3.3 Data Collection and Measurement
Qualitative data and quantitative data differs from each other. In our research, we planned to collect quantitative data aligned with the research we are going to carry out. Quantitative data represent data taken in the form of numbers and can be measured (Bennett and Jessani, 2011). There are a few reasons which support us from choosing quantitative data. First of all, since quantitative data is in numerical form, therefore, it can be analysed using precise statistical method (Bennett and Jessani, 2011) and from here, a summary of data can be generalized across groups of people (Babbie, 2010; Mujis, 2010). Besides, as compared to qualitative data, quantitative data is more efficient in terms of time saving. Since we have time constrain, quantitative data is more appropriate for our research. Furthermore, quantitative data is both definite and objective (unbiased) as it is numerical in nature (Center for Innovation in Research and Teaching [CIRT], n.d.) and it seek rigorous measurement instead of individual interpretations. Last but not least, quantitative data is more highly structured and we know clearly on what we are looking for in advance (CIRT, n.d.). This is due to quantitative data are usually gathered through highly structured methods such as questionnaires, survey, etc (CIRT, n.d.) and normally includes closed-ended questions and responses that provide measurable answer (CIRT, n.d.; Creswell, 2014).
In order to collect primary data, a quantitative collection method of survey questionnaire is chosen. Survey questionnaires are popular data collection instrument for acquiring information from respondents which consists of a list of questions and other prompts (Bird, 2009). In this research, survey questionnaire will be carried out by posting on a website and respondents can assess it through internet. This type of questionnaire is also known as Internet survey (Zikmund, Babin, Carr & Griffin, 2013). The use of Internet survey brought us a few benefits. First of all, Internet survey allow us to reach a large number of respondents quickly and cost-effectively (Zikmund et al., 2013). This is due to the use of Internet has been developed aggressively over the past years, therefore, this ease the presentation of questionnaire and with this computerized self-administered questionnaire, it will eliminate the cost of paper, postage, and data entry as well as other administrative costs. Besides, since the whole process of answering questionnaire will be self-administered, thus, the questionnaire will be constructed precisely to be completed by a respondent without interference of the researcher. With this, the risk of interviewer influence the answer can be avoided and the final data we obtained will be objective and unbiased (Zikmund et al., 2013). Besides avoiding risk, as it is a self-administered questionnaire, we can save time when collecting data because we do not need to be present and supervise when the survey is conducted (Zikmund et al., 2013).
In this research, TARUC Penang Branch students are asked to volunteer to complete a questionnaire. The questionnaire is divided into two main parts. The first part of questionnaire will be used to collect respondent’s basic information such as gender, age, education level and experience using online banking. The respondent’s basic information will be measured based on a nominal scale. While for the second part, the questionnaire will be established based on the variables. There are a total of five independent variables which are perceived usefulness, perceived ease of use, security and privacy, social influence as well as cost saving. Each of the variables will be measured by five point Likert scales ranging from “strongly disagree” (1) to “strongly agree” (5). There will be four questions for each variables and it is compulsory to answer before the respondents can proceed to the next variables.
No.
Measurement Item
Scale of Measurement
Source
1.
Perceived usefulness.
• I think that paying tuition fees online can save my time and enable me to accomplish my tasks more quickly.
• I think that paying tuition fees online is more convenient than traditional ways of paying.
• I think paying tuition fee online increase my productivity.
• I think that paying tuition fees online is useful.
1-5
Lee (2008) (1) ; Salciuviene, Auruskevicience and Ivanauskiene (2014) (3); Amin (2009) (5); Chong, Ooi, Lin and Tan (2010) (8)
2.
Perceived ease of use
• I find that it is easy to pay tuition fees online.
• I think it is easy to learn how to pay tuition fees online.
• I think paying tuition fee online makes it easier for me to conduct payment transactions.
• I find paying tuition fees online easy to use.
1 – 5
(3), (8), Lim, Wu and Tran (2014) (last)
3.
Security and privacy
• I believe that the transaction of paying tuition fees through online is secure and private.
• I believe the paying tuition fees through internet banking channel will be processed securely.
• I believe online banking has sufficient technical capacity to ensure that no others will get my personal information on the Internet.
• I feel paying tuition fees online is really secure to use.
1 – 5
(8),(5), (3)
4.
Social influence
• My family influence me to pay tuition fees through online banking.
• My friends influence me to pay tuition fees through online banking.
• Most people around me thinks that paying tuition fees online is a good idea.
• If I pay tuition fees online, most of the people will regard me as clever.
1 – 5
(1), (last), Amin (5)
5.
Cost saving
• I think paying tuition fees online will help me save time.
• I think that the service charge is acceptable when I use to online banking to pay tuition fees.
• I think the fee I pay to use the internet connection is affordable when I pay tuition fees through online banking.
• I think that paying tuition fees online allow me to enjoy cost saving.
1 – 5
Data Analysis Methods
In analysing the data collected, Pearson Correlation Test which is a method of descriptive analysis and multiple regression will be used.
Descriptive analysis
Descriptive analysis is the most elementary statistical analysis which convert raw data into a form that illustrate the basic characteristics such as central tendency, distribution and variability (Zikmund et al., 2013). In descriptive analysis, mean, medians, mode, variance, range and standard deviation are broadly employed. In this research, a large number of respondents will be measured and therefore with the use of descriptive analysis, it will helps to streamline the large amounts of data in a sensible way by providing a summary which describes the basic properties of a variable and can be arranged in tables, graphs or charts.
Pearson Correlation Test
The Pearson Correlation Test, also known as Pearson product-moment correlation coefficient or simple as the correlation coefficient. It is used to measure the strength of the linear relationship between two variables (Zikmund et al., 2013). The symbol for Pearson’s correlation will be differ when measure in population and sample. For population, the symbol is “ρ” while for sample, the symbol is “r”. In this research, since we are measuring sample, therefore the symbol of “r” will be used.
The formula for Pearson correlation is
Pearson correlation (r) can be ranged from -1 to +1 in which -1 represents a perfect negative or inverse linear relationship, an r of +1 indicates a perfect positive linear relationship while for an r of 0 represents there is no correlation between the two variables (Zikmund et al., 2013). The reason of using Pearson correlation test is that it is simple to calculate and easy to interpret. With the use of this test, we can easily identify the relationship between the two variables after calculating the correlation coefficient (McDonald, 2015).
Multiple Regression
In accordance to Zikmund et al. (2013), regression analysis is one of the tools that can be used to measure the linear association between a dependent variable and independent variable. Multiple regression analysis is a development of simple linear regression analysis, which measure two or more independent variables on a single dependent variables (Zikmund et al., 2013). The formula for multiple regression analysis is Y= b0+b1X1+b2X2+b3X3+…+bnXn+ei in which Y represent dependent variable while X represent independent variable. The reason that multiple regression used is because there are more than two measurement variables. In our research, there are six measurement variables in total including one dependent variable and five independent variables, thus this method is used. Besides, with the use of multiple regression, it helps us to understand the relationships between dependent and independent variables and help us to identify the causes which may affect the variation in the dependent variable.