Conclusion
This analysis was essential in order to resume the results of
the questionnaire and to prepare the hypotheses testing. This would allow the
reader to get an overview of the respondents' environmental knowledge, green
purchase behaviour and their profiles; before entering in the details with the
hypotheses.
In the second part, all the different hypotheses were test each
by one in order to see if it could be validated or not, with the actual
sample.
3.3 Hypotheses Testing
3.3.1 Data cleaning and normality testing
The data were already gathered in an excel file and pasted on
SPSS.
Firstly, the researcher has cleaned all errors and mistakes in
the questionnaire. Some responses were out of range, logically inconsistent or
had extreme values. This kind of data is not admissible in the analysis.
Moreover, some responses were missing, ambiguous or not properly recorded.
After cleaning the data, a normality test was conducted in order
to see if the variables are well distributed or not.
Table 3.4 Normality Testing
Tests de normalité*
|
Kolmogorov-Smirnova
|
Shapiro-Wilk
|
|
Statistique
|
ddl
|
Signification
|
Statistique
|
ddl
|
Signification
|
|
knowledge_
|
,235
|
150
|
,000
|
,861
|
150
|
,000
|
|
intention
|
,120
|
150
|
,000
|
,969
|
150
|
,002
|
|
living
|
,208
|
150
|
,000
|
,845
|
150
|
,000
|
|
socio
|
,126
|
150
|
,000
|
,967
|
150
|
,001
|
|
green_consump
|
,177
|
150
|
,000
|
,924
|
150
|
,000
|
a. Correction de signification de Lilliefors
*Board explanation: Normality tests to see the
results for Kolmogorov-Smirnova and Shapiro-Wilk in term of
statistic and signification.
Normality testing: The normality testing is used in
order to see if each variable are well distributed. A normal distribution is a
theoretical frequency distribution that is bell-shaped and
symmetrical, with tails extending indefinitely either side of the
centre. The mean, median and mode coincide at the centre. (Hun Myoung Park,
Ph.D. 2008)
For this study, the data don't seem to follow a normal
distribution as the significations for each is lower than 0,005. However, most
of the time, data appear to not follow a normal distribution and, as this
doesn't have a serious impact of the rest of the analysis, the researcher has
decided to not transform the data.
3.3.2 Regression analysis
After gathering those different results, simple regression was
established in order to test the formulated hypotheses. In order to test those
hypotheses, as it was explained in the methodology part, the researched has mad
simple linear regression in order to see if the independent variable permitted
to explain the dependent variable.
3.3.2.1 Theoretical review
Definition: Simple linear regression permits to measure
the linear relationship between two variables, as the correlation, but it gives
a direction the relationship: in other words it permits to assess how much the
independent variable (IV) is explaining the variation of the dependent variable
(DV). (O. Renaud and G. Pini 2005)
Null hypothesis: In the case of regression, the null
hypothesis is that there is no relationship between the dependent variable and
the independent variable, so the independent variable does not predict the
dependent variable. The alternative hypothesis is that it is possible to
predict the dependent variable from the independent variable. Eric Yergeau.
(2007)
For all the following hypotheses:
- The Significance Level is set has: á = 0.05 and,
- If p-value (Sig) < á the regression line fits the
data better than a flat line; the relationship is significant. (UCLA University
2008)
| 
