The main findings of this research are divided into 5 sections. The first section shows the results found when including and analyzing the whole sample. The second section shows the results found when analyzing the Spanish sample. The third section shows the results extracted from the Taiwanese sample. The fourth section shows the comparative results between the Spanish and the Taiwanese samples. While fifth section presents the results obtained from the comparison between different rank employees
Main Sample Results
The main study is conducted with the 181 participants (employees, middle managers and top managers) that compose the sample included in the research.
In the Table 5.1 are gathered the validity and reliability tests run with N=181, which includes and combines the participants from Spain and Taiwan.
For the main research, an EFA was calculated once again due to the inclusion of a new dimension within the equation model. As appreciate in Table 5.1, factor analysis revealed the acceptance of the statistics from the KMO and the Bartlett Test’s of Sphericity which statistics between .5 and 1.0 for each of the dimensions in the case of KMO and the significance of the Bartlett’s Test confirm the division of the factors adopted within the dimensions of the Glocal CSR model.
Moreover, Cronbach’s Alpha statistics of the inner dimensions of the model are all higher than 0.7, indicating a high level of internal consistency and reliability.
Meanwhile, Composite Reliability statistics offered the confirmation of the high reliability of the model with all its values being higher than 0.7.
Furthermore, the Average Variance Extracted (AVE) determines the variance captured by the indicators relative to the measurement error. Given that all the values are greater than .50, it can be concluded that the data meet the convergent validity (Hair, Tatham & Anderson, 2006).
Regarding the factor loadings, individual item reliabilities were computed by examining this test. According to previous researchers and as expressed Nguyen (2014),
“loadings of 0.5 or greater are acceptable if there are additional indicators for describing
74
the latent construct. Therefore, items with loadings of 0.5 or greater are maintained”. In our main research, all the factors presented values higher than 0.5.
Table 5.1
Main Sample Validity and Reliability Results
KMO
Note: *** Significance level p < .01
According to results from SPSS and PLS tests presented in Table 5.1, all the minimum requirement of reliability and validity tests were met, making the Glocal CSR model valid and reliable for the research purposes of this study.
75
• From financial perspective 0.755 (19.782)***
• From customer perspective 0.778 (18.566)***
• From internal business process perspective 0.868 (46.532)***
• From learning & growth perspective 0.852 (41.942)***
Corporate Social Responsibility R
²
=64.4%• Vision, values and strategy 0.885(59.188)***
• Management systems, organization and processes 0.882(50.371)***
• Products and services 0.856(40.914)***
• Resource and environmental management 0.631(12.638)***
• Communication & Reporting 0.865(46.864)***
• Stakeholder management 0.810(32.267)***
Figure 5.1 Main study Glocal CSR model results
Note: Path Coefficients & T-values: * p < .1, ** p < .05, *** p < .01
76
To evaluate the significance of path coefficients, a bootstrapping 50x procedure was conducted obtaining t-values for each of the paths, in order to assess the significance of the relation between dimensions.
Moreover, to evaluate the statistical significance of the factor loadings the same bootstrapping 50x method was conducted, as expressed in Chapter 3, testing the significance of the factor loadings included in the model. The results of the multivariate test of the Glocal CSR structural equation model are presented in Figure 5.1 and Tables 5.1 and 5.2.
The research showed that all the path coefficients and factor loadings are significant at a 99% level of confidence.
Main Sample Most & Less Dominant Factors
Figure 5.1 gathers all the factors and theirs loading values after the Glocal CSR Structural Equation Model was analyzed as a whole. For each of the dimensions, the most dominant and less dominant factors are going to be highlighted and have into consideration for the theoretical and managerial implications that will be expressed in Chapter 6.
United Nations global compact
Within this dimension, the Environment factor (loading = 0.887***) it has risen as the most dominant and consequently as the one which has the most positive influence over CSR.
On the other side, the Labour Rights factor (loading = 0.544***) showed up as the less dominant and therefore the one with the lowest impact on CSR.
Local influences
In the case of the Local Influences and regulations, the Employees (loading = 0.838***) turned to be the most dominant factor, being the main local force for organizations in their process to engage CSR.
Contrary, the concern for the Stakeholders and suppliers (loading = 0.809***) is the less influential factor affecting CSR.
77 Corporate social responsibility
For the central dimension, the factor that leads CSR is Vision, values and strategy (loading = 0.885***), it has risen as the most dominant among the 6 factors included.
On the opposite side, Resource and Environmental Management factor (loading
= 0.631***) appeared to be as the less dominant, and therefore where organizations need to improve to achieve higher CSR status.
Business performance
In the last dimension of the model, The Internal Business Performance (loading
= 0.868***) established as the most dominant factor.
Meanwhile, The Financial Business Performance factor (loading = 0.755***) has risen has the less significant one.
Main Sample Path Coefficients & Explained Variances
As the results included in the Figure 5.1 expresses, the main study found significant and positive effects between the dimensions included. Therefore, for the purpose of test the hypothesis based on the literature review, the research used T-values after bootstrapping of each of the path coefficients and checked their level of significance.
Table 5.2 shows the path coefficients of the variables as well as their level of significance.
Following this results, we can affirm that the Variable United Nations Global Compact (ß = .443, t = 9.959, p < .01) has positive and significant effects over Corporate Social Responsibility and therefore, we reject Null Hypothesis 1 (Ho1)
The research also shows the positive and significant effect of the Local Influences variable (ß = .472, t = 9.135, p < .01) on Corporate Social Responsibility, and consequently rejecting the Null Hypothesis 2 (Ho2).
78
For the case of the Null Hypothesis 3 (Ho3), the study revealed the positive and significant effects of CSR (ß = .660, t = 9.124, p < .01) over the Business Performance of the surveyed organizations, what led us to reject Ho3.
From the sample N=181 the alternatives hypotheses were accepted concluding that:
Ha1. The variable UNGC (X1) has significant effects on CSR (Y1).
Ha2. The Variable LI(X2) has significant effects on CSR (Y1).
Ha3. The variable CSR (X3) has significant effects on BP (Y2).
Table 5.2
Main Study path Coefficient Analysis Results
Path Coefficients & T-values: * p < .1, ** p < .05, *** p < .01
Moreover, when analyzing the path coefficients, the statistics from the model showed the indirect effects that the independent variables UNGC (0,292) and the LI (0,311) have over the Business Performance as is presented in Table 5.3.
Table 5.3
Main Study Indirect Effects Results
Variables 1 2 3 4
1 UNGC - 0,292
2 LI - 0,311
3 CSR -
4 BP -
Path Hypothesis β-path Adj.
t-value Sig. Direction Null Hypotheses
UNGCCSR Ho1 0.443 9.959 *** + Rejected
LICSR Ho2 0.472 9.135 *** + Rejected
CSRBP Ho3 0.660 9.124 *** + Rejected
79
After having all the statistics and having rejected the null hypothesis, we determined the percentage of variance of the dependent variables that is explained by the independent variables. Table 5.4 presents the R-square and adjusted R-square given by the Glocal CSR model.
Table 5.4
Main Study R square & Adjusted R Square
Variables R Square Adj. R Square
CSR 64.8% 64.4%
BP 43.5% 43.2%
As Table 5.4 shows, UNGC and LI explained 64.4% of the variance of CSR, what gives us a high grade of accuracy when predicting CSR.
In the case of Business Performance, CSR directly and UNGC and LI in an indirect way explain 43.2% of its variability.
Main Study Discussion
From the international rights promoted by the UNGC, we appreciate that environmental concern settled as the main factor of influence for CSR while, the labour rights are the less significant. The structural equation model the positive relation of the UNGC towards CSR.
However, when comparing the international values of UNGC with local influences, the local dimension showed to present greater effects on CSR, with employees as the driven force of companies when engaging CSR.
Regarding CSR dimension, the dominant factor vision, values and strategy represent the milestone on implementing and enhancing CSR among the whole organization structure and operations. This result is also supported by employee concerns and vision about CSR, which are leading CSR engagement.
However it is also seemed that Resource and Environmental is the weakest factor for CSR.
With these results we can suggest that even organizations which have in their vision and strategies the concern for environment and sustainable future, in reality, they
80
are still adopting measures and regulations that will help them to achieve more innovative and sustainable operations in the long-term.
Analyzing the effects of CSR on Business Performance, the results showed the positive influence of the first one over the second. Within business performance, the internal business performance resulted in being the factor with the greatest influence from CSR and once again, indicates the restructuration stage that business are facing when engaging CSR.
On the opposite side, the financial performance turned as the weakest factor, and therefore the one in which CSR has the least impact, which supports what stated by UNGC (2014a) saying that businesses are moving forward, and therefore, this changing state need the financial resources to be implemented.
Spain’s Sample Results
The research conducted with Spanish participants is composed by 81 employees, middle managers and top managers of manufacturing organizations.
In order not repeat the explanation given in Main Sample section, the validity and reliability results are going to be shown in Table 5.5, and only special considerations, if any, will be highlighted.
According to results from SPSS and PLS tests presented in Table 5.5, the minimum requirement of reliability and validity tests were met, making the Glocal CSR model valid and reliable for conducting the research among the Spanish sample.
However, in the UNGC dimensions, the factor Labour Right dropped till (0.370), indicating that in further analyses this factor should be revised and it might be modified for fitting better with the other constructs included in this dimension. In the present research, we maintained the Labour Rights factor due to the fact that it doesn’t suppose statistical problems for the Main or Taiwanese samples studies.
Following the same pattern as in the main sample study, and in order to evaluate the statistical significance of the factor loadings and the path coefficients, bootstrapping 50x method was conducted and applied to the Glocal CSR model.
81
The results obtained from the Structural Equation Model run by SmartPLS software are presented in Figure 5.2 and Table 5.6.
Table 5.5
Spain’s Sample Study Validity and Reliability Results
KMO
Note: *** Significance level p < .01
82
Figure 5.2 Spain’s sample Glocal CSR model results
Note: Path Coefficients & T-values: * p < .1, ** p < .05, *** p < .01
• From financial perspective 0.792 (13.994)***
• From customer perspective 0.772 (10.868)***
• From internal business process perspective 0.877 (29.057)***
• From learning & growth perspective 0.866 (33.510)***
Corporate Social Responsibility R
²
=75.5%• Vision, values and strategy 0.921 (65.440)***
• Management systems, organization and processes 0.904(49.858)***
• Products and services 0.869(32.679)***
• Resource and environmental management 0.670(12.304)***
• Communication & Reporting 0.842(23.092)***
• Stakeholder management 0.770(16.879)***
83
Spain’s Sample Most & Less Dominant Factors
Figure 5.2 contains all the factors and theirs loading values obtained after running the Glocal CSR Structural Equation Model using the Spanish Sample N=81.
As done with the main sample study, within each of the dimensions the most dominant and less dominant factors are going to be highlighted. Those with the most and less relevance will have special consideration in Chapter 6.
United Nations global compact
As shown in the main sample study, the Spanish sample also revealed the Environment factor (loading = 0.949***) as the most dominant factor within this dimension. We can conclude that Environment is the factor which influences the most at the time of Spanish manufacturing organizations engaging into CSR.
On the other side, the Labour Rights factor (loading = 0.370***) was once again the less dominant factor and therefore the one with the lowest impact on CSR. In this case, special consideration for further and future research will be taken in order to fit with the statistical minimums.
Local influences
In the case of the Local Influences and regulations, the Media & Lobby Groups (loading = 0.841***) turned to be the most dominant factor, being the main Spanish local driven force and source of pressure for organizations in their process to engage CSR.
On the opposite side, the Employees (loading = 0.808***) happen to be the factor with less influence and therefore with the lowest repercussion for the CSR status in Spanish manufacturing organizations.
Corporate social responsibility
Within CSR, the factor that leads this dimension in the Spanish context is Vision, values and strategy (loading = 0.921***), which as happen in the main sample study, has risen as the most dominant among the 6 factors included in the CSR dimension.
Oppositely, the Resource and Environmental Management factor (loading = 0.670***) showed again as the weakest factor, and as expressed in the main sample
84
study, where organizations need to improve to achieve higher levels of CSR and consequently of Business Performance.
Business performance
The Internal Business Performance factor (loading = 0.877***) has risen as the most significant and the factor that is affected the most by the influence that CSR has over the Business Performance of Spanish organizations.
Within Business Performance, The Customer Perspective Business Performance (loading = 0.772***) is the weakest factor among the rest, result that supports what Hertz (2004) expressed about the image and levels of trust and confidence from the European population regarding corporations.
Spain’s Sample Path Coefficients & Explained Variances
Indicated in Figure 5.2, the Spanish sample study also found significant and positive effects between the dimensions included in the Glocal CSR model. Table 5.6 compiles the path coefficients of the variables as well as their level of significance.
According to these results, the research supports the alternative hypothesis (Ha.
1) that states that the Variable United Nations Global Compact (ß = .448, t = 7.239, p
< .01) has positive and significant effects over Corporate Social Responsibility and therefore, the findings reject Null Hypothesis 1 (Ho1)
For the Spanish context, the research also shows the positive and significant effect of the Local Influences variable (ß = .530, t = 7.142, p < .01) on Corporate Social Responsibility, and consequently we reject the Null Hypothesis 2 (Ho2).
Results also showed how CSR has positive and significant effects (ß = .660, t = 9.124, p < .01) over the Business Performance of the Spanish manufacturing surveyed organizations. With this results the Null Hypothesis 3 (Ho3) was rejected as it happens in the main sample study.
The statistic results obtained from testing the hypotheses are summarized and presented in Table 5.6.
85 Table 5.6
Spain’s Sample Path Coefficient Results
Note: Path Coefficients & T-values: * p < .1, ** p < .05, *** p < .01
From the Spanish sample (N=81) the alternatives hypotheses were also accepted concluding that:
Ha1. The variable UNGC has significant effects on CSR in Spanish organizations.
Ha2. The Variable LI has significant effects on CSR in Spanish organization.
Ha3. The variable CSR has significant effects on the BP of Spanish organizations.
But apart from the direct effects expressed by the path coefficients indicated in figure 5.2 and table 5.6, the structural equation model also revealed the existence of indirect effects from the United Nations Global Compact and the Local Influences on the depended variable Business Performance. The values from the indirect effects are collected in Table 5.7.
Table 5.7
Spain’s Sample Study Indirect Effects Results
Variables 1 2 3 4
1 UNGC - .300
2 LI - .355
3 CSR -
4 BP -
Path Hypothesis β-path Adj.
t-value Sig. Direction Null Hypotheses
UNGCCSR Ho1 0.427 5.134 *** + Reject
LICSR Ho2 0.432 4.610 *** + Reject
CSRBP Ho3 0.660 9.124 *** + Reject
86
From the indirect effects results it is denoted that the variables Business Performance is affected by other variables not including in the research and that as will be expressed in Chapter 6, future researchers might need to establish additional control variables to obtain a more accurate and direct influence of CSR over Business Performance.
Once again, the percentage of variance from the dependent variables CSR and BP that is explained by the independent variables UNGC and LI and CSR respectively are presented in Table 5.8, indicating the R-square and adjusted R-square showed by the Glocal CSR model.
Table 5.8
Spain’s Sample R Square & Adjusted R Square
Variables R Square Adj. R Square
CSR 76.1% 75.5%
BP 44.8% 44.1%
As Table 5.8 indicates, the percentage of variance of CSR which is explained by UNGC and LI is equal to 75.5%. With this statistic the research and Glocal CSR model determined a high grade of predictability for CSR, considering this number, as a success of the model.
In the case of Business Performance, its percentage of variance explained by CSR (44.1%) which is considered as a high statistic, however, it also indicates that BP of Spanish manufacturing organizations is influenced by different factors that have not been included in this research, and therefore further researches can improve the model.
Spain’s Study Discussion
The Glocal CSR model conducted in Spain rejected the null hypotheses stated in Chapter 3. From the results, it is obtained that the Environmental factor postulates as the dominant value from those promoted by the UNGC globally.
In this context, Spanish organizations seem to be perceiving higher influences from the local influences composed by the employees, stakeholders and the local media and lobby groups. As the results show, the local media and lobby groups are considered the most influential factor for Spanish manufacturing organizations what implies that
87
union and other external pressure groups execute the main maximum effects over Spanish CSR.
On the opposite side, employees were found as the weakest factor and therefore where Spanish business can work to foment CSR.
For CSR, the Spanish sample showed the same results as found in the main study.
Lastly regarding Business Performance, it is shown the internal business perspective as the dominant factor, while the customer perspective appeared as the weakest factor reminding the low confidence that European and most of the Spanish citizens have regarding organizations, due to all the recent corruption scandals splashing political parties and national institutions.
Taiwan’s Sample Results
The research conducted in Taiwan is composed by N=100. The descriptive and demographic characteristics of the sample were introduced in Chapter 4 Tables 4.6 &
4.7.
As analyzed for the main sample and the Spanish sample, for Taiwan also validity and reliability tests were run in order to confirm the results found in the previous samples, and thereby assure the consistency of the Glcocal CSR model. For this purpose, Table 5.9 shows the validity and reliability statistic gotten from this sample.
Once the validity and reliability analyses were successfully run, the Glocal CSR structural equation model was tested running SmartPLS software. Its results are shown in Figure 5.3 and Tables 5.10, 5.11 and 5.12.
88 Table 5.9
Taiwan’s Sample Study Validity and Reliability Results KMO
With the results included in Table 5.9, we can affirm the validity and reliability of the measurement instrument as well as the Glocal CSR model. All the statistics are higher than the minimums required for accepting the conditions of validity and reliability.
89
Figure 5.3 Taiwan’s sample Glocal CSR model results
Note: Path Coefficients & T-values: * p < .1, ** p < .05, *** p < .01
• From financial perspective 0.716 (13.790)***
• From customer perspective 0.802 (18.843)***
• From internal business process perspective 0.863 (40.504)***
• From learning & growth perspective 0.834 (32.005)***
Corporate Social Responsibility R
²
=53.4%• Vision, values and strategy 0.854(35.821)***
• Management systems, organization and processes 0.851(24.231)***
• Products and services 0.847(24.785)***
• Resource and environmental management 0.627(7.061)***
• Communication & Reporting 0.897(38.767)***
• Stakeholder management 0.852(26.134)***
UN Global Compact
• Human Rights 0.781 (17.592)***
• Labour Rights 0.731 (11.995)***
• Environment 0.788 (17.977)***
90
Taiwan’s Sample Most & Less Dominant Factors
In Figure 5.3 are presented all the factors and theirs loading values obtained after running the Glocal CSR Structural Equation Model with the Taiwanese Sample (N=81).
For fitting the comparative purposes of this research, the results are also
For fitting the comparative purposes of this research, the results are also