Review Reports - Leaderships That Emerge: Profile and Performance of Promoters in Farmer-to-Farmer Processes in Casanare, Colombia

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper explores the initial performance of using the CaC model to train rural extension agents in the province of Casanare, Colombia, a less organized region, and the topic selection has important practical and theoretical value. However, there are still some deficiencies in the following aspects:
1. Insufficient theoretical dialogue. The introduction combs the core principles of the CaC model, but does not clarify the differentiated contributions of the paper and existing research. Existing research has focused on the effectiveness of the CaC model (such as Cuba and Central America), and the paper focuses on the characteristics of promoters in areas with weak organization, but does not highlight the theoretical expansion in this scenario. Core concepts such as promoter performance and cultural belonging are only operationalized through evaluation indicators without explanation combined with relevant theories, resulting in insufficient theoretical depth of research.
2. The sample size is too small, and the representativeness and robustness of the conclusions are insufficient. Only eight promoters were selected as the research objects, which makes it difficult to support conclusions such as no association between promoter type division (three clusters) and socio-demographic variables, and there may be a risk that extreme cases dominate the results. Multiple correspondence analysis (MCA) and K-means clustering are based on small samples, and the statistical robustness of the results is questionable. At the same time, in a close community environment, participants' evaluations, especially those of their fellow villagers, may be more positive, which may be one reason for the small variance of community evaluations.
3 The study does not address the core issues raised. The article proposes to address how CaC approaches can start in regions that lack organizational foundations. The introduction makes it clear that Casanare province lacks the ripe conditions for the CaC model (no organized network, no historical tradition), which is precisely the value of the study. However, the results section only reports "promoters perform well" and "split into three categories", and the discussion section only explains "why different evaluators give different scores".
4. The analysis of results is not in-depth enough. The paper mentions that 87.5% are male promoters, but this is only attributed to structural and symbolic barriers, without further analysis: is there a sampling bias in which female promoters are omitted? Are there implicit differences between the performance characteristics of existing female promoters and those of men? This paper points out that the promoters are mainly middle-aged and elderly people, and the participation of young people is insufficient, but it does not analyze the specific impact of the absence of youth on the sustainability of CaC model.

Author Response

We thank you for your detailed and constructive comments on ‘Leadership that Emerges: Profile and Performance of Promoters in Farmer-to-Farmer Processes in Casanare, Colombia’. All comments were carefully considered and addressed in the revised manuscript.

The revision involved four main areas of change: (1) reframing the study as a descriptive case study, replacing inferential analyses with non-parametric descriptive statistics; (2) strengthening the theoretical framing and research questions in the introduction; (3) enriching the promoter characterization with the IPPTA index; and (4) updating the reference list and expanding the conclusions with practical implications and a dedicated limitations section. Detailed responses to each comment, with references to the relevant lines and sections of the revised manuscript, are provided in the point-by-point document.

As the manuscript was submitted for major revision, we were unable to use track changes in the revised file. Instead, each point-by-point response references the specific lines, paragraphs, and/or sections of the revised manuscript where the corresponding changes can be found. Also, text with main changes are highlighted in the adjusted manuscript.

Comment 1:

Insufficient theoretical dialogue. The introduction combs the core principles of the CaC model, but does not clarify the differentiated contributions of the paper and existing research. Existing research has focused on the effectiveness of the CaC model (such as Cuba and Central America), and the paper focuses on the characteristics of promoters in areas with weak organization, but does not highlight the theoretical expansion in this scenario. Core concepts such as promoter performance and cultural belonging are only operationalized through evaluation indicators without explanation combined with relevant theories, resulting in insufficient theoretical depth of research.

Response 1:

We agree with this observation. Much of the original introduction described CaC methodology principles without positioning the study relative to the existing literature or clarifying its theoretical contribution. To address this, we made two additions to the introduction: (1) we incorporated evidence from studies conducted in diverse global regions documenting FtF effectiveness across different agricultural and organizational contexts [lines 33-37], and noted how these contexts differ from the nascent conditions found in southern Casanare [lines 37-41]; and (2) we grounded the study in social learning theory as an analytical framework, clarifying that while farmer promoters are not equivalent to conventional extension agents, dimensions of their knowledge-exchange performance can be conceptualized and assessed through this theoretical ground using the proposed matrix [lines 71-81].

Comment 2:

The sample size is too small, and the representativeness and robustness of the conclusions are insufficient. Only eight promoters were selected as the research objects, which makes it difficult to support conclusions such as no association between promoter type division (three clusters) and socio-demographic variables, and there may be a risk that extreme cases dominate the results. Multiple correspondence analysis (MCA) and K-means clustering are based on small samples, and the statistical robustness of the results is questionable. At the same time, in a close community environment, participants' evaluations, especially those of their fellow villagers, may be more positive, which may be one reason for the small variance of community evaluations.

Response 2:

We agree. The original manuscript made inferential claims not warranted by the sample size, and the use of MCA and k-means clustering on eight cases lacked statistical justification. We addressed this by: (1) reframing all findings around the specific promoters studied, avoiding generalizations beyond the cases [across the introduction and conclusions sections]; (2) replacing multivariate analyses with descriptive statistics—medians and interquartile ranges (Table 5), a composite agroecological transition index [lines 240-256], a global performance index [lines 263-264], and tercile-based grouping for acquaintance time [lines 273-277]; and (3) acknowledging the potential influence of prior social ties on community perception scores, incorporated as an exploratory analysis [lines 273-278; lines 446-449]. The reduced sample size is addressed as inherent to the case study design adopted [lines 201-213].

Comment 3:

The study does not address the core issues raised. The article proposes to address how CaC approaches can start in regions that lack organizational foundations. The introduction makes it clear that Casanare province lacks the ripe conditions for the CaC model (no organized network, no historical tradition), which is precisely the value of the study. However, the results section only reports 'promoters perform well' and 'split into three categories', and the discussion section only explains 'why different evaluators give different scores'.

Response 3:

We agree with this point. The revised manuscript addresses this gap more directly. The discussion now connects the promoter profiles and performance findings to the conditions of emergent FtF contexts, arguing that the diversity of profiles across agroecological transition levels and training competencies is meaningful evidence of how FtF processes can develop in territories without established organizational foundations [discussion paragraphs on Groups 1, 2, and 3]. The conclusions section was also revised to draw practical implications for facilitators and coordinators working in nascent FtF settings, and to note how differentiated promoter profiles can inform targeted accompaniment and the gradual consolidation of agroecological processes [lines 578-590].

Comment 4:

The analysis of results is not in-depth enough. The paper mentions that 87.5% are male promoters, but this is only attributed to structural and symbolic barriers, without further analysis: is there a sampling bias in which female promoters are omitted? Are there implicit differences between the performance characteristics of existing female promoters and those of men? This paper points out that the promoters are mainly middle-aged and elderly people, and the participation of young people is insufficient, but it does not analyze the specific impact of the absence of youth on the sustainability of CaC model.

Response 4:

We agree that these dimensions warranted deeper analysis. Regarding gender, the single female promoter (P1, Group 1) obtained the highest global performance score (4.50), which is reported in the results [lines 335-342]. However, a systematic comparison by sex is not statistically meaningful with this sample. In the discussion, we clarified that male predominance likely reflects structural barriers to female leadership in rural Colombia rather than sampling bias—participation was self-selected from an existing project cohort [lines 240-244]—and we framed this as both a process weakness and a strategic opportunity for future FtF design [lines 244-250]. Regarding youth, the discussion was expanded to link low youth engagement not only to structural migration pressures [lines 252-260] but also to the implications for the long-term sustainability of FtF processes, noting that the absence of intergenerational knowledge transfer poses a challenge for consolidating agroecological movements in nascent contexts [lines 260-264].

First, regarding gender composition, it is important to clarify that the study did not rely on convenience sampling or any directed selection process that could introduce bias. Participation emerged from an open and voluntary process within the existing social base of the SEGALI project (40 producers), from which only a small subset expressed willingness to assume the promoter role, resulting in eight participants, mostly men. This configuration was not methodologically induced but reflects the actual participation dynamics within the project context, as clarified in the revised manuscript [lines 116–123]. In this sense, the low participation of women is interpreted not as a methodological artifact but as an empirical finding that points to structural and socio-cultural constraints limiting women’s access to leadership roles in rural contexts. Consistently, the only female promoter obtained the highest global performance score, suggesting that the observed gap is not related to differences in capacity but to barriers in access, an interpretation further developed in the discussion [lines 467–475].

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

I am pleased to review this manuscript. My reviews are as follows:

This manuscript contains a problem of citing outdated literature. Most of the cited studies are relatively old and lack recent research.
The authors adopted convenience sampling for participant selection. This may lead to a lack of representativeness and reduce the external validity of the research results. In addition, although the authors claim that previously validated assessment tools were used, they did not provide relevant references, nor did they conduct reliability or validity analyses. Therefore, it is not possible to confirm whether the assessment tools possess good measurement quality.
The study sample includes only eight promoters. Such a small sample size severely limits the reliability of statistical inference. In addition, the authors did not explain whether the data distribution follows a normal distribution, yet confidence intervals were directly used for estimation. The methodological appropriateness of this approach is questionable.
The sample size of this study includes only eight promoters, yet the authors still conducted multivariate statistical analyses (Multiple Correspondence Analysis and k-means clustering). Conducting clustering analysis with such a small sample size lacks statistical meaning.
This manuscript uses Likert scales as evaluation indicators, but the authors report the results using means, standard deviations, and 95% confidence intervals. Given the extremely small sample size and the ordinal nature of the data, such statistical estimations lack methodological justification.
The authors claim that “validated instruments” were used, but they did not provide the source of the scales and did not conduct reliability or validity tests. The authors even explicitly state that Cronbach’s alpha was not calculated. This approach does not demonstrate that the measurement instruments are reliable.
The expert evaluation was conducted by only two evaluators, and no assessment of rating consistency was reported. Therefore, the evaluation results may be highly subjective.
The study lacks a clear theoretical framework or research model. The Introduction mainly provides conceptual descriptions and does not establish testable research hypotheses or an analytical framework.
The study analyzes only eight farmer promoters, yet the conclusions extend to broader issues such as “agroecological transition” and “sustainable rural development.” These claims lack sufficient empirical support and therefore constitute overgeneralization.
The numbering in the Methods section contains clear errors.

This manuscript presents several methodological and analytical limitations, including an extremely small sample size, inappropriate statistical analyses, a lack of measurement validation, and an inadequate theoretical framework.

Author Response

Comment 1:

This manuscript contains a problem of citing outdated literature. Most of the cited studies are relatively old and lack recent research.

Response 1:

We partially agree. The FtF literature on the pedagogical dimensions of promoter performance remains limited and necessarily draws on foundational works. Nevertheless, we updated the reference list by adding recent studies documenting FtF effectiveness in diverse global contexts, including sustainable land management, agroforestry, and agricultural innovation diffusion [references 5-9, 12-14, introduced between lines 33-37]. We also incorporated more recent methodological references into the materials and methods section [references 29, 42, 43, 44, 45, 46, 47].

Comment 2:

The authors adopted convenience sampling for participant selection. This may lead to a lack of representativeness and reduce the external validity of the research results. In addition, although the authors claim that previously validated assessment tools were used, they did not provide relevant references, nor did they conduct reliability or validity analyses. Therefore, it is not possible to confirm whether the assessment tools possess good measurement quality.

Response 2:

We agree. The sampling procedure is now described as purposive rather than convenience sampling, arising from the existing social base of the SEGALI project [lines 131-138]. The scope of findings was revised throughout to reflect the case study design and avoid claims of external representativeness. Regarding instrument validation, the manuscript describes the content validation procedure (academic panel review, semantic review, and pilot test) [lines 152-162], supported by references on instrument validation in applied research [32]. A clarification was also added on why Cronbach's alpha was not used as the primary reliability criterion for this multidimensional instrument, drawing on literature that supports content validity and dimensional coherence for composite Likert-based rubrics in applied evaluation contexts [references 44, 45].

Comment 3:

The study sample includes only eight promoters. Such a small sample size severely limits the reliability of statistical inference. In addition, the authors did not explain whether the data distribution follows a normal distribution, yet confidence intervals were directly used for estimation. The methodological appropriateness of this approach is questionable.

Response 3:

We agree. The analysis was revised to rely exclusively on descriptive statistics: medians and interquartile ranges for performance indicators (Table 5), composite indices for agroecological transition [lines 240-256] and global performance [lines 263-264], and tercile-based groupings for acquaintance time [lines 273-277]. Confidence intervals and normality assumptions are no longer reported anywhere in the manuscript. The case study framing is made explicit in Section 2.5, where the small sample is justified as inherent to the analytical design [lines 201-213].

Comment 4:

The sample size of this study includes only eight promoters, yet the authors still conducted multivariate statistical analyses (Multiple Correspondence Analysis and k-means clustering). Conducting clustering analysis with such a small sample size lacks statistical meaning.

Response 4:

We agree. MCA and k-means clustering have been removed. Promoter grouping is now based on a descriptive, case-oriented typification approach using the agroecological transition index and the global performance index as primary criteria [lines 201-213; Table 6], consistent with small-n comparative case study frameworks [references 41, 42].

Comment 5:

This manuscript uses Likert scales as evaluation indicators, but the authors report the results using means, standard deviations, and 95% confidence intervals. Given the extremely small sample size and the ordinal nature of the data, such statistical estimations lack methodological justification.

Response 5:

We agree. Means, standard deviations, and confidence intervals have been removed from all results. It is important to clarify that means were used as an integral part of the global and dimensional assessment index, considering the multidimensional nature of the matrix. The manuscript now reports medians and interquartile ranges throughout (Table 5), which are appropriate non-parametric descriptors for ordinal Likert data and small samples. This is consistent with references on the use of composite Likert scales in applied evaluation research [references 44, 45], and is applied consistently across the results and discussion sections.

Comment 6:

The authors claim that 'validated instruments' were used, but they did not provide the source of the scales and did not conduct reliability or validity tests. The authors even explicitly state that Cronbach's alpha was not calculated. This approach does not demonstrate that the measurement instruments are reliable.

Response 6:

We acknowledge the concern. The evaluation matrix was not adapted from a pre-existing standardized scale but constructed for this study, with each indicator grounded in cited conceptual references (Table 1). The validation procedure is described in Section 2.4 [lines 152-162]: content review by two academic panelists, semantic revision, and a pilot test with rural promoters not included in the final sample [reference 32]. Regarding reliability, the manuscript clarifies that Cronbach's alpha was not calculated because the instrument is a multidimensional rubric with conceptually distinct dimensions rather than a unidimensional psychometric scale. For such instruments, content validity and dimensional coherence—supported by expert review—are considered appropriate reliability indicators [lines 163-167; references 44, 45]. Nevertheless, as a complementary verification, Cronbach’s alpha was estimated, yielding a value of 0.78, which indicates an acceptable level of internal consistency. This result was not included in the manuscript to avoid misinterpreting the instrument as a unidimensional scale, but it is provided here as additional evidence supporting the robustness of the measurement approach.

Comment 7:

The expert evaluation was conducted by only two evaluators, and no assessment of rating consistency was reported. Therefore, the evaluation results may be highly subjective.

Response 7:

We acknowledge this limitation. To reduce subjectivity, a calibration session was conducted with both evaluators before data collection, in which each indicator's working definition was reviewed and discussed to align interpretation and minimize the halo effect [lines 174-183]. Their scores were subsequently integrated using equivalent weighting to obtain a consolidated rating per indicator. We recognize that formal inter-rater reliability measures (e.g., Cohen's kappa or intraclass correlation) were not calculated, and this is now acknowledged explicitly among the study's limitations, alongside the small sample size and single-encounter design [lines 567-578]. However, the use of clear operational definitions, prior calibration, and the integration of evaluations support a consistent application of the instrument within the study context.

Comment 8:

The study lacks a clear theoretical framework or research model. The Introduction mainly provides conceptual descriptions and does not establish testable research hypotheses or an analytical framework.

Response 8:

We agree. The revised introduction grounds the study explicitly in social learning theory as the analytical framework for understanding and evaluating promoter performance in FtF settings [lines 71-81]. Two research questions are stated clearly [lines 88-91], and the study's contribution is positioned in relation to the gap between FtF research in consolidated versus nascent organizational contexts [lines 33-41]. As a case study, the work does not test inferential hypotheses, but the research questions provide a clear analytical orientation addressed systematically in the results and discussion.

Comment 9:

The study analyzes only eight farmer promoters, yet the conclusions extend to broader issues such as 'agroecological transition' and 'sustainable rural development.' These claims lack sufficient empirical support and therefore constitute overgeneralization.

Response 9:

We agree. The conclusions and discussion sections were revised to adjust all statements to the specific promoters studied, avoiding generalizations to broader populations or processes. Where concepts such as agroecological transition are included, they serve to contextualize the findings rather than as empirically supported claims derived from the sample. The revised conclusions frame implications for practitioners and researchers in similar nascent FtF contexts as propositions that warrant further investigation.

Comment 10:

The numbering in the Methods section contains clear errors.

Response 10:

We agree. Subsection numbering in the Materials and Methods section was corrected throughout the manuscript.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

I thank the authors and the editor for the opportunity to review this paper. I appreciated both the research topic and the methodological approach. Below, I have highlighted a few minor aspects that could help enhance the manuscript.

1) I suggest to highlight the research questions

2) The I suggest to highlight the limitation of this study and the further development of the research

3) I also recommend strengthening the Conclusion section by emphasizing the key findings and their implications for the different stakeholders.

4) Finnaly, in the references, I suggest including this recent contribution on the same topic of this paper

Peter Michael Rosset, Ivanete Ferreira Fernandes, Lia Pinheiro Barbosa, Cosma dos Santos Damasceno, Weeraboon Wisartsakul, Unlearning the green revolution: Inventory of agroecological practices in Ceará, Brazil, an instrument for decolonizing territory and (re)valuing peasant knowledge,
Environmental Science & Policy, Volume 165, 2025, 104022, ISSN 1462-9011,
https://doi.org/10.1016/j.envsci.2025.104022.

Author Response

Comment 1:

I suggest to highlight the research questions.

Response 1:

We agree. The two research questions are now stated explicitly at the end of the introduction [lines 88-91]: (1) How do participants evaluate promoter performance along the pedagogical dimensions of FtF exchange processes? and (2) What is the relationship between promoter profile attributes and perceived performance?

Comment 2:

I suggest to highlight the limitation of this study and the further development of the research.

Response 2:

We agree. A dedicated paragraph on limitations and future research directions was added as the second-to-last paragraph of the discussion section [lines 567-578], covering the small sample size, the single-encounter design, the absence of inter-rater reliability measures, and the need for longitudinal and multi-cycle FtF studies.

Comment 3:

I also recommend strengthening the Conclusion section by emphasizing the key findings and their implications for the different stakeholders.

Response 3:

We agree. The conclusions were revised to highlight implications for farmer promoters, facilitators, and coordinators in FtF processes, drawing on the four-group typology proposed by ANAP process and identified its relevance for promoter accompaniment.

Comment 4:

Finally, in the references, I suggest including this recent contribution on the same topic of this paper.

Response 4:

We agree. The suggested reference (Rosset et al., 2025) was incorporated in the discussion section [reference 62, lines 567-574], where it is used to situate the present study within the early pedagogical stages of the FtF methodology, as described by Rosset et al. in their inventory of agroecological practices in Ceará, Brazil.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Overall, the authors have adequately addressed my previous comments and have made substantial revisions to the manuscript. In its current form, the manuscript generally meets the standards for publication.

My only remaining concern relates to the very small sample size (n = 8). In general, studies with such limited samples are more commonly approached using qualitative methods. The use of quantitative analysis in this context may raise concerns regarding generalizability. Although certain statistical techniques can accommodate small sample sizes, there remains a risk of substantial bias in the results.

I suggest that the authors, in their future research, provide a clearer explanation for the use of a small sample and the adoption of quantitative methods. In addition, potential issues arising from this approach should be explicitly noted to inform future researchers.

Author Response

We thank the editorial team and the reviewer for the continued engagement with our manuscript and for the constructive feedback provided during this second round of review. The remaining comments have been considered, and the manuscript has been revised accordingly. We believe the changes further strengthen the methodological transparency of the work.

Below we present our point-by-point response. All modifications are indicated in the manuscript using Track Changes mode.

Comment 1:

Response 1:

We agree with this concern and appreciate the reviewer's recognition of the revisions made in the previous round. The treatment of small sample sizes using quantitative descriptors is a genuine methodological consideration, and we have taken additional steps to address it transparently.

In Section 2.5 (Information Analysis), the treatment of sample size within the case study framework was revised to make explicit that the findings should be interpreted in a non-inferential manner. The relevant passage now reads [lines 205–207]:

"...population samples. In this sense, although the small number of participants is consistent with the case study design, it constitutes a limitation for broader inference and should be considered when interpreting the results [42]."

Additionally, the discussion of limitations was expanded in the manuscript. A brief mention of sample size that appeared at line 576 was removed and replaced in the following paragraph in an expanded version [lines 583–591], where we clarify the inferential constraints derived from the sample size of eight promoters. To provide a balanced perspective, we also note the contribution of this type of study to the development of participatory extension processes, particularly in contexts—such as nascent FtF settings in under-documented territories—where small-n, case-based approaches are a common and methodologically justified starting point for building understanding.

Author Response File: Author Response.pdf