The homogeneity of education

This study divides education into 3 levels: senior high, college, and above research institute. The result of homogeneity test shows that there is no significant difference between brand image (F=.002, p> .05), brand awareness (F=.535, p> .05), country of origin image (F=.107, p> .05), perceived quality (F=.188, p> .05), and purchase intention (F=.453, p> .05), indicating that variables would not have different effects due to education. The homogeneity exists (see Table 4-9).

Table 4-9 The Homogeneity Test Result of Education

Variables Mean F value

Senior high (n=43)

College (n=188)

Research institute (n=37)

Brand image 5.63 5.62 5.62 .002

Brand awareness 5.37 5.37 5.57 .535

Country of origin 5.70 5.71 5.78 .107

Perceived quality 5.63 5.66 5.76 .188

Purchase intention 5.56 5.69 5.76 .453

*p< .05, **p< .01, ***p< .001

4.4 Path analysis

Path analysis is used to test the causal models by examining the relationships among variables. This study uses Partial Least Squares (PLS) Algorithm to examine the path coefficient (β) and coefficients of determination(R²) that are tested by t-test through SmartPLS. Path coefficient (β) is examining the possible causal linkage among variables and is tested by t-value; coefficients of determination, R² value, is the strength and direction of the linear relationship between two variables and their predicted values, which ranges from 0 to 1. 0 indicates that the proposed model does not improve prediction over the mean model while 1 means perfect prediction (K. G.

Martin, 2005). It can be used to assess the goodness of fit of the model. The R² in

63

perceived quality is .718 and purchase intention is .724, which indicates that this model has 71.8% and 72.4% explained variance. Fig. 4-1 presents the path analysis of purchase intention as dependent variable. The standardized path coefficients are all larger than .05, which indicates that each relationship between two dimensions is different from zero (Stevens, 2009).

The results of the hypothesis tests are verified and explained as following and the model is provided in Fig. 4-1. The hypothesis 1 is that green product’s brand image is positively related to perceived quality. The path coefficient (β) between brand image of green products and perceived quality is .119 (t=.288, p> .05), which has an insignificant influence, indicating that the relationship between brand image of green products and consumers perceived quality is insignificant. Thus, the results do not support H1. To verify H2 that green product’s brand image is positively related to purchase intention, this study examines the path coefficient (β) between brand image of green products and purchase intention. The result of path coefficient (β) is .224 (t=2.586, p< .01) that has a positive influence. Thus, the results support H2 that the better brand image of green products, the higher intention consumers would purchase.

Moreover, the path coefficient (β) between brand awareness of green products and perceived quality (H3) is .970 (t=2.159, p< .05), which has a positive influence, showing H3 that green product’s brand awareness is positively related to perceived quality is supported. The stronger brand awareness of green products is, the higher products quality consumers would perceive.

The H4 is that green product’s brand awareness is positively related to purchase intention. The path coefficient (β) between brand awareness of green products and purchase intention is .305 (t=3.208, p< .01), which has a positive influence, indicating that the stronger brand awareness of green products, the higher intention consumers would purchase. Thus, the results support H4. In addition, the H5 that country of

64

origin image moderates the relationships between green product’s brand image and perceived quality is not supported because the path coefficient (β) of country of origin image to green product’s brand image and perceived quality is .796 (t=1.054, p> .05), which is not statistically significant. Thus, country of origin image does not moderate the relationships between green product’s brand image and perceived quality.

Furthermore, the path coefficient (β) of country of origin image to green product’s brand awareness and perceived quality (H6) is -1.314 (t=1.855, p> .05) that has an insignificant result. Thus, H6 is not supported, showing that country of origin image does not moderate the relationships between green product’s brand awareness and perceived quality. The H7 is that green product’s perceived quality is positively related to purchase intention. The path coefficient (β) between perceived quality and purchase intention is .417 (t=4.406, p< .001), which has a positive influence, indicating that the better quality green products is, the higher intention consumers would purchase. Thus, the results support H7.

65

Fig. 4-1 Path Analysis Model and t-values Note: Standardised estimates with t-values in brackets

*p< .05, **p< .01, ***p< .001

In addition to examining causal relationship and direct effect between independent variable and dependent variable, independent variable may also have indirect effect on dependent variable through other variables. Therefore, this study further compares the effects among variables to understand their relationships. The direct effect is the correlation from one variable to another, whereas indirect effect is one variable affecting another which in turn affects a third variable and the total effects is the total of direct and indirect effects (direct + indirect). In this study, in order to verify hypothesis 8 and hypothesis 9 that whether perceived quality has mediating role on the relationship between green products’ brand image and brand awareness and purchase intention, the direct and indirect effect among variables would be considered. The Table 4-10 shows the direct, indirect, and total effects of the variables.

66

Table 4-10 Path Analysis of Direct, Indirect, and Total Effects

Path Direct effect Indirect effect Total effects

Brand image Perceived

In variable of perceived quality, the effect of brand awareness is higher (total effect= .970), presenting that brand awareness of green products influences consumers perceived quality more than brand image. In variable of purchase intention, the effect of perceived quality is the highest (total effect= .417), showing that perceived quality is the factor that influences consumers purchase intention of green products the most in the model.

Besides, this study investigates whether brand image of green products will directly influence purchase intention. The results provide sufficient evidence to examine hypothesis 8 that perceived quality mediates the effect of green product’s brand image on purchase intention. The direct effect of brand image of green products on purchase intention is .224, and indirect effect is .050. The direct effect is larger than indirect effect, suggesting that perceived quality does not have mediating effect.

Hence, the hypothesis 8 is not supported.

This study further examines hypothesis 9 that perceived quality mediates the effect of green product’s brand awareness on purchase intention. The results show that the path coefficient of direct effect is .305 and indirect effect is .404, which is

67

significant, indicating that there is a mediating influence between the two variables and perceived quality is partial mediation. Therefore, the hypothesis 9 is supported.

Besides, it can be noted the indirect effect ( .404) is larger than direct effect ( .305), showing that the effect of green product’s brand awareness on purchase intention through perceived quality is more significant than the effect of green product’s brand awareness on purchase intention. This study concludes the result of hypotheses as shown in Table 4-11.

Table 4-11 Results of Hypothesis Testing

Hypothesis β-value Result

H1: Brand image of green products positively related to perceived quality

.119 Not supported H2: Brand image of green products positively related to

purchase intention

.224*** Supported H3: Brand awareness of green products positively related

to perceived quality

.970* Supported H4: Brand awareness of green products positively related

to purchase intention

.305** Supported H5: Country of origin image moderates the relationships

between green product’s brand image and perceived quality

.796 Not supported

H6: Country of origin image moderates the relationships between green product’s brand awareness and perceived quality

-1.314 Not supported

H7: Perceived quality of green products positively related to purchase intention

.417*** Supported H8: Perceived quality mediates the effect of green

product’s brand image on purchase intention

.050 Not supported H9: Perceived quality mediates the effect of green

product’s brand awareness on purchase intention

.404*** Supported

*p< .05, **p< .01, ***p< .001

68