Information Acquisition and Green Technology Adoption Among Chinese Farmers: Mediation by Perceived Usefulness and Moderation by Digital Skills

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article employs empirical research to elucidate how information acquisition facilitates the adoption of green production technologies by enhancing farmers' perceptions of their utility. Additionally, it affirms the pivotal role of digital skills in reinforcing this process. While the study presents several innovative aspects, there are notable areas requiring enhancement regarding structure, formatting, language, and symbol consistency:

1. The second section, "Theoretical Analysis and Research Hypothesis," contains content that overlaps with both the introduction and the materials and methods sections. It is advisable to consolidate this section into Parts 1 and 2 for improved clarity and coherence.
2. It is recommended to conduct a third-order interaction effect analysis within the model. Specifically, this analysis should assess the combined impact of information acquisition, digital skills, and social network embeddedness on technology adoption.
3. The article treats diverse green production technologies as a homogeneous category, neglecting the potential systematic effects that specific attributes of these technologies (such as complexity, compatibility, cost, and the timeline for realizing benefits) may have on the research model. This issue should be addressed in the discussion section to elucidate the conditions under which this approach is valid.
4. The discussion section should more explicitly engage with the literature gaps identified in the introduction. It is essential to clearly articulate how the findings of this study confirm, extend, or challenge existing research. For example, consider addressing how your results relate to Bambio et al. (2022) regarding information quality, Davis's (1989) work on perceived usefulness, and van Dijk's (2005) exploration of the digital divide. By establishing these connections, you will effectively highlight the contributions of your research to the existing body of knowledge.
5. The formatting is inadequate and requires significant enhancement. For instance, the statements in lines 309, 324, 326, and 364 should begin on separate lines. Furthermore, all formulas should be numbered for clarity. Additionally, the titles of the tables should only capitalize the first word, and there is insufficient spacing between adjacent tables.
6. There is a lack of standardization in the references, necessitating comprehensive improvements. For example, in line 667, the first reference does not include abbreviated author names. In line 705, the journal is missing the year of publication and does not adhere to the correct format. In line 750, there is inconsistency in font size. It is essential to carefully verify that all in-text citations and reference list entries conform to the target journal's style guide.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript investigates the multidimensional structure of information acquisition–deconstructed into channel diversity, content quality, and source credibility–and its impact on the adoption of green production technologies, based on micro–survey data from 574 grain–growing farmers in Hebei Province. Overall, the research addresses a highly relevant topic aligned with the global trend of agricultural green transformation and the “Dual Carbon” goals, offering contemporary significance and policy relevance. The empirical approach is methodologically sound, the data analysis techniques are transparent, and the study makes a valuable contribution to understanding the micro-mechanisms of agricultural technology diffusion. The policy recommendations provided are well-targeted. Therefore, I recommend acceptance of this manuscript pending minor revisions.

Specific revisions are suggested as follows: (1) In the introduction, the global context of the agricultural green transition should be briefly supplemented to strengthen the international perspective of the research foundation. (2) The research objectives should more explicitly articulate the theoretical contributions of this study. (3) The methods section needs to include a clear statement regarding the official references or criteria used for selecting the sample counties, to enhance the persuasiveness of the sample’s representativeness and credibility. (4) In the discussion, a summarizing statement on the significance of the mediating effect of perceived usefulness should be added to more effectively frame the subsequent analysis. (5) When mentioning “two-way targeted” interventions in the policy recommendations, briefly illustrating with an example of a potential policy mix would make this suggestion more concrete. (6) A thorough check for consistency in reference formatting throughout the manuscript is necessary, ensuring uniformity in citation style, especially between Chinese and international references. (7) The conclusions section should include a concise summary emphasizing the value of the integrated “information acquisition–perceived usefulness–technology adoption” theoretical framework, thereby reinforcing the scholarly contribution of the findings. These adjustments will further enhance the paper’s rigor and readability.

 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

The manuscript, titled "Impact of Information Acquisition on Farmers' Adoption of Green Production Technologies: The Mediating Role of Perceived Usefulness and the Moderating Role of Digital Skills," which you submitted to Sustainability, addresses a topic of significant current and scientific interest: the adoption of green technologies in agriculture. This topic is linked to the ecological transition and global climate goals, and thus contributes to the debate on transitions to sustainable agricultural practices in contexts characterized by rapid digitalization. Reading the manuscript, I identified several strengths, which I summarize (briefly) as follows: a solid theoretical foundation that integrates the Technology Acceptance Model and Digital Divide Theory, providing an original and well-argued approach; a rigorous and replicable quantitative methodology, with a large sample (574 farms), an excellent response rate (95.7%), and rigorous data quality controls; a multidimensional analysis that, by breaking down information acquisition into three dimensions (channel diversity, content quality, and source credibility), represents an advance over previous studies that treated this variable unidimensionally; and the robustness of the results, which appear consistent with the hypotheses formulated. However, despite the highlighted merits, the manuscript presents some significant critical issues that limit its scientific impact. In particular, in my opinion, the aspects that deserve further attention include:

1. Causality issues: The use of cross-sectional data severely limits the ability to infer causal relationships. While you acknowledge this limitation, conclusions are often formulated in causal terms without adequate caution.
2. Unaddressed endogeneity: Strategies are not implemented to address potential endogeneity issues (instrumental variables, matching, etc.), which are particularly critical when analyzing relationships between digital skills and adoption behaviors.
3. Limited generalizability: The exclusive focus on Hebei Province (China) limits the generalizability of the findings to other geographic, cultural, and economic contexts.
4. Measurement limitations: Some key variables (e.g., digital skills, perceived usefulness) are based exclusively on self-reports, potentially affected by social desirability bias.

Furthermore, in my opinion, the manuscript is very long and dense: some passages could be streamlined and made more concise. The discussion, while sound, could be enriched with greater comparison with international studies beyond the Chinese context. Finally, although some recent bibliographic references are present, it would be helpful to include additional European and North American contributions to increase the manuscript's international appeal.

Below are some suggestions, differentiated by section of the manuscript, which I hope will help you identify corrections/additions that need to be made to the manuscript before it can be published.

Review the editing of the entire manuscript in accordance with the Sustainability guidelines.

Review punctuation and spaces.

Tables and figures should be separated from the text by a blank line (both before and after).

Check/correct the dates (it should probably be 2024 instead of 2025). On line 197, it says "July-August 2025," but we are only in late September/early October 2024. This could raise questions about: When was the research actually conducted? Was the paper prepared in advance with fictitious dates? Are there other inaccurate data?

Number the equations (on the right side) and leave them on their own line (the text should start on the next line).

Title: The title is descriptive and clearly communicates the main variables and framework of the study. However, it is excessively long (24 words) and does not mention the specific geographical context (China/Hebei). I suggest considering simplifying it to make it more concise. For example, "Information Acquisition, Digital Skills, and Green Technology Adoption among Farmers: Evidence from China," or "Information Acquisition and Green Technology Adoption among Chinese Farmers: Mediation by Perceived Usefulness and Moderation by Digital Skills."

Abstract: Although it appears well-structured, highlighting objectives, methodology, results, and implications, it is very dense, does not clearly specify the geographical context until the end, lacks an explicit statement on the cross-sectional nature of the data and its causal limitations, and the final recommendations are generic and could be more specific. I would suggest a stylistic revision to make it more concise and direct, improving readability (simplify sentences and more clearly highlight novel elements and specific contributions). Furthermore, I would like to offer these additional suggestions: begin by specifying: "Based on survey data from 574 grain farmers in Hebei Province, China..."; add the caveat: "Cross-sectional results suggest that..." instead of direct causal statements; and finally, make the recommendations more actionable.

Keywords: These could be simplified by avoiding repeating words already included in the article title.

Introduction: The introduction effectively contextualizes the problem with updated statistical data (FAO 2023, Chinese data 2015-present) and highlights the gap in the existing literature. Furthermore, it presents a good logical progression from the general (global problem) to the specific (Chinese context) and effectively articulates the study's theoretical and empirical contributions. In my opinion, however, it could be streamlined (it is currently excessively long and contains text that could easily be moved to the literature review section), and some concepts (e.g., digital divide, perceived usefulness) are repeated without adding new information (I suggest eliminating these redundant sections). Line 33: "Because of multiple pressures" should be "Due to multiple pressures," I believe. Line 55 states that insufficient information is "one key factor" but is not sufficiently supported by preliminary empirical evidence. Furthermore, a short paragraph describing the paper's structure should be added at the end of the introduction.

Literature: Integrate more international references (Europe, North America, Africa) to strengthen the comparative dimension.

Theoretical Analysis and Research Hypothesis: Provides an accurate and well-structured theoretical framework with clear derivations of the hypotheses, consistent with the literature. There is good integration between the Technology Acceptance Model and the Digital Divide Theory; the six hypotheses are logically derived and testable, and the conceptual diagram (Figure 1) is clear and effective. However, as already mentioned, some citations focus on Chinese studies; it would be useful to integrate contributions from other geographical areas to provide broader comparative breadth. Furthermore:

1. Underdeveloped Hypotheses H5 and H6: The hypotheses relating to policy perception and social networks are introduced more briefly than the others but are not adequately tested in the empirical results;
2. Lack of alternative hypotheses: Possible alternative relationships or competition between theoretical models are not discussed;
3. Premature operational definitions: Some operational definitions of the variables are anticipated here but belong in the Methods section.

I therefore suggest further developing the theory underlying H5 and H6, or considering removing them if they are not central to the analysis; adding a discussion of possible nonlinear or conditional effects; and finally, separating theory and operationalization more clearly. Line 119: "Trialability" is a technical term that could be better explained.

Materials and Methods: This section provides a detailed and transparent description of the sampling process, an excellent stratification strategy based on the Digital Village Index, and good quality control measures (telephone verification of 15% of the sample, 95.2% consistency). Furthermore, it presents a clear description of the variables with a well-structured Table 1 and, finally, a rigorous mathematical specification of the statistical models that makes the analysis methodology easily replicable. However, the text is very technical and at times verbose, with several critical issues:

1. Validity of the measures:
• The construct validity of the scales (e.g., Cronbach's alpha, factor analysis) is not discussed;
• The digital skills scale is self-reported without objective validation;
• Perceived usefulness measures future expectations, not actual experience.
2. Missing data handling: The management of missing data beyond the exclusion of invalid questionnaires is not discussed.
3. Brant test: Only briefly mentioned in the Results section (line 381) but not described in the Methods.
4. Sample size justification: No a priori power analysis is provided to justify n=574.
5. Errors and inaccuracies:
• "mu" (Chinese unit)
• Bootstrap with 1,000 iterations is standard but could be increased to 5,000-10,000 for greater stability.

First, I would suggest including a figure/diagram summarizing the variables and models for easier reading. Next:

• Add a subsection on the validation of measurement scales; • Include analysis of the distribution of missing data; • Move the Brant test description to the Methods; • Provide power analysis or at least justification for the sample size; • Since "mu" is a Chinese unit of measurement (explained in the footnote), it could create difficulties for international readers, so it would be appropriate to always also report the equivalent in hectares (or use hectares directly as the unit of measurement); • Bootstrap with 1,000 iterations is standard but could be increased to 5,000-10,000 for greater stability (consider whether or not to do so).

Estimation Results and Analysis: In general, these provide a systematic, well-structured, and organized presentation of the results, accompanied by clear and informative tables with appropriate statistical tests and precise interpretations. Furthermore, in my opinion, the progressive analysis from main effects to mediation, moderation, and heterogeneity also appears good; there is good use of descriptive statistics (OR values, confidence intervals); and the robustness checks appear appropriate. However, some tables are too numerous and detailed; I would suggest moving some (robustness and sub-analyses) to the appendix. Other critical issues identified include:

1. Excessive causal interpretation:
• Repetitive causal language ("promotes," "enhances," "drives") not justified by cross-sectional design;
• Lines 389-391: "increases the probability" implies causality but is only association.
2. Effects of H5 and H6 not reported: Are the hypotheses on policy perception and social network embeddedness tested? The results are not presented in the tables.
3. Lack of diagnostics:
• VIFs for multicollinearity are not reported;
• No discussion of outliers or influential cases;
• Residual plots or goodness-of-fit beyond Pseudo R² are missing.
4. Tables 3 and 4: The breakdown of indirect effects by PU size (Table 4) is interesting but:
• The statistical method used is unclear;
• "Explained Variance Ratio" is not defined;
• There may be problems with Type I error inflation from multiple testing.
5. Heterogeneity analysis:
• The dichotomization of digital skills at the median is arbitrary;
• The Chow test is applied to nonlinear models without discussing the assumptions;
• Plain vs. mountainous areas: differences in socioeconomic characteristics are missing.
6. Limited robustness checks:
• Only the model type and variables change; endogeneity is not addressed;
• Sensitivity to outliers or different model specifications is not tested.

In relation to this, I suggest:

• Reformulate everything in terms of associations, not causality; • Add results for H5 and H6 or remove these hypotheses; • Include complete diagnostics in the appendix; • Clarify decomposition methods in Table 4 and apply correction for multiple testing; • Justify the cutoff for digital skills or use continuous analysis; • Add robustness checks for endogeneity (e.g., instrumental variables, Heckman correction); • Provide comparative descriptives for groups in heterogeneity analysis.

Discussion: This section provides an appropriate summary of the main results consistent with the hypotheses. An attempt is made to connect the findings to the existing literature, explicitly acknowledging three limitations of the study and reflecting on the theoretical significance of the results. The gaps identified in this section concern:

1. Repetitiveness: Much of the content repeats what was already stated in the Results without adding new interpretations.
2. Underestimated Limitations:
• Acknowledged limitations are treated superficially;
• Causality and endogeneity issues are downplayed;
• Not mentioned: self-report bias, social desirability, and common method variance issues.
3. Lack of Critical Comparison:
• Little discussion of results that contradict or confirm specific previous studies;
• Discrepancies with international literature are not discussed.
4. Unexplored Causal Mechanisms:
• It is stated that "yield confidence" is higher for channel diversity, but the reason for this is not explained;
• Deep psychological or behavioral explanations are lacking.
5. General practical implications: The policy implications are vague and not sufficiently rooted in the specific findings.

In this case, my suggestions are:

• Strengthen the comparison with international studies on the role of digital skills and technology transfer; • Reduce repetition and focus on novel interpretations; • Significantly expand the limitations section, including:

• Common method bias and how it might affect the results;
• Selection bias in sampling;
• External validity concerns;
• Self-reported measures and their biases. • Add a subsection dedicated to systematic comparisons with previous studies; • Develop theoretical causal mechanisms more deeply; • Make practical implications more specific and actionable, with concrete examples (if possible); • Line 592: "relational society" needs further explanation for international readers.

Conclusions and Recommendations: Overall, these provide a clear and appropriate summary of the main contributions and appear well-connected to results and policy implications (there is a good explicit link between findings and policy implications). They provide five detailed policy recommendations. Overall, in my opinion, they are slightly wordy; it would be helpful to more clearly distinguish between theoretical, practical, and policy implications. Specifically, the critical issues concern:

1. Conclusions are too definitive: Given the cross-sectional nature, the conclusions are formulated too assertively (e.g., "significantly promote," "enhance");
2. Generic recommendations:
• The five policy recommendations are reasonable but not sufficiently specific;
• There is a lack of guidance on priorities, sequencing, costs, and implementation responsibilities;
• They do not indicate how to overcome identified barriers.
3. Lack of future research: No specific research agenda is proposed to overcome the identified limitations.
4. No reflection on trade-offs: The recommendations do not discuss possible conflicts between objectives or resource constraints.
5. Generalizability not discussed: There is no mention of how the recommendations could/should be adapted to other contexts.

In my opinion, the need for longitudinal studies (panel data, field experiments) and a deeper understanding of the role of social networks could be emphasized. Therefore, as a future perspective, I would strongly recommend the use of panel data or field experiments for future research. Furthermore:

• Temper conclusions with appropriate caveats (e.g., "Results suggest...," "Associations indicate..."); • Make SMART recommendations (Specific, Measurable, Achievable, Relevant, Time-bound):

• Instead of "enhance information content quality" → "Develop standardized video tutorials in local dialects demonstrating each technology with measurable outcomes";
• Instead of "digital skills training" → "Implement 3-month pilot programs targeting farmers aged 50+ with evaluation metrics"; • Add a "Future Research Directions" section with 3-4 specific directions; • Discuss trade-offs (e.g., training costs vs. expected benefits, prioritization among target groups); • Add a section on how to adapt findings to other contexts.

References: Although relatively up-to-date (includes many sources from 2022-2025) and with good literature coverage, it is only partially adequate. The identified gaps concern:

1. Dominance of Chinese contextual literature: Many sources (about 15-20) are specific studies on the Chinese context, but there is insufficient balance with classical international theoretical frameworks.
2. Theoretical gaps:
• Seminal references on the Technology Acceptance Model are missing (Davis 1989 is cited in the text but not in the bibliography);
• Digital Divide Theory: van Dijk is missing (cited in discussion but not in the reference list);
• Rogers' Diffusion of Innovations (cited but not in the bibliography).
3. Formatting inconsistencies:
• Inconsistent citation styles (some with DOI, others with URL);
• Author names: some with initials, others without;
• Journal titles: inconsistent abbreviations.
4. Gray sources: Lack of transparency on cited government reports (e.g., China Agricultural and Rural Informatization Development Report 2024, Hebei Province Digital Village Development Report 2024).

I suggest:

• Add all theoretical references cited in the text to the bibliography; • Completely standardize formatting according to Sustainability guidelines; • For government reports, provide permanent URLs or DOIs when available; • Consider adding additional references on:

• Methodologies for managing endogeneity in cross-sectional studies;
• Common method bias;
• Validation of psychometric scales.

Comments on the Quality of English Language

Although the manuscript is written in generally good English, it contains several grammatical and stylistic errors that compromise the readability and professional quality of the publication. Additionally, some redundant or imprecise expressions require revision.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author has completed the revision.

Author Response

please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for your revisions. I agree to publish this manuscript.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

I acknowledge that in this second version of the paper, you have made a significant effort to address the requests from the first round of review, implementing more than half of the suggested changes. The manuscript has undergone substantial improvements in key areas such as model diagnostics, descriptive statistics, handling of missing data, and policy recommendations. However, in my view, several important issues remain that must be addressed before the paper can be considered for publication—particularly concerning date inconsistencies, excessive causal language, and the omission of key theoretical references.

Specifically, I would like to highlight the following points:

1. Excessive causal language: Despite the explicit request, the manuscript continues to use causal language that is inappropriate for cross-sectional data. Problematic examples include:

• Line 15: "significantly promote technology adoption"
• Lines 383–384: "increases the probability"
• Line 421: "high-quality information can most effectively translate into"
• Line 460: "digital skills significantly enhance"
• Line 591: "significantly and positively correlated" (this is correct, but an isolated case)

I recommend a systematic revision of the manuscript to replace:

• "promotes/enhances/drives/increases" → "is associated with" / "is correlated with"
• "influences" → "is related to"
• "Results indicate that" → "Results suggest that" / "Cross-sectional associations indicate that"

Specific suggestion: Add a methodological disclaimer at the beginning of the Results section (e.g., "Given the cross-sectional nature of our data, all reported effects represent associations rather than causal relationships.")

1. Missing key theoretical references: These are cited in the text but still not included in the bibliography:

• Davis, F.D. (1989) – Technology Acceptance Model (cited in lines 22, 245, 558)
• van Dijk, J.A.G.M. (2005) – The Deepening Divide: Inequality in the Information Society (cited in the Discussion)
• Rogers, E.M. – Diffusion of Innovations (cited in lines 118, 386)

Please ensure that these references are properly added to the bibliography.

1. Psychometric validation of the scales: The manuscript does not discuss the validity and reliability of the measurement scales. I suggest including a brief discussion in the Methods section (or Appendix) addressing:

• Cronbach's alpha for perceived usefulness and digital skills
• Confirmatory factor analysis (if applicable)
• Convergent and discriminant validity

If these tests have not been conducted, please acknowledge this as a limitation in the Discussion section.

More generally, I would also note the following issues:

• The abstract does not clearly specify the cross-sectional nature of the study or its implications;
• The introduction does not conclude with a paragraph outlining the structure of the paper;
• The Brant test is discussed in the Results section but should be described in the Methods;
• Some tables are not clearly separated from the text by blank lines before or after them;
• While endogeneity is acknowledged as a limitation (lines 543–544), it is not addressed empirically.

Comments on the Quality of English Language

Although the manuscript is written in generally good English, it contains several grammatical and stylistic errors that compromise the readability and professional quality of the publication. Additionally, some redundant or imprecise expressions require revision.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf