CHAPTER 3: RESEARCH METHODOLOGY

3.1 INTRODUCTION

The previous chapter provides theories related to e-commerce adoption among

SMEs. This chapter presents the proposed model framework, the development

of hypotheses, as well as the selection of measures and questionnaire design.

The sampling design and data collection procedure, and the analysis technique

used in this study are also explained.

3.2 DEVELOPMENT OF HYPOTHESES

Many past studies have advocated that the use of e-commerce positively

influences an organisation’s performance, specifically in terms of growth

(Raymond et al. 2005), financial gain (Johnston et al., 2007), and competitive

advantage (Teo, 2007). However this study is more interested in how the SMEs

perceive those benefits. Examples of benefits typically associated with e-

commerce adoption are: increase of revenues and profits, reduction in costs,

improvement in customer service, development of new market segments and

the streamline of business operations. Therefore, hypothesis 1 reflects the

relationship described:

24
H1: The perceived relative advantage is positively related to adoption of e-

commerce

In this study, perceived compatibility refer to how well e-commerce fits with the

current business processes as well as the suppliers and customers. According

to Grandon and Pearson (2004), compatibility between e-commerce and firm’s

culture, values and preferred work practices is an important factor in determining

adoption. Consequently, the lack of organisational compatibility may impose

constraints on the level of e-commerce used (Gibbs and Kramer, 2004).

Therefore, Hypothesis 2 reflects the level of perceived compatibility and e-

commerce adoption:

H2: The perceived compatibility is positively related to adoption of e-commerce.

Perceived complexity is defined as the degree to which a firm perceives the

adoption of e-commerce would be complicated. Many studies have found

significant relationship between complexity and ecommerce adoption (e.g Huy

and Filiatraut 2006; Joen et al. 2006). SMEs might perceive e-commerce as

something which is complex and not applicable to their current level of business.

Since SMEs are known to have low level of managerial and technical skills, they

might perceive implementation of e-commerce as very challenging (Hussin et

al., 2008). Hence, Hypothesis 3 was derived:

25
H3: The perceived complexity is negatively related to adoption of e-commerce.

In the context of Malaysia, among the major barriers to e-commerce adoption in

Malaysia are: not having e-commerce knowledge, lack of e-commerce skills,

and lack of skilled workers to operate e-commerce (Sulaiman and Jaafar 2003).

Therefore, Hypothesis 4 is derived:

H4: The e-commerce knowledge and expertise is positively related to adoption

of e-commerce.

According to Saffu et al. (2008), top management’s enthusiasm to adopt

information technology is one of the factors that contribute towards adopting of

e-commerce. Hypothesis 5 is postulated to reflect this relationship:

H5: The management attitude towards e-commerce is positively related to

adoption of e-commerce

As described in the literature review section, the external change agents that

might influence e-commerce adoption include the organisation advisors, the

government, external consultants, and also e-commerce solution providers.

Hence, Hypothesis 6 is derived:

H6: The adoption of e-commerce is positively related to external change agents.

26
A number of recent studies have discussed how the interactions with other

organisations in a global economy have enforced these SMEs to adopt e-

commerce technologies (e.g Hinson and Abor, 2005; Sarosa and Underwood,

2005; Nirjar and Tylecote, 2005; Jaw and Chen, 2005; Hinson and Sorensen,

2006). This shows that the pressure from trading partners also affect the

whether e-commerce is adopted. Hence, Hypotheses 7 reflects the relationship

postulated:

H7: The adoption of e-commerce is positively related to pressure from trading

partners

Relationships among players within the same industry also affect the overall

industry structure (Gregor and Johnston, 2000). These relationships determine

the extent of competition and rivalry within the industry and play a role in e-

commerce adoption. Therefore, the last hypothesis in this study reflects this

scenario:

H8: The adoption of e-commerce is positively related to pressure from

competitors.

27
 As describe in the literature review and reflected in the research model, below is

the summary of the hypotheses developed for this study. There are altogether

eight hypotheses.

Table 3.1: Summary of Hypotheses

No Hypothesis

Technological Context

H1 Perceived relative advantage is positively related to adoption of e-commerce.

H2 Perceived compatibility is positively related to adoption of e-commerce.

H3 The perceived complexity is negatively related to adoption of e-commerce.

Organisational Context

H4 E-commerce knowledge and expertise is positively related to adoption of e-commerce.

Management attitude towards e-commerce is positively related to adoption of e-

H5
commerce.

Environmental Context

H6 External change agent is positively related to adoption of e-commerce

H7 Pressure from trading partners is positively related to adoption of e-commerce.

H8 Pressure from competitors is positively related to adoption of e-commerce.

28
3.3 THEORETICAL FRAMEWORK

After reviewing all the related literature and generating the hypotheses for this

study, the following conceptual framework is illustrated in Figure 2.1 below. The

framework consists of 8 independent variables, all of which are postulated to

affect the same dependent variable, e-commerce adoption.

Perceived
Relative Advantage

Technological Perceived
Context Compatibility

Perceived Complexity

E-commerce
Knowledge & Expertise

Organizational Management Attitude E-Commerce

Context towards E-Commerce Adoption

External Change
Agents

Pressures from
Trading Partners
Environmental
Context Pressures from
Competitors

Figure 3.1: Theoretical framework

29
3.4 SELECTION OF MEASURES

Items used to measure the constructs to confirm content validity were adapted

from previous studies (Chan and Hui, 2001, Gengatharen and Standing, 2004,

Scupola, 2003, Scupola, 2004) that have incorporated the Tornatzky and

Fleischer’s (1990) model. These measures are the independent variables of the

conceptual model and are categorized in Table 3.2.

Table 3.2: Measures of e-commerce adoption factors

Factor Items Source

Useful in job Chan and Hiu

(2001)
Enhance effectiveness on the job Chan and Hiu
(2001)
Perceived Relative Enables to accomplish specific tasks Chan and Hiu
Advantage quickly (2001)
Increase productivity Chan and Hiu
(2001)
Improve organization performance Chan and Hiu
(2001)
With nature of our operations Gengatharen and
Standing (2004)
With organization culture Chan and Hiu
Perceived (2001)
Compatibility With organization values Chan and Hiu
(2001)
With preferred work practices Chan and Hiu
(2001)
Difficult to use Scupola (2003)

Learning to operate would be hard Scupola (2003)

Perceived
Complexity Interaction would be confusing Scupola (2003)

Take a long time to use successfully Scupola (2003)

30
 Table 3.2 continued: Measures of e-commerce adoption factors

Fully aware of e-commerce Gengatharen and

capabilities Standing (2004)
Has the right people to initiate Gengatharen and
E-commerce
Standing (2004)
Knowledge and
Know how e-commerce can help Gengatharen and
Expertise
Standing (2004)
Know what e-commerce activities to Gengatharen and
be adopted Standing (2004)
Regards e-commerce as a high Chan and Hiu
priority (2001)
Enthusiastic about the adoption Chan and Hiu
Management Attitude (2001)
towards E-commerce Always supports any e-commerce Gengatharen and
initiatives Standing (2004)
Willing to invest in e-commerce Gengatharen and
initiatives Standing (2004)
Recommended by advisors Scupola (2003)

Pressured by the government Scupola (2003)

External Change
Agents Demonstrated by external Scupola (2003)
consultants
Convinced by e-commerce solution Scupola (2003)
providers
Depend on firms already using e- Chan and Hiu
commerce (2001)
Pressure from Pressured by trading partners Chan and Hiu
Trading Partners (2001)
Will affect relationship with trading Chan and Hiu
partners (2001)
Competition is a factor in adopting Chan and Hiu
(2001)
Pressure from Pressured by industry players Chan and Hiu
Competitors (2001)
Will affect competitive advantage Chan and Hiu
(2001)

The dependent variable, extent of e-commerce adoption was computed by

adding the number of e-commerce applications adopted by the SMEs.

31
3.5 SAMPLING DESIGN

In order to collect the necessary response from a large number of SMEs, survey

questionnaires were used. To generalize the results, a sample size of 300 was

targeted. Convenient sampling is used to select the data for this research. A list

of SMEs was obtained from SME Corp. The initial plan was to randomly select

700 SMEs from the Klang Valley region from the list. The respondents would

then be contacted, either via email or post, to the take part in the survey.

3.6 DATA COLLECTION PROCEDURE

3.6.1 Questionnaire

A quantitative survey was used to collect the primary data, by using a structured

questionnaire. The questionnaire was designed based on past researches on e-

commerce (Lewis and Cockrill, 2002). The questionnaire was divided into four

sections, as show below:

• Section A: Demographics Profile

• Section B: Organisation Profile

• Section C: E-commerce Adoption

• Section D :E-commerce Adoption Factors

32
The first section (Section A) asked for the respondents’ demographics, such as

gender, age, education level and current position within the organization. This

section also helped to filter out respondents whose current position is not

executive level or above.

The second section (Section B) asked about the nature of the organization, such

as the total number of to employees and yearly revenue. The corresponding

answers are then used to verify that the respondents’ company is indeed an

SME, according to the SME definition.

The third section (Section C) determined the current e-commerce adoption level

within the organization. Finally, the fourth section (Section D) contained

questions that represent the e-commerce adoption determinants, as described

in the research framework.

The multiple-item methods were used in the final two sections (Section C and

D). Each item was measured based on a seven-point Likert scale from strongly

disagree (1) to strongly agree (7). The odd number rating scale was adopted as

it allows the respondent to ‘sit on the fence’ by selecting a neutral statement.

The final version of the questionnaire can be viewed in Appendix 1.

33
3.6.2 Pilot Test

Before the commission of the survey, a pilot test of the structured questionnaire

was conducted, involving several respondents. The questionnaire was assessed

in terms of ease of understanding, logical consistencies, sequence of the items,

contextual relevance and so to enable the questionnaire to be further refined.

26 respondents from various SMEs participated in the pilot test. Personal visits

were made to managers of selected SMEs, and they were requested to

participate in the survey. Cooper and Schindler (2003) stated that the pilot test

group size may range from 25 to 100 respondents and do not have to be

statistically selected.

The pilot test revealed that the average time duration to complete the

questionnaire is only 5.5 minutes, which is appropriate, considering the number

of questions in the questionnaire. All the pilot test respondents also claimed that

the questions were clear and comprehensible, indicating that the survey had a

relatively low level of difficulty.

34
3.6.3 Data Collection

Based on the sampling list, the survey was made available to the shortlisted

respondents. The questionnaire was made into an online form and its URL was

sent to each of the 500 respondents that have email addresses. Another 200

printed version of the survey was posted to the rest.

However, it was discovered that unlike medium-sized enterprises, many of the

small-sized enterprises did not take part in survey, thus resulting in an overall

low response rate.

Therefore, to reach out to small-sized enterprises, physical questionnaires were

then distributed in person to a number of SME’s in the Klang Valley. Field

assistants were used to carry out the survey and collect the necessary data. The

locations visited include Mara Headquarter Building in Kuala Lumpur, Kompleks

PKNS Shah Alam, Kompleks PKNS Bangi, and Diamond Square Gombak. For

each company approached, the personnel was interviewed and assisted in filling

up the questionnaire. This approach was adopted since many pilot test

respondents indicated that they prefer someone to help write down their

responses.

35
3.7 DATA ANALYSIS TECHNIQUES

3.7.1 Descriptive Analysis

Descriptive statistics were applied to summarize the demographic profile of the

respondents and the mean of each of the factors. Variance and standard

deviation were obtained to measure variability around the mean of a distribution.

3.7.2 Reliability Analysis

The reliability of the instruments used was one of the concerns in this study. The

reliability scale text would be utilized to determine the instruments validity.

According to Sage et al. (1980), measurements are reliable if the true aspects of

the trait measured, instead of the chance aspects. Therefore, the instrument

must be reliable to the extent that the scores by the respondents should be

almost equal if measurement is repeated.

In order to examine the reliability of the instruments, the Cronbach’s Alpha was

used to measure reliability of the underlying dimensions. The Cronbach’s Alpha

estimate indicates how highly the items in the questionnaire are interrelated in

order to determine the instrument reliability. According to Nunnaly (1978), the

Alpha which is more than 0.7 indicates a high reliability.

36
3.7.3 Correlation Analysis

Correlation analysis was used to examine whether there was any association

between each of the factors and the extent of e-commerce adoption. The extent

of e-commerce adoption was computed by adding the number of e-commerce

applications being adopted by the respondents, which means that the higher the

score, the higher the adoption.

In determining whether the factors have any influence on the extent of e-

commerce adoption, that is to test the hypotheses formulated earlier, regression

analysis was used. For the purpose of regression analysis, the extent of e-

commerce adoption is the dependent variable while the factors are the

independent variables. It was performed using the SPSS software.

The hypotheses and their relationships were tested through correlation by

individually measuring e-commerce adoption in relation to each of the factors.

Correlation analysis indicates if a linear relationship exists between two

variables. Although correlation does not indicate the reasons the relationship is

present, the correlation coefficient does tell if the relationship is significant or

not. A perfect positive correlation has a coefficient of 1.0, no correlation has 0

and a perfect negative coefficient has 1.0.

37
Multiple Regression

Multiple regression was employed to assess the significant level of each of the

eight determinants. Regression is an analysis of a predictor variable’s

relationship to a dependent variable. In this study, multiple regression via

stepwise analysis was employed. In stepwise regression, the order of entry for

variables is based strictly on statistical criteria with an objective of determining

which independent variables best predict the dependent variable (Tabachnick

and Fidell, 1989). The SPSS program selects the variables in the order of their

ability to contribute to the overall prediction. Hair et al. (2006) claimed that

stepwise estimation is a powerful and popular approach to variable selection

since it assesses the contribution of each independent variable to the regression

model, based on the greatest contribution.

3.8 CONCLUSION

This chapter has presented the fundamental structure of the research

methodology used in this study. The proposed research framework was adopted

from previous researcher’s research model. Eight hypotheses were developed

and will be answered in the next chapter.

38