Analyse the Respondents Based from the Questionnaire or Survey Form

3.5 Method of Analysis

In this research, to analyse the respondents based from the questionnaire or survey form, there are few data analysis methods that we can choose to use according to the studies of the purposes are Preliminary Analysis, Descriptive Analysis, Scale Measurement and Inferential Analysis. Under preliminary analysis, there are includes Non-response Bias Analysis, Common Method Variances Analysis, Multicollinearity Analysis and Normality Test Analysis (Malhotra and Birks, 2003). These are for the researchers to develop the quality of responses and they can find out the errors by inspecting from the questionnaire or survey data files (Alvin C. Burns & Ronald F. Bush, 2012). For the descriptive analysis, it is the elementary transformation of data in a method to describe a basic characteristic which are central tendency, variability and distribution (Zikmund et al., 2013). While the reliability and validity test was conducted to ensure the measure’s internal consistency and construct validity which is for the scales measurement (Zikmund et al., 2013). While according to inferential analysis was used to investigate the research question, the theoretical framework, and the hypothesis (Chiew et al., 2014).

3.5.1 Preliminary Analysis 

Non Reponse Bias Analysis

Common Method Variance Analysis

Multicollinearity Analysis

Normality Test Analysis

3.5.2 Descriptive Analysis

This analysis is the elementary transformation of data in a method to describe a basic characteristic which are central tendency, variability and distribution. The most common used of descriptive statistic is central tendency. It is used to measures the location with mean, mode, and median (Malhotra et al.2008). For the measures of variability which indicate the distribution dispersion (Malhotra, 2012). It usually used to calculate on interval or ratio data which includes the range, variance and standard deviation (Malhotra, 2012). Under distribution where includes histogram, bar chart and pie chart.

According to this research will be use distribution to analyse for the purpose of collection the reliable data. For histograms are one of the useful distribution for analyse any type of data. But with continuous variable such as interval or ration, it is useful for providing fast assessment on the distribution of data (Zikmund et al., 2013). For bar chart, it may be the most flexible among any types of graph, this is because any data shown on any graph also can show on the bar chart. Additionally, if anything which can’t be shown, the bar chart can be readily describe (McDaniel and Gates, 2013). Besides, pie chart is more appropriate for displaying the result of market research in wide range of circumstances (McDaniel and Gates, 2013).

3.5.3 Reliability and Validity Tests

Both Reliability and Validity tests are important concepts in research which are used for improving the accuracy of the assessment and evaluation in the research process (Tavakol and Dennick, 2011). Reliability refers to the results which are consistency, stability as well as repeatability. If the results have been received in the same situations but different circumstances, the results can be considered by the researchers are reliable (Twycross and Shields, 2005). Validity refers to the extent to which the conclusions drawn from the research design such the true experiment (Hair et al., 2013).

3.5.4 Inferential Analysis

From this research, hypothesis will be examined by Multiple Regression Analysis. It is a simple regression analysis allowing a metric dependent variable to be estimated by multiple independent variables (Zikmund et al., 2013). It is used to estimate the value of a variables based on the value of two or more than other variables. Based on the research, there can be a linear simple regression analysis (Zikmund et al., 2013) with one Dependent Variable (DV) - Consumer Purchasing Decision on Organic Foods is described by one Independent Variable (IV) - Media Channel. Alternately, use this regression analysis to understand whether the consumers purchasing decision on organic foods can predicted based on gender, age group, and current educational level. By using SPSS statistics, interpret and report the results from the test, the guideline of “quick start” will show how to carry out multiple regression analysis. However, have to understand the different assumptions that the research data must meet before introduce to this procedures in order to give a valid result in multiple regression (Hair et al., 2013).

Mediated Regression Analysis (Mediator)

Based from Baron and Kenny (1986) proposed the mediated regression analysis will mediating effect will be taking place which is the model was tested and reflect the four steps of procedures are fulfilled. According Baron & Kenny's first step of procedures involve the foundation of an effect which may be mediated. Second step is involves handling the mediator variable as an outcome variable. Third step is the initial variable must be controlled while forming the correlation between the two other variables. And for the final step is the establishment can only be achieved if the effect of the initial variable over the outcome variable while controlling for mediator variable is zero. If these four steps of procedures are met, then the research data is consistent with the mediation hypothesis.  However, if only the first three steps are satisfied, then the partial mediation is observed in the research data.  If one or more of these relationships are non-significant, the researchers usually will conclude that mediation is not possible (MacKinnon, Fairchild, & Fritz, 2007).