5.3.1. Path Coefficient
The structural equation model analysis was conducted to verify the difference between path coefficients statistically and to examine whether the hypothesis was satisfied (Fig. 2).
As a result of the path coefficient verification analysis (Table 4), first the relationship between authenticity and brand advertisement (Coefficient = 0.029, β = 0.040) shows that authenticity value does not have a significant positive effect on brand advertisement. Thus, hypothesis 1 set in this study was rejected. However, the relationship between relevance and brand advertisement (Coefficient = 0.464, β = 0.543), and that between sustainability and brand advertisement (Coefficient = 0.162, β = 0.208) show that relevance and sustainability values have a significant positive effect on brand advertisement. Thus, hypotheses 2 and 3 set in this study were supported.
Second, the relationship between authenticity and brand preference (Coefficient = 0.080, β = 0.096) shows that authenticity value and brand preference have a significant positive effect on brand advertisement. Thus, hypothesis 4 set in this study was supported. However, the relationship between relevance and brand preference (Coefficient = 0.031, β = 0.031), and that between sustainability and brand preference (Coefficient = 0.080, β = 0.090) show that relevance and sustainability values do not have a significant positive effect on brand preference. Thus, hypotheses 5 and 6 set in this study were rejected.
Third, the relationship between brand advertisement and brand preference shows a significant positive effect on brand preference (Coefficient = 0.575, β = 0.505). Thus, hypothesis 7 set in this study was supported.
Table 4
Results of path coefficient verification
Path | Coefficient | β | S.E | C.R | Supported or Rejected |
authenticity → brand advertisement | 0.029 | 0.040 | 0.035 | 0.818 | Rejected (H1) |
relevance → brand advertisement | 0.464 | 0.543 | 0.085 | 5.485*** | Supported (H2) |
sustainability → brand advertisement | 0.162 | 0.208 | 0.076 | 2.138* | Supported (H3) |
authenticity → brand preference | 0.080 | 0.096 | 0.036 | 2.213* | Supported (H4) |
relevance → brand preference | 0.031 | 0.031 | 0.090 | 0.341 | Rejected (H5) |
sustainability → brand preference | 0.080 | 0.090 | 0.075 | 1.062 | Rejected (H6) |
brand advertisement → brand preference | 0.575 | 0.505 | 0.091 | 6.322*** | Supported (H7) |
**p < .01, ***p < .00 |
5.3.2. Mediating Effect Verification
In this study, AMOS 24.0 program and bootstrapping method were used to verify the significant mediating effect of brand advertisement that is set as a parameter. The analysis results from Table 5 are summarized as follows.
Table 5
Results of mediating effect verification
Path | Direct effect | Indirect effect | Total effect |
Authenticity → brand advertisement → brand preference | - | - | - |
Relevance → brand advertisement → brand preference | 0.029 | 0.040 | 0.035 |
Sustainability → brand advertisement → brand preference | 0.464 | 0.543 | 0.085 |
*p < .05, **p < .01, ***p < .001 |
First, brand advertisement did not appear to have an indirect effect in the correlation between authenticity value and brand preference, was and was therefore excluded from the mediating effect test. Second, as a result of examining the mediating effect of brand advertisement on the relationship between relevance value and brand preference, the indirect effect of relevance value on brand preference through brand advertisement was .274, which was statistically significant at the significance level p < .001. Therefore, brand advertisement has a complete mediating effect between relevance value and brand preference. Third, as a result of examining the mediating effect of brand advertisement on the relationship between sustainability value and brand preference, the indirect effect of sustainability value on brand preference through brand advertisement was .105, which was statistically significant at the significance level p < .001. It means that brand advertisement did act as a completed mediator in the relationship between between sustainability value and brand preference. Thus, hypothesis 8 set in this study was supported.
In addition, the results of mediating analysis using the PROCESS macro in SPSS are shown in Table 6 below.
Table 6
Results of mediating effect verification using the PROCESS macro
Path | Direct effect B (95% CI) | Indirect effect B (95% CI) | Total effect B (95% CI) |
authenticity → brand advertisement → brand preference | .082 (.007–.157) | .026 (− .006–.064) | .108 (.025–.192) |
relevance → brand advertisement → brand preference | .068 (− .051–.187) | .215 (.131–.295) | .283 (.165–.401) |
sustainability → brand advertisement → brand preference | .070 (− .027–.168) | .066 (.020–.153) | .136 (.030–.243) |
The direct and indirect effects, and the total effect size were confirmed through bootstrapping proposed by Hayes, and the 95% confidence interval confirmed the direct and indirect effects, and statistical significance of the total effect. As a result, the indirect effect size of authenticity on brand performance was .026, but the 95% confidence interval was − .006–.064, which was not significant. However, the direct effect was .082, and the 95% confidence interval was .007–.157. In other words, authenticity directly increased the brand performance without mediating the brand advertisement. The indirect effect of relevance on brand performance was .215 by mediating brand advertisement, and 95% confidence interval was .131–.295, which did not include 0. However, as direct effect was not significant, the relevance did not affect the direct brand performance, but the brand performance was significantly increased only through brand advertisement. The indirect effect size of sustainability on brand performance was .066, and the 95% confidence interval was .020–.153, which did not include 0. However, the direct effect was not significant, and therefore, the sustainability did not affect the direct brand performance, and was significantly increased by brand advertisement.