How Do Italian Consumers Value Sustainable Certifications on Fish?—An Explorative Analysis

Round 1

Reviewer 1 Report

This is an interesting article that, with some modifications, can contribute to extending the literature on consumers' acceptance and preference for fish attributes.

Indeed, while I appreciate the introduction (informative and concise) and the literature review (comprehensive and informative), I believe that the focus on sustainable certification of the fish product is then very nuanced, almost getting lost along the paper. In fact, starting from the title and the introduction, the reader would expect a clear focus on this dimension/attribute of the product (i.e. sustainable certification or sustainable production) along the entire study, but as one proceeds with the reading, the same review (although interesting) focuses on a myriad of different attributes that take the reader far away from the previously stated aim (to extend the literature on the importance that consumers attach to sustainability certifications for fish product). Similarly, the attributes’ choice for the best-worst methodology goes in this sense (the sustainable production represents one attribute out of the 9 considered), little to justify the title and the research objective as currently stated. Finally, although interesting, the implications and suggestions described in the conclusions discuss the concept of sustainable production above all. Due to this, I think that the focus on sustainable production/certification has to be clarified throughout the paper more consistently.

The methodological choice appears to be suitable for the objective of the investigation and the investigation is correctly carried out, even if a better description of the experimental design used and on how the latent class clustering model was carried out is required. Authors are therefore advised to describe the methodology in more detail.

To conclude, the first part of the conclusions appears too concise, considering the large number of attributes analyzed and the variables used to characterize the clusters. Therefore it is advisable to better discuss the main evidence (lines 353-358). I appreciate the description of the limitations of the study, which prevents any generalization of the results.

 

Author Response

Let us first thank the reviewer for the time spent on revising the paper as well as for the comments. The suggestions were very useful in revising the manuscript and have been incorporated into the revised paper version. Reviewer's comments helped us to understand the weak points of the paper and how to strengthen it. As a result, we believe the quality of the revised paper has been significantly improved thanks to your suggestion.

Reviewer’s original comments are reported below in italics and underlined while the answers to reviewer’s comments follow each comment directly and are prefaced by the word ANSWER

This is an interesting article that, with some modifications, can contribute to extending the literature on consumers' acceptance and preference for fish attributes.

ANSWER We thank the reviewer for the kind words of appreciation of our work.

Indeed, while I appreciate the introduction (informative and concise) and the literature review (comprehensive and informative), I believe that the focus on sustainable certification of the fish product is then very nuanced, almost getting lost along the paper. In fact, starting from the title and the introduction, the reader would expect a clear focus on this dimension/attribute of the product (i.e. sustainable certification or sustainable production) along the entire study, but as one proceeds with the reading, the same review (although interesting) focuses on a myriad of different attributes that take the reader far away from the previously stated aim (to extend the literature on the importance that consumers attach to sustainability certifications for fish product).

ANSWER We thank the reviewer for the comment which give us the possibility to clarify and better frame our research question. As the referee correctly points out,  the introduction section mainly focuses on the fish sustainable certifications and sustainable production process. The introduction ends by suggesting the importance to explore the consumer's preferences for such product feature, which is the goal of our study. For sake of completeness, one cannot neglect that several other product features, potentially available on fish (see Table 1 in our study), may jointly affect consumers' preferences for sustainable fish products. Each consumer ranks the sustainable fish "dimension" at a different importance level when he/she compares it with other fish characteristics by preferring some (e.g., product origin and taste) over others (e.g., sustainably and price). Thus, one has to account for the existence of other product features on fish to correctly: i) assess the importance of sustainable certifications/sustainable production process for fish consumers; as well as, ii) to characterize the sustainable fish consumer. For what we said above, we had to discuss the many attributes that can affect consumer acceptance of and preference for fish, besides the sustainable certification or production process. Hope we have clarified the referee's concern. Please referee can see Page 2, line 73 to 81

 

Similarly, the attributes’ choice for the best-worst methodology goes in this sense (the sustainable production represents one attribute out of the 9 considered), little to justify the title and the research objective as currently stated.

ANSWER We thank the referee for his/her comment. This comment is connected with the one above. In detail, sustainable certification is available on fish products in combination with other product features (e.g., country of origin, price, taste, experience, fish species, production process, and storage process), thus researchers who want to assess the relevance of the sustainable certification for consumers have to take into account all the attributes that potentially affect consumers’ preferences and attributes combined effects. The best-worst methodology allows assessing to what extent one specific product attribute (e.g., sustainable certification) values for the consumers when compared with other product features are available on the product. Hope we have clarified the referee’s concern as well as we stress this point in the modified version as well as we stress this point in the modified version of the manuscript at page 2, lines 73-81. The title has been modified as “How do Italian consumers value sustainable certifications on fish ?– an explorative analysis.

Finally, although interesting, the implications and suggestions described in the conclusions discuss the concept of sustainable production above all. Due to this, I think that the focus on sustainable production/certification has to be clarified throughout the paper more consistently.

ANSWER We thank the referee for the comment. We stress the focus on sustainable certification at Page 11, lines 345-348 and lines 357-360, and Page 12, lines 371-374 and lines 385-388. Also, more discussion on the main evidence is added in the conclusion section. Please see the revised version of the manuscript at Page 12, lines 391-413

The methodological choice appears to be suitable for the objective of the investigation and the investigation is correctly carried out, even if a better description of the experimental design used and on how the latent class clustering model was carried out is required. Authors are therefore advised to describe the methodology in more detail.

ANSWER We thank the referee for his/her comment. We expand the discussion of the methodological approach. Please the referee can see Page 6, lines 207 to 218 and Page 9 lines 297-309, as well as Table A in appendix A.

To conclude, the first part of the conclusions appears too concise, considering the large number of attributes analyzed and the variables used to characterize the clusters. Therefore it is advisable to better discuss the main evidence (lines 353-358). I appreciate the description of the limitations of the study, which prevents any generalization of the results.

ANSWER We thank the referee for his/her comment. As mentioned in our comment above, we expanded the discussion on the main evidence in the conclusion section. Please the referee can see Page 12, lines 391 to 413.

 

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Authors

This is an interesting and important topic. First of all, the chapter should not start with a table as it is in section 3. Besides, section 3 should be separated into a description of results and a discussion of results. This will make the analyzed data more transparent. Please indicate how the respondents were selected, when the survey was conducted, what was the structure of the questionnaire, was it legitimate. Is this sample size appropriate for this type of study? What were the limitations of this study? More broadly, the results should be discussed in the context of comparisons with other surveys and general consumer trends for certified products.

Author Response

Let us first thank the reviewer for the time spent on revising the paper as well as for the comments. The suggestions were very useful in revising the manuscript and have been incorporated into the revised paper version. Reviewer's comments helped us to understand the weak points of the paper and how to strengthen it. As a result, we believe the quality of the revised paper has been significantly improved thanks to your suggestion.

Reviewer’s original comments are reported below in italics and underlined while the answers to reviewer’s comments follow each comment directly and are prefaced by the word ANSWER

 

Dear Authors

This is an interesting and important topic.

 

ANSWER We thank the reviewer for the kind words of appreciation of our work.

 

First of all, the chapter should not start with a table as it is in section 3.

 

ANSWER We thank the referee for his/her comment. Table 3 was moved to the previous section avoiding having a table at the beginning of section 3. Please see the revised version of the manuscript at Page 7, line 247.

 

Besides, section 3 should be separated into a description of results and a discussion of results. This will make the analyzed data more transparent.

ANSWER We thank the referee for his/her comment. We respectfully disagree with the referee as we kept results and discussion tight together in a sole section to not hinder the paper’s readability given the several attributes/consumer groups that we discuss. Also, the MPDI author's guidelines allow to have a "Results and Discussion" section in which authors point out the results of their analysis and discuss them compared to the existing literature.

Please indicate how the respondents were selected, when the survey was conducted, what was the structure of the questionnaire, was it legitimate.

 

ANSWER We thank the referee for his/her comment. Data collection ran from March 2021 to May 2021. The questionnaire was created on Google form whose link was shared via social media (e.g., WhatsApp, Facebook, LinkedIn) to collect a convenience sample of fish consumers' responses. The revised manuscript version has been expanded and now it encloses such information. Please referee can see page 6, lines 225 to 234. The structure of the questionnaire is described at page 6, line 235 to 244. The questionnaire was legitimate, please see page 6, lines 227 to 229.

 

Is this sample size appropriate for this type of study?

ANSWER We thank the referee for letting us clarify this point. Related to the sample size, the answer to the question is always “more is better”. However, Nylund-Gibson and Choi (2018), reviewing numerous studies, suggest that the optimal sample size to apply the Best-Worst method and a latent class analysis is of 300 or more cases, although smaller samples may be adequate with simpler models and “well-separated” classes (Nylund-Gibson & Choi, 2018). Thus, our work with a sample size of 312 observations is appropriate for performing such analysis.

Nylund-Gibson, K., & Choi, A. Y. (2018). Ten frequently asked questions about latent class analysis. Translational Issues in Psychological Science, 4(4), 440-461. https://doi.org/10.1037/tps0000176

Also, several studies from the consumer marketing field and focusing on agri-food products use samples of comparable size, we list some studies here for your convenience:

• Perito, M. A., Sacchetti, G., Di Mattia, C. D., Chiodo, E., Pittia, P., Saguy, I. S., & Cohen, E. (2019). Buy local! Familiarity and preferences for extra virgin olive oil of Italian consumers. Journal of Food Products Marketing, 25(4), 462-477. Sample size: 179 observations.
• de-Magistris, T., Gracia, A., & Barreiro-Hurle, J. (2017). Do consumers care about European food labels? An empirical evaluation using best-worst method. British Food Journal, 119, 2698–2711. Sample size: 549 observations.
• Mueller, S.and Rungie, C. (2009). Is there more information in best‐worst choice data? : Using the attitude heterogeneity structure to identify consumer segments". International Journal of Wine Business Research, 21 (1), 24-40. Sample size: 304 observations.

 

What were the limitations of this study?

 

ANSWER We thank the referee for the comment. The study limitations are listed at Page 13, lines 452 to 463.

 

More broadly, the results should be discussed in the context of comparisons with other surveys and general consumer trends for certified products.

 

ANSWER We thank the referee for the comment. We would like to point out results discussion has been done comparing our results with those obtained from survey-based studies investigating consumer and certified agri-food products. See references number 17,18 and 46-54

Author Response File: Author Response.docx

Reviewer 3 Report

This study assessed consumers’ preference for sustainable fish. It seems interesting but many details are missing to be able to assess its quality. Following I summarize my most relevant comments:

• The authors should explain why they chose this list of attributes. I mean that they should justify the use of each attribute.
• Why authors chose a convenience sample?
• The authors should explain why the price is the least important attribute in purchasing fish. Is it in line with previous studies?
• The authors should justify why they opted for a LCM. A LCM fits the data better than an MNL?
• Authors should report the values of the BIC and LL to understand how they chose the optimal number of classes.
• Authors should report the significance of the parameters.
• The authors should clarify why the model in Table 5 has not a constant.
• Authors should estimate the WTP of the different classes for sustainable certification

Author Response

Let us first thank the reviewer for the time spent on revising the paper as well as for the comments. The suggestions were very useful in revising the manuscript and have been incorporated into the revised paper version. Reviewer's comments helped us to understand the weak points of the paper and how to strengthen it. As a result, we believe the quality of the revised paper has been significantly improved thanks to your suggestion.

Reviewer’s original comments are reported below in italics and underlined while the answers to reviewer’s comments follow each comment directly and are prefaced by the word ANSWER

 

This study assessed consumers’ preference for sustainable fish. It seems interesting but many details are missing to be able to assess its quality.

Following I summarize my most relevant comments:

The authors should explain why they chose this list of attributes. I mean that they should justify the use of each attribute.

ANSWER We thanks the referee for his/her comment. The attributes listed in Table 1 were identified on the base of the literature review on consumer acceptance of and preferences for fish attributes in section 2 “Literature review”. To the best of our knowledge, we include in our analysis all the attributes investigated in the literature reviewed. Literature sources related to the attributes selection criteria have been added in Table 1 of the revised version of the manuscript, placed side by side of the attribute description. Please see the revised version of Table 1. Hope we have clarified the referee’s concern.

 

Why authors chose a convenience sample?

ANSWER We use a convenience sample as less expensive compared to other sampling methods, also the questionnaire is easily accessible from a broader audience of consumers. Lastly, collecting data from a convenience sample requires less time for collecting data compared to the time needed by other types of samples techniques. We include such explanation in the revised manuscript version. Please referee can see page 6, lines from 230 to 234. Also, we modify the title to stress the explorative aim of our study, please referee can see page 1, line 1 The new title is now: “How do Italian consumers value sustainable certifications on fish – an explorative analysis”.

 

The authors should explain why the price is the least important attribute in purchasing fish. Is it in line with previous studies?

ANSWER We thank the referee for his/her comment. Product’s price was the least important attribute in purchasing fish products in the average sample. This result is likely due to the fact that fish consumers overall are already willing to give away part of their income to get a pricy product as fish to which are potentially attached many features in which consumers are interested in (e.g., taste, origin, sustainable certification). Thus, such a result suggest that sampled consumers are overall not particularly price-sensitive in selecting fish, although as confirmed in the literature and already described in the original version of the paper (please, see page 11, line 357 to line360), there exists a share of consumers, in our case the Careless consumer, that place relevant importance to fish price in purchasing the product. We expand the discussion on the relevance of price attribute for the average consumer in the revised version of the text. Referee can see page 12, lines 408 to 412.

 

The authors should justify why they opted for a LCM. A LCM fits the data better than an MNL?

ANSWER We thank the referee for the comment. We used a Latent Class Clustering Model, or Latent Class Analysis (LCA), to clustering the sample and further analyses the heterogeneity underlying attribute importance among respondents as LCA is a semi-parametric variant of the MNL which does not require specific assumptions about the distributions of parameters across individuals and overall perform better than the latter (Hensher et al., 2005). Also, in the Latent Class Clustering Model, consumers are assumed to belong to different segments, each of it characterized by unique class-specific utility parameters. In other words, within each segment, consumer preferences are homogeneous, but preferences vary between segments, allowing for a more in-depth understanding of consumer heterogeneity (Hensher et al., 2005).

Hensher, D. A., Rose, J. M., Rose, J. M., & Greene, W. H. (2005). Applied choice analysis: a primer. Cambridge university press.

 

Authors should report the values of the BIC and LL to understand how they chose the optimal number of classes.

ANSWER We thank the referee for the comment. The revised manuscript encompasses a table reporting the Log-Likelihood (LL) value function, the Bayesian information criterion (BIC) and the Akaike information criterion (AIC) for evaluating how well the model fits the data it was generated from. The table is reported in the manuscript’s appendix A. The optimal numbers of classes was assessed by using Log-likelihood,  Akaike information criterion (AIC) and Bayesian information criterion (BIC) (values reported in Table A in Appendix A). For this work, all indicators improved as more clusters were added, supporting the presence of multiple segments in the sample. Although indicators further improved as more cluster were added, the changes were much smaller from a five- to a six-cluster model compared to the move from a four- to a five-cluster model. Also, the five-cluster model provided the best interpretability compared to the six-cluster one. Thus, as the model interpretability is as important as the statistical tests, the five-cluster model was selected for our analysis. We added more discussion related to this point in the main revised text. The referee can see page 12, lines 408 to 413.

Authors should report the significance of the parameters.

ANSWER To explore respondent heterogeneity, we used the average BW score and latent class cluster analysis to identify homogeneous groups of consumers with similar preferences for fish attributes. Hence, clusters were generated using solely attribute BW scores as dependent variables under the LCA assumption that data is generated by a probability distribution that defines a certain number of latent clusters. All segments differ significantly (p-value < 0.01) from each other with respect to nine quality attributes tested (country of origin, price, taste, past experience, fish species, production process, storage process, sustainable certification, nutritional content). Then, each sub-group generated by LCA were then characterized in terms of demographic characteristics to identify market segments. We added further discussion related to this point at page 9, line 297 to 309.

The authors should clarify why the model in Table 5 has not a constant.

ANSWER Please the referee can see the comment above in which we clarify that clusters were generated via LCA using solely attribute BW scores as dependent variables to explore respondents’ heterogeneity in preferences for fish attributes as customary in many studies. Please referee can see some recent studies which use a similar approach:

- Pomarici, E., Lerro, M., Chrysochou, P., Vecchio, R., & Krystallis, A. (2017). One size does (obviously not) fit all: Using product attributes for wine market segmentation. Wine Economics and policy, 6(2), 98-106.

- De Gennaro, B. C., Roselli, L., Bimbo, F., Carlucci, D., Cavallo, C., Cicia, G., Del Giudice, T., Lombardi, A, Paparella, A. & Vecchio, R. D. (2021). Italian consumers value health claims on extra-virgin olive oil. Journal of Functional Foods, 81, 104461.

- de-Magistris, T., Gracia, A., & Barreiro-Hurle, J. (2017). Do consumers care about European food labels? An empirical evaluation using best-worst method. British Food Journal, 119(12), 2698-2711.

- Perito, M. A., Sacchetti, G., Di Mattia, C. D., Chiodo, E., Pittia, P., Saguy, I. S., & Cohen, E. (2019). Buy local! Familiarity and preferences for extra virgin olive oil of Italian consumers. Journal of Food Products Marketing, 25(4), 462-477.

 

Authors should estimate the WTP of the different classes for sustainable certification

ANSWER  We thank the reviewer for the comment which give us the possibility to clarify and better frame our research question.

First, we are not interested, at least in this study explorative in nature, in sizing the consumers’ WTP for sustainable certification across consumers' classes. We will address this point in future research, as we point out in the revised version of the conclusion section. Please see the page 13, lines 459 to 460.

Second, we want to stress the need and urgency of this work as existing studies have explored the preferences for fish attributes, including for sustainable certification, in combination with a few other products features by failing to account for the multiple product’s attributes potentially available on fish and that jointly affect consumers preferences (e.g., taste, price, production process, storage process, experience) (see Table 1 in our study). This calls for a more complex research design allowing to rank consumers' preferences for the multiple product features available on the fish product, including the sustainable certification (see Table 1 in our study), as well as to size and the characterize "sustainable" consumers using their socio-economic characteristics. This, was the main goal of our study. Hope we have clarified referee’s concern. Referee can see the modified version of the manuscript at page 2, lines 73-81 in which we stress the aim of the work.

Author Response File: Author Response.docx

Reviewer 4 Report

Dear Authors,

Thank you for submitting a very relevant paper. Here are my comments which should support you in the publishing process:

1. Before line 73 on page 2, please define the research gap referring to the literature review. The research gap should serve as a basis for the aim of the study in lines 73-75.
2. I would recommend renaming 1.1 into section "2. Literature Review".
3. In the Literature Review, it would be beneficial for the reader to have a table, presenting in a very transparent manner how the three factors (search, experience, and credence attributes) are composed.
4. In the paragraph (lines 168-181) please explain what were the alternatives to performing a Best-Worst Method.
5. In lines 218-219 you write: "Data were collected through an online survey by collecting a convenience sample of 312 Italian household fish shoppers responsible". Please explain the target sample, how you approached them and what are their socio-demographic characteristics.
6. At the end of Section 3, it would be good to have a table depicting all five consumer groups with their main characteristics.

I wish you good luck with the publishing process.

Best regards,

The Reviewer

Author Response

Let us first thank the reviewer for the time spent on revising the paper as well as for the comments. The suggestions were very useful in revising the manuscript and have been incorporated into the revised paper version. Reviewer's comments helped us to understand the weak points of the paper and how to strengthen it. As a result, we believe the quality of the revised paper has been significantly improved thanks to your suggestion.

Reviewer’s original comments are reported below in italics and underlined while the answers to reviewer’s comments follow each comment directly and are prefaced by the word ANSWER

Dear Authors,

Thank you for submitting a very relevant paper.

ANSWER We thank the reviewer for the kind words of appreciation of our work.

 

Here are my comments which should support you in the publishing process:

1. Before line 73 on page 2, please define the research gap referring to the literature review. The research gap should serve as a basis for the aim of the study in lines 73-75

ANSWER We thank the referee for his/her suggestion. As suggested by the referee, the revised manuscript version points out the existing research gap compared to the extant studies that we address in our work. Please referee can see page 2 from line 73 to 81.

 

1. I would recommend renaming 1.1 into section "2. Literature Review".

ANSWER We rename subsection 1.1 into section 2. Many thanks for the comment.

 

 

1. In the Literature Review, it would be beneficial for the reader to have a table, presenting in a very transparent manner how the three factors (search, experience, and credence attributes) are composed.

ANSWER We thank the referee for the suggestion. We modify Table 1 by including a first column that reports whether the product attribute, listed and described in the second and third columns, is a search, experience, or a credence attribute. We hope that this help to clarify the referee’s concern.

 

1. In the paragraph (lines 168-181) please explain what were the alternatives to performing a Best-Worst Method.

ANSWER The alternative methods to Best-Worst that can be used as well as their drawbacks already indicated in the original manuscript version at page 2, line 87 to 92 and page 4 lines 170-175.

 

1. In lines 218-219 you write: "Data were collected through an online survey by collecting a convenience sample of 312 Italian household fish shoppers responsible". Please explain the target sample, how you approached them and what are their socio-demographic characteristics.

ANSWER We thank the referee for the comment. Data collection ran from March 2021 to May 2022. The questionnaire was created on Google form whose link was shared via social media (e.g., WhatsApp, Facebook, LinkedIn) to collect a convenience sample of fish consumers' responses and whose socio-demographic characteristics are reported in Table 3. We expanded the data collection description in the revised manuscript, referee can see Page 6, lines 225 to 235.

 

1. At the end of Section 3, it would be good to have a table depicting all five consumer groups with their main characteristics.

I wish you good luck with the publishing process.

Best regards,

The Reviewer

 

ANSWER We thank the referee for the comment. The five consumer groups characteristics are reported in Table 6.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments have been addressed appropriately, and the previously vague points which needed to be clarified have been resolved. The revised manuscript looks now more focused. The paper is publishable in its current version.

Reviewer 3 Report

Thank you to the authors for their efforts.