CHAPTER 3 3.1 Introduction
3.5 Data analysis
Statistical Package for the Social Sciences, also known as SPSS is a software platform that provides highly developed statistical analysis, a large library of machine learning algorithms, text analysis, open data extensibility, big data integration, and seamless application deployment. SPSS is accessible to users of all skill levels due to its ease of use, flexibility, and scalability. Furthermore, it is adequate for projects of all sizes and levels of complexity, and it can assist the study in identifying new opportunities, improving efficiency, and reducing risk (George & Mallery, 2021).
SPSS requires the implementation of a set of variables, followed by the creation of cases based on appropriate data input into these variables. The SPSS datasheet primarily
contains four types of variables which includes independent variables, dependent variables, intervening variables, and moderator variables (Bala, 2016). Independent variable is a cause while the dependent variable effect whose value is affected by changes in the independent variable. An intervening variable, also known as a mediating variable, is a hypothetical variable that is typically used to explain causal links between other variables in research.
Follow up with the moderating variable, it can change the relationship between independent and dependent variables (Rahman & Muktadir, 2021).
According to Kulas et al. (2021), through SPSS, researchers are able to analyze the study with the six steps of hypothesis testing process and SPSS output which includes the descriptive statistics, hypothesis testing, Z- and T- Tests, Anovas, correlation and regression (Inferential Analyses). The latest version of it is IBM SPSS 28 which is the most stable SPSS version as new features meta-analysis, power analysis, ratio statistics, relationship maps, statistics workbook, search, table side-pane editor, and high contrast support are provided (IBM, 2021).
3.5.1 Descriptive analysis
Descriptive analysis, as known as descriptive statistics, which is to show the properties of a group of observations in a descriptive way (Marshall & Jonker, 2010), also assists researchers in making sense of enormous amounts of data. Descriptive statistics is the scientific approach for gathering, organising, analysing, and interpreting data for the purposes of description and decision-making (Kaushik & Mathur, 2014). This analysis will look at the relationship between each of the age groups in this study, which are assumed to have a relationship with perceived usefulness and perceived ease of use. Descriptive analysis is dealing with the presentation of numerical facts, or data, in the form of tables or graphs, as
well as the methodology for data analysis (Kaushik & Mathur, 2014). Therefore, descriptive analysis will be used in this study.
3.5.2 Regression analysis
Researchers can use regression analysis to determine the effect of one independent variable on a dependent variable while controlling for any number of other independent variables or controls (Kremelberg, 2011). There are several types of regression. Linear regression, the most fundamental types of regression in machine learning. It consists of a predictor variable and a dependent variable that are related to each other in a linear fashion;
Logistic regression, it will be used when there is binary dependent variable (Menard, 2002);
Ordered logistic regression, it sub from logistic regression. The specific of it is dependent variable more than two levels; Multinomial logistic regression is the simple extent from binary logistic regression that allows for the dependent or outcome variable to have more than two categories (Starkweather & Moske, 2011); Negative binomial regression or Poisson regression, which used in explaining over-dispersion in multivariable count data (Kremelberg, 2011).
Linear regression will be practised in this study as the best strategy for estimating the causal effects of buying decision on binary outcomes. It was shown when linear regression is safer compared to the others as correlation values are directly interpretable in terms of probabilities, and when correlations or fixed effects are included (Gomila, 2021). Under linear regression, there are two terms likely to meet in quantitative method studies. A regression model with a single independent variable is referred to as simple linear regression, whereas a regression model with two or more independent variables is referred to as multiple linear regression (Kremelberg, 2011).
In this study, multiple linear regression, also known as classical linear regression model is being chosen to determine numerical targets and one or more numerical predictors
(Stehlik-Barry & Babinec, 2017). The variable to be predicted (dependent variable) has a linear relationship with the independent variable in the multiple linear regression model (Perdana et al., 2021). By using multiple linear regression, this study ought to know the influence of each variable, including how customer trust and the system quality of E- commerce platforms affect perceived usefulness and perceived ease of use, and affect their attitude toward using E-commerce platforms, behavioural intention to use, and, ultimately, their buying decision. Instead, perceived ease of use will affect the perceived usefulness as well.
3.5.3 ANOVA
ANOVA is similar to the t-tests which is used to compare two groups of means scores of independent variables. The difference is that ANOVA can compare more than two sets of scores which t-tests cannot. ANOVA is focused on the differences between means rather than differences between variances. To determine whether the means differ, the technique of variance is applied (Woodrow, 2014). In this study, ANOVA examines the means within an age group or between age groups to see if the means of the groups differ significantly. Under ANOVA types, one-way ANOVA is frequently used (Woodrow, 2014), testing several independent variables (IV) with three or above three groups and a continuous dependent variable (DV) (Allen et al., 2009). The reason for using this test is to determine whether there are significant differences in the buying decisions among the group of the independent variables. Thus, it examines the null hypothesis that samples from various groups were drawn from the same population. Accompanying the computing of an F-statistic, comparison group means will be done as the variability between groups is compared to the variability within the groups (Verma, 2012).