Between Commitment and Practice—Sustainability Attitudes and Behaviors in Spain—A Mixed-Methods Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Required improvements:

• Abstract: Make clearer the specific research gap and theoretical contribution of the study and explain briefly why the Spanish context and mixed-methods design add novelty.
• Introduction: Make the extensive literature review simpler by highlighting the gap that remains unresolved. Ensure that the study's original contribution to sustainability behavior, credibility, and feasibility debates are expressed clearly early on in the section.
• Methodology: To increase clarity, it's important to explicitly justify the mixed-methods sequencing, distinguish exploratory versus explanatory.
• Results: Try to tighten narrative explanations around figures and tables and highlight key effect sizes and patterns instead of repeating all descriptive statistics.
• Discussion and Contribution: Make the discussion more effective by directly linking findings and explaining how the study enhances knowledge about feasibility and credibility for sustainability.
• Future Research: Expand the future research and make clearer pathways for policy and intervention research can be achieved through behavioral or purchasing data, longitudinal designs, or cross-country comparisons.
• Tables and Figures: Enhance tables by minimizing visual opacity, ensuring consistent labeling and scale interpretation, and adding brief interpretive notes to emphasize why every table is theoretically relevant instead of just descriptive.

In general, the paper has a well-organized structure.  

Author Response

Thank you for your careful reading and constructive feedback. We revised the manuscript substantially to improve structure, clarify the research gap and contribution, strengthen theoretical grounding in pro-environmental behavior, and increase mixed-methods transparency. Key revisions include: (i) restructuring the Introduction into thematic subsections with appropriate subheadings; (ii) making research questions and hypotheses explicit; (iii) clarifying the sequential mixed-methods rationale and the differing sampling logic of the qualitative and quantitative phases; (iv) improving figure readability by removing slanted x-axis labels; (v) removing redundant presentation of sample structure (keeping Table 2 as the sole summary and removing the redundant age-distribution figure); (vi) tightening terminology to avoid implying latent constructs without factor analysis; and (vii) strengthening mixed-methods integration through a new triangulation table (Table 4).

Point-by-point response to Reviewer 1:

1) Abstract: Clarify the research gap and theoretical contribution; explain novelty (Spain + mixed-methods).

Response: We rewrote the Abstract to foreground the endorsement–enactment gap and to frame feasibility and credibility as conditioning mechanisms, highlighting novelty of the Spanish context and sequential mixed-methods design (p. 1, l. 6–18).

2) Introduction: Simplify and highlight contribution early; clarify unresolved gap and contribution.

Response: We reorganized the Introduction into thematic subsections with explicit transitions and clearer progression from literature to gap to aims (p. 2, l. 16 to p. 8, l. 23). We also made research questions and hypotheses explicit at the end of the Introduction (p. 8, l. 23–40).

3) Methodology: Justify mixed-method sequencing; clarify exploratory vs explanatory roles.

Response: We clarified the sequential design logic and integration rationale in Section 2.1 (p. 9, l. 15–30) and added an explicit rationale for differing target groups between phases in Section 2.2.2 (p. 10, l. 7–16).

4) Results/Discussion: Tighten narrative and strengthen integration between qualitative and quantitative evidence.

Response: We strengthened integration by adding Table 4 (Mixed-methods triangulation) mapping focus-group themes to the corresponding survey indicators and indicating where each pattern is reported (p. 30, l. 5–30).

5) Tables/Figures: Improve clarity and consistency.

Response: We improved figure readability by removing slanted x-axis labels (e.g., Figure 2 p. 18, l. 7–12; Figure 7 p. 27, l. 12–18). We also removed redundant presentation of sample structure by retaining Table 2 as the sole summary (p. 16, l. 28–40).

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for the opportunity to review the manuscript. The study addresses an important issue concerning the gap between sustainability attitudes and behaviours among consumers. The topic is interesting; however, the manuscript requires further development, as outlined in the comments below.

Good luck with the revision.

Reviewer

 

General comments:

1/ The objective of the paper should be presented more precisely. At present, four research questions are stated, but the research problem and the research gap are not clearly defined.

2/ The literature review is currently included in the introduction. It is recommended to create a separate section for the literature review and to ground the research problem and research questions more firmly in the existing literature.

3/ The analytical part is relatively simple. The analysis mainly presents means and frequencies, and the only more advanced method applied is a simple linear model with a low R-squared value. It would be advisable to consider presenting more advanced analytical results.

 

Specific comments and suggestions

1/ Abstract

Suggested change/comment: please present the research gap and the research objective in the abstract, as these components are currently missing.

 

2/ Line 126

There is: “ESD research argues…”

Suggested change/comment: ESD acronym should be explained

 

3/ Line 281-288

There is: 2.2.3. Procedure and Moderation

Suggested change/comment: please indicate when FGI was conducted

 

4/ Line 315-319

There is: 2.3.2. Survey Administration and Platform

Suggested change/comment: please indicate when (for how long) survey was conducted

 

5/ Line 247-260

There is: 2.1. Research Design and Rationale

Suggested change/comment: in the qualitative part, the target audience (TG) consists of students, whereas in the quantitative part the survey is conducted among a broader group of consumers (mean age approximately 42 years). There is no explanation provided for why the target groups differ between the qualitative and quantitative parts. Please add your comments.

 

6/ Line 363-368

There is: 2.5. Quantitative Analysis Strategy; “For selected 0–10 items, we additionally report an NPS-style descriptive grouping to facilitate interpretation: detractors (0–6), passives (7–8), and promoters (9–10). This categorization is used as a heuristic descriptive summary and is not interpreted as an official Net Promoter Score”.

Suggested change/comment: this analytical decision and strategy should be explained and justified more carefully. Is there any example in the literature of such approach, what are the benefits?

Author Response

Thank you for your careful reading and constructive feedback. We revised the manuscript substantially to improve structure, clarify the research gap and contribution, strengthen theoretical grounding in pro-environmental behavior, and increase mixed-methods transparency. Key revisions include: (i) restructuring the Introduction into thematic subsections with appropriate subheadings; (ii) making research questions and hypotheses explicit; (iii) clarifying the sequential mixed-methods rationale and the differing sampling logic of the qualitative and quantitative phases; (iv) improving figure readability by removing slanted x-axis labels; (v) removing redundant presentation of sample structure (keeping Table 2 as the sole summary and removing the redundant age-distribution figure); (vi) tightening terminology to avoid implying latent constructs without factor analysis; and (vii) strengthening mixed-methods integration through a new triangulation table (Table 4).

Point-by-point response to Reviewer 2:

General comment 1: Objective and gap are not clearly presented.

Response: We clarified the research problem/gap and formalized the contribution through explicit research questions and hypotheses in Section 1.9 (p. 8, l. 23–40).

General comment 2: Separate literature review from the introduction / improve structure.

Response: To address the monolithic structure, we reorganized the Introduction into thematic subsections with subheadings (Sections 1.1–1.9) (p. 2, l. 16 to p. 8, l. 23).

General comment 3: Quantitative analysis is simple; consider more advanced analyses.

Response: We agree that correlation screening and multiple regression would strengthen inference about simultaneous predictors. However, the response-level dataset could not be retrieved, preventing additional modeling during revision. We therefore (i) frame associations as exploratory, (ii) strengthen integration via Table 4 (p. 30, l. 5–30), and (iii) explicitly state this limitation and prioritize multivariate modeling in future replications (Section 4.5, p. 32, l. 31–45).

Specific comment: Explain ESD acronym.

Response: We expanded the acronym to “Education for Sustainable Development (ESD)” in the Introduction (p. 2, l. 16 to p. 8, l. 23).

Specific comment: Indicate when the focus group was conducted and the survey fieldwork period.

Response: We included fieldwork timing information for both the focus group (Section 2.2.3, p. 10, l. 20–30) and the survey (Section 2.3.2, p. 11, l. 25–33).

Specific comment: Explain differing target groups (students vs broader consumers).

Response: We added a clear rationale for the different target groups in Section 2.2.2 (p. 10, l. 7–16).

Specific comment: Justify NPS-style grouping.

Response: We clarified that the detractor/passive/promoter grouping is a descriptive heuristic for 0–10 distributions (not an official NPS and not inferential) in Section 2.5 (p. 14, l. 14–20).

Reviewer 3 Report

Comments and Suggestions for Authors

In this paper authors in the introduction section authors mainly focused on sustainable development targets, but this paper is related to the pro-environmental behaviour. Thus, I missed the psychology part and about main motives. For me it is not clear to what background authors referred performing their analysis. For me also was not clear how the questionnaire was constructed. The analysis of data also is very poor. Authors presented only the descriptive analysis, but the deep statistical analysis was not performed. Thus, it is only the presented raw results without deep analysis searching the motives of pro-environmental behaviour. 

 

Author Response

Comment: Psychology/motives are missing; clarify theoretical background; questionnaire construction unclear; statistical analysis too descriptive.

Response: We strengthened the pro-environmental behavior framing by integrating consumer-psychology mechanisms and explicit links to TPB/VBN (Sections 1.3–1.5 within the restructured Introduction, p. 2, l. 16 to p. 8, l. 23). We clarified questionnaire development and pretesting within the Methods (Section 2.3, including instrument-development details, around p. 11, l. 25–33). We also strengthened mixed-methods integration via Table 4 (p. 30, l. 5–30). Regarding deeper statistical modeling, we explain the dataset constraint and specify multivariate modeling as a priority for future work (Section 4.5, p. 32, l. 31–45).

Reviewer 4 Report

Comments and Suggestions for Authors

The article submitted for review addresses a very important topic. However, it does not meet the requirements of a scientific article. Therefore, it must be revised before publication.

1. The introduction is a very monolithic text. Please divide it into subsections with appropriate subheadings. The introduction should also include research questions.
2. Construct is a concept derived from factor or confirmatory analysis, which the authors do not use. The use of the word "construct" in Table 1 is therefore misleading to the reader.
3. Sometimes the authors repeat some information (e.g. they present the same information in Figure 1 and Table 2)
4. Results:
   1. Figure 3 is unsightly (slanted labels on the OX axis).
   2. The analysis of the results is very simple and based solely on descriptive statistics. I believe the authors should have examined the relationships between the variables studied using statistical tests.
   3. Panel C. Focal associations (bivariate models): The authors should have performed a multiple regression analysis instead of a simple regression analysis. The R2 coefficients presented in the table indicate that the authors obtained very weak models. A multiple regression analysis could have yielded better results. It would also show significant and nonsignificant predictors. However, before performing regression analysis, authors should also conduct correlation analysis.
   4. Social and labor dimensions: A dimension is a complex statistical construct that authors do not calculate. Therefore, they should not use the word "dimension"
   5. Summary of results: Such a short summary of the research looks bad in an article. I believe the authors should abandon it.
5. In conclusion, the authors should recall the research questions posed in the introduction and provide the answers to these questions.

Author Response

1) Introduction is monolithic; divide into subsections; include research questions.

Response: We reorganized the Introduction into subsections with subheadings (Sections 1.1–1.9, p. 2, l. 16 to p. 8, l. 23) and included explicit research questions and hypotheses in Section 1.9 (p. 8, l. 23–40).

2) “Construct” is misleading without factor/confirmatory analysis.

Response: We replaced “Construct” with “Measure” and adjusted related wording to avoid implying latent-variable modeling (Table 1, p. 12, l. 28–45).

3) Repetition (Figure 1 vs Table 2 show the same information).

Response: We removed the redundant age-distribution figure and retained Table 2 as the sole summary of sample characteristics (Table 2, p. 16, l. 28–40). Figure 1 is now used for sustainability awareness/exposure results (p. 17, l. 29–40).

4) Figure is unsightly due to slanted x-axis labels.

Response: We reformatted the relevant figures to avoid slanted labels and improve readability (e.g., Figure 2 p. 18, l. 7–12; Figure 7 p. 27, l. 12–18).

5) Analysis too simple; add correlation screening and multiple regression; report significant/non-significant predictors.

Response: We agree these analyses would strengthen inference. Because the response-level dataset could not be retrieved, we cannot estimate additional models during revision. We therefore (i) avoid claims requiring multivariate estimation, (ii) strengthen interpretive rigor via Table 4 (p. 30, l. 5–30), and (iii) explicitly state the limitation and prioritize correlation/MLR for future replications (Section 4.5, p. 32, l. 31–45).

6) Do not use “dimension” without statistical computation.

Response: We replaced “dimension(s)” with “domain(s)” where applicable to avoid implying computed latent dimensions (Measures/Results terminology; see Table 1, p. 12, l. 28–45).

7) Summary of results is too short; abandon it.

Response: We reframed this section as a synthesis and strengthened integration through Table 4 (p. 30, l. 5–30).

8) Conclusions should recall and answer the research questions.

Response: We expanded the Conclusions to explicitly answer each research question (Section 5, p. 34, l. 33–45).

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

My comments are the same as in previous version. This paper is not serious scientific paper, it looks like decsriptive analysis with wrong statistical analysis.

Author Response

Dear Reviewer 3,

Thank you for your careful reading of our manuscript and for reiterating your concerns about (i) the theoretical background and psychological mechanisms underlying pro-environmental behavior, (ii) transparency in survey construction, and (iii) the depth and appropriateness of the statistical analysis. We took these comments seriously and substantially revised the manuscript to address each point directly. Below we respond point-by-point and indicate where changes were implemented.

1) “The introduction focuses mainly on sustainable development targets, but the paper is related to pro-environmental behavior. I missed the psychology part and main motives.”

Response: We agree that a manuscript on sustainability enactment must be grounded in behavioral and psychological mechanisms (not only in SDG-oriented institutional framing). In the revised version, we reframed the introduction to explicitly situate the study in the pro-environmental behavior literature and consumer psychology, and we clarified the mechanisms that motivate or constrain enactment.

What we changed:

• We added a consumer-psychology framing of the attitude–behavior/commitment–practice gap early in the Introduction, explicitly discussing interacting determinants (values, norms, perceived behavioral control, habits, social norms/identity, trade-offs, and credibility/trust in claims).
• We added a dedicated theoretical framing section drawing on Theory of Planned Behavior (TPB) and Value–Belief–Norm (VBN) theory to clarify “motives” and mechanisms and to explain why favorable attitudes may not translate into action under constraints.
• To ensure the psychological framing is not only “in the introduction” but also explicitly tied to our qualitative evidence, we inserted a bridging sentence in Results 3.1 stating how the focus-group themes map onto TPB (perceived behavioral control/constraints) and VBN (normative endorsement vs contingent enactment).

Result: The paper now has a clear behavioral/psychological backbone: feasibility and perceived control, credibility/greenwashing skepticism, and normative endorsement as mechanisms conditioning enactment.

2) “It is not clear to what background authors referred performing their analysis.”

Response: We clarified the theoretical and conceptual background that guides the analysis (psychological mechanisms and the commitment–practice gap literature), and we made explicit how this background maps onto research questions and hypotheses.

What we changed:

• We articulated the research gap, research questions (RQ1–RQ5) and hypotheses (H1–H5) as derived from the TPB/VBN and feasibility/credibility literatures, making the analytic logic explicit rather than implicit.
• We further clarified the mixed-method integration rationale and how qualitative meaning-making informs the quantitative measurement strategy, while remaining cautious about generalization given non-probability sampling.

3) “It was not clear how the questionnaire was constructed.”

Response: We fully agree that questionnaire construction must be transparent and replicable. We therefore expanded the Methods section to clearly describe instrument development, thematic dimensions, item formats, and derived variables, and we provide the full instrument.

What we changed:

• We rewrote the quantitative methods section to explicitly describe how the survey instrument was developed from the qualitative phase, and we structured the instrument into five thematic dimensions aligned with the study’s research questions (salience/knowledge; endorsement–enactment gap; feasibility constraints; credibility interpretations; attitude/knowledge–behavior associations).
• We specified response formats and operationalization: 0–10 ratings, categorical items, practice checklist, practice-count derivation, willingness-to-change ordered item, and willingness-to-pay (0–100%) plus scaled 0–10 presentation for comparability in selected summaries.
• We clarified the conceptual knowledge item (Brundtland-type definition) and how it is coded and used (correct/incorrect indicator; not an exclusion criterion).
• We provide the full survey instrument and response options in Appendix A to support replication.

4) “The analysis is very poor / only descriptive; deep statistical analysis was not performed.”

Response: We expanded the quantitative analysis substantially beyond descriptive reporting and ensured that the statistical methods match the measurement level of our variables (ordinal 0–10 indicators; discrete practice count; ordered willingness-to-change outcome). We also reorganized Results to emphasize interpretable associations and mechanism-consistent patterns.

What we changed:

• Rank-based association analysis: We now report Spearman’s ρ for main attitudinal, behavioral, and perception indicators, consistent with ordinal measurement.
• Complete correlation matrix: We provide the full correlation matrix as Supplementary Table S2, and we explicitly cite it in the Results (Section 3.4), so it is not “hidden” or unanchored.
• Non-parametric group comparisons: We provide exploratory subgroup comparisons by gender and age group using appropriate non-parametric tests, reported as Supplementary Table S1 and referenced in the manuscript.
• Multivariable models: We added two complementary multivariable analyses:
• A Poisson count model for practice count (reporting IRR with robust 95% CI),
• An ordered logistic regression for willingness to change (reporting OR with 95% CI),
  both reported in Supplementary Table S3 and explicitly referenced in the Results (Sections 3.4 and 3.7).
• Within-respondent test for the social/labor gap: For the paired 0–10 items (worker wellbeing vs labor purchasing criterion), we added a Wilcoxon signed-rank test (paired, non-parametric) to show that the difference reflects a systematic within-respondent divergence rather than noise.

Result: The analysis is no longer “raw results only.” It includes bivariate associations, non-parametric group comparisons, and multivariable models that align with the measurement structure and the paper’s theoretical mechanisms.

5) “Wrong statistical analysis / not serious scientific paper.”

Response: We respectfully disagree with the characterization, but we recognize that earlier versions could be interpreted as relying too heavily on descriptive summaries and on conventions that might be confused with marketing metrics. In response, we made methodological choices more conservative and transparent, and we removed terminology that could be misleading.

What we changed:

• We removed NPS terminology and any “promoters/detractors” labels. Distributional summaries are now described neutrally as low/medium/high ranges and explicitly framed as descriptive heuristics (not inferential metrics).
• We ensured that the manuscript’s main association results are expressed consistently via Spearman’s ρ and ordinal-appropriate methods, and we anchored all supplementary analyses (S1–S3) to the relevant text sections (Methods and Results), so readers can verify the claims.
• We clarified the exploratory nature of inference given non-probability sampling and self-report measures, and we explicitly interpret models as associations rather than causal effects.

6) “The manuscript does not search motives of pro-environmental behavior.”

Response: We address motives/mechanisms through two complementary components:

1. Qualitative interpretive frames (conditional endorsement; feasibility trade-offs; credibility discounting) supported by translated verbatim quotations, and explicitly mapped onto TPB/VBN.
2. Quantitative indicators that operationalize these mechanisms (concern, self-perceived behavior, practice count, willingness to change, credibility framings, WTP, knowledge), examined via rank correlations and multivariable models.

Where this is reflected:

• Qualitative mechanism framing and quotations: Section 3.1.
• Quantitative association + multivariable evidence: Section 3.4 and Supplementary Tables S2–S3.
• WTP as normative indicator conditioned by feasibility/credibility, linked to S3: Section 3.7.

Closing

We appreciate the reviewer’s insistence on theoretical grounding, transparency, and appropriate statistical analysis. We believe the revised manuscript now provides: (i) a clear behavioral/psychological framework (TPB/VBN + credibility/feasibility mechanisms), (ii) full instrument transparency (Appendix A), and (iii) analyses that go beyond descriptives and are matched to variable measurement levels (Spearman correlations, non-parametric tests, and multivariable Poisson/ordered-logit models with results in Supplementary Tables S1–S3). We hope these revisions address the concerns and strengthen the manuscript’s scientific rigor.

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have revised the article according to the sugestions. 

Author Response

Thank you for your kind words. 

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

My comments are the same, I did not notice any improvement of the methodology, it looks like a descriptive analysis. 

Author Response

Dear Reviewer 3,

Thank you for your continued review. We are concerned that our second-round revisions may not have been visible in the version you consulted, because the methodological and analytical changes directly address your original points and go beyond descriptive reporting. Below we list the concrete changes with precise locations in the revised manuscript.

1) Theoretical background and “psychology part” (motives and mechanisms)

You noted that the paper should be grounded in pro-environmental behavior and psychological motives rather than primarily in SDG framing. In response, we substantially revised the Introduction and Results to make the behavioral framework explicit. Specifically:

• We added and clarified a behavioral/psychological framing of selective sustainability enactment and the attitude–behavior/value–action gap, linking it to established mechanisms (feasibility constraints, perceived control, habits, norms, and credibility/trust). See Section 1 (especially [Sections 1.3–1.6]).
• We explicitly anchor our interpretation in Theory of Planned Behavior (TPB) and Value–Belief–Norm (VBN) theory (perceived behavioral control/constraints; normative endorsement vs constrained enactment). See [Section 1.6 / Discussion 4.2–4.4].
• We added an explicit “bridge” statement in the qualitative results showing how the focus-group themes map onto TPB/VBN mechanisms. See Section 3.1.

2) Questionnaire construction and methodological transparency

You noted that the questionnaire construction and the background guiding the analysis were unclear. We therefore expanded the Methods and added full replication materials:

• We rewrote the survey methods to clearly describe instrument development, item formats, and derived variables (e.g., practice-count construction; knowledge coded correct/incorrect). See Section 2.3 (Survey instrument, measures, and derived variables).
• We provide the full survey instrument and response options for replication. See Appendix A.
• We also strengthened qualitative-method reporting (sampling, procedure, and analysis with Braun & Clarke thematic analysis). See Section 2.2.

3) “Descriptive analysis” / “wrong statistical analysis”

Your central concern was that the analysis was only descriptive and not statistically rigorous. We addressed this by (i) using methods appropriate to measurement level (ordinal 0–10 items, count outcomes, ordered outcomes), and (ii) adding multivariable modeling and reporting full results.

Key changes (all new in the revised manuscript):

• Rank-based associations (Spearman’s ρ) for ordinal 0–10 indicators and WTP, with the full matrix provided. See Section 2.5 and [Table S2 / Supplementary Table S2], referenced in Section 3.4.
• Non-parametric subgroup comparisons (gender and age group), reported transparently. See Section 2.5 and [Table S1 / Supplementary Table S1], referenced in Section 3.8.
• Multivariable models that go beyond descriptives:
• A Poisson count model for practice count (reporting IRR with 95% CI; robust standard errors; negative binomial as sensitivity if needed),
• An ordered logistic regression for willingness to change (reporting OR with 95% CI).
  These results are reported in [Table 4 in the main text] (full outputs previously provided as S3; now integrated into the manuscript body). See Section 2.5 (analysis strategy) and Section 3.4 / 3.7 where the models are cited.
• We also added an appropriate paired non-parametric test (Wilcoxon signed-rank) for the worker-wellbeing vs labor-purchasing divergence to demonstrate a systematic within-respondent shift rather than a simple difference in means. See Section 3.5.

Collectively, these additions mean the manuscript now includes inferential analysis at three levels: rank-based bivariate associations, non-parametric subgroup tests, and multivariable regression models appropriate to outcome distributions. The analysis is therefore not limited to descriptive reporting.

If, despite these changes, you still consider the methodology insufficient, we would be grateful if you could specify which specific methodological component remains missing (e.g., measurement design, model choice, robustness checks, reporting standards), so that we can address it concretely.

Sincerely,
The Authors