Dynamic Capabilities, Environmental Management Capabilities, Stakeholder Pressure and Eco-Innovation of Chinese Manufacturing Firms: A Moderated Mediation Model

Round 1

Reviewer 1 Report

The research work on Dynamic Capabilities, Environmental Management Capabilities,  Stakeholder Pressure, and Eco-innovation: A Moderated Mediation  Model is very interesting to read and good work is carried out. However, i have few observations;

(1) The title can be better written including the demography of the work carried out.

(2) literature review can concern stakeholder and stakeholder pressure framework as explained by Dwivedi and Momaya (2003)...

Dwivedi, R. and Momaya, K., 2003. Stakeholder flexibility in e-business environment: A case of an automobile company. Global Journal of Flexible Systems Management, 4(3), pp.21-32.

 

(2) The method of the moderated mediation model is very well explained with the detailed method and which is missing in the work..therefore suggest you read the improved method part..

Yoshikuni, A.C. and Dwivedi, R. (2023), "The role of enterprise information systems strategies enabled strategy-making on organizational innovativeness: a resource orchestration perspective", Journal of Enterprise Information Management, Vol. 36 No. 1, pp. 172-196. https://doi.org/10.1108/JEIM-10-2021-0442

 

 

(3) 181 sample size is enough to carry out the research. However, it must be supported using pilot study and standard deviation and margin of error. Because 181 is not enough sample considering the population.

 

(4) The most important issue is observed which is multicollinearity and not reported in the paper. please include the multicollinearity and its explanation. 

 

All the best

Moderate english editing is required for improving the work..

Author Response

Thank you so much for providing constructive comments and giving us the opportunity for further revisions! We believe that we have been able to address your comments and suggestions, and that our paper has been substantially improved thanks to you. Below, we indicate how we respond to each of your comments.

 

 (1) The title can be better written including the demography of the work carried out.

Thank you so much for this advice! We have changed the title into “Dynamic Capabilities, Environmental Management Capabilities, Stakeholder Pressure and Eco-innovation of Chinese Manufacturing Firms: A Moderated Mediation Model” to include the demography of the work.

 

(2) literature review can concern stakeholder and stakeholder pressure framework as explained by Dwivedi and Momaya (2003)...

Dwivedi, R. and Momaya, K., 2003. Stakeholder flexibility in e-business environment: A case of an automobile company. Global Journal of Flexible Systems Management, 4(3), pp.21-32.

Thank you so much for this comment! We have read the recommended article and added related discussion of stakeholder pressure in Section 3.4, which makes our literature review more comprehensive.

 

(2) The method of the moderated mediation model is very well explained with the detailed method and which is missing in the work..therefore suggest you read the improved method part..

Yoshikuni, A.C. and Dwivedi, R. (2023), "The role of enterprise information systems strategies enabled strategy-making on organizational innovativeness: a resource orchestration perspective", Journal of Enterprise Information Management, Vol. 36 No. 1, pp. 172-196. https://doi.org/10.1108/JEIM-10-2021-0442

Thank you so much for this advice! The recommended article provides a good example of moderated mediation model using PLS-SEM method. It is embarrassing to admit that we are not skilled at using the PLS-SEM method and afraid to make mistakes in the results, since the time given for this revision is quite limited. We had tested the mediation effects, moderation effects and moderated mediation effects using ordinary least squares regression, specifically the bootstrap resampling method recommended by Preacher and Hayes (2004) and Hayes (2015) using PROCESS macro for SPSS, which is also popular among researchers (e.g., Elidemir, Ozturen and Bayighomog, 2020; Sun et al., 2020; Xiao, Wang and Guo, 2022 all published on Sustainability). Besides, Hayes, Montoya and Rockwood (2017) illustrate that the results of PROCESS and SEM are largely identical. Therefore, we believe that the method that we implemented is adequate to test the hypotheses. We will keep on learning the PLS-SEM method and use it in our future studies.

 

(3) 181 sample size is enough to carry out the research. However, it must be supported using pilot study and standard deviation and margin of error. Because 181 is not enough sample considering the population.

The sample size is relatively small but enough for our research. As we noted in Section 5.2, the sample size exceeded the minimum value of 150 suggested by Anderson and Gerbing (1988) for structural equation modeling, and also exceeded the minimum sample size required to achieve the desired power of 0.8 (MacCallum, Browne and Sugawara, 1996). The sample size is also sufficient for multiple regression analysis according to the rule-of-thumb which suggests that the ratio of participants to variables should be at lease 10:1 (Van Voorhis and Morgan, 2007).

However, we didn't conduct pilot study and we will pay more attention to research design and sample size in our further studies. We're not sure about the standard deviation and margin of error that you mentioned and we're sorry about that. In papers utilizing PROCESS for hypotheses testing, it's common to report standard error and bootstrap confidence intervals (e.g., Elidemir, Ozturen and Bayighomog, 2020; Sun et al., 2020; Xiao, Wang and Guo, 2022). We reported these data in Table 5, Table 6 and Table 7.

 

(4) The most important issue is observed which is multicollinearity and not reported in the paper. please include the multicollinearity and its explanation. 

Thank you so much for this comment. Pearson correlation coefficients in Table 2 are all smaller than 0.7, and the greatest correlations exist between two dimensions of dynamic capabilities (AC and RC, 0.493) and between two dimensions of eco-innovation (PSI and PDI, 0.566). Also, as is shown in Section 5.3, we have mean centered the variables and interaction items to avoid potential multicollinearity. Besides, we have reported the values of variance inflation factor (VIF), which showed that the maximum VIF value is 1.416, much lower than the threshold of 10 suggested by Cohen et al. (2003), indicating that multicollinearity is not a severe problem.

 

Comments on the Quality of English Language

Moderate english editing is required for improving the work..

We have checked the manuscript thoroughly to improve our English writing.

 

References

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior research methods, instruments, & computers, 36, 717-731.

Hayes, A. F. (2015). An index and test of linear moderated mediation. Multivariate behavioral research, 50(1), 1-22.

Elidemir, S. N., Ozturen, A., & Bayighomog, S. W. (2020). Innovative behaviors, employee creativity, and sustainable competitive advantage: A moderated mediation. Sustainability, 12(8), 3295.

Sun, H., Rabbani, M. R., Ahmad, N., Sial, M. S., Cheng, G., Zia-Ud-Din, M., & Fu, Q. (2020). CSR, co-creation and green consumer loyalty: Are green banking initiatives important? A moderated mediation approach from an emerging economy. Sustainability, 12(24), 10688.

Xiao, Z., Wang, Y., & Guo, D. (2022). Will Greenwashing Result in Brand Avoidance? A Moderated Mediation Model. Sustainability, 14(12), 7204.

Hayes, A. F., Montoya, A. K., & Rockwood, N. J. (2017). The analysis of mechanisms and their contingencies: PROCESS versus structural equation modeling. Australasian Marketing Journal, 25(1), 76-81.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130-149.

Van Voorhis, C. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for determining sample sizes. Tutorials in quantitative methods for psychology, 3(2), 43-50.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied multiple regression/correlation analysis for the behavioral sciences. Routledge.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments on sustainability-2366986:

 

I generally have three concerns about the current paper. First, the writing style is too casual to be acceptable. Second, the data and numerical implementation (e.g., working codes) should be shared online. Third, the data and results are not properly discussed and investigated, which requires further intelligent efforts. The current version is absolutely unacceptable and significant works are expected in at least one round of major revision. Below are detailed comments.

 

Severe abbreviation issue:

RBV in abstract should be defined.

NRBV defined in section 3.1 is never cited and should not be defined at all.

EMS is defined multiple times in the main text.

The abbreviation CR is defined multiple times.

Multiple definitions of many other abbreviations (e.g., AC) should also be avoided.

The authors should check their manuscript thoroughly. Such a huge amount of writing problems should not exist in a submitted paper.  

 

The 21-page length seems too long and many unimportant technical details are included in the main article. The authors should only include crucial insights and necessary discussions in the main text, while moving the technical details to the supplementary material.

 

The data presentation could be bettered by using plots rather than tables. For example, Table 1 could be transformed to distribution or pie chart for better visualization.

 

To avoid confusion, I would suggest to explicitly define at least the full names of RMSEA, CFI and many other metrics.

 

In section 5.2 (line 530-531), CFI is slightly low <0.95 and RMSEA is slightly large >0.6.

 

What is the correlation coefficient presented in Table 2? A lot of candidates could be imagined with only this keyword.

 

I neither do not understand why the authors are explicitly defining confidence interval as CI while leave the SD abbreviation undefined.

 

Data availability is rather poor in this paper. As an academic contribution, the authors should upload their dataset to some online repository with free access to ordinary researchers (peers). The working code should also be shared along with the data.

 

The regressed model has small R2 (e.g., <0.2 in Table 4), which requires some discussion from an unbiased perspective.

 

 

 

Must be improved. Such careless writing should be avoided in any publication. 

Author Response

Thank you so much for providing constructive comments and giving us the opportunity for further revisions! We believe that we have been able to address your comments and suggestions, and that our paper has been substantially improved thanks to you. Below, we indicate how we respond to each of your comments.

 

1. The writing style is too casual to be acceptable.

1) Severe abbreviation issue: 

RBV in abstract should be defined. 

NRBV defined in section 3.1 is never cited and should not be defined at all. 

EMS is defined multiple times in the main text. 

The abbreviation CR is defined multiple times. 

Multiple definitions of many other abbreviations (e.g., AC) should also be avoided. 

The authors should check their manuscript thoroughly. Such a huge amount of writing problems should not exist in a submitted paper.

2) I neither do not understand why the authors are explicitly defining confidence interval as CI while leave the SD abbreviation undefined. 

3) To avoid confusion, I would suggest to explicitly define at least the full names of RMSEA, CFI and many other metrics. 

4) What is the correlation coefficient presented in Table 2? A lot of candidates could be imagined with only this keyword. 

Thank you so much for you detailed advices on our writing style. We have checked the manuscript carefully and fixed writing problems, including those that you kindly pointed out. Specifically, we checked abbreviation for variables (AC, RC, EMS, PSI, PDI, SP), theories (RBV, NRBV) and metrics (CR, SD, SE, RMSEA, CFI, GFI) to avoid multiple definitions, lack of definitions and unnecessary abbreviation. Also, to avoid confusion, we have made it clear that the correlation coefficients in Table 2 refers to Pearson correlation coefficients among variables.

 

2. The data and numerical implementation (e.g., working codes) should be shared online.

Data availability is rather poor in this paper. As an academic contribution, the authors should upload their dataset to some online repository with free access to ordinary researchers (peers). The working code should also be shared along with the data. 

Thank you so much for this advice. To improve the availability of our data, we have uploaded the data as supplementary materials, which will be available online if this manuscript is to be published. Data was analyzed using SPSS and AMOS software, therefore we didn't develop any working code.

 

3. The data and results are not properly discussed and investigated, which requires further intelligent efforts.

Thank you so much for this comment. We have rewritten Section 6.1 to make further discussion of our results, including statistically significant results as well as insignificant results to provide more insights.

 

4. Others.

1) In section 5.2 (line 530-531), CFI is slightly low <0.95 and RMSEA is slightly large >0.6. 

Although Hu and Bentler (1999) suggest more rigorous criteria in evaluating model fit, others argue that "conventional CFA goodness of fit criteria are too restrictive when applied to most multifactor rating instruments" (Marsh, Hau and Wen, 2004; Marsh, Hau and Grayson, 2005). Also, they indicate that Hu and Bentler's approach has potential problems since Type 1 error rates are larger at smaller sample size while the likelihood of correctly rejecting a false model decreases as sample size becomes larger. In our manuscript, CFI (0.927) is slightly below 0.95, CFI above 0.90 is also widely considered acceptable (e.g., Yang et al., 2011; Cheng, Yang and Sheu, 2014; Yousaf, 2021; Wu et al., 2023). The value of RMSEA (0.068) is between 0.05 and 0.08, which indicates fair fit according to Browne and Cudeck (1992) and MacCallum, Browne and Sugawara (1996). Therefore, the fit indices of our measurement model are acceptable.

The aforementioned articles are listed in the end.

 

2) The regressed model has small R2 (e.g., <0.2 in Table 4), which requires some discussion from an unbiased perspective. 

Thank you so much for this comment. Although the R² of the regression model are smaller than 0.2 in Table 4, the value of R²improves significantly as the independent variables, mediation variable, moderation variable and interaction items enter the model (except the R² change between Model 7 and Model 10) (please see the table below).

Model	R2	R2 change	F change	Significance of F change
Dependent variable: EMS
Control variables	.100	.100	6.119	.001
Model 1	.191	.091	9.181	.000
Model 8	.348	.157	12.838	.000
Dependent variable: PSI
Control variables	.022	.022	1.228	.301
Model 2	.110	.089	8.119	.000
Model 4	.191	.080	16.108	.000
Model 9	.232	.042	2.867	.038
Dependent variable: PDI
Control variables	.003	.003	.167	.918
Model 5	.053	.050	4.297	.015
Model 7	.095	.042	7.514	.007
Model 10	.110	.015	.880	.453

 

We also notice that many studies on drivers of eco-innovation using survey data have small R², because there are various antecedents of eco-innovation and the factors included in one study are just a small portion of them. For instance, R² values range from .015 to .292 in Cai and Zhou's (2014) research, range from .157 to .242 in Long and Liao's (2021) research, and range from .165 to .450 in Chang and Gotcher's (2020) research.

The aforementioned articles are listed in the end.

 

3) The 21-page length seems too long and many unimportant technical details are included in the main article. The authors should only include crucial insights and necessary discussions in the main text, while moving the technical details to the supplementary material. 

Thank you so much for this advice. We have moved the measurement items and item loading in Table 3 into Appendix A to make the main text shorter. We have also moved description of EMS measurement in Section 2.3 into Section 4.2.3 and deleted repetitive sentences.

 

4) The data presentation could be bettered by using plots rather than tables. For example, Table 1 could be transformed to distribution or pie chart for better visualization. 

Thank you so much for this advice. We have transformed Table 1 into pie charts (Figure 2 and Figure 3) to better represent the characteristics of sample firms.

 

Comments on the Quality of English Language

Must be improved. Such careless writing should be avoided in any publication. 

We have checked the manuscript thoroughly to improve our English writing.

 

References

Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural equation modeling: a multidisciplinary journal, 6(1), 1-55.

Marsh, H. W., Hau, K. T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler's (1999) findings. Structural equation modeling, 11(3), 320-341.

Marsh, H. W., Hau, K. T., & Grayson, D. (2005). Goodness of fit in structural equation models. In Contemporary psychometrics: A festschrift for Roderick P. McDonald. Lawrence Erlbaum Associates Publishers.

Yang, M. G. M., Hong, P., & Modi, S. B. (2011). Impact of lean manufacturing and environmental management on business performance: An empirical study of manufacturing firms. International Journal of production economics, 129(2), 251-261.

Cheng, C. C., Yang, C. L., & Sheu, C. (2014). The link between eco-innovation and business performance: A Taiwanese industry context. Journal of cleaner production, 64, 81-90.

Yousaf, Z. (2021). Go for green: Green innovation through green dynamic capabilities: Accessing the mediating role of green practices and green value co-creation. Environmental science and pollution research, 28(39), 54863-54875.

Wu, R., Zhang, J., Yu, Y., Jasimuddin, S. M., & Zhang, J. Z. (2023). The Impact of Value Cocreation on CSR Innovation and Economic Performance. Sustainability, 15(5), 4008.

Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological methods & research, 21(2), 230-258.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1, 130-149.

Cai, W. G., & Zhou, X. L. (2014). On the drivers of eco-innovation: empirical evidence from China. Journal of Cleaner Production, 79, 239-248.

Long, S., & Liao, Z. (2021). Are fiscal policy incentives effective in stimulating firms' eco‐product innovation? The moderating role of dynamic capabilities. Business Strategy and the Environment, 30(7), 3095-3104.

Chang, K. H., & Gotcher, D. F. (2020). How and when does co-production facilitate eco-innovation in international buyer-supplier relationships? The role of environmental innovation ambidexterity and institutional pressures. International Business Review, 29(5), 101731.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper presents results from survey of companies in Zhiejiang, China on their perception of what drives eco-invention. There are several points that the authors should consider:

1. The list of the questions used in the survey should be given as a supplementary material.

2. Lines 264-265 "results becomes" should be "results become".

3. Line 292 "what really matter" should be matters, but this expression can also be deleted.

4. Line 379 "to response" should be "to respond"

5. On Table 1 "Age" should be clarified that this is the age of the company and not the person surveyed. The age of the person surveyed, if available, should also be given in this table.

There are some minor points that authors should consider with correction of the language. Some examples are listed in the comments above.

Author Response

Thank you so much for providing constructive comments and giving us the opportunity for further revisions! We believe that we have been able to address your comments and suggestions, and that our paper has been substantially improved thanks to you. Below, we indicate how we respond to each of your comments.

 

1. The list of the questions used in the survey should be given as a supplementary material.

Thank you so much for this advice. We have moved the list of the questions and their factor loadings into Appendix A to make Table 3 shorter and make the main text more concise and easier to read.

 

2. Lines 264-265 "results becomes" should be "results become".
3. Line 292 "what really matter" should be matters, but this expression can also be deleted.
4. Line 379 "to response" should be "to respond"

Thank you so much for your detailed advice. We have amended the typing errors that you kindly pointed out and also checked the manuscript thoroughly to avoid such problems.

 

5. On Table 1 "Age" should be clarified that this is the age of the company and not the person surveyed. The age of the person surveyed, if available, should also be given in this table.

Thank you so much for this advice. We have made it clear that "age" referred to firm age. Also, we transformed Table 1 into pie charts (Figure 2 and Figure 3) to better visualize the characteristics of sample firms.

 

Comments on the Quality of English Language

There are some minor points that authors should consider with correction of the language. Some examples are listed in the comments above.

We have checked the manuscript thoroughly to improve our English writing.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments on sustainability-2366986.R1:

 

The manuscript has been revised carefully and the current version seems fine. However, there are still some language and abbreviation issues remaining. I would generally recommend acceptance after removing these nonscientific problems. Below are some remaining problems I identified in a glance.

 

RBV defined in the abstract is not cited. Thus, it should not be defined.

 

The EMS abbreviation is defined twice in the main text, first on page 2 and second at the end of page 18 (Appendix A).

 

 

Improvable points still exist. 

Author Response

Thank you so much for providing further comments and advices! We believe that we have been able to address your comments and suggestions, and that our paper has been further improved thanks to you. Below, we indicate how we respond to your comments.

 

The manuscript has been revised carefully and the current version seems fine. However, there are still some language and abbreviation issues remaining. I would generally recommend acceptance after removing these nonscientific problems. Below are some remaining problems I identified in a glance.

RBV defined in the abstract is not cited. Thus, it should not be defined.

The EMS abbreviation is defined twice in the main text, first on page 2 and second at the end of page 18 (Appendix A).

Thank you so much for your detailed advice! We have deleted the abbreviation of RBV in the abstract, and the definition of EMS in Appendix A. Besides, we have checked the manuscript thoroughly to make it more comprehensible.

 

Comments on the Quality of English Language

Improvable points still exist.

We have checked the manuscript again to improve our English writing and make it easier to understand.

Author Response File: Author Response.pdf