inter-item covariance among the items

V-bar = the average variance

The formula shows that if the number of items increases, the Cronbach’s alpha

increases. If the average inter-item correlation is low, Cronbach’ alpha will be low.

Cronbach’s alpha increases when the average inter-item correlation increases. If the


- 164 -

inter-item correlations are high, then it shows that the items are measuring the same

underlying construct. This is the meaning of high reliability of the data and is

referring to how well the items measure a single unidimensional latent construct

(UCLA 2009a).

The constructs derived from factor analysis in this research were tested on their

reliability using Cronbach’s alpha coefficient. This was done by selecting scale and

reliability analysis options from SPSS and the Alpha model was selected. The

reliability of the constructs is reported in the factor analysis section.

4.4 Factor Analysis

Zikmund (2000) described the purpose of factor analysis as to summarise the

information contained in a large number of variables into a smaller number of

constructs (Zikmund 2000). Hair (1998) describes factor analysis as a multivariate

technique that can be used to accommodate multiple variables and to understand their

complex relationships (Hair et al. 1998).

Factor analysis takes thousands of measurements and observations and resolves them

into regular patterns that can be interpreted by the researcher. (Rummel 1967)

In this study, factor analysis is used to filter out the significant constructs from large

groups of variables. The factor scores created will be used as the basis for further

regression analysis. Factor analysis was conducted for two groups of variables; the

dependent variables and independent variables. The data were initially screened using


- 165 -

a correlation matrix method and results were generated through a Principle

Component Analysis (PCA) technique

4.4.1 Factor Analysis of Independent Variables

Correlation Matrix

The first step of factor analysis is to study the correlation matrix of the variables. A

correlation matrix is simply a rectangular array of numbers which gives the

correlation coefficients between a single variable and every other variable in the

investigation. The correlation matrix contains useful knowledge and the researcher

can use it to study the relationships among the variables. Rummel defined the

coefficients of correlation as to express the degree of linear relationship between the

row and column variables of the matrix. If the coefficient is closer to zero, the lesser

the relationship between the variables; the closer to one, the stronger the relationship

between the variables. A negative sign indicates that the variables are inversely

related (Rummel 1967).

The correlation matrix made up of the coefficient values and significance test values

is illustrated in Table 4.1. The top half of this table contains the Pearson correlation

coefficient between all variables, and the bottom half of the table contains the one-

tailed significance of these coefficients. Significance test values less than 0.05

suggest that we can be 95 percent confident that the two variables are correlated.

The values in table 4.1 for the independent variables are scanned through to identify

those variables with significance test values that are all greater than 0.05 (Field 2005).


- 166 -

This would indicate that a particular variable was not significantly correlated to any

others. All of the independent variables have numerous instances where the

significance test value is less than 0.05. All variables are therefore accepted for

further analysis.


- 167 -

Table 4-1 Correlation matrix for independent variables

Relative Advantage

1

Relative Advantage

2 Compatibility1 Compatibility2 Complexity1 Complexity2 trailability1 trailability2 observability1 observability2Relative Advantage1 1.000 0.549 0.260 0.221 0.452 0.622 0.369 0.385 0.379 0.368

Relative Advantage2 0.549 1.000 0.298 0.217 0.370 0.387 0.117 0.093 0.254 0.256

Compatibility10.260 0.298 1.000 0.841 0.267 0.252 0.031 0.034 0.355 0.330

Compatibility20.221 0.217 0.841 1.000 0.282 0.237 0.012 0.038 0.382 0.381

Complexity10.452 0.370 0.267 0.282 1.000 0.695 0.329 0.335 0.454 0.479

Complexity2 0.622 0.387 0.252 0.237 0.695 1.000 0.497 0.445 0.451 0.458

trailability1 0.369 0.117 0.031 0.012 0.329 0.497 1.000 0.861 0.363 0.346

trailability20.385 0.093 0.034 0.038 0.335 0.445 0.861 1.000 0.382 0.368

observability10.379 0.254 0.355 0.382 0.454 0.451 0.363 0.382 1.000 0.932

observability20.368 0.256 0.330 0.381 0.479 0.458 0.346 0.368 0.932 1.000

optional decision1 0.176 0.173 0.098 0.103 0.152 0.098 0.142 0.180 0.149 0.155


collective decision1 0.112 0.143 -0.087 -0.128 0.032 0.057 0.179 0.193 0.174 0.188

collective decision2 0.148 0.125 -0.119 -0.091 0.058 0.133 0.287 0.203 0.158 0.181

authority decision1 0.059 0.157 0.215 0.206 0.056 0.023 0.079 0.089 0.294 0.282


mass media10.397 0.162 -0.145 -0.126 0.263 0.379 0.285 0.289 0.122 0.117

mass media20.358 0.196 -0.035 -0.061 0.219 0.313 0.181 0.212 0.095 0.099

interpersonal10.191 0.146 -0.093 -0.161 0.122 0.181 0.228 0.204 -0.028 -0.054

interpersonal20.224 0.149 -0.126 -0.124 0.209 0.253 0.279 0.268 0.139 0.146

norm1 0.302 0.073 0.155 0.174 0.141 0.246 0.342 0.352 0.264 0.260norm2 0.065 0.239 0.354 0.380 0.083 -0.009 -0.088 -0.026 0.214 0.183degree of network1 0.257 0.075 0.082 0.101 0.176 0.257 0.320 0.334 0.216 0.244

degree of network2 0.302 0.090 0.129 0.125 0.268 0.320 0.348 0.369 0.240 0.251

Change Agent Efforts1 0.315 0.087 -0.019 0.003 0.291 0.402 0.392 0.334 0.155 0.164

Change Agent Efforts2 0.243 0.148 0.051 0.069 0.281 0.223 0.269 0.328 0.209 0.203


Perceived usefulness1 0.400 0.283 0.227 0.215 0.506 0.505 0.326 0.294 0.402 0.425



Loyalty1 0.461 0.365 0.423 0.450 0.418 0.430 0.262 0.275 0.489 0.477Loyalty2 0.279 0.396 0.401 0.404 0.365 0.375 0.121 0.163 0.397 0.408Loyalty3 0.283 0.277 0.264 0.299 0.327 0.299 0.115 0.136 0.300 0.285Confidence1 0.531 0.321 0.082 0.123 0.470 0.580 0.384 0.401 0.370 0.384

Confidence20.485 0.260 0.101 0.185 0.456 0.521 0.388 0.414 0.332 0.368

Confidence30.251 0.358 0.378 0.416 0.352 0.271 0.069 0.070 0.369 0.399

Relative Advantage1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Relative Advantage2 0.000 0.000 0.000 0.000 0.000 0.038 0.079 0.000 0.000

Compatibility10.000 0.000 0.000 0.000 0.000 0.319 0.305 0.000 0.000

Compatibility20.000 0.000 0.000 0.000 0.000 0.426 0.284 0.000 0.000

Complexity10.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Complexity20.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

trailability10.000 0.038 0.319 0.426 0.000 0.000 0.000 0.000 0.000

trailability2 0.000 0.079 0.305 0.284 0.000 0.000 0.000 0.000 0.000

observability1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

observability20.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000



collective decision1 0.044 0.015 0.093 0.026 0.315 0.195 0.003 0.002 0.004 0.002

collective decision2 0.012 0.028 0.035 0.085 0.189 0.022 0.000 0.001 0.008 0.003



mass media1 0.000 0.007 0.014 0.027 0.000 0.000 0.000 0.000 0.032 0.038

mass media20.000 0.001 0.296 0.178 0.000 0.000 0.003 0.001 0.074 0.067

interpersonal10.002 0.013 0.078 0.007 0.031 0.003 0.000 0.001 0.334 0.208

interpersonal20.000 0.012 0.028 0.030 0.001 0.000 0.000 0.000 0.017 0.013

norm1 0.000 0.134 0.009 0.004 0.016 0.000 0.000 0.000 0.000 0.000norm2 0.161 0.000 0.000 0.000 0.103 0.448 0.091 0.349 0.001 0.003degree of network1 0.000 0.128 0.106 0.063 0.004 0.000 0.000 0.000 0.000 0.000

degree of network2 0.000 0.085 0.025 0.029 0.000 0.000 0.000 0.000 0.000 0.000







Loyalty1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Loyalty2 0.000 0.000 0.000 0.000 0.000 0.000 0.033 0.006 0.000 0.000Loyalty3 0.000 0.000 0.000 0.000 0.000 0.000 0.041 0.019 0.000 0.000Confidence1

0.000 0.000 0.108 0.030 0.000 0.000 0.000 0.000 0.000 0.000

Confidence2 0.000 0.000 0.062 0.002 0.000 0.000 0.000 0.000 0.000 0.000

Confidence30.000 0.000 0.000 0.000 0.000 0.000 0.148 0.143 0.000 0.000

Correlation

Sig. (1-tailed)

Correlation Matrix(a)

a. Determinant = 3.82E-012


- 168 -

optional decision1

optional decision2

collective decision1

collective decision2

authority decision1

authority decision2

mass media1

mass media2 interpersonal1 interpersonal2

Relative Advantage1 0.176 0.170 0.112 0.148 0.059 0.072 0.397 0.358 0.191 0.224Relative Advantage2 0.173 0.116 0.143 0.125 0.157 0.197 0.162 0.196 0.146 0.149Compatibility1 0.098 0.157 -0.087 -0.119 0.215 0.007 -0.145 -0.035 -0.093 -0.126Compatibility2 0.103 0.130 -0.128 -0.091 0.206 0.004 -0.126 -0.061 -0.161 -0.124Complexity1 0.152 0.191 0.032 0.058 0.056 0.115 0.263 0.219 0.122 0.209Complexity2 0.098 0.155 0.057 0.133 0.023 0.128 0.379 0.313 0.181 0.253trailability1 0.142 0.028 0.179 0.287 0.079 0.197 0.285 0.181 0.228 0.279trailability2 0.180 0.054 0.193 0.203 0.089 0.170 0.289 0.212 0.204 0.268observability1 0.149 0.270 0.174 0.158 0.294 0.236 0.122 0.095 -0.028 0.139observability2 0.155 0.253 0.188 0.181 0.282 0.251 0.117 0.099 -0.054 0.146optional decision1 1.000 0.545 0.258 0.238 0.205 0.155 0.088 0.101 0.229 0.139optional decision2 0.545 1.000 0.231 0.291 0.141 0.169 0.190 0.201 0.181 0.155collective decision1 0.258 0.231 1.000 0.712 0.382 0.454 0.202 0.120 0.286 0.387collective decision2 0.238 0.291 0.712 1.000 0.271 0.368 0.173 0.143 0.268 0.415authority decision1 0.205 0.141 0.382 0.271 1.000 0.660 0.046 -0.079 0.193 0.162authority decision2 0.155 0.169 0.454 0.368 0.660 1.000 0.240 0.074 0.182 0.252mass media1 0.088 0.190 0.202 0.173 0.046 0.240 1.000 0.579 0.321 0.303mass media2 0.101 0.201 0.120 0.143 -0.079 0.074 0.579 1.000 0.401 0.283interpersonal1 0.229 0.181 0.286 0.268 0.193 0.182 0.321 0.401 1.000 0.622interpersonal2 0.139 0.155 0.387 0.415 0.162 0.252 0.303 0.283 0.622 1.000norm1 0.179 0.079 0.293 0.286 0.219 0.220 0.289 0.125 0.273 0.345norm2 0.233 0.206 0.218 0.166 0.280 0.164 -0.055 0.080 0.122 0.086degree of network1 0.187 0.194 0.370 0.427 0.158 0.243 0.247 0.195 0.394 0.441degree of network2 0.248 0.134 0.366 0.378 0.132 0.169 0.218 0.122 0.390 0.468Change Agent Efforts1 0.161 0.251 0.197 0.293 0.060 0.199 0.407 0.311 0.275 0.286Change Agent Efforts2 0.250 0.266 0.225 0.230 0.320 0.333 0.290 0.200 0.319 0.257Change Agent Efforts3 0.227 0.209 0.232 0.244 0.307 0.301 0.181 0.092 0.317 0.236Perceived usefulness1 0.159 0.260 0.225 0.270 0.258 0.271 0.233 0.175 0.235 0.278Perceived usefulness2 0.185 0.218 0.163 0.172 0.271 0.257 0.215 0.161 0.218 0.272Perceived usefulness3 0.166 0.250 0.182 0.294 0.104 0.196 0.334 0.229 0.213 0.339Loyalty1 0.202 0.237 0.109 0.135 0.260 0.182 0.195 0.250 0.062 0.128Loyalty2 0.105 0.217 0.107 0.066 0.187 0.145 -0.037 0.112 0.016 0.138Loyalty3 0.060 0.030 0.019 0.036 0.154 0.084 0.136 0.140 -0.015 0.028Confidence1 0.140 0.123 0.213 0.262 0.049 0.098 0.404 0.253 0.166 0.307Confidence2 0.066 0.218 0.224 0.256 0.055 0.142 0.356 0.288 0.230 0.338Confidence3 0.184 0.149 0.203 0.125 0.260 0.172 0.020 0.056 -0.058 0.099Relative Advantage1 0.004 0.005 0.044 0.012 0.185 0.137 0.000 0.000 0.002 0.000Relative Advantage2 0.004 0.039 0.015 0.028 0.008 0.001 0.007 0.001 0.013 0.012Compatibility1 0.068 0.008 0.093 0.035 0.000 0.459 0.014 0.296 0.078 0.028Compatibility2 0.059 0.024 0.026 0.085 0.001 0.474 0.027 0.178 0.007 0.030Complexity1 0.010 0.002 0.315 0.189 0.197 0.040 0.000 0.000 0.031 0.001Complexity2 0.069 0.009 0.195 0.022 0.362 0.025 0.000 0.000 0.003 0.000trailability1 0.015 0.336 0.003 0.000 0.114 0.001 0.000 0.003 0.000 0.000trailability2 0.003 0.205 0.002 0.001 0.087 0.005 0.000 0.001 0.001 0.000observability1 0.012 0.000 0.004 0.008 0.000 0.000 0.032 0.074 0.334 0.017observability2 0.009 0.000 0.002 0.003 0.000 0.000 0.038 0.067 0.208 0.013optional decision1 0.000 0.000 0.000 0.001 0.009 0.092 0.063 0.000 0.017optional decision2 0.000 0.000 0.000 0.016 0.005 0.002 0.001 0.003 0.009collective decision1 0.000 0.000 0.000 0.000 0.000 0.001 0.034 0.000 0.000collective decision2 0.000 0.000 0.000 0.000 0.000 0.004 0.015 0.000 0.000authority decision1 0.001 0.016 0.000 0.000 0.000 0.240 0.116 0.002 0.007authority decision2 0.009 0.005 0.000 0.000 0.000 0.000 0.132 0.003 0.000mass media1 0.092 0.002 0.001 0.004 0.240 0.000 0.000 0.000 0.000mass media2 0.063 0.001 0.034 0.015 0.116 0.132 0.000 0.000 0.000interpersonal1 0.000 0.003 0.000 0.000 0.002 0.003 0.000 0.000 0.000interpersonal2 0.017 0.009 0.000 0.000 0.007 0.000 0.000 0.000 0.000norm1 0.003 0.115 0.000 0.000 0.000 0.000 0.000 0.028 0.000 0.000norm2 0.000 0.001 0.000 0.006 0.000 0.006 0.203 0.111 0.031 0.095degree of network1 0.002 0.001 0.000 0.000 0.008 0.000 0.000 0.001 0.000 0.000degree of network2 0.000 0.021 0.000 0.000 0.022 0.005 0.000 0.032 0.000 0.000Change Agent Efforts1 0.007 0.000 0.001 0.000 0.183 0.001 0.000 0.000 0.000 0.000Change Agent Efforts2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001 0.000 0.000Change Agent Efforts3 0.000 0.001 0.000 0.000 0.000 0.000 0.003 0.082 0.000 0.000Perceived usefulness1 0.008 0.000 0.000 0.000 0.000 0.000 0.000 0.004 0.000 0.000Perceived usefulness2 0.002 0.000 0.006 0.004 0.000 0.000 0.000 0.007 0.000 0.000Perceived usefulness3 0.006 0.000 0.003 0.000 0.057 0.001 0.000 0.000 0.001 0.000Loyalty1 0.001 0.000 0.049 0.020 0.000 0.003 0.001 0.000 0.173 0.026Loyalty2 0.056 0.000 0.052 0.159 0.002 0.014 0.286 0.045 0.403 0.018Loyalty3 0.182 0.325 0.384 0.291 0.009 0.101 0.019 0.016 0.410 0.335Confidence1 0.017 0.031 0.001 0.000 0.227 0.068 0.000 0.000 0.006 0.000Confidence2 0.159 0.000 0.000 0.000 0.204 0.015 0.000 0.000 0.000 0.000Confidence3 0.002 0.011 0.001 0.029 0.000 0.004 0.381 0.199 0.192 0.067

Correlation

Sig. (1-tailed)


- 169 -

norm1 norm2degree

of network1degree

of network2Change Agent

Efforts1Change Agent

Efforts2Change Agent

Efforts3Perceived

usefulness1Perceived

usefulness2Perceived

usefulness3Relative Advantage1 0.302 0.065 0.257 0.302 0.315 0.243 0.203 0.400 0.412 0.426Relative Advantage2 0.073 0.239 0.075 0.090 0.087 0.148 0.182 0.283 0.308 0.139Compatibility1 0.155 0.354 0.082 0.129 -0.019 0.051 0.146 0.227 0.261 0.107Compatibility2 0.174 0.380 0.101 0.125 0.003 0.069 0.180 0.215 0.272 0.158Complexity1 0.141 0.083 0.176 0.268 0.291 0.281 0.265 0.506 0.536 0.415Complexity2 0.246 -0.009 0.257 0.320 0.402 0.223 0.251 0.505 0.509 0.435trailability1 0.342 -0.088 0.320 0.348 0.392 0.269 0.277 0.326 0.313 0.294trailability2 0.352 -0.026 0.334 0.369 0.334 0.328 0.331 0.294 0.305 0.306observability1 0.264 0.214 0.216 0.240 0.155 0.209 0.189 0.402 0.398 0.302observability2 0.260 0.183 0.244 0.251 0.164 0.203 0.192 0.425 0.415 0.298optional decision1 0.179 0.233 0.187 0.248 0.161 0.250 0.227 0.159 0.185 0.166optional decision2 0.079 0.206 0.194 0.134 0.251 0.266 0.209 0.260 0.218 0.250collective decision1 0.293 0.218 0.370 0.366 0.197 0.225 0.232 0.225 0.163 0.182collective decision2 0.286 0.166 0.427 0.378 0.293 0.230 0.244 0.270 0.172 0.294authority decision1 0.219 0.280 0.158 0.132 0.060 0.320 0.307 0.258 0.271 0.104authority decision2 0.220 0.164 0.243 0.169 0.199 0.333 0.301 0.271 0.257 0.196mass media1 0.289 -0.055 0.247 0.218 0.407 0.290 0.181 0.233 0.215 0.334mass media2 0.125 0.080 0.195 0.122 0.311 0.200 0.092 0.175 0.161 0.229interpersonal1 0.273 0.122 0.394 0.390 0.275 0.319 0.317 0.235 0.218 0.213interpersonal2 0.345 0.086 0.441 0.468 0.286 0.257 0.236 0.278 0.272 0.339norm1 1.000 0.379 0.582 0.541 0.352 0.251 0.287 0.331 0.276 0.385norm2 0.379 1.000 0.315 0.236 0.007 0.113 0.162 0.174 0.170 0.133degree of network1 0.582 0.315 1.000 0.782 0.304 0.331 0.351 0.333 0.366 0.462degree of network2 0.541 0.236 0.782 1.000 0.340 0.275 0.291 0.336 0.362 0.445Change Agent Efforts1 0.352 0.007 0.304 0.340 1.000 0.620 0.601 0.407 0.360 0.438Change Agent Efforts2 0.251 0.113 0.331 0.275 0.620 1.000 0.838 0.414 0.377 0.335Change Agent Efforts3 0.287 0.162 0.351 0.291 0.601 0.838 1.000 0.413 0.398 0.322Perceived usefulness1 0.331 0.174 0.333 0.336 0.407 0.414 0.413 1.000 0.824 0.655Perceived usefulness2 0.276 0.170 0.366 0.362 0.360 0.377 0.398 0.824 1.000 0.672Perceived usefulness3 0.385 0.133 0.462 0.445 0.438 0.335 0.322 0.655 0.672 1.000Loyalty1 0.299 0.381 0.310 0.290 0.232 0.314 0.303 0.496 0.497 0.479Loyalty2 0.166 0.316 0.186 0.239 0.150 0.171 0.232 0.357 0.400 0.265Loyalty3 0.062 0.274 0.137 0.120 0.030 0.176 0.171 0.210 0.198 0.190Confidence1 0.403 0.043 0.346 0.386 0.383 0.254 0.273 0.363 0.311 0.379Confidence2 0.334 -0.025 0.422 0.371 0.387 0.284 0.297 0.391 0.444 0.458Confidence3 0.237 0.390 0.210 0.216 0.008 0.123 0.150 0.277 0.272 0.165Relative Advantage1 0.000 0.161 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000Relative Advantage2 0.134 0.000 0.128 0.085 0.094 0.012 0.003 0.000 0.000 0.017Compatibility1 0.009 0.000 0.106 0.025 0.384 0.219 0.013 0.000 0.000 0.053Compatibility2 0.004 0.000 0.063 0.029 0.480 0.147 0.003 0.000 0.000 0.008Complexity1 0.016 0.103 0.004 0.000 0.000 0.000 0.000 0.000 0.000 0.000Complexity2 0.000 0.448 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000trailability1 0.000 0.091 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000trailability2 0.000 0.349 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000observability1 0.000 0.001 0.000 0.000 0.009 0.001 0.002 0.000 0.000 0.000observability2 0.000 0.003 0.000 0.000 0.006 0.001 0.002 0.000 0.000 0.000optional decision1 0.003 0.000 0.002 0.000 0.007 0.000 0.000 0.008 0.002 0.006optional decision2 0.115 0.001 0.001 0.021 0.000 0.000 0.001 0.000 0.000 0.000collective decision1 0.000 0.000 0.000 0.000 0.001 0.000 0.000 0.000 0.006 0.003collective decision2 0.000 0.006 0.000 0.000 0.000 0.000 0.000 0.000 0.004 0.000authority decision1 0.000 0.000 0.008 0.022 0.183 0.000 0.000 0.000 0.000 0.057authority decision2 0.000 0.006 0.000 0.005 0.001 0.000 0.000 0.000 0.000 0.001mass media1 0.000 0.203 0.000 0.000 0.000 0.000 0.003 0.000 0.000 0.000mass media2 0.028 0.111 0.001 0.032 0.000 0.001 0.082 0.004 0.007 0.000interpersonal1 0.000 0.031 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.001interpersonal2 0.000 0.095 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000norm1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000norm2 0.000 0.000 0.000 0.457 0.043 0.007 0.004 0.005 0.022degree of network1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000degree of network2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Change Agent Efforts1 0.000 0.457 0.000 0.000 0.000 0.000 0.000 0.000 0.000Change Agent Efforts2 0.000 0.043 0.000 0.000 0.000 0.000 0.000 0.000 0.000Change Agent Efforts3 0.000 0.007 0.000 0.000 0.000 0.000 0.000 0.000 0.000Perceived usefulness1 0.000 0.004 0.000 0.000 0.000 0.000 0.000 0.000 0.000Perceived usefulness2 0.000 0.005 0.000 0.000 0.000 0.000 0.000 0.000 0.000Perceived usefulness3 0.000 0.022 0.000 0.000 0.000 0.000 0.000 0.000 0.000Loyalty1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Loyalty2 0.006 0.000 0.002 0.000 0.011 0.004 0.000 0.000 0.000 0.000Loyalty3 0.173 0.000 0.018 0.034 0.327 0.004 0.004 0.001 0.001 0.002Confidence1 0.000 0.255 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Confidence2 0.000 0.353 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000Confidence3 0.000 0.000 0.001 0.000 0.454 0.031 0.011 0.000 0.000 0.006

Correlation

Sig. (1-tailed)


- 170 -

Loyalty1 Loyalty2 Loyalty3 Confidence1 Confidence2 Confidence3Relative Advantage1 0.461 0.279 0.283 0.531 0.485 0.251Relative Advantage2 0.365 0.396 0.277 0.321 0.260 0.358Compatibility1 0.423 0.401 0.264 0.082 0.101 0.378Compatibility2 0.450 0.404 0.299 0.123 0.185 0.416Complexity1 0.418 0.365 0.327 0.470 0.456 0.352Complexity2 0.430 0.375 0.299 0.580 0.521 0.271trailability1 0.262 0.121 0.115 0.384 0.388 0.069trailability2 0.275 0.163 0.136 0.401 0.414 0.070observability1 0.489 0.397 0.300 0.370 0.332 0.369observability2 0.477 0.408 0.285 0.384 0.368 0.399optional decision1 0.202 0.105 0.060 0.140 0.066 0.184optional decision2 0.237 0.217 0.030 0.123 0.218 0.149collective decision1 0.109 0.107 0.019 0.213 0.224 0.203collective decision2 0.135 0.066 0.036 0.262 0.256 0.125authority decision1 0.260 0.187 0.154 0.049 0.055 0.260authority decision2 0.182 0.145 0.084 0.098 0.142 0.172mass media1 0.195 -0.037 0.136 0.404 0.356 0.020mass media2 0.250 0.112 0.140 0.253 0.288 0.056interpersonal1 0.062 0.016 -0.015 0.166 0.230 -0.058interpersonal2 0.128 0.138 0.028 0.307 0.338 0.099norm1 0.299 0.166 0.062 0.403 0.334 0.237norm2 0.381 0.316 0.274 0.043 -0.025 0.390degree of network1 0.310 0.186 0.137 0.346 0.422 0.210degree of network2 0.290 0.239 0.120 0.386 0.371 0.216Change Agent Efforts1

0.232 0.150 0.030 0.383 0.387 0.008

Change Agent Efforts20.314 0.171 0.176 0.254 0.284 0.123


Perceived usefulness1 0.496 0.357 0.210 0.363 0.391 0.277

Perceived usefulness20.497 0.400 0.198 0.311 0.444 0.272


Loyalty1 1.000 0.754 0.587 0.419 0.419 0.513Loyalty2 0.754 1.000 0.526 0.340 0.451 0.493Loyalty3 0.587 0.526 1.000 0.390 0.248 0.344Confidence1 0.419 0.340 0.390 1.000 0.664 0.467Confidence2 0.419 0.451 0.248 0.664 1.000 0.391Confidence3 0.513 0.493 0.344 0.467 0.391 1.000Relative Advantage1 0.000 0.000 0.000 0.000 0.000 0.000Relative Advantage2 0.000 0.000 0.000 0.000 0.000 0.000Compatibility1 0.000 0.000 0.000 0.108 0.062 0.000Compatibility2 0.000 0.000 0.000 0.030 0.002 0.000Complexity1 0.000 0.000 0.000 0.000 0.000 0.000Complexity2 0.000 0.000 0.000 0.000 0.000 0.000trailability1 0.000 0.033 0.041 0.000 0.000 0.148trailability2 0.000 0.006 0.019 0.000 0.000 0.143observability1 0.000 0.000 0.000 0.000 0.000 0.000observability2 0.000 0.000 0.000 0.000 0.000 0.000optional decision1 0.001 0.056 0.182 0.017 0.159 0.002optional decision2 0.000 0.000 0.325 0.031 0.000 0.011collective decision1 0.049 0.052 0.384 0.001 0.000 0.001collective decision2 0.020 0.159 0.291 0.000 0.000 0.029authority decision1 0.000 0.002 0.009 0.227 0.204 0.000authority decision2 0.003 0.014 0.101 0.068 0.015 0.004mass media1 0.001 0.286 0.019 0.000 0.000 0.381mass media2 0.000 0.045 0.016 0.000 0.000 0.199interpersonal1 0.173 0.403 0.410 0.006 0.000 0.192interpersonal2 0.026 0.018 0.335 0.000 0.000 0.067norm1 0.000 0.006 0.173 0.000 0.000 0.000norm2 0.000 0.000 0.000 0.255 0.353 0.000degree of network1 0.000 0.002 0.018 0.000 0.000 0.001degree of network2 0.000 0.000 0.034 0.000 0.000 0.000Change Agent Efforts1

0.000 0.011 0.327 0.000 0.000 0.454


Change Agent Efforts3 0.000 0.000 0.004 0.000 0.000 0.011



Perceived usefulness3 0.000 0.000 0.002 0.000 0.000 0.006

Loyalty1 0.000 0.000 0.000 0.000 0.000Loyalty2 0.000 0.000 0.000 0.000 0.000Loyalty3 0.000 0.000 0.000 0.000 0.000Confidence1 0.000 0.000 0.000 0.000 0.000Confidence2 0.000 0.000 0.000 0.000 0.000Confidence3 0.000 0.000 0.000 0.000 0.000

Correlation

Sig. (1-tailed)



- 171 -

KMO and Bartlett’s Test

The next test of factor analysis is the Kaiser-Meyer-Olkin (KMO) and Bartlett's test.

The Kaiser-Meyer-Olkin (KMO) is used to test the adequacy of using factor analysis

for the variables (Pallant 2007). A value close to 0 indicates that the correlations of

the variables are far apart and not appropriate for factor analysis. A value close to 1

indicates that the correlations of the variables are relatively compact and deemed

appropriate for factor analysis. The KMO value should be greater than 0.5 for a

satisfactory factor analysis to proceed. (Field 2005).

For the independent variables collected for this research, the KMO value is 0.814

(See Table 4.2) which is close to 1 and indicates that the correlations of the variables

are relatively compact and deemed to be fit for factor analysis (Field 2005).

The Barlett’s test coefficient is to test the significant of the correlations of the

variables. If the significance test value is less than 0.01, it indicates that the

correlations of variables are significant. In this case, the value is 0.00 (P < 0.01) and

hence indicates that factor analysis is appropriate and significant (Field 2005).


- 172 -

Table 4-2 KMO and Bartlett’s Test for dependent variables

KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of

Sampling Adequacy. .814

Approx. Chi-

Square

5735

.485

df 630

Bartlett's Test

of Sphericity

Sig. .000


Communalities Testing

The next item from the factor analysis output is a table of communalities which

shows how much of the variance in the variables has been accounted for by the

extracted factors. Communality is the proportion of common variance within a

variable. The extraction method is based on the Principal Component Analysis which

is based on the assumption that all variance is common and communalities value

should be 1 before extraction (Field 2005). For extraction values lesser than 0.3, they

are considered to not fit well with other items in its component. A low value lesser

than 0.3 indicates that the variable should be omitted from the analysis (Pallant 2007).

All the values for the independent variables are above 0.3 (see Table 4.2) and none of

the variables need be excluded.


- 173 -

Table 4-3 Communalities for independent variables

Communalities

Initial Extraction Relative Advantage1 1.000 .637 Relative Advantage2 1.000 .553 Compatibility1 1.000 .772 Compatibility2 1.000 .762 Complexity1 1.000 .614 Complexity2 1.000 .730 trailability1 1.000 .788 trailability2 1.000 .790 observability1 1.000 .770 observability2 1.000 .776 optional decision1 1.000 .668 optional decision2 1.000 .806 collective decision1 1.000 .725 collective decision2 1.000 .667 authority decision1 1.000 .774 authority decision2 1.000 .732 mass media1 1.000 .601 mass media2 1.000 .635 interpersonal1 1.000 .666 interpersonal2 1.000 .581 norm1 1.000 .637 norm2 1.000 .650 degree of network1 1.000 .764 degree of network2 1.000 .747 Change Agent Efforts1 1.000 .698 Change Agent Efforts2 1.000 .854 Change Agent Efforts3 1.000 .849 Perceived usefulness1 1.000 .805 Perceived usefulness2 1.000 .839 Perceived usefulness3 1.000 .727 Loyalty1 1.000 .720 Loyalty2 1.000 .673 Loyalty3 1.000 .587 Confidence1 1.000 .729 Confidence2 1.000 .656 Confidence3 1.000 .620

Extraction Method: Principal Component Analysis.



- 174 -

Total Variances

The next item of the factor analysis shows all of the factors extractable from the

analysis along with their Eigenvalues, the percent of variance attributable to each

factor, and the cumulative variance of the factor and the previous factors.

The total variances are explained using the Principal Components Analysis extraction

method for the independent variables. Principal Components Analysis found nine

components with Eigenvalues that are greater than 1 (See Table 4.4). If nine factors

were extracted, then 71 percent of the variance would be explained.

Table 4-4 Total variance explained for independent variables

Total Variance Explained

10.325 28.681 28.681 10.325 28.681 28.681 3.847 10.686 10.6863.808 10.578 39.259 3.808 10.578 39.259 3.530 9.805 20.4922.690 7.473 46.732 2.690 7.473 46.732 3.309 9.191 29.6831.758 4.884 51.617 1.758 4.884 51.617 3.080 8.555 38.2381.694 4.706 56.323 1.694 4.706 56.323 2.603 7.230 45.4681.606 4.460 60.783 1.606 4.460 60.783 2.526 7.016 52.4841.327 3.685 64.468 1.327 3.685 64.468 2.450 6.806 59.2901.266 3.517 67.986 1.266 3.517 67.986 2.440 6.777 66.0671.127 3.132 71.118 1.127 3.132 71.118 1.818 5.051 71.118

.998 2.773 73.891

.928 2.578 76.469

.838 2.329 78.797

.736 2.045 80.842

.706 1.961 82.804

.628 1.746 84.549

.598 1.662 86.212

.523 1.452 87.664

.504 1.400 89.064

.411 1.143 90.207

.389 1.079 91.286

.346 .961 92.247

.329 .915 93.162

.313 .870 94.032

.275 .764 94.796

.257 .714 95.510

.238 .661 96.171

.231 .642 96.813

.213 .593 97.406

.185 .513 97.919

.157 .435 98.354

.134 .373 98.727

.123 .341 99.068

.106 .295 99.362

.096 .266 99.628

.078 .217 99.845

.056 .155 100.000

Component123456789101112131415161718192021222324252627282930313233343536

Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %Initial Eigenvalues Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings




- 175 -

Scree Plot

The Eigenvalues extraction in the previous section indicated that nine factors have

been identified and this can be further verified by using a Scree Plot in which the

Eigenvalues are plot from the largest value to the smallest value. In this Scree Plot,

there seemed to be a few bends, the bends occurred at component three, six and seven.

However, as shown in chart 4.7, the slope seemed to stop changing significantly after

the seven components. The Scree Plot does not appear to be very useful for

interpretation, so it was decided to use nine components for further factor analysis

testing as indicated by choosing factors with an Eigenvalue greater than 1. Some

additional factor analysis tests were done using three factors, six factors and seven

factors extraction, some failed to have coverage in rotation, and some did not provide

results that are meaningful in the context of the analysis. As a result, the nine factors

extraction fell into a pattern that has the most meaningful interpretation. However,

two of the components in the theoretical model have broken into two factors and are

explained in the next section.


- 176 -

Chart 4-7 Scree Plot for independent variables


Rotated Factor Matrix

A common method of rotation is Varimax, which is an orthogonal method of rotation,

in which the loadings for the factors are given by the projections of each plotted point

onto the new axes specified by the rotation. The Varimax criterion essentially drives

squared loadings towards the end of the range 0 to 1, and negative loadings towards -

1, 0 or 1 and away from intermediate values (Hair et al. 1998). In an orthogonally

rotated factor matrix, factors are uncorrelated (Rummel 1967). This makes them

more suitable for use as a basis for creating independent variables for use in multiple


- 177 -

regression analysis, as an underlying assumption in regression analysis is that the

independent variables are not correlated with each other. When any independent

variables in a multiple regression are highly correlated a potential problem of

muticollinearity exists (Zikmund 2000). In this research, orthogonally rotation is

selected as an appropriate method. Factor scores are to be used to create construct

variables for regression and it is desirable that these new construct variables should

not be correlated with each other. Regression variable created from factor scores

computed from factor loadings on each orthogonal component resolve the issues of

multicollinearity in multiple regeression (SPSS 2000).

This research will suppress the variables with absolute values less than 0.5 and those

sorted for earlier interpretation (Field 2005). The loadings of the factor analysis are

sorted and the report has suppressed the output with values less than 0.5. The

suppression of the output with values less than 0.5 has facilitated a better

interpretation of the results.

Table 4.5 shows a Varimax rotated factor matrix. The items in the matrix have been

sorted so that items (variables) that load onto a factor are grouped together, listed in

decreasing order of factor loading scores. All factor loading scores less than 0.5 have

been suppressed from this table to enable ease of factor interpretation.


- 178 -

Table 4-5 Rotated component matrix for independent variables

Rotated Component Matrixa

.739

.722

.704

.671

.601

.511 .818 .799 .706 .542 .823 .793 .714 .534 .796 .792 .608 .599 .763 .670 .580 .527 .817 .772 .524 .844 .843 .669 .823 .812 .578 .856 .759

Loyalty2Loyalty3Confidence3Loyalty1Confidence1Confidence2degree of network1degree of network2norm1interpersonal2collective decision2Perceived usefulness2Perceived usefulness1Perceived usefulness3Complexity1trailability1trailability2observability1observability2Complexity2mass media2mass media1interpersonal1Relative Advantage1Relative Advantage2Compatibility1Compatibility2norm2Change Agent Efforts3Change Agent Efforts2Change Agent Efforts1authority decision1authority decision2collective decision1optional decision2optional decision1

1 2 3 4 5 6 7 8 9Component

Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.

Rotation converged in 8 iterations.a.


We can see from Table 4.5 that nine factors have been extracted as a result of the

rotation of the variables using Principal Component Analysis. The variables which

lead onto factor 1 are Loyaly1, Loyalty2, Loyalty3, Confidence1, Confidence2 and


- 179 -

Confidence3. This construct is closely related to the consumer loyalty variable in the

theoretical model described in Chapter Two and is associated to hypothesis 7.

Factor 2 consists of the variables: Degree of network1, Degree of network2, Norm1

and Interpersonal2. This is similar to the Nature of the Social System variable but

with the addition of the Interpersonal component that can be found in the

Communication channels variable of the proposed model. It can be expected that

Social System is closely related to Communication channel and so it is reasonable

that this variable may cross over. Thus, this can be considered as a revision to the

proposed model and is related to hypothesis 4.

In the factor 3, the variables Perceived usefulness1, Perceived usefulness2, Perceived

usefulness3 and Complexity1 are extracted. The construct is closely related to the

Technology Acceptance Model variable with the addition of the Complexity1

variable and linked to hypothesis 6. The question linked to Complexity 1, “Easier to

access investment information online” (See Appendix C) is closely related to the

Perceived usefulness variable.

For factor 4, the extracted variables are Trailability1, Trailability2, Observability1

and Observability2. This construct can be linked to the Perceived attribute of

innovations variable but did not include all the components in hypothesis one. Thus,

hypothesis one is divided into two portions and a newly named hypothesis, 1a, is

related to this factor.

The variables Mass Media1, Mass Media2, Interpersonal1 and Relative advantage1

loaded onto factor 5. This is similar to the Communication channels but with the

addition of the Relative advantage1 element of the Perceived attributes of innovations


- 180 -

variable. The question associated to Relative advantage1, “Quicker to trade online”

(See Appendix C) could be viewed as an element closely related to Communication

channels. The communication channel could act as a catalyst in Quicker to trade

online. This factor is thus linked to hypothesis 3.

In factor 6, the variables are Compatibility1, Compatibility2 and Norm2. These

variables can be found in the group of Perceived attributes of innovations variables

with the addition of the variable Norm2 of the Nature of the social system variable.

The question linked to variable Norm2 “I feel left out if I do not sign up to Online

Securities Trading” (See Appendix C) can be viewed as an element related to

Perceived attributes of innovations. Thus, this factor forms the second portion of the

hypothesis 1 and is named as Hypothesis1b.

The variables Change agent efforts1, Change agent effort2 and Change agent effort3

load onto factor 7. This construct is closely related to the Extent of change agent’s

promotion efforts variable in the model which is also hypothesis 5 in this study.

In factor 8, the variables are Authority decision1, Authority decision2 and Collective

decision1. This construct has some variables as the Type of Innovation-Decision

variable of the model but did not consist of all the variables and is related to

hypothesis 2 in this study. Thus, hypothesis 2 has been divided in two portions and

this construct is named as Hypothesis 2a.

The last extracted factor 9 consists of the variables Optional decision1 and Optional

decision2. Similar to component 8, this construct is related to the Type of Innovation-

Decision variable in the model as described in Chapter Two. This formed the second

portion of hypothesis 2 and was renamed as Hypothesis 2b.


- 181 -

The extracted nine components in the rotated factor matrix in Table 4.5 overall can

be accepted as the revised theoretical model and as a basis for regression analysis in

order to test the relationship between these nine independent factors and the

dependent variable which is the post-adoption usage behaviour of trading stock

online. The nine extracted components are closely related to the seven hypotheses in

Chapter Two with the exception that hypothesis 1 and hypothesis 2 have been broken

into two portions respectively.

4.4.2 Reliability Testing of Factors (Independent Variables)

This section explores the variables extracted in component 1 and whether each could

form a meaningful construct in this study. As construct variables are to be created for

each extracted factor, the reliability of each factor is assessed using Cronbach’s alpha.

Factor 1 extracted consisted of the variables: Loyalty2, Loyalty3, Confidence3,

Loyalty1, Confidence1 and Confidence2 as shown in Table 4.6. This has been

labelled as FS1-H7 as it is closely related to the consumer loyalty construct as in

Hypothesis 7 indicated in the literature review chapter.

Reliability Test of FS1-H7

As described in section 4.3.1, Cronbach’s alpha is a coefficient of reliability and it

indicates how well a set of variables measures a single unidimensional latent

construct. If Cronbach’s alpha is high, it means that the reliability of the factor is high.

FS1-H7: Consumer Loyalty


- 182 -

As shown in Table 4.6, the Cronbach’s alpha result of the FS1-H7 construct is 0.840.

This indicates that the reliability of the component for this independent variable is

good as it well exceeds the acceptable value of 0.70 (Hair et al. 1998).

Table 4-6 Reliability test of FS1-H7

Reliability Statistics

.840 6

Cronbach'sAlpha N of Items


The extracted Factor 2 is made of the variables: degree of network1, degree of

network2, norm1 and interpersonal2 (see Table 4.6). The factor score label created

for this component is FS2-H4 as it is closely related to the nature of the social system

construct as in Hypothesis 4 stated in the literature review chapter.


As shown in Table 4.7, the Cronbach’s alpha result of the FS2-H4 construct is 0.818,

indicating that the reliability of the component well exceeded the acceptable value of

0.70 (Hair et al. 1998).

FS2-H4: Nature of the Social System


- 183 -



Cronbach's Alpha N of Items .818 4


The variables Perceived usefulness2, Perceived usefulness1, Perceived Usefulness3

and Complexity1 make up factor 3 (see Table 4.5). The factor score label created for

this component is FS3-H6 since it is associated with Hypothesis 6, the Perceived

usefulness construct in the literature review chapter. In this context, ‘Complexity’ is

referring to the low level of complexity of using online securities trading.



again indicating the reliability of the factor as it well exceeded the acceptable value

of 0.70 (Hair et al. 1998).





FS3-H6: Perceived Usefulness and Complexity


- 184 -

The variables Trialability1, Trialability2, Observability1 and Observability2 are

found in extracted factor 4 (see Table 4.5), labelled as FS4-H1a as it consisted of two

variables found in the Perceived Attributes of Innovations construct found in

Hypothesis 1 as mentioned in Chapter Two. The construct does not contain all of the

components in the original Hypothesis 1 but only two components and is thus

labelled as H1a.

Reliability Test of FS4-H1a

As shown in Table 4.9, the Cronbach’s alpha result of the FS4-H1a construct is 0.823,

indicating a reliable variable can be constructed as it well exceeded the acceptable

value of 0.70 (Hair et al. 1998).

Table 4-9 Reliability test of FS4-H1a




FS4-H1a: Perceived Attribute of Innovations


- 185 -

The extracted factor 5 has the variables Mass media2, Mass media1, Interpersonal1

and Relative Advantage1 (see Table 4.5). This is closely related to the

Communication Channel construct in the theoretical model found in the literature

review chapter. The Communication Channel construct relates to Hypothesis 3 in the

model and hence the factor score for factor 5 is labelled as FS5-H3.


As shown in Table 4.10, the Cronbach’s alpha result of the FS5-H3 construct is 0.704.

It indicated that the reliability of the factor is acceptable as it just exceeded the

acceptable value of 0.70 (Hair et al. 1998).





FS5-H3: Communication Channel


- 186 -

In factor 6, the variables extracted are Compatibility1, Compatibility2 and Norm2

(see Table 4.5). Two of these variables are elements found in the Perceived Attributes

of Innovations construct in the Hypothesis 1 of the theoretical model stated in

Chapter Two. The factor score for component 6 is labelled as FS6-H1b. The

construct does not contain all of the components in the original Hypothesis 1 but only

one component and another closely related component and is thus labelled as H1b.

Reliability Test of FS6-H1b

As shown in Table 4.11, the Cronbach’s alpha result of the FS6-H1b construct is

0.755. It indicated that the reliability of the component is good as it exceeded the


Table 4-11 Reliability test of FS6-H1b




FS6-H1b: Perceived Attribute of Innovations


- 187 -

The variables Change Agent Efforts3, Change Agent Efforts2 and Change Agent

Efforts1 make up factor 7 (see Table 4.5). The factor score label created for this

component is FS7-H5 as it is associated with the Extent of Change Agent’s

Promotion Effects construct in Hypothesis 5 of the theoretical model in chapter two.



suggesting reliability of the component is good as it well exceeds the acceptable






FS7-H5: Extent of Change Agent’s Promotion Effects


- 188 -

In the component 8, the extracted variables are Authority decision1, Authority

decision2 and Collective decision1 (see Table 4.6). This construct is highly related to

the Type of Innovation-Decision construct in Hypothesis 2 of the theoretical model.

However, the extracted construct does not consist of all the variables in the construct

of Hypothesis 2, hence the factor score created is labelled as FS8-H2a. The factor

score for component 8 is labelled as FS6-H1b. The construct does not contain all of

the variables in the original Hypothesis 2 but only two components and is thus

labelled as H2a.

Reliability Test of FS8-H2a

As shown in Table 4.13, the Cronbach’s alpha result of the FS8-H2a construct is

0.746. It indicates that the reliability of the component is acceptable as it exceeds the


Table 4-13 Reliability test of FS8-H2a




FS8-H2a: Type of Innovation-Decision


- 189 -

In the last extracted component, the two variables extracted are Optional decision2

and Optional decision1 (see Table 4.5). This component consists of the variables in

the Type of Innovation-Decision construct which relate to Hypothesis 2, similar to

the case of component 8. The label created for component 9 is hence FS9-H2b. The

construct does not contain all of the variables in the original Hypothesis 2 but only

the Optional decision components and is thus labelled as H2b.

Reliability Test of FS9-H2b

As shown in Table 4.14, the Cronbach’s alpha result of the FS9-H2b construct is

0.701. The reliability of the component is acceptable as it exceeds the acceptable


Table 4-14 Reliability test of FS9-H2b




The summary of the labelled factor scores derived from the extracted components are

summarised in Table 4.15. All of the extracted constructs are deemed to be reliable

from the results of the reliability tests.

FS9-H2b: Type of Innovation-Decision


- 190 -

Table 4-15 Summary of extracted components

Component Variable Associated Hypothesis in Chapter Two

Factor Score Label

Cronbach’s Alpha

1 Loyalty2 Loyalty3 Confidence3 Loyalty1 Confidence1 Confidence2

Hypothesis 7 FS1-H7 0.840

2 Degree of network1 Degree of network2 Norm1


3 Perceived usefulness2 Perceived usefulness1 Perceived usefulness3 Complexity1


4 Trialability1 Trialability2 Observability1 Observability2

Hypothesis 1a FS4-H1a 0.823

5 Mass media2 Mass media1 Interpersonal1 Relative Advantage1


6 Compatibility1 Compatibility2 Norm2

Hypothesis 1b FS6-H1b 0.755

7 Change Agent Efforts3 Change Agent Efforts2 Change Agent Efforts1


8 Authority decision1 Authority decision2 Collective decision1

Hypothesis 2a FS8-H2a 0.746

9 Optional decision2 Optional decision1

Hypothesis 2b FS9-H2b 0.701



- 191 -

4.4.3 Factor Analysis for Dependent Variables

The previous section described the factor analysis of the independent variables, and

nine components have been extracted to form nine constructs that matched the

hypotheses of the theoretical model in this study. The next step was to conduct factor

analysis for the dependent variables.

Correlation Matrix

Pallent suggested that the first step to do for the factor analysis results is to

investigate the inter-correlation between variables (Pallant 2007). The top half of the

table contains the Pearson correlation coefficient between all variables whereas the

bottom half contains the one-tailed significance of the coefficients.

The loadings in table 4.16 for the dependent variables are scanned to identify those

variables with significance test values that are greater than 0.05 (Field 2005). As all

the dependent variables are less than 0.05, there is no need to consider eliminating

any variables at this stage. Furthermore, this initial scan shows that all seven

variables are significantly correlated to all 6 other variables.


- 192 -

Table 4-16 Correlation matrix for dependent variables

Correlation Matrixa

1.000 .469 .433 .409 .586 .600.469 1.000 .422 .258 .452 .492.433 .422 1.000 .457 .413 .402.409 .258 .457 1.000 .512 .463.586 .452 .413 .512 1.000 .848

.600 .492 .402 .463 .848 1.000

.000 .000 .000 .000 .000.000 .000 .000 .000 .000.000 .000 .000 .000 .000.000 .000 .000 .000 .000.000 .000 .000 .000 .000

.000 .000 .000 .000 .000

frequencyVolumeTypelocationinvestment tips exchangeinvestment informationfrequencyfrequencyVolumeTypelocationinvestment tips exchangeinvestment informationfrequency

Correlation

Sig. (1-tailed)

frequency Volume Type locationinvestment

tips exchange

investmentinformationfrequency

Determinant = .059a.


KMO and Bartlett’s Test

As mentioned in previous sections, the Kaiser-Meyer-Olkin (KMO) is being used to

test the adequacy of using factor analysis for the variables (Pallant 2007). A value

close to 1 indicates that the correlations of the variables are relatively compact and

deemed appropriate for factor analysis (Field 2005). For the data collected for the

dependent variables in this research, the KMO value is 0.810 (see Table 4.17) and

hence the data are fit for factor analysis.

The Barlett’s test coefficient is to test the significance of the correlations of the

variables. If the value is less than 0.01, it indicates that the correlations of variables

are significant. In this case, the value is 0.00 (P < 0.01) and hence indicates that

factor analysis is appropriate and significant (Field 2005).


- 193 -

Table 4-17 KMO and Bartlett’s Test for dependent variables


.810

645.32515

.000

Kaiser-Meyer-Olkin Measure of SamplingAdequacy.

Approx. Chi-SquaredfSig.

Bartlett's Test ofSphericity


Communalities Testing

As mentioned in earlier sections, communality is the proportion of common variance

within a variable. The extraction method is based on the Principal Component

Analysis which is based on the assumption that all variance is common and the

communalities value should be 1 before extraction (Field 2005). For extraction

values less than 0.3, they are considered as to not fit well with other items in its

component. A low value less than 0.3 indicates that the variable should be omitted

from the analysis (Pallant 2007). In this case, all the values are above 0.3 (see Table

4.18) and none of the dependent variables are necessary to be excluded.

Table 4-18 Communalities for dependent variables

Communalities

1.000 .6061.000 .4481.000 .4451.000 .4501.000 .747

1.000 .745

frequencyVolumeTypelocationinvestment tips exchangeinvestment informationfrequency

Initial Extraction




- 194 -

Total Variances

Table 4.19 shows the variance explained by the 6 dependent variables following

extraction of components using the Principal Components Analysis extraction

method. On the basis of Eigenvalues > 1, one single factor would seem to be

optimum here. Each additional factor after 1 has an Eigenvalue of less than 1.

Table 4-19 Total variance explained based on Eigenvalues


3.442 57.366 57.366 3.442 57.366 57.366.766 12.761 70.127.737 12.277 82.403.472 7.871 90.274.435 7.258 97.532.148 2.468 100.000

Component123456

Total % of Variance Cumulative % Total % of Variance Cumulative %Initial Eigenvalues Extraction Sums of Squared Loadings



Scree Plot

Chart 4.8 shows a Scree Plot of Eigenvalues and a very clear indication that after one

factor, little is gained by having a greater number factors. In the Scree Plot, one looks

for a ‘bend’ in the plot line to determine where to stop the retention of factors. As

shown in Chart 4.8, the slope of the plot stops changing significantly after the bend at

two components, so it is suggested that the resulted loading with only one extracted

factor is appropriate.


- 195 -

Chart 4-8 Scree Plot for dependent variables

Component Number654321

Eige

nval

ue

4

3

2

1

0

Scree Plot


Initial Factor Matrix

Since only one component is extracted for the six dependent variables, rotation is not

completed as all the dependent variables can be combined as one single construct.

The component score is labelled as FS-Dep Var as the dependent variable for

regression analysis.


- 196 -

Table 4-20 Initial component matrix

Component Matrix(a)

Component

1 Investment tips exchange .864 Investment information Frequency .863

Frequency .779 Location .671 Volume .670 Type .667


a 1 components extracted.

Rotated Component Matrix(a)

a Only one component was extracted. The solution cannot be rotated.


4.4.4 Reliability Testing of Factor (Dependent Variable)

As shown in Table 4.21, the Cronbach’s alpha reliability statistic result for the

dependent variables is 0.837. This indicates that the reliability of the component to be

used for the dependent variable is good as it well exceeds the acceptable value of

0.70 (Hair et al. 1998).

Table 4-21 Reliability statistics


.837 6

Cronbach'sAlpha N of Items



- 197 -

4.5 Multiple Regression Analysis

A standard multiple linear regression analysis was conducted to test the hypotheses

of this research. The multiple linear regression analysis was adopted in this research

as it is a statistical approach that can be utilised to analyse the relationship between

many independent variables and a single dependent variable (Hair et al. 1998). The

variables are based on the factors derived from factor analysis, and they are arranged

as nine independent variables, as detailed in Table 4.15, and one dependent variable

which is the post-adoption behaviour of online securities trading by the retail

investors. The values of the variables are computed by summing the factor scores of

the components extracted in factor analysis.

4.5.1 Multiple Regression Model

As identified in the Chapter 3, the multiple regression model can be presented as the

following equation:

Y = α + β1X1 + β 2X2 + β 3X3 + β 4X4 + β 5X5 + β 6X6 + β 7X7 + β 8X8 + β 9X9 + ε

For this research, Y is the dependent variable, is labelled as “FS-Dep Var” and it is

made up of the variables “frequency”, “volume”, “type”, “location”, “investment tips

exchange” and “investment information frequency”. The “FS-Dep Var” was created

through factor analysis as it is the sum of the factor scores for the dependent variable

components. That is, for each respondent case, the dependent variable value will be

the sum of the scores on each of the 7 dependent variables that make up the construct.

These scores are the values indicated by the respondents in their survey response.


- 198 -

There are nine groups of independent variables in the multiple regression model,

which are used to create the construct variables. A potential problem with mutiple

regression analysis is that of muticollinearity amongst independent, or predictive,

variables. Pallant (2007) states that multicollinearity exists when independent

variables are highly correlated and this does not contribute to a good multiple

regression model.

The independent regression variables X1 to X9 are derived from the sums of factor

scores in the same way as the dependent variable described above. The factors were

derived using an orthogonal rotation. Factors thus created are designed to be

uncorrelated and so the variables derived from their factor scores are also not

correlated (Rummel 1967). It is assumed, therefore, that an otherwise potential

problem of muticollinarity in mutiple regression analysis will not be a problem in this

analysis and the mutiple regression model and variables are deemed to be appropriate.

To conduct the regression analysis in SPSS, the construct variable for the combined

dependent variables was selected in the ‘Dependent’ box of linear regression. All of

the nine construct variables for the independent variables were selected in the

‘Independent(s)’ box. Other statistical options of regression analysis were also

selected in the testing and are explained in the following sections.

The multiple regression model of this research is presented as follows:

FS-Dep Var = α + β 1(FS1-H7) + β 2(FS2-H4) + β 3(FS3-H6) + β 4(FS4-H1a) +

β 5(FS5-H3) + β 6(FS6-H1b) + β 7(FS7-H5) + β 8(FS8-H2a) + β 9(FS9-H2b) + ε

where;


- 199 -

FS-Dep Var = dependent variable (post-adoption behaviours)

FS1-H7 = independent variable #1 (consumer loyalty)

FS2-H4 = independent variable #2 (Nature of the Social System and Interpersonal)

FS3-H6 = independent variable #3 (Perceived usefulness and Complexity)

FS4-H1a = independent variable #4 (Perceived Attributes of Innovation: Trialability

and Observability)

FS5-H3 = independent variable #5 (Communication channels and Relative

Advantage)

FS6-H1b = independent variable #6 (Perceived Attributes of Innovation:

Compatibility and Norm)

FS7-H5 = independent variable #7 (Change Agent)

FS8-H2a = independent variable #8 (Type of Innovation-Decision: Authority

Decision & Collective Decision)

FS9-H2b = independent variable #9 (Type of Innovation-Decision: Optional

Decision)

α = constant for the intercept value of Y axis

β1 to β9 = Regression coefficient for each independent variable

ε = Random error


- 200 -

4.5.2 Model Summary – R Square

The first output of the multiple regression analysis is the model summary and the

main focus is the R Square value which represents the percentage of the dependent

variable explained by the independent variables. In this case, 51 percent of the

variation in the dependent variable FS1-Dep Var (post adoption behaviour) can be

explained by the independent variables.

The Durbin-Watson statistic has also been included in the model summary and

informs us about whether the assumption of independent errors is well founded (Field

2005). It is better for the value to be closer to 2 and in this case, the value is 2.081 as

presented in Table 4.22 which meant that the independent errors assumption has been

met.

Table 4-22 Model summary

Model Summary(b)

Model R R Square Adjusted R Square

Std. Error of the Estimate Durbin-Watson

1 .727(a) .529 .510 .69996262 2.081 a Predictors: (Constant), FS9-H2b, FS8-H2a, FS7-H5, FS6-H1b, FS5-H3, FS4-

H1a, FS3-H6, FS2-H4, FS1-H7

b Dependent Variable: FS-Dep Var



- 201 -

4.5.3 ANOVA Table

The next output of the regression test is the ANOVA table shown in Table 4.23.

SPSS allows specification of multiple models in a single regression command. The

number under the ‘Model’ column indicates the number of the model being reported.

As shown in Table 4.23, there is only 1 model reported in this regression test. The

total variance in the regression test can be explained by the independent variables

(Regression) and the variance which is not explained by the independent variables

(Residual). The sums of squares for the regression and residual add up to the ‘Total’.

The degree of freedom is labelled as df in the ANOVA table as shown in Table 4.23.

The Mean Square shown in the table is the Sum of Squares divided by the respective

df. The F-value in the table is the Mean Square divided by the Mean Square Residual,

yielding 27.72 as shown in Table 4.23. The p-value associated with this F-value is

very small (0.000). Typically, if the p-value is less than 0.05, it implies that the

independent variables reliably predict the dependent variable (UCLA 2009b) (Field

2005).

Table 4-23 ANOVA table

ANOVA(b)

Model Sum of Squares df Mean Square F Sig.

1 Regression 122.232 9 13.581 27.720 .000(a) Residual 108.768 222 .490 Total 231.000 231

a Predictors: (Constant), FS9-H2b, FS8-H2a, FS7-H5, FS6-H1b, FS5-H3, FS4-

H1a, FS3-H6, FS2-H4, FS1-H7

b Dependent Variable: FS-Dep Var



- 202 -

4.5.4 Model Parameters

The last part of the multiple regression output is the parameters of the regression

model. In multiple regression analysis, the model takes the form of an equation that

contains a coefficient for each predictor, or independent variable. The model

parameter table presents the values of the coefficients that indicates the individual

contribution of each predictor to the changes in the dependent variable (Field 2005).

The coefficient values tell us about the relationship between post-adoption behaviour

and each predictor. If the t statistic value is large (t > 1) and the associated Sig. value

is less than 0.05 (p < 0.05) then the predictor is making a significant contribution to

the model (Field 2005).

The results in Table 4.24 summarised the findings of the relationship between the

dependent variable FS-Dep Var (Post adoption behaviour) and the nine independent

variables. Further discussion of the results of the analysis is presented to show the

significance of the hypotheses of this study. The constant value is insignificant (p >

0.05) as shown in Table 4.24 and thus is removed from the regression model. In this

research, it has no significance in determining the value of post-adoption behaviour at

the Y-intercept when the independent variable is at zero value. The regression test

has been re-run without the constant and the result is shown in Table 4.25. It has no

significant difference as compared to the regression test with the Constant included.

The revised formula for the regression model without the Constant is as follows:

FS-Dep Var = α + β 1(FS1-H7) + β 2(FS2-H4) + β 3(FS3-H6) + β 4(FS4-H1a) +

β 5(FS5-H3) + β 6(FS6-H1b) + β 7(FS7-H5) + β 8(FS8-H2a) + β 9(FS9-H2b) + ε


- 203 -

Table 4-24 Coefficients table

Coefficients(a)

Model Unstandardised Coefficients

Standardised Coefficients t Sig.

B Std. Error Beta B Std. Error 1 (Constant) 1.08E-016 .046 .000 1.000 FS1-H7 .332 .046 .332 7.209 .000 FS2-H4 .439 .046 .439 9.535 .000 FS3-H6 .344 .046 .344 7.475 .000 FS4-H1a .118 .046 .118 2.554 .011 FS5-H3 .079 .046 .079 1.706 .089 FS6-H1b .184 .046 .184 3.994 .000 FS7-H5 .171 .046 .171 3.716 .000 FS8-H2a .035 .046 .035 .766 .444 FS9-H2b .152 .046 .152 3.304 .001

a Dependent Variable: FS-Dep Var


Table 4-25 Coefficients table (without constant)

Coefficientsa,b

.332 .046 .332 7.225 .000

.439 .046 .439 9.557 .000

.344 .046 .344 7.492 .000

.118 .046 .118 2.560 .011

.079 .046 .079 1.710 .089

.184 .046 .184 4.003 .000

.171 .046 .171 3.724 .000

.035 .046 .035 .768 .443

.152 .046 .152 3.312 .001

FS1-H7FS2-H4FS3-H6FS4-H1aFS5-H3FS6-H1bFS7-H5FS8-H2aFS9-H2b

Model1

B Std. Error

UnstandardizedCoefficients

Beta

StandardizedCoefficients

t Sig.

Dependent Variable: FS-Dep Vara.

Linear Regression through the Originb.



- 204 -

4.6 Hypotheses Testing

In this section, the hypotheses are tested and discussed based on the outputs from the

multiple regression analysis. There are seven null hypotheses in the original

theoretical model derived from the seven research propositions. Two of these were

expanded, and so here nine hypotheses will be tested to be accepted or rejected in the

following sections. Statistical analyses for the multiple regression model to test the

hypotheses were conducted through the use of SPSS version 15.0 for Windows. The

interpretations of the hypotheses are presented below.

4.6.1 Test of Hypothesis 1a

(H1a)0: There is no correlation between FS4-H1a (Trialability and Observability) and

post-adoption behaviour of retail investors using online securities trading.

Table 4.25 shows that for the coefficient of the construct variable FS4-H1a, the t

statistic value is bigger than 1 and the Sig. value is less than 0.05 (p < 0.05), so the

null hypothesis is rejected and the alternative hypothesis is accepted (see Table 4.26).

This means that the independent variables, Trialability and Observability have a

positive and significant impact on the overall post-adoption behaviour of the retail

investors.

4.6.2 Test of Hypothesis 1b

(H1b)0:There is no correlation between FS6-1b (Compatibility and Norm) and post-

adoption behaviour of retail investors using online securities trading.

For the coefficient of construct FS6-1b, the t statistic value is bigger than 1 and the

Sig. value is less than 0.05 (p < 0.05) and so the null hypothesis is rejected and the


- 205 -

alternative hypothesis is accepted (see Table 4.26). This means that the independent

variables, Compatibility and Norm have a positive and significant influence on the

dependent variable.

It is concluded that the null hypothesis 1 is rejected and the alternative hypothesis is

accepted. The researcher concludes that the independent construct variable, Perceived

Attributes of Innovation has a positive and significant influence on the post-adoption

behaviour of retail investors using online securities trading. The construct contains

the variables Compatibility and Norm that were included in the original Hypothesis 1

of the theoretical model.

4.6.3 Test of Hypothesis 2a

(H2a)0: There is no correlation between FS8-H2a (Authority decision and Collective

decision) and post-adoption behaviour of retail investors using online securities

trading.

For construct FS8-H2a, the t statistic value is 0.768 which is smaller than 1 and the

Sig. value is 0.443 (p > 0.05) (see Table 4.26). This means that the null hypothesis

for the independent variables, authority decision and collective decision is accepted.

The construct FS8-H2a, that contains the independent variables authority decision

and collective decision, has no significant influence on the dependent variable in the

regression model.


- 206 -

4.6.4 Test of Hypothesis 2b

(H2b)0:There is no correlation between FS9-H2b(Optional decision) and post-

adoption behaviours of retail investors using online securities trading.

The finding from the construct FS9-H2b shows that the t value is bigger than 1 and

the Sig. value is 0.001 (p < 0.05) (see Table 4.32). This means that the null

hypothesis for this construct is rejected and the alternative is accepted. The

independent variables, Optional decision1 and Optional decision2 indeed have a

positive and significant impact on the post-adoption behaviour.

In the original theoretical model, the Type of Innovation Decision variable includes

Optional decision, Collective decision and Authority decision. However, in the new

model, the Type of Innovation Decision variable has been divided into two construct

variables. The construct variable FS8-H2a which includes the Authority decision and

Collective decision has accepted the null hypothesis (See Table 4.26). The construct

variable FS9-H2b which includes the variable Optional decision1 and Optional

decision2 has rejected the null hypothesis (See Table 4.26). It can be concluded from

the findings that only the construct variable which consists of the Optional decision

variable has an influence on post-adoption behaviour of the retail investors.

4.6.5 Test of Hypothesis Three

(H3)0: There is no correlation between FS5-H3 (Communication channels) and post-

adoption behaviour of retail investors using online securities trading.

As shown in Table 4.25, the t statistic value for FS5-H3 is 1.710 and the Sig. value is

0.089 (p > 0.05).


- 207 -

It is suggested that the predictor FS5-H3 has no significant association to the

dependent variable of the regression model. The null hypothesis is accepted and the

alternative hypothesis is rejected. This means that there is no significant correlation

between the pre-adoption communication channels and the post-adoption behaviour

of retail investors using online securities trading.

4.6.6 Test of Hypothesis Four

(H4)0: There is no correlation between FS2-H4 (Nature of the social system) and


As shown in Table 4.25, the t value for the coefficient of this construct is 9.557 (t >

1) and the Sig. value is less than 0.05 (p < 0.05) and hence the predictor FS2-H4 has

significant association with the dependent variable of the regression model. The null

hypothesis is rejected and the alternate hypothesis is accepted. It is suggested that the

nature of the social system construct is positively influencing the dependent variable


4.6.7 Test of Hypothesis Five

(H5)0: There is no correlation between FS7-H5 (Extent of change agent’s promotion

efforts) and post-adoption behaviour of retail investors using online securities trading.

As shown in Table 4.25, the t statistic value for this construct is 3.724 (t > 1) and

the Sig. value is less than 0.05 (p < 0.05) and hence the predictor FS7-H5 has a

positive and significant impact on the dependent variable of the regression model.

The null hypothesis is rejected and the alternative hypothesis is accepted. We can

conclude that the extent of change agent’s promotion efforts has a positive and


- 208 -

significance influence on the post-adoption behaviour of retail investors using online

securities trading.

4.6.8 Test of Hypothesis Six

(H6)0: There is no correlation between FS3-H6 (Perceived usefulness and

Complexity) and post-adoption behaviour of retail investors using online securities

trading.

As shown in Table 4.25, the t statistic value for this construct is 7.492 (t > 1) and

the Sig. value is less than 0.05 (p < 0.05) and hence the predictor FS3-H6 has a

positive and significant impact on the dependent variable of the regression model.

The null hypothesis is rejected and the alternate hypothesis is accepted, that is, that

there is a positive correlation between the construct variable FS3-H6, made up of the

variables Perceived usefulness and Complexity or the low level of complexity, and


4.6.9 Test of Hypothesis Seven

(H7)0: There is no correlation between FS1-H7 (consumer loyalty) and post-adoption

behaviour of retail investors using online securities trading.

As shown in Table 4.25, the value of the t-statistic (7.225) is significant at 0.000 (p <

0.05).

This means that the construct FS1-H7 has a positive and significant impact on the

dependent variable of the regression model. Thus, the null hypothesis is rejected and

the alternative hypothesis is accepted. The researcher concludes that consumer


- 209 -

loyalty has a positive and significant influence on the post-adoption behaviour of


4.6.10 Summary of Hypotheses Testing

From the result of testing all the nine hypotheses as shown in Table 4.26; only the

null hypotheses for H2a and H3 were accepted. The research findings suggested that

FS4-H1a (Trialability and Observability), FS6-H1b (Compatibility and Norm), FS9-

H2b (Optional decision), FS2-H4 (Nature of the social system), FS7-H5 (Extent of

change agent’s promotion efforts), FS3-H6 (Perceived usefulness and Complexity)

and FS1-H7 (Consumer loyalty) are the influential factors in the post-adoption

behaviour of retail investors using online securities trading.

The research findings are somewhat different from the theoretical model in Chapter

Two as in H1, the perceived attributes of innovations has been divided into two

Hypotheses, H1a and H1b based on the factor analysis results. However, both H1a

and H1b are rejected, giving support to the original theoretical model.

The Hypothesis 2 in Chapter Two, Type of Innovation-Decision has also been

divided into two Hypotheses, H2a and H2b. H2a consists of the Authority decision

and Collective decision variables while H2b consists of the Optional decision

variables as a result of the factor analysis. H2a is accepted and H2b is rejected,

suggesting that the model in Chapter Two must be adapted to reflect the influence of

optional decisions only.

The hypotheses testing has also accepted (H3)0. This means that Communication

channels have no significant impact on the post-adoption behaviour of the retail

investors using online securities trading.


- 210 -

Table 4-26 Summary of the results of hypotheses testing

Null Hypothesis Parameter Result (H1a)0: FS4-H1a (Trialability and Observability) has no association with post-adoption behaviour.

Significant at 0.011 Reject the null hypothesis

(H1b)0: FS6-H1b (Compatibility and Norm) has no association with post-adoption behaviour.


(H2a)0: FS8-H2a (Authority decision and Collective decision) has no association with post-adoption behaviour.

Significant at 0.443 Accept the null hypothesis as insufficient significant variables.

(H2b)0: FS9-H2b (Optional decision) has no association with post-adoption behaviour.


(H3)0: FS5-H3 (Communication channels) has no association with post-adoption behaviour.

Significant at 0.089 Accept the null hypothesis

(H4)0: FS2-H4 (Nature of the social system) has no association with post-adoption behaviour.


(H5)0: FS7-H5 (Extent of change agent’s promotion efforts) has no association with post-adoption behaviour.


(H6)0: FS3-H6 (Perceived usefulness and Complexity) has no association with post-adoption behaviour.


(H7)0: FS1-H7 (Consumer loyalty) has no association with post-adoption behaviour.



Figure 4.2 below shows the revised theoretical model with the significant

independent variables of Perceived attributes of innovations, Type of innovation-

decision, Nature of the social system, Extent of change agent’s promotion efforts,

Perceived usefulness, Consumer loyalty and the dependent variable, post-adoption


- 211 -

behaviour of the retail investors using online securities trading. It is somewhat

different from the conceptual theoretical model described in Chapter Two (see Figure

2.15) as the Communication channels variable has been removed, as well as some

components of other variables. The relative advantage component of the Perceived

attributes of innovations variable has been removed in the revised model. The

Perceived attributes of innovations variable in the original conceptual model is now

divided into two independent variables as a result of the factor analysis. The Type of

innovation-decision in the original model is now left with the Optional decision

variable as shown in Figure 4.2. The Interpersonal component is now merged with

the Nature of the social system independent variable since the component could be

considered as part of the social system. The variable Complexity is referring to a low

level of complexity, or the use of online securities trading being effort-free in this

research context. The Complexity component is now merged with the Perceived

usefulness variable as Complexity is closely related to the Perceived ease of use

component in the Technology Acceptance Model described in Chapter Two. There

are no changes in the Extent of change agent’s promotion efforts variable; Consumer

loyalty variable and post-adoption usage behaviour dependent variable as compared

to the conceptual model developed in Chapter Two shown in Figure 2.17.


- 212 -

Figure 4-2 Revised theoretical model

Hypothesis 7

Perceived Attributes of Innovations (H1a)

- Trailability- Observability

Extent of Change Agent’s Promotion Efforts

- Perceived Usefulness

Customer Loyalty

- Loyalty- Confidence

Independent variables Dependent variables

Hypothesis 1a

Hypothesis 2b

Hypothesis 4

Hypothesis 5

Nature of the Social System- Norms- Degree of network interconnectedness

Type of Innovation-Decision- Optional decision

Hypothesis 6

Post-Adoption Usage Behaviour of Online Securities Trading

-

- Type of shares- Location- Volume- Frequency- Investment tips exchange- Investment information frequency

Perceived Attributes of Innovations (H1b)

- Compatibility- Norm

Hypothesis 1b

- Interpersonal

- Complexity



- 213 -


The chapter started with presenting the data profile for the data collected through the

online survey. The demographic profiles of the sample have been described. The

majority of the respondents are of the age group of 21 to 30, with a Bachelor degree

and equivalent, working as professionals with an income level of $35,000 to $50,000,

and are single males, though a wide cross section of occupations and income brackets

is represented by the sample.

The survey collected data on a large number of variables reflecting factors

influencing the adoption of online security trading. Factor analysis was undertaken

on these many variables to reduce the data to nine common factors. A single factor

representing post-adoption security trading was also derived. Testing was conducted

on these factors to ensure reliability. Factor scores were then used to create nine

independent construct variables and one dependent construct variable.

To test the hypotheses that these nine independent pre-adoption variables had a

significant influence on post adoption behaviour, a regression model was devised.

Standard linear multiple regression was conducted with nine independent variables.

Hypothesis testing was undertaken by testing the statistical significance of the

coefficient associated with the nine independent construct variables. Of the nine null

hypotheses, seven were rejected and two accepted. Thus, the researcher concluded

that seven of the construct variables had a positive and significant influence on post

adoption behaviour of retail security trading.


- 214 -

A revised theoretical model, adapted from the one derived in Chapter Two is

presented, incorporating the findings of the statistical analysis.

The interpretation of the results will be discussed in the conclusion, Chapter Five.

The chapter also will summarise this research and its limitations to provide

recommendations for future research.

Chapter Five: Conclusions and implications

- 215 -

Chapter 5 Conclusions and implications

5.1 Introduction to Chapter Five

In this final chapter, the conclusions and implications of the research findings and the

implications derived from the thesis are discussed. The chapter is organised into eight

sections and concludes with a final statement of the thesis. Following the introduction

is a restatement of the research problem and issues. In section three, results of

hypotheses testing are summarised. The theoretical implications of the research

findings are described in section four with a statement of what the findings have

contributed to the body of knowledge. Managerial implications are presented in

section five. Section six acknowledges the limitations of this research. In section

seven, recommendations for future research, based on the current study and findings,

are presented. The last section concludes the findings and summarises this research.

All the sections mentioned are summarised in the Figure 5.1.


- 216 -

Figure 5-1 Overview of Chapter Five

5.1 Introduction to Chapter Five

5.2 Restatement of the research problem and hypotheses

5.3 Conclusion about the hypotheses and research problem5.3.1 Conclusion for hypothesis one5.3.2 Conclusion for hypothesis two5.3.3 Conclusion for hypothesis three5.3.4 Conclusion for hypothesis four5.3.5 Conclusion for hypothesis five5.3.6 Conclusion for hypothesis six5.3.7 Conclusion for hypothesis seven

5.4 Contributions to the body of knowledge5.4.1 Consequences of Innovations5.4.2 Technology Acceptance Model5.4.3 Consumer loyalty

5.6 Limitations of the research5.5.1 Scope limitation5.5.2 Geographical limitation5.5.3 Online questionnaire

5.7 Recommendations for future research5.6.1 Other products5.6.2 Other geographical regions5.6.3 Other factors affecting post-adoption consequences5.6.4 Other post-adoption consequences

5.8 Research conclusion

5.5 Managerial Implications



- 217 -

5.2 Restatement of the research problem and hypotheses

The main objective of this research is to investigate the extent to which pre-adoption

factors influence the consequential behaviour of retail investors using online

securities trading. The following research issues were addressed:



RI2) Does Perceived usefulness affect the post-adoption usage behaviour of the

retail investors using online securities trading?



The following null hypotheses were tested. These are adapted from the literature

review chapter and modified after preliminary data analysis using the methods

described in Chapter Three - Methodology.

Nine hypotheses were tested. They are restated below.

(H1a)0: There is no correlation between the construct variable FS4-H1a (Trialability

and Observability) and post-adoption behaviours of retail investors using online

securities trading.

(H1b)0: There is no correlation between construct variable FS6-H1b (Compatibility

and Norm) and post-adoption behaviours of retail investors using online securities

trading.


- 218 -

(H2a)0: There is no correlation between construct variable FS8-H2a (Authority

decision and Collective decision) and post-adoption behaviours of retail investors

using online securities trading.

(H2b)0: There is no correlation between construct variable FS9-H2b (Optional

decision) and post-adoption behaviours of retail investors using online securities

trading.

(H3)0: There is no correlation between construct variable FS5-H3 (Communication

channels) and post-adoption behaviours of retail investors using online securities

trading.

(H4)0: There is no correlation between construct variable FS2-H4 (Nature of the

social system) and post-adoption behaviours of retail investors using online securities

trading.

(H5)0: There is no correlation between construct variable FS7-H5 (Extent of change

agent’s promotion efforts) and post-adoption behaviours of retail investors using


(H6)0: There is no correlation between construct variable FS3-H6 (Perceived

usefulness and Complexity) and post-adoption behaviours of retail investors using


(H7)0: There is no correlation between construct variable FS1-H7 (Consumer

loyalty) and post-adoption behaviours of retail investors using online securities

trading.


- 219 -

5.3 Conclusion about the hypotheses and research

problem

The theoretical model of this research was developed and discussed in Chapter Two

(see Table 2.17). From the research propositions formed in Chapter Two, hypotheses

were established and tested using the methodology outlined in Chapter Three. The

research results are detailed in Chapter Four and the conclusions of the hypotheses

test are presented in the following section. A summary of the test results are listed in

Table 5.1

Table 5-1 Summary of results from testing of hypotheses

Null Hypothesis Result Interpretation (H1a)0: FS4-H1a (Trialability and Observability) has no association with post-adoption usage behaviour.

Reject the null hypothesis

Trialability and Observability positively influence post-adoption usage behaviour.

(H1b)0: FS6-H1b (Compatibility and Norm) has no association with post-adoption usage behaviour.


Compatibility and Norm positively influence post-adoption usage behaviour.

(H2a)0: FS8-H2a (Authority decision and Collective decision) has no association with post-adoption usage behaviour.

Accept the null hypothesis There is no influence of Authority decision and Collective decision on post-adoption usage behaviour.

(H2b)0: FS9-H2b (Optional decision) has no association with post-adoption usage behaviour.


Optional decision factor positively influence post-adoption usage behaviour.

(H3)0: FS5-H3 (Communication channels) has no association with post-adoption usage behaviour.

Accept the null hypothesis There is no influence of Communication channels on post-adoption usage behaviour.

(H4)0: FS2-H4 (Nature of the social system) has no association with post-adoption usage behaviour.

Reject the null hypothesis Nature of the social system factor positively influences post-adoption usage behaviour.

(H5)0: FS7-H5 (Extent of change agent’s promotion efforts) has no association with post-adoption usage behaviour.

Reject the null hypothesis Extent of change agent’s promotion efforts positively influences post-adoption usage behaviour.


- 220 -

(H6)0: FS3-H6 (Perceived usefulness and Complexity) has no association with post-adoption usage behaviour.

Reject the null hypothesis Perceived usefulness and complexity (ease of use) positively influence post-adoption usage behaviour.

(H7)0: FS1-H7 (Consumer loyalty) has no association with post-adoption usage behaviour.

Reject the null hypothesis Consumer loyalty positively influences post-adoption usage behaviour.


5.3.1 Conclusion for hypothesis one

As detailed in Chapter Four, hypothesis 1 was found to comprise two common

factors. These were used as the basis to construct two construct variables. The first is

named Trialability and Observability, the second is named Compatibility and Norm.

Hypotheses 1a and 1b were then treated separately.

(H1a)0: There is no correlation between the construct variable FS4-H1a (Trialability

and Observability) and post-adoption behaviours of retail investors using online

securities trading.

This hypothesis was rejected and it was concluded that the Trialability and

Observability variables are related positively with post-adoption usage behaviour of

retail investors. This suggests that the level, or degree, of Trialability and

Observability prior to adoption of online securities trading influences the degree or

patterns of using the product by the retail investors. In other words when online

purchasers have a higher chance of testing the product and have a higher degree of

Observability, the retail investors also have higher usage of online securities trading.

This result is supported by the previous research of Karahana et al (1999) who


- 221 -

concluded Trialability and Observability are factors that influence post-adoption

beliefs (Karahanna, Straub & Chervany 1999).

(H1b)0: There is no correlation between construct variable FS6-H1b (Compatibility

and Norm) and post-adoption behaviours of retail investors using online securities

trading.

This hypothesis was also rejected, leading to the conclusion that Compatibility and

Norm variables are related positively with post-adoption usage behaviour of retail

investors. The variable Compatibility can be viewed as highly related to the variable

Norm. The results suggest that when online purchasers felt that the product is of

higher compatibility to their previous mode of trading and in line with what is used

by other retail traders, the retail investors also have higher usage of the online

securities trading. This result is supported by previous research of Karahana et al

(1999) who concluded that Compatibility and Norm are factors that influence post-

adoption beliefs (Karahanna, Straub & Chervany 1999).

5.3.2 Conclusion for hypothesis two

Based on the findings from the data analysis chapter, the Type of innovation-decision

was found to comprise two factors. Thus, two constructs and two hypotheses have

been developed. Authority decision and Collective decision make up one factor while

Optional decision is another factor. However, both constructs are related to the Type

of innovation-decision variable as stated in the literature review chapter.


- 222 -

(H2a)0: There is no correlation between construct variable FS8-H2a (Authority

decision and Collective decision) and post-adoption behaviours of retail investors

using online securities trading.

This hypothesis was accepted, so it is concluded that Authority decision and

Collective decision variables have no significant influence on post-adoption usage

behaviour of retail investors. This provides evidence that the external decision factors

might be more relevant to pre-adoption decisions and not to post-adoption usage.

The researcher is not aware of any other research testing the relationship between

external decision factors and post-adoption behaviour.

(H2b)0: There is no correlation between construct variable FS9-H2b (Optional

decision) and post-adoption behaviours of retail investors using online securities

trading.

This hypothesis was rejected. Optional decision variables are related positively with

post-adoption usage behaviour of retail investors. This suggests that with a higher

degree of personal choice in decision making by retail investors, there will be a

tendency towards higher usage of online securities trading. As with the Authority and

Collective decision variables, the researcher is not aware of prior research testing a

relationship between Optional decision factors in pre-adoption and post-adoption

usage.


- 223 -

5.3.3 Conclusion for hypothesis three

(H3)0: There is no correlation between construct variable FS5-H3 (Communication

channels) and post-adoption behaviours of retail investors using online securities

trading.

Hypothesis three was accepted, with the conclusion that Mass media and

Interpersonal variables (Communication channels) have no significance influence on

post-adoption usage behaviour of retail investors. These two related variables might

be quite important to pre-adoption decisions but have no significant influence on the

retail investors’ usage behaviour after they adopted online securities trading. This

seems consistent with the findings about Authority and Collective decisions, where

outside influence on pre-adoption decisions have little or no influence on post-

adoption behaviour.

5.3.4 Conclusion for hypothesis four

(H4)0: There is no correlation between construct variable FS2-H4 (Nature of the

social system) and post-adoption behaviours of retail investors using online securities

trading.

Rejection of hypothesis four and acceptance of the alternative leads to the conclusion

that the Nature of the social system variable is related positively with post-adoption

usage behaviour of retail investors.

The Nature of the social system is an element in the diffusion process that determines

the rate of adoption of innovations. Rogers (2003) defined the social system as “a set

of interrelated units engaged in joint problem solving to accomplish a common goal”


- 224 -

(Rogers 2003, p. 23). Rogers (2003) claimed that the nature of the social system

affects individuals’ innovativeness, and in deciding whether or not to adopt an

innovation, the individuals depend mainly on the communicated experience of others

much like themselves who have already adopted the new idea.

In this research context, the individual is also part of the social system using online

securities trading. The closer the linkage or behaviour of the individual retail investor

with other online retail investors positively influences the usage of trading stock

online. Thus, it is consistent that the Nature of the social system variable not only

affects the rate of adoption but also the post-adoption usage.

5.3.5 Conclusion for hypothesis five

(H5)0: There is no correlation between construct variable FS7-H5 (Extent of change

agent’s promotion efforts) and post-adoption behaviours of retail investors using


This hypothesis is rejected. The researcher therefore concludes that the extent of the

change agent’s promotion efforts is related positively with post-adoption usage

behaviour of retail investors. While external influences on pre-adoption such as

Collective and Authority decision have no influence, Change agent efforts on

promotion do have a positive effect on the degree or patterns of using the product by

the retail investors. It may be that, like the social system, the effect of the change

agent promotion efforts become internalised for the users, or at least has some lasting

effect on usage.


- 225 -

Rogers (2003) claimed that Change agents’ success in securing the adoption of

innovations by clients is positively related to the extent of change agent effort in

contacting clients. The main objective of the pre-adoption phase of most change

agents is to secure the adoption of new ideas by their clients. This hypothesis shows

that the Change Agent’s promotion effects not only influence the adoption of the new

ideas but also influence the post-adoption usage. Thus, it could be an indication for

brokerage firms to encourage change agents to contact clients regularly even if they

have subscribed to the online securities trading by the firm. This will positively

increase the usage of trading the stock online and hence bring more revenue to the

brokerage firms through higher volume of transaction commissions. Based on the

finding, it is suggested that the brokerage firms should continue to promote online

securities trading to the existing clients to enhance usage.

5.3.6 Conclusion for hypothesis six

(H6)0: There is no correlation between construct variable FS3-H6 (Perceived

usefulness and Complexity) and post-adoption behaviours of retail investors using


Hypothesis six, drawn from Davis’ (1989) Technology Acceptance Model, is rejected.

The research results would therefore support Davis’ model. Variables, Perceived

usefulness and Complexity, or lack of complexity, is related positively with post-

adoption usage behaviour of retail investors. Davis defined Perceived usefulness as

the degree to which a person believes that using a particular system would enhance

his or her job performance (Davis, F. D. 1989). In this hypothesis context, the

variable Complexity is referring to the level of complexity in accessing investment


- 226 -

information through online securities trading. In other words, the lack of effort in

using an online securities trading tool is a similar construct as Perceived ease of use

in Davis’ Technology Acceptance Model. Davis defined Perceived ease of use as the

degree to which the person believes that using a particular system would be free from

effort (Davis, F. D. 1989).

This hypothesis suggested that the level of Perceived usefulness in online securities

trading and lack of complexity, or perceived ease of use, influenced the degree or

patterns of using the product by the retail investors. In other words, if the retail

investor believes that the online securities trading system is useful and easy to use, it

affects positively on the post-adoption usage of the system.

This result is supported by previous researchers such as Naidoo and Leonard who

concluded that Perceived usefulness significantly influences post-adoption usage

(Naidoo & Leonard 2007). Bhattacherjee also found in his research that Perceived

usefulness is positively influencing the continued use of information systems

(Bhattacherjee 2001). Gefen and others also concluded in their study that perceived

usefulness is a strong predictor for repeat consumers (Gefen, Karahanna & Straub

2003).

5.3.7 Conclusion for hypothesis seven

(H7)0: There is no correlation between construct variable FS1-H7 (Consumer

loyalty) and post-adoption behaviours of retail investors using online securities

trading.


- 227 -

The final hypothesis was also rejected. Consumer loyalty is related positively with

post-adoption usage behaviour of retail investors. This hypothesis suggested that the

degree of consumer loyalty towards online securities trading influenced the degree or

patterns of using the product by the retail investors. This result supports Oliver’s

interpretation on consumer loyalty as a pattern of repeat purchasing. Intuitively we

would expect that consumer loyalty is a significant factor influencing increased usage

by retail investors of online securities trading (Oliver 1999).

The results of the hypothesis study have revealed that consumer loyalty is an

important element to be considered by the brokerage firms to ensure the continuing

usage of online securities trading by the retail investors. After getting customers to

subscribe to the online securities trading service, the brokerage firms should spend

efforts in after-sales service in order to maintain the customer loyalty and therefore

an increased probability of higher usage of the service provided. Facing the strong

competition in the brokerage market and low transaction commissions, customer

loyalty is a critical factor to retain the customers and to encourage higher usage and

thus more revenue from the online securities transactions.

5.4 Contributions to the body of knowledge

The previous sections discussed the research problem and the conclusions reached

from testing of the hypotheses. The findings and conclusions reached by this research

resulted in several significant contributions to the disciplines discussed in the

literature review of Chapter Two that are: Consequences of innovations; Technology

Acceptance Model and Consumer loyalty, in a number of ways.


- 228 -

5.4.1 Consequences of Innovations

The findings and conclusions in this research have made significant contributions to

the areas of post-adoption usage behaviour, and add to the limited investigation of

consequences of innovations. This research found that several pre-adoption factors

have a significant influence on the post-adoption usage behaviour of the retail

investors using online securities trading which are the consequences of adopting the

innovation. It showed that Trailability, Observability, Compatibility, Complexity

(lack of complexity or ease of use), Optional decision, Social system, Change agent’s

efforts, Perceived usefulness and Consumer loyalty all have a significant influence on

the post-adoption usage behaviour of the retail investors using online securities

trading.

The above discussion implies that brokerage firms should not only focus on getting

customers to subscribe to online securities trading, they need to understand the

influence between the pre-adoption factors and the post-adoption usage to ensure the

continuing usage of the product by the retail investors. Knowing all these pre-

adoption factors, they should consider enhancing the product or the services provided

to the retail investors. For example, continued promotion efforts could lead to higher

usage of the product.

5.4.2 Technology Acceptance Model

The findings and conclusions in this research have reinforced the importance of the

Perceived usefulness variable in the Technology Acceptance Model and have

proposed that there is significant influence of perceived usefulness on post-adoption

usage behaviour. The results of this study have also shown that a lower level of


- 229 -

complexity or Perceived ease of use by the retail investors has a positive influence on

the post-adoption usage.

This confirms past research and reinforces the literature on the Technology

Acceptance Model to show the Perceived usefulness and lack of complexity

(Perceived ease of use) are strong predictors of post-adoption usage.

It implies that brokerage firms could consider building the Perceived usefulness and

low level of complexity (Perceived ease of use) features into the online securities

trading service. This would create a higher chance of getting customers to subscribe

to the product as well as to increase the usage of the product. In summary, retail

investors trading online need to believe that the product they are using is useful and

free of effort in order to continue using the product.

5.4.3 Consumer loyalty

The findings and conclusions in this research have strengthened the past research and

conclude that there is a significant influence of consumer loyalty on post-adoption

usage behaviour. This supports Oliver’s finding as he stated that loyalty is a pattern

of repeat purchase from the consumer (Oliver 1999). The research findings and

results contribute to the literature in consumer loyalty and could recommend further

research on consumer loyalty and post-adoption usage behaviour. A deeper

understanding of consumer loyalty could imply that strengthening the loyalty and

confidence factors of the consumers might increase the usage of a product. Thus,

other researchers could conduct further investigation on what factors could increase

consumer loyalty and confidence and hence increase the product usage. This is

beyond the scope of this study.


- 230 -

The above discussion implies that brokerage firms could spend more effort in

promoting customer loyalty programs, and in building higher confidence in the retail

investors using the online securities trading product, as these will enhance the

probability of higher usage of the product and hence increase revenue through the

transaction commissions. The brokerage firms could investigate further on what

factors affect customer loyalty and confidence in the online securities trading service

and thus increase the usage.

5.5 Managerial Implications

The dependent variable of this research study, post-adoption usage behaviour of retail

investors using online securities trading is of critical importance to the financial

industry especially the brokerage firms. Online securities trading is a different

marketing product compared to other once-off sales products. Most brokerage firms

offer a minimum subscription fee or do not charge customers signing up to the online

securities trading service. The brokerage firms only earn the transaction commissions

when the customers use the online tool to track stocks. The brokerage commissions

for online securities trading is usually very low and thus the brokerage firms are

required to encourage the customers to have a higher usage of trading stock online

and thus bring in more earnings. The findings of this study indicated that there are

significant factors among the diffusion pre-adoption attributes, Perceived usefulness

and Consumer loyalty which influence the post-adoption usage behaviour. Thus, the

findings might provide more understanding of the reasons for continuing with online

securities trading and increasing the volume or frequency of usage. This will have a

significant impact on the brokerage firm, since the commission earnings are directly


- 231 -

related to the extent of use. The brokerage firms would be better informed by

understanding the factors affecting usage of the online securities trading so as to

maintain a sustainable financial advantage. Continuing usage or higher usage of

online securities trading by the retail investors is definitely a financial advantage to

the brokerage firms.

Based on the understanding of this relationship of the pre-adoption factors,

Technology acceptance factor and Consumer loyalty factor, brokerage firms should

build in all these features to enhance the online securities trading service. Brokerage

firms should make the product less complex and build in more features that retail

investors believe are useful to them in trading stock online. These should be

considered in the product development strategy by the brokerage firms when

designing newer version of online securities trading products. In addition, brokerage

firm should also focus on after-sales service and their marketing communication

strategy to continue to promote the online securities trading to the customers by

informing of any new update in the features, or by constantly contacting customers

about any special discount in transaction commission fees. All of these efforts might

be able to increase the probability of higher usage of trading stock online by the retail

investors.


- 232 -

5.6 Limitations of the research

This research study has several limitations that need to be acknowledged. Although

this research has provided relevant study and understanding of post-adoption usage

behaviour of retail investors using online securities trading, it is important to address

the limitations to provide awareness for future researchers.

5.6.1 Scope limitation

This research is only limited to the study of pre-adoption factors affecting the post-

adoption consequences of online securities trading and not any other online

technologies. There are online tools offering many other functions like online

banking services, online foreign exchange trading, online insurance, online portfolio

management and online financial information. Conclusions may be drawn that

similar effects may apply in these online environments, but further research into these

specific areas would be needed to confirm these conclusions.

5.6.2 Geographical limitation

The sampling for this research is limited to retail investors in Singapore only and

does not include other retail investors from other countries. The trading behaviours

and usage of the online securities trading tool could be different for retail investors

from other countries although the research should lead to reasonably general

conclusions that would apply in other countries. The pre-adoption factors influencing

the retail investors from other countries could be different as compared to the retail

investors in Singapore though there would be many similarities. Thus, the pre-


- 233 -

adoption factors might have a different degree of influence on post-adoption usage

behaviour of retail investors.

5.6.3 Online questionnaire

The use of an online questionnaire limits the pool for sampling to Internet users only.

However, those people who do not have access to the Internet are not the target

sampling for this research. Non-Internet users are not likely to be traders online, so

this limitation of the sample pool is not considered a significant delimitation. The

research is only limited to experienced Internet users who are comfortable in

responding to the online questionnaire, based on the assumption that they are capable

of accessing the Internet since they are adopters of Internet based online securities

trading which is also based on Internet and web browser technologies.

5.7 Recommendations for future research

5.7.1 Other products

The replication of this research can be applied to other online products, for example

online banking or online financial investment. Another possible product for further

research on pre-adoption influence on post-adoption usage is the trading of foreign

exchange currencies online. Future research could also be carried out in other

industries beside financial, for example, online education, or online purchases of

consumer goods, or online purchases of business inputs products. The theoretical

model in this research can be a recommended model to investigate how pre-adoption

factors and Technology acceptance factors influence post-adoption usage behaviour

of other innovations, products or services.


- 234 -

5.7.2 Other geographical regions

The replication of this research could be applied in other geographical regions. This

research is focused on the retail investors in Singapore and using the online securities

trading tool based in Singapore. The findings might be different if the study was to be

applied in other countries which would add more value to the area of study.

5.7.3 Other factors affecting post-adoption consequences

The research has investigated pre-adoption factors as well as Technology acceptance

factors that affect the post-adoption consequences of online securities trading,

however there may be many other factors that influence trading behaviour. Future

research should build on the findings of this study and should try to identify

additional factors associated with post-adoption consequences of online securities

trading. For example, the latest technology infrastructure and enhanced product

functionality might be of interest to future studies on post-adoption behaviour. The

concerns of security threats in trading stock online might have an influence on post-

adoption behaviour as well.

5.7.4 Other post-adoption consequences

One might also attempt to include other post-adoption behaviour or consequences of

using online securities trading by the retail investors. For example, one might use the

model to study whether there are impacts of pre-adoption factors on financial returns

in using online securities trading. Another possible post-adoption behaviour could be

the trading strategy or patterns of trading by the retail investors trading online.


- 235 -

5.8 Research conclusion

In this chapter, the final conclusions and implications of this research are presented.

Firstly, the research problem and the related research issues of this research are

restated.

Many researches in the past have been conducted in understanding the factors

affecting the adoption of new innovations. However, few studies have been

conducted in the pre-adoption factors and Technology Acceptance Model factors

influencing the post-adoption usage of an innovation. This study found that there are

positive influences of pre-adoption factors, Technology Acceptance Model factors

and Consumer loyalty on the post-adoption usage of online securities trading by the

retail investors in Singapore. The research also found that external factors like

Authority decision, Collective decision and Communication channels have no

significant influence on post-adoption usage of trading stock online. These imply that

brokerage firms should perhaps focus on internal factors affecting the retail investors,

to increase the probability of higher usage of the online securities trading and thus

generate higher transaction commissions for the company.

Implications were drawn on how this research contributed to the existing body of

knowledge and the field of online securities trading. This was followed by a

discussion of the research limitations, and recommendations for future research were

explored. By understanding which pre-adoption factors influence post-adoption usage

behaviour, online brokerage companies can focus their product development and

marketing communication strategies to increase the use of online securities trading


- 236 -

and thus expand their businesses. Based on the current theoretical model, future

researchers could investigate how pre-adoption factors influence post-adoption usage

behaviour of other similar products and services.

- 237 -

Bibliography and References 1

Ajzen, I 1985, 'From intentions to actions: A theory of planned behavior.' In J.Kuhi & J. Beckmann (EDs.), Action - control: From cognition to behavior, pp. 11-39.

Ajzen, I & Fishbein, M 1980, Understanding Attitudes and Predicting Social

Behavior., Prentice-Hall, New Jersey. AMInvestment 2007, Fraser Direct Online Securities, <http://www.ambg.com.my/>. Antonides, G & Raaij, WFv 1998, Consumer Behaviour A European Perspective,

First Edition edn, Wiley. Bass, FM 1969, 'A New Product Growth For Model Consumer Durables',

Management Science, vol. 15, no. 5, p. 215. Bhattacherjee, A 2001, 'Understanding Information Systems Continuance: An

Expectation Confirmation Model', MIS Quarterly, vol. 25, no. 3, pp. 351-70. Bryman, A 1984, 'The Debate about Quantitative and Qualitative Research: A

Question of Method or Epistemology?' The British Journal of Sociology, vol. 35, no. 1, pp. 75-92.

CDP 2008, Securities Account for an Individual 2008, <http://www.cdp.com.sg/account/create_individual.html>. Chea, S & Luo, MM 2007, 'Cognition, Emotion, Satisfaction, and Post-Adoption

Behaviors of E-Service Customers', paper presented to 40th Annual Hawaii International Conference, Hawaii.

Cheung, CM, Chan, GW & Limayem, M 2005, 'A Critical Review of Online

Consumer Behaviour: Empirical Research', Journal of Electronic Commerce in Organizations, vol. 3, no. 4, pp. 1-19.

Chua, HH 2008, 'Power failure down broker's trading system', Straits Times, 16 Jan

2008. Coakes, sJ, Steed, L & Price, J 2008, SPSS Version 15.0 for Windows: Analysis

without Anguish, First edition edn, Wiley. Davis, F, Bagozzi, RP & Warshaw, PR 1989, 'User Acceptance of Computer

Technology: A Comparison of Two Theoretical Models', Management Science, vol. 35, no. 8, p. 982.

- 238 -

Davis, FD 1989, 'Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology', MIS Quarterly, vol. 13, no. 3, p. 319.

Dearing, J & Singhal, A 2006, Communication of Innovations: A Journey with Everett

M. Rogers., Sage, London. DeForge, K 2001, Getting Started in Online Broker, 1 edn, Wiley. Deshpande, R 1983, 'Paradigms Lost: On Theory and Method in Research in

Marketing', Journal of Marketing, p. 101. Dillon, A & Morris, MG 1996, 'User Acceptance of Information Technology:

Theories and Models', Annual Review of Information Science and Technology, vol. 31, pp. 3 - 32.

Donio, J, Massari, P & Passiante, G 2006, 'Customer satisfaction and loyalty in a

digital environment: an empirical test', Journal of Consumer Marketing, vol. 23, no. 7, pp. 445-57.

Easterby-Smith, M, Thorpe, R & Lowe, A 1991, The Philosophy of Research Design,

Sage, London. EL-Gayar, O & Moran, M 2007, 'Examing Students' Acceptance of Tablet PC using

TAM', Issues in Information Systems, vol. VIII, no. 1. Field, A 2005, Discovering statistics using SPSS 2nd edition edn, Sage Publications,

London. Fishbein, M & Ajzen, I 1975, Belief, attitude, intention, and behavior: An

introduction to theory and research, Addison-Wesley. Floh, A & Treiblmaier, H 2006, 'WHAT KEEPS THE E-BANKING CUSTOMER

LOYAL?' Journal of Electronic Commerce Research, vol. 7, no. 2. Formsite 2008, Formsite website, viewed 2008 <Error! Hyperlink reference not

valid.. Garber, LL & Dotson, MJ 2002, 'A method for the selection of appropriate business-

to-business integrated marketing communications mixes', Journal of Marketing Communications, vol. 8, pp. 1-17.

Gatignon, H & Robertson, TS 1985, 'A Propositional Inventory for New Diffusion

Research', Journal of Consumer Research, vol. 11, no. 4, p. 849.

- 239 -

---- 1989, 'Technology Diffusion: An Empirical Test of Competitive Effects', Journal of Marketing, vol. 53, pp. 35-49.

Gefen, D, Karahanna, E & Straub, DW 2003, 'TRUST and TAM in online shopping:

An Integrated Model', MIS Quarterly, vol. 27, no. 1, p. 51. Gharavi, H, Love, PED & Cheng, EWL 2004, 'Information and communication

technology in the stockbroking industry: and evolutionary approach to the diffusion of innovation', Industrial Management & Data Systems, vol. 104, no. 9, pp. 756-65.

Glaser, M 2003a, Online Broker Investors: Demographic Information, Investment

Strategy, Portfolio Positions, and Trading Activity, Universitat Mannheim, 2003.

---- 2003b, Online Broker Investors: Demographic Information, Investment Strategy,

Portfolio Positions, and Trading Activity, University of Mannheim. Guba, EG & Lincoln, YS 1994, 'Competing Paradigms in Qualitative Research', in

NK Denzin & YS Lincoln (eds), Handbook of Qualitative Research, Sage, Thousand Oaks, pp. 105-17.

Hair, JF, Anderson, RE, Tatham, RL & Black, WC 1998, Multivariate Data Analysis,

Fifth Edition edn, Prentice Hall. Hanekom, J 2006, 'A theoretical framework for the online consumer response process',

Master of Arts thesis, University of South Africa. Healy, M & Perry, C 2000, 'Comprehensive criteria to judge validity and reliability

of qualitative research within the realism paradigm', Qualitative Market Research, vol. 3, no. 3, pp. 118-26.

Hennig-Thurau, T, Gwinner, KP & Gremler, DD 2002, 'An Integration of Relational

Benefits and Relationship Quality', Journal of Service Research, vol. 4, no. 3, pp. 230-47.

Iyengar, JV 2004, 'A Discussion of current and potential issues relating information

security for Internet communications', Competitiveness Review, vol. 14, no. 1/2, p. 90.

Jacoby, J & Kyner, DB 1973, 'Brand Loyalty Vs. Repeat Purchasing Behavior',

Journal of Marketing Research, vol. 10, pp. 1-9.

- 240 -

Jasperson, JS, Carter, PE & Zmud, RW 2005, 'A Comprehensive Conceptualization Of Post-Adoptive Behaviors Associated With Information Technology Enabled Work Systems', MIS Quarterly, vol. 29, no. 3, p. 525.

Johnson, RB & Onwuegbuzie, AJ 2004, 'Mixed Methods Research: A Research

Paradigm Whose Time Has Come', Educational Researcher, vol. 33, no. 7, p. 14.

Karahanna, E, Straub, DW & Chervany, NL 1999, 'Information technology adoption

across time: a cross-sectional comparison of pre-adoption and post-adoption beliefs', MIS Quarterly, vol. 23, no. 2, p. 183.

Kendall, J, Tung, LL, Chua, KH, Ng, DCH & Tan, SM 2001, 'Electronic Commerce

Adoption by SMEs in Singapore', paper presented to 34th Hawaii International Conference on System Sciences, Hawaii.

Kong, W 1999, 'Internet Trading: A New Singapore Passtime?' Financial Planner,

July 1999, pp. 10-21. Kotler, P 2000, Marketing Management, International Edition edn, Prentice Hall. Krauss, SE 2005, 'Research Paradigms and Meaning Making: A Primer', The

Qualitative Report, vol. 10, no. 4, pp. 758-70. Kurnia, S & Chien, A-WJ 2003, 'The Acceptance of Online Grocery Shopping', paper

presented to 16th Bled eCommerce Conference, Slovenia, 9-11 June 2003. Lau, A, Yen, J & Chau, PYK 2001, 'Adoption of On-line Trading in the Hong Kong

Financial Market', Journal of Electronic Commerce Research, vol. 2, no. 3, p. 2001.

Lavidge, RJ & Steiner, GA 1961, 'A Model For Predictive Measurements of

Advertising Effectiveness', Journal of Marketing, pp. 59-62. Liew, A 2006, 'SGX StockWhiz Online Trading Challenge: "Seize The Day!"', paper

presented to SGX StockWhiz, Singapore, 6 December 2006. Luo, MM-L, Remus, W & Chea, S 2006, 'Technology Acceptance of Internet-based

Information Services: An Integrated Model of TAM and U&G Theory', paper presented to Twelfth Americas Conference on Information Systems, Mexico, August 4 2006.

MAS 2001, Securities and Futures Act (Chapter 289).

- 241 -

Mason, RD, Lind, DA & Marchal, WG 2000, Statistical Techniques in Business and Economics, Tenth Edition edn, McGraw Hill.

McKechnie, S, Winklhofer, H & Ennew, C 2006, 'Applying the technology

acceptance model to the online retailing of financial services', International Journal of Retail & Distribution Mangement, vol. 34, no. 4/5.

McMullan, R & Gilmore, A 2008, 'Customer loyalty: an empirical study', European

Journal of Marketing, vol. 42, no. 9/10, pp. 1084-904. Naidoo, R & Leonard, A 2007, 'Perceived usefulness, service quality and loyalty

incentives: Effects on electronic service continance', South African Journal of Business Management, vol. 38, no. 3.

Neal, C, Del, P & Haskins, Q 2004, Consumer Behaviour - Implications for

Marketing Strategy 4th Edition edn, McGraw Hill. Oliver, RL 1999, 'Whence Consumer Loyalty?' Journal of Marketing, vol. 63, p. 33. Pallant, J 2007, SPSS Survival Manual, 3rd Edition edn, A&U. Parhasarathy, M & Bhattacherjee, A 1998, 'Understanding Post-Adoption Behavior in

the Context of Online Services', Information Systems Research, vol. 9, no. 4. Park, K 1998, 'INNOVATIVE PRODUCT USAGE BEHAVIOR IN THE POST-

ADOPTION PROCESS', Asia Pacific Advances in Consumer Research vol. 3. Parthasarathy, M & Bhattacherjee, A 1998, 'Understanding Post-Adoption Behavior

in the Context of Online Services ', Information Systems Research, vol. 9, no. 4. Pelham, BW & Blanton, H 2003, Conducting Research in Psychology, Second

Edition edn, Thomson. Perry, C 2002, A Structured Approach To Presenting Theses: Notes For Students And

Their Supervisors. Perry, C, Riege, A & Brown, L 1999, 'Realism's Role Among Scientific Paradigms in

Marketing Research', Irish Marketing Review, vol. 12, no. 2, p. 16. POEMS-Seminars 2008, <http://www.poems.com.sg/Seminar/seminar.asp>. POEMS 2008.

- 242 -

Rigopoulos, G & Askounis, D 2007, 'A TAM Framework to Evaluate Users' Perception towards Online Electronic Payments', Journal of Internet Banking and Commerce, vol. 12, no. 3.

Robertson, TS & Gatignon, H 1986, 'Competitive Effects on Technology Diffusion',

Journal of Marketing, vol. 50, no. 3, p. 1. ---- 1991, 'How Innovators Thwart New Entrants into Their Market', Planning Review,

vol. 19, no. 5, p. 4. Rogers, EM 2003, Diffusion of Innovations, Fifth Edition edn, Free Press. Rummel, RJ 1967, 'Understanding Factor Analysis', The Journal of Conflict

Resolution, vol. 11, no. 4, pp. 444-80. Schiffman, L, Bednall, D, Cass, AO, Paladino, A & Kanuk, L 2005, Consumer

Behaviour, 3rd Edition edn, Pearson. SGX 2006, Annual Report, Singapore. ---- 2008, 2008, <Error! Hyperlink reference not valid.. Shankar, GD 2002, 'Impact of e-Business on stock broking', paper presented to

Australian Computer Society Australia, 26 Jan 2002. Sharma, MK & Bingi, P 2000, 'The Growth Of Web-Based Investement', Information

Systems Management. Sharma, P & Maleyeff, J 2003, 'Internet education: potential problems and solutions',

The International Journal of Educational Management, vol. 17, no. 1, pp. 19-25.

Shorter-Judson, BG 2000, 'Adoption and Diffusion of Innovations in the Airline

Industry: An Investigation of Consumer Preference for Alternative Ticketing Methods (The Case of Electronic Ticketing)', PhD thesis, Golden Gate University.

Smart-Investor 2000, 'e-Trading: Phase Two', Smart Investor, pp. 50-6. Sohn, SY & Ahn, BJ 2003, 'Multigeneration diffusion model for economic

assessment of new technology', Technological Forecasting & Social Change, vol. 70, pp. 251–64.

Solomon, MR 1999, Consumer Behaviour, Fourth Edition edn, Prentice Hall.

- 243 -

SPSS 2000, Advanced Statistical Analysis Using SPSS, SPSS Inc. Strong, EK 1925, The Psychology of Selling, McGraw Hill, New York. Talyor, S & Todd, P 1995, 'Decomposition and crossover effects in the theory of

planned behavior: A study of consumer adoption intentions', International Journal Of Research in Marketing, vol. 12, no. 2, p. 137.

Tan, A 2000, 'Phillip's Poems puts the writing on the wall', Business Times, p. 3. Tan, M & Teo, TSH 2000, 'Factors Influencing the Adoption of Internet Banking',

Journal of the Association for Information Systems, vol. 1. Teo, TSH, Tan, M & Peck, SN 2004, 'Adopters and Non-Adopters of Internet Stock

Trading in Singapore', Behaviour and Information Technology, vol. 23, no. 3, pp. 211-23.

Teweles, RJ & Bradley, ES 1998, The Stock Market, Seventh Edition edn, Wiley. TSI 2006a, Online Stock Trading in Asia. ---- 2006b, 'Internet Trading'. UCLA 2009a, What does Cronbach's alpha mean?, UCLA Academic Technology

Services, viewed 24th Jan 2009 2009, <http://www.ats.ucla.edu/stat/spss/faq/alpha.html>.

---- 2009b, Annotated SPSS Output - Regression Analysis, UCLA, viewed 24th Jan

2009 2009, <http://www.ats.ucla.edu/stat/Spss/output/reg_spss.htm>. Venkatesh, V, Morris, MG, Davis, GB & Davis, FD 2003, 'User Acceptance of

Information Technology: Toward A Unified View', MIS Quarterly, vol. 27, no. 3, pp. 42-478.

Vitartas, P, Jayne, N, Ellis, A & Rowe, S 2007, 'Student adoption of web base video

conferencing software: A comparison of three student discipline groups', paper presented to ascilite Singapore 2007, Singapore, 2007.

Wells, WD & Presky, D 1996, Consumer Behavior, First Edition edn, Wiley. Werner, M & Murphy, R 2007, 'On-Line Business: Is There Loyalty?' The Business

Review, vol. 9, no. 1, p. 250.

- 244 -

Zhu, K & Kraemer, KL 2005, 'Post-Adoption Variations in Usage and Value of E-Business by Organizations: Cross-Country Evidence from the Retail Industry', Information Systems Research, vol. 16, no. 1, pp. 61-84.

Zhu, K, Dong, S, Xu, SX & Kraemer, KL 2006, 'Innovation diffusion in global

contexts: determinants of post-adoption digital transformation of European companies', European Journal of Information Systems, vol. 15, pp. 601-16.

Zikmund, WG 2000, Business Research Methods, Sixth Edition edn, Dryden.

- 245 -

Appendix A Letter of ethic approval form

- 246 -

- 247 -

Appendix B Letter of Introduction

- 248 -

Appendix C Questionnaire

Online Securities Trading Survey

Please take some time to fill out this survey about online securities trading in

Singapore. Please tick the appropriate answers for the questions below.

PART A: Online Securities Trading Opinions (Please tick the answer according to the scale indicated)

1. To what extent do you agree with the following statements about Online Securities Trading as compared to trading via the broker?

Strongly Disagree

Disagree Slightly Disagree

Neutral Slightly Agree

Agree Strongly Agree

(a) Quicker to trade online (b) Cheaper to trade online (c) Process is not much different from calling the broker

(d) Trading information is not much different from calling the broker

(e) Easier to access investment information online

(f) Easier to trade using Online Securities Trading

(g) Able to do a trial trade which is not possible via the broker

(h) Easier to obtain online demonstration and explanation

(i) More investors signed up for Online Securities Trading system

(j) More investors started to trade using Online Securities Trading

- 249 -

2. To what extent do you agree with the following statements about your selection of Online Securities Trading? Strongly

Disagree Disagree Slightly

Disagree Neutral Slightly

Agree Agree Strongly

Agree (a) Have considered other options like automatic voice trading or WAP trading via phone?

(b) Have considered online trading as an additional method of trading

(c) Have consulted other investors using Online Securities Trading

(d) Have consulted my friends using Online Securities Trading

(e) Have been advised by investment experts to sign up Online Securities Trading

(f) Stock Exchange has liberalized brokerages' commission rate for Online Securities Trading

3. To what extent do you agree that exposure to Online Securities Trading are through the following means: Strongly




Agree (a) Online advertisements like Internet or email

(b) Mass media like TV or newspaper advertisements

(c) Broker's explanation and demonstration

(d) Friends' and other investors' advice

4. To what extent do you agree with the following statements about your relationship with other Online Securities Trading users? Strongly




Agree (a) Know more friends who used Online Securities Trading

(b) I feel left out if I do not sign up Online Securities Trading

(c) Consult or discuss with friends or other investors when I trade online

(d) Exchange investment information with friends or other investors using Online Securities Trading

- 250 -

5. To what extent do you agree with the following statements about the promotional efforts of Online Securities Trading? Strongly




Agree (a) Constantly received Online Securities Trading information

(b) Satisfaction with the broker's promotional efforts on Online Securities Trading

(c) Satisfaction with the brokerage firm's promotional efforts on Online Securities Trading

6. How long did you take to adopt the Online Securities Trading system counting from the day you are aware of it?

less than a week about a month 3 to 6 months about a year

more than a year

7. To what extent do you agree with the usefulness of Online Securities Trading? Strongly




Agree (a) Trading Online increases my trading profit

(b) The system facilitates diversification of my portfolio

(c) I can react to the stock market quicker

8. To what extent do you agree with the following statements about your loyalty to Online Securities Trading? Strongly




Agree (a) Online Securities Trading will be my major method to trade

(b) I will introduce Online Securities Trading to non-users

(c) I will not consider other new methods of trading in near future

- 251 -

9. To what extent do you agree with the following statements about your confidence of Online Security Trading? Strongly




Agree (a) It is a highly secured system (b) It is easily accessible (c) It is very reliable

PART B: Online Securities Trading Usage

10. How often do you trade online in the last 12 months? Daily Weekly Every 2 Weeks Monthly Every 3 months

Every 6 months or more

11. To what extent do you agree with the following statements about using Online Securities Trading as compared to trading via the broker? Strongly




Agree (a) I trade more frequently now (b) I trade in smaller lot sizes now (c) I buy certain categories of stock (eg. High Tech Stocks)frequently

(d) I trade from more locations like office, Cybercafé and overseas in addition to home

(e) I exchange investment tips with other online investors easily

(f) I check investment information frequently

12. Which of the following information you will usually used to conduct your online trading?

Live prices Historical prices Stock charts Company information

Financial News Portfolio Management Newsgroup discussion

Financial planning Trading (buy / sell)

- 252 -

PART C: Demographic Section 13. What is your age range?

21 – 30 31 – 45 46 – 55 56 & up

14. What is your highest education level? Secondary & below Junior College & equivalent

Diploma & equivalent

Bachelor degree & equivalent Master degree & above

15. What is your current occupation? Executive and Managerial Professional

Technical, Production & related

Administration, Sales & Services Self-employed

Others (Please specify)___________________

16. What is your average annual income range in SGD? 20,000 & below 21,000 - 35,000 35,001 - 50,000

50,001 - 65,000 65,001 - 80,000 80,001 - 100,000 100,001 & above

17. What is your marital status? Single Married Others (Please specify)

__________________________

18. What is your gender? Male Female

** End of Survey **

- 253 -

Appendix D SPSS Outputs

Factor Analysis

Notes

Output Created 21-OCT-2008 20:48:33 Comments Input Data

C:\DBA\Anthony's thesis\Chapter 4 Data Analysis\20102008\20102008.sav

Active Dataset DataSet1 Filter <none> Weight <none> Split File <none> N of Rows in

Working Data File 232

Missing Value Handling

Definition of Missing MISSING=EXCLUDE: User-defined missing values are treated as missing.

Cases Used LISTWISE: Statistics are based on cases with no missing values for any variable used.

Syntax FACTOR /VARIABLES freq volume type location investtips investinfo /MISSING LISTWISE /ANALYSIS freq volume type location investtips investinfo /PRINT INITIAL CORRELATION SIG DET KMO EXTRACTION ROTATION FSCORE /FORMAT SORT BLANK(.50) /PLOT EIGEN /CRITERIA MINEIGEN(1) ITERATE(25)

- 254 -

/EXTRACTION PC /CRITERIA ITERATE(25) /ROTATION VARIMAX /SAVE REG(ALL) /METHOD=CORRELATION .

Resources Elapsed Time

0:00:00.39

Maximum Memory Required 5928 (5.789K) bytes

Processor Time 0:00:00.35

Variables Created FAC1_2 Component score 1

- 255 -

Correlation Matrix(a)

frequency Volume Type location investment tips exchange

investment information frequency

frequency 1.000 .469 .433 .409 .586 .600 Volume .469 1.000 .422 .258 .452 .492 Type .433 .422 1.000 .457 .413 .402 location .409 .258 .457 1.000 .512 .463 investment tips exchange .586 .452 .413 .512 1.000 .848

Correlation

investment information frequency .600 .492 .402 .463 .848 1.000

frequency .000 .000 .000 .000 .000 Volume .000 .000 .000 .000 .000 Type .000 .000 .000 .000 .000 location .000 .000 .000 .000 .000 investment tips exchange .000 .000 .000 .000 .000

Sig. (1-tailed)

investment information frequency .000 .000 .000 .000 .000

a Determinant = .059

- 256 -


Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .810

Approx. Chi-Square 645.325 df 15

Bartlett's Test of Sphericity

Sig. .000

Communalities

Initial Extraction frequency 1.000 .606 Volume 1.000 .448 Type 1.000 .445 location 1.000 .450 investment tips exchange 1.000 .747 investment information frequency 1.000 .745


- 257 -


Component Initial Eigenvalues Extraction Sums of Squared Loadings Total % of Variance Cumulative % Total % of Variance Cumulative % 1 3.442 57.366 57.366 3.442 57.366 57.366 2 .766 12.761 70.127 3 .737 12.277 82.403 4 .472 7.871 90.274 5 .435 7.258 97.532 6 .148 2.468 100.000


- 258 - Component Number

654321

Eigenvalue

4

3

2

1

0

Scree Plot

- 259 -

Component Matrix(a)

Component

1 investment tips exchange .864 investment information frequency .863

frequency .779 location .671 Volume .670 Type .667


a 1 components extracted.

Rotated Component Matrix(a)

a Only one component was extracted. The solution cannot be rotated.

- 260 -

Component Score Coefficient Matrix

Component

1 frequency .226 Volume .195 Type .194 location .195 investment tips exchange .251 investment information frequency .251


Rotation Method: Varimax with Kaiser Normalization.

Component Scores.

- 261 -

Component Score Covariance Matrix

Component 1 1 1.000


Rotation Method: Varimax with Kaiser Normalization.

Component Scores.

- 262 -

Regression

Notes

Output Created 22-OCT-2008 20:35:48 Comments

Data C:\DBA\Anthony's thesis\Chapter 4 Data Analysis\22102008\22102008.sav

Active Dataset DataSet1 Filter <none> Weight <none> Split File <none>

Input

N of Rows in Working Data File 232

Definition of Missing User-defined missing values are treated as missing.

Missing Value Handling

Cases Used Statistics are based on cases with no missing values for any variable used.

Syntax REGRESSION /MISSING LISTWISE /STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT FAC1_2 /METHOD=ENTER FAC1_1

- 263 -

FAC2_1 FAC3_1 FAC4_1 FAC5_1 FAC6_1 FAC7_1 FAC8_1 FAC9_1 /SCATTERPLOT=(*ZRESID ,*ZPRED ) /RESIDUALS DURBIN NORM(ZRESID) /SAVE MAHAL .

Elapsed Time

0:00:00.86

Memory Required 6340 bytes Additional Memory Required for Residual Plots

504 bytes

Resources

Processor Time 0:00:00.79

Variables Created or Modified

MAH_5 Mahalanobis Distance

- 264 -

Variables Entered/Removed(b)

Model Variables Entered

Variables Removed Method

1 FS9 VR - H2b, FS8 VR - H2a, FS7 VR - H5, FS6 VR - H1b, FS5 VR - H3, FS4 VR - H1a, FS3 VR - H6, FS2 VR - H4, FS1 VR - H7(a)

. Enter

a All requested variables entered.

b Dependent Variable: FS VR - Dep Var

- 265 -

Model Summary(b)

Model R R Square Adjusted R Square

Std. Error of the Estimate Durbin-Watson

1 .727(a) .529 .510 .69996262 2.081 a Predictors: (Constant), FS9 VR - H2b, FS8 VR - H2a, FS7 VR - H5, FS6 VR - H1b, FS5 VR - H3, FS4 VR - H1a, FS3 VR - H6, FS2 VR

- H4, FS1 VR - H7


ANOVA(b)

Model Sum of Squares df Mean Square F Sig.

Regression 122.232 9 13.581 27.720 .000(a) Residual 108.768 222 .490

1

Total 231.000 231 a Predictors: (Constant), FS9 VR - H2b, FS8 VR - H2a, FS7 VR - H5, FS6 VR - H1b, FS5 VR - H3, FS4 VR - H1a, FS3 VR - H6, FS2 VR

- H4, FS1 VR - H7


- 266 -

Coefficients(a)

Model Unstandardized Coefficients Standardized Coefficients t Sig.

B Std. Error Beta B Std. Error 1 (Constant) -1.08E-016 .046 .000 1.000 FS1 VR - H7 .332 .046 .332 7.209 .000 FS2 VR - H4 .439 .046 .439 9.535 .000 FS3 VR - H6 .344 .046 .344 7.475 .000 FS4 VR - H1a .118 .046 .118 2.554 .011 FS5 VR - H3 .079 .046 .079 1.706 .089 FS6 VR - H1b .184 .046 .184 3.994 .000 FS7 VR - H5 .171 .046 .171 3.716 .000 FS8 VR - H2a .035 .046 .035 .766 .444 FS9 VR - H2b .152 .046 .152 3.304 .001

a Dependent Variable: FS VR - Dep Var

- 267 -

Residuals Statistics(a)

Minimum Maximum Mean Std. Deviation N Predicted Value -

2.2308719 1.8464117 .0000000 .72742092 232

Std. Predicted Value -3.067 2.538 .000 1.000 232 Standard Error of Predicted Value .067 .289 .140 .039 232

Adjusted Predicted Value -1.9453162 1.8262173 -.0001311 .72520194 232

Residual -2.73056602

1.65274191 .00000000 .68619152 232

Std. Residual -3.901 2.361 .000 .980 232 Stud. Residual -4.112 2.402 .000 1.007 232 Deleted Residual -

3.03341293

1.75201023 .00013110 .72387277 232

Stud. Deleted Residual -4.268 2.429 -.002 1.015 232 Mahal. Distance 1.141 38.394 8.961 5.904 232 Cook's Distance .000 .188 .006 .016 232 Centered Leverage Value .005 .166 .039 .026 232

a Dependent Variable: FS VR - Dep Var

- 268 -

Charts

Normal P-P Plot of Regression Standardized Residual

Dependent Variable: FS VR - Dep Var

Observed Cum Prob1.00.80.60.40.20.0

Expected Cum Prob

1.0

0.8

0.6

0.4

0.2

0.0

- 269 -

Regression Standardized Predicted Value20-2-4

Regression Standardized Residual2

0

-2

-4

Scatterplot

Dependent Variable: FS VR - Dep Var

- 270 -

Coefficients(a) without Constant

Coefficientsa,b

.332 .046 .332 7.225 .000

.439 .046 .439 9.557 .000

.344 .046 .344 7.492 .000

.118 .046 .118 2.560 .011

.079 .046 .079 1.710 .089

.184 .046 .184 4.003 .000

.171 .046 .171 3.724 .000

.035 .046 .035 .768 .443

.152 .046 .152 3.312 .001

FS1-H7FS2-H4FS3-H6FS4-H1aFS5-H3FS6-H1bFS7-H5FS8-H2aFS9-H2b

Model1

B Std. Error

UnstandardizedCoefficients

Beta

StandardizedCoefficients

t Sig.

Dependent Variable: FS-Dep Vara.

Linear Regression through the Originb.

- 271 -

~ End of Thesis ~

Documents

Factors influencing the post-adoption consequences of online