Overall, the objective of the research analysis is to try to answer the research questions formulated in Chapter 1. Accordingly, a conceptual framework and propositions have been developed from these questions, and justified in the literature review section. Therefore, the data analysis will be pursued to gather evidence pertaining to the several concepts identified. Quantitative data from the questiormaire survey will be analyzed by utilizing the Statistical Package for Social Science version 22 (SPSS V.22) and SmartPLS v.3. These software packages have been widely used by researchers in various disciplines. Below are the steps that will be followed in the data analysis process.
3.12.1: Analysis of Survey response …show more content…
3.12.3: Analysis of Response bias
Secondly response biases will be determined. Response bias is the effect of non-responses on survey estimates (Fowler, 2002). Non-response bias is a major source of bias in survey research. If it is not addressed properly, it can lead to conclusions that differ systematically from the actual situation in the population. To determine this, the researcher will conduct a t-test to compare the early response group and the late response group for their responses on dependent and demographic variables. This will constitutes a respondent-nonrespondent check for response bias. (Saunders, 2009)
3.12.4: Normality …show more content…
Confirmatory factor analysis will be done using Partial Least Squares (PLS-SEM) with the aid of SmartPLS v.3.software. PLS-SEM path modeling using SMARTPLS is appropriate to carry on the confirmatory factor analysis which is more reliable and valid (Wan-Afthanorhan, 2013). PLS will be chosen mainly because it allows latent constructs to be modeled as either formative or reflective indicators. Reflective indicators reflect an unmeasured latent construct that is deemed to exist before it is measured and account for the observed variances and covariance. Formative indicators are used to form a superordinate construct where the individual indicators are weighted according to their relative importance in forming the construct (Chin, 1998). Moreover, the normality of data distribution not assumed, thus data with non-normal distributions can be conducted in structural equation modeling since its application is performed in a non parametric way. PLS is also recommended when either cross-sectional, survey, or quasi-experimental research designs are used; when a large number of manifest and latent variables are modeled or when too many or too few cases are available (Falk, 1992).These conditions apply to this study because it will adopt a survey design; the sample size in this study is relatively small (126) and the Likert- scale used in this study normally do not