Methodology: E-commerce adoption in SMEs

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.



The chapter begins with the research design and continues with the population and sampling procedures. Data collection method and the criteria of the method are explained. Then, the variables and measurements are illustrated. The different test can be used for data analyses is explained.

3.1Research Design

The purpose of this research study is analysing the determinants that influence the level of e-commerce adoption in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs. Following the past study, it was highlighted that five of the perceived attributes are relative advantage, compatibility, complexity, trialability and, observability. Researcher chooses the relative advantage, complexity and, observability because those determinants have the significant result on the past studies. However, past studies only focus on the organization internal factors and fail to analyse the external factors as well, such as consumer behaviour and government policy. Hence, these two external factors are included in this research study. This survey is considered as the cross-sectional as it examines the relationship between those five determinants and adoption e-commerce in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs. Researcher chooses the primary data collection method which is survey questionnaire for the study because it can gather the data directly from the population and has high reliability (Sindhu, 2012). The unit of analysis for this study is SMEs industry in Selangor. The method of data collection is self-administered questionnaires through mail and hand over face-to-face. SMEs e-mail address are tracked from the official website of SMEs.

3.2Population and Sampling Procedures

3.2.1 Target Population

Our research focuses on the SMEs in Selangor and Wilayah Persekutuan Kuala Lumpur only because they are big cities in Malaysia. The total number of SMEs in Malaysia in 2010 was 645,136. There were 125,904 small medium enterprises in Selangor and 84,261 in Wilayah Persekutuan Kuala Lumpur area which is more than 32% of the total. So, the target population is those enterprises managers or owners in Selangor and Wilayah Persekutuan Kuala Lumpur (Table A3: Principal statistics of SMEs by state, 2010).

3.2.2 Sampling Elements

Sampling Elements means that a single member or unit of target population about which information will be obtained and it is often an individual but can also be an organization, group or household (Centre for Micro Finance, 2012). In this research, the sampling elements are those respondents who are in the top management of the small medium enterprise in the Selangor and Wilayah Persekutuan Kuala Lumpur such as the manager who are also the owner of the company. The person who have the power to manage and control the company will clearly know that whether the company adopting e-commerce during running operation.

3.2.3 Sampling Location

In this research, primary data is being used in this study. Therefore, the questionnaires will be distributed at the small medium enterprises which are located in Selangor and Wilayah Persekutuan Kuala Lumpur.

3.2.4 Sampling Techniques

Sampling techniques have separated into two methods which are the probability sampling and non-probability sampling. Probability sampling is relying only on random or chance selection so it also called random sampling (Centre for Micro Finance, 2012). Probability sampling can increase sample’s representativeness of the population and decrease the sampling error and sampling bias. Probability sampling which are included simple random sampling, systematic sampling, stratified random sampling, cluster sampling and multi-stage sampling. Non-probability sampling is a sampling technique where the samples are gathered in a process that does not give all the individuals in the population equal chances of being selected (Joseph, 2009). For non-probability sampling is included quota sampling, purposive or judgmental sampling, snowball sampling, self-selection sampling and convenience sampling.

In this research, researcher is using the probability sampling which applied the simple random sampling. Probability sampling has been used for this research due to this sampling technique suitable for the selected sample from the targeted population. Advantages of this method are that it eases the assembling of sample from given population since every person is given equal opportunities of being selected and less expensive. Otherwise, there are also limitation of this method which is it is hard to complete all list of company usually choose the sample size from the target population (Joseph, 2009). Thus, the questionnaire will be distributed to the each company which is located at the Selangor and Wilayah Persekutuan Kuala Lumpur SMEs. Since population of SMEs Selangor is 125,904 and Wilayah Persekutuan Kuala Lumpur is 84,261, the researcher would like to choose the sample based on the sample size.

3.2.5 Sample Size

According to Green’s (1991), he estimated the sample size formula which is the n = 50 + 8(m). “n” is the sample size and the “m” is the number of independent variables. In this research study there are five independent variables which are relative advantages, complexity, observability, consumer behaviour and government policy. So, n=50 + 8 (5) = 90, which means that our sampling size are at least 90 target respondents. In order to get the confidence level of 95% from the population (210,166), an accurate survey must have complete the 383 respondents (Surveygizmo, 2012).

3.3Data Collection Method

Data collection methods have two types which is primary data and secondary data. The method chosen depends on the researcher’s need.

3.3.1 Primary Data

This research was conducted using primary data collection method. The primary data collection was done by distributing the survey questionnaires. Furthermore, Siniscalco and Auriat (2005) stated that, the information collected by means of a questionnaire are facts, activities, level of knowledge, opinions, expectations and aspirations, membership of various groups, and attitudes and perception. Thus, researchers choose questionnaire to identify the determinants that influence the level of e-commerce adoption of SMEs in Malaysia. The survey questionnaire would be used by web-based and face-to-face delivery to company in order to achieve the response rate.

In the past studies, researchers found that the limitation of the data collection method is the questionnaire only written in English, some of the response group prefer written in Chinese. To eliminate this problem, we distributed questionnaires consists of two language which are written in English and Chinese according to the respondents convenient.

3.4Research Instrument

Pilot test has been conducted before the survey questionnaires are being distributed to targeted respondents who are the managers or owners of SMEs in Selangor and Wilayah Persekutuan Kuala Lumpur. A total of 20 survey questionnaires have been distributed to the lecturers and tutors of Universiti Tunku Abdul Rahman as sample for Pilot test. The questionnaires are given face-to-face with the targeted people in order to ease the data gathering and time saving. The data gathered are analysed using SAS software to get confirmation on the survey questionnaire questions are according to the core of the study. The results of the Pilot test are illustrated in the table below.

Table 3.1: Summary of Reliability Statistics for Pilot Test


Cronbach’s Alpha

Number of Items

Relative Advantages (IV1)



Complexity (IV2)



Observability (IV3)



Government Policy (IV4)



Trend of Consumer Behavior (IV5)



Adoption of E-commerce (DV)



Table 3.2: Summary of Normality Statistics for Pilot Test




Relative Advantages









Government Policy



Trend of Consumer Behaviour



Adoption of E-commerce



Source: Developed for the research

3.5Variables and Measurement

3.5.1 Scale of Measurement

The survey questionnaires are designed based on the variables. According to Sekaran (2003), the measurements of the variables are important in order to test the hypotheses and obtain answers to complex issues. The scaling techniques be used in the research was nominal scale, ordinal scale, ratio scale and interval scale

Table 3.3: Definition of Dependent and Independent Variable

Dependent Variable


Adoption of E-commerce amongst Selangor and Wilayah Persekutuan Kuala Lumpur SMEs

Adoption as a decision to make full use of an innovation as the best course of action whereas rejection is a decision not to adopt available adoption.

Independent Variables


Relative Advantage

The degree to which an innovation is perceived as being a better idea than the idea it supersedes


The degree to which an innovation is perceived as relatively difficult to understand and use.


The degree to which the result of an innovation are visible to others.

Government Policy

Government guided in its management of public affair, or the legislature in its measure

Trend of Consumer behaviour

The attitudes behaviour of consumer or customer towards a new invention or technology.

Table 3.4: Measurement Used For Each Variable



Scale of Measurement


SMEs sector



Size of SMEs



Company Internet Connection



Computer skill of employees



How long adopt E-commerce in SMEs



Dependent Variable

Adoption of e-commerce in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs


5-point Likert Scale

Independent Variables

Relative Advantage


5-point Likert Scale



5-point Likert Scale



5-point Likert Scale

Government Policy


5-point Likert Scale

Trend of Consumer Behaviour


5-point Likert Scale

3.6Data Analysis

Data analysis is defined as the process of evaluating data using analytical and logical reasoning to examine each component of the data provided as how they are being collected, processed, analysed and interpreted accordingly to suit the aim of the study. This form of analysis is just one of the many steps that must be completed when conducting a research experiment. Data from various sources are gathered, reviewed and then analysed to form findings or conclusions. In this research, the data was collected from the survey conducted through questionnaire.

3.6.1 Descriptive Analysis

Descriptive analysis provides brief summaries about the sample and the observations that have been made. It is used to explain and describe the information of sample collected and summarizes a given data set, which can either be a representation of the entire population or a sample. The measures used to describe the data set are measures of central tendency and measures of variability or dispersion. Measures of central tendency include the mean, median and mode, while measures of variability include the standard deviation or variance. In this research, descriptive analysis is used in part A of demographic profile and part B of general information. Frequency Distribution

Frequency distribution is a representative which is either used in a graphical or tabular format. It is used to obtain a count of the number of responses or observation within a given interval associated with different values of one variable and to express them into percentage terms. The interval must be mutually exclusive and exhaustive. Frequency distribution is used to analyze respondents demographical profile in part A such as gender, age, living states, occupation, income level and education level as well as general information in part B. The mean and average are measures of central tendency which are used to analyze data collected in the part C of the questionnaire.

3.6.2 Scale Measurement

Scale measurement refers to how variables are measured. It is used primarily to verify the quality of the data collected and this can be determined by the reliability level of the data. There are four types of scale measurement which is nominal, ordinal, interval and ratio. Normality Test

Normality test is used to determine whether the variables or input data are normally distributed. Kolmogorov-Smirnov test is used in this research to analyze the data as it is more suitable for larger sample size, which is an n > 50 sample (Fasano and Franceschini, 1987). In accordance to Saunders et al. (2009) data are considered normal if the p-value is more than 0.05. Besides, histogram and normal probability plot (P-P plot) are used to justify whether the data is normally distributed. If the data inspected are normal distribution, the histogram should be represented by a bell-shaped curve while P-P plot should produce a straight line. Generally, significant value from the Kolmogorov-Smirnov tests is as follow:

Sig. value ≤ 0.05: data is not normally distributed

Sig. value > 0.05: data is normally distributed Reliability Test

Reliability test is used to describe the overall consistency and stability of a measure with which the research instrument measures the constructs (Malhorta and Peterson, 2006). A measure is said to have a highreliabilityif it produces similar results under consistent conditions. For this research, reliability test is carried out to verify whether the items in the questionnaire are related to each other. Cronbach’s Alpha reliability test is most commonly used by averaging the coefficient varies from 0 to 1. The following table shows the level of reliability:

Table 3.5: Rules of Thumb about Cronbach’s Alpha Coefficient Size

Alpha Coefficient Range

Strength of Association

< 0.6


0.6 to < 0.7


0.7 to < 0.8


0.8 to < 0.9

Very Good



Adapted from: Hair, J. F., Money, A. H., Samouel, P. & Page, M. (2007). Research Methods for Business (2nd Ed.). Chichester, West Sussex, UK: John Wiley & Sons Ltd

Based on the rule of thumb about Cronbach – Alpha, coefficient alpha is range from 0 to 1. A score less than 0.6 are considered as poor and a score of over 0.8 is considered good. Researchers normally consider an alpha of 0.7 as a minimum; although lower coefficients might be acceptable depending on the research objectives (Sekaran & Bougie, 2010). An acceptable of reliability shows respondents are answering the questions of survey in a consistent manner (Hair et al., 2010). Multicollinearity Analysis

Multicollinearity analysis is the occurrence of several independent variables in a multiple regression model which are closely correlated to one another. Multicollinearity can cause strange results when attempting to study how well individual independent variables contribute to an understanding of the dependent variable. In general, multicollinearity can cause wide confidence intervals and strange P values for independent variables. It has been carried out to test whether there is any correlation between the four independent variables in this research. Hair et al. (2010) stated that a multicollinearity problem exist when independent variables are strongly correlated to each other’s. The correlation value that more than 0.9 is considered high relation.

3.6.3 Inferential Analysis

Inferential analysis is an analysis of a set of data used to investigate a specific hypothesis or assumption. We use inferential statistics to try to infer from the sample data what the population might think or to make judgments of the probability that an observed difference between groups is a dependable one or one that might have happened by chance in this study. Thus, we use inferential statistics to make inferences from our data to more general conditions. Correlation indicated was used for the inferential analysis of this research to investigate the relationship between complexity, government policy, observation, consumer behavior and relative advantages and adoption of e-commerce in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs. Pearson’s Correlation Analysis

Pearson’s correlation analysis is used to evaluate the direction and strength of relationship between the two variables. In this research, this analysis is chosen to measure the co-variation between the five independent variables and adoption of e-commerce in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs.

The coefficient (r) indicates both the direction of the relationship and magnitude of the linear relationship. The correlation coefficient ranges from +1.0 is considered perfect positive relationships to -1.0 which indicates perfect negative relationships and signifies that the independent variable has the direct relationship with the dependent variable and vice versa. While value of 0 shows no linear relationship between the two variables. Correlation coefficient value range from 0.10 to 0.29 is deemed to be weak, from 0.30 to 0.49 is regarded as medium and from 0.50 to 1.0 is believed to be strong (Cohen, 1988). However, to avoid multicollinearity problem among independent variables, this value should not go further than 0.9 (Hair et al., 2007). Multiple Linear Regressions

Multiple linear regressions (MLR) are a technique thatuses several explanatory variables to predict the outcome of a response variable. It attempts to study the relationship between two or more independent variables and a dependent variable by fitting a linear equation to observed data (Malhorta & Peterson, 2006). In this study, multiple regression equation is used to answer certain basic equation between dependent variable adoption of e-commerce in Selangor and Wilayah Persekutuan Kuala Lumpur SMEs and independent variables including complexity, government policy, observation, consumer behavior and relative advantages on whether the relationship exists, how strong is the relationship; and whether the relationship is positively or negatively skewed. To observe the relationship between the variables, it will be estimated by the following equation,

Yi= α + β1X1 + β2X2 + β3X3 + β4X4 + β5X5

Whereby, Yi = Adoption of E-commerce in Selangor SMEs

α = Constant

X1 = Complexity

X2 = Government policy

X3 = Observation

X4 = Consumer Behavior

X5 = Relative Advantages


The chapter shows the flow of methodology from the beginning stage which is research design, population and sampling, data collection method, variables and measurement and finally method of data analyses. SAS software is used to process the data gained. The next chapter will present the result from the data gained according to descriptive and inferential analysis.