Sustainable Food Consumption: Demand for Local Produce in Singapore

Round 1

Reviewer 1 Report

The paper “Local produce and food security under climate change in Singapore” presents a survey study on local food consumers and their willingness to pay for local products. The topic is interesting and important, but in the moment the paper has many short comings. There is not a real analysis of the collected data present, only the results of the survey is shown. There is an attempt to interpret the results but no real discussion or a connection to other studies. The survey in itself has its flaws and there are some unanswered questions regarding the data collection. I would also like to see the connection to a sustainable food chain or sustainable food consume? Otherwise, the paper might be better of in another journal.

Overall, I do not think, the paper is ready to be considered for publication.

 

Comments:

The literature review should be a part of the introduction to summarize the main background for the paper.

 

3.1. Data: How were the participants chosen? It is not clear to me what group of people took part of the study. Were the participants invited? Was it advertised? – these information are important to estimate the limitations of the study and if all groups in a society are covered.

 

3.2 should be structures differently and separate between the data sampling and steps of analysis. It would be easier for the reader to follow the study if it is clearly stated, 1. What are the questions in the survey and possible outcomes (data collection), and then, 2. how the author used the regression model.

 

What is the difference between equation 1 and 3? They seem identical.

L245 – 245 there is a reference to a much later table. Usually, it the tables should be presented in the order, they are mentioned in the text.

 

L271 As not everyone will know the currency and how it relates to internationally currency, it would be good to give an estimate in US$ as well, maybe in brackets behind original currency.

 

L267 – 274 – The ethnic distribution is related to the state distribution but not the other categories. Would there be data that can be related to the survey data to show if the study group is representative for the whole population?

 

It is a bit confusing to have the groups Singaporean and Chinese – could the author please clarify if this refers to the where they were born, the passport they hold or how they feel themselves? Are there other groups that took part in the study but could not be matched to one of the two?

 

Figure 1 to 4 show a very similar picture and distribution with always 80 – 90% of people agreeing (3-5) more or less with the statement. That makes me wonder, if the way the questions are asked are bias. There should be some questions included that would increase the participants to think the other way around, so that they do not automatically click on the agree buttons.

Also, was there an analysis done, if the same group of people always disagreed? Is there a correlation with the answers to income, age group, etc.?

 

Willingness to pay: I am a little confused with Table 5. The % for not willing to pay more should sit a bit more separate, otherwise it looks like the whole table presents the % of people that are not willing to pay.

 

L339-340 This statement needs a reference.

L342 also needs a reference

L347-348 Another statement and the attempt of an explanation, however, this should be backed up by the literature.

These is a trend through the whole paper. If the author makes a statement it should be presented with a reference and discussed further. Please, revise accordingly.

Also, the reference list looks short and there could be more recent studies referenced.

 

There are quite big tables presented, that could be moved to a supplementary. The result section could be condensed to the main and important results and outcomes. I do not see a proper analysis on the findings, only the presentation of the survey.

 

The beginning of the discussion repeats parts of the introduction. The first 2 paragraph add nothing new and could be delated.

 

Other studies are mentioned in the discussion but not referenced.

 

Overall, there is not really a discussion present in the paper.

 

 

Author Response

I am extremely grateful for all the constructive comments for this paper. Please find my responses to your comments below. I hope the revisions I have made is satisfactory.

The paper “Local produce and food security under climate change in Singapore” presents a survey study on local food consumers and their willingness to pay for local products. The topic is interesting and important, but in the moment the paper has many short comings. There is not a real analysis of the collected data present, only the results of the survey is shown. There is an attempt to interpret the results but no real discussion or a connection to other studies. The survey in itself has its flaws and there are some unanswered questions regarding the data collection. I would also like to see the connection to a sustainable food chain or sustainable food consume? Otherwise, the paper might be better of in another journal.

Overall, I do not think, the paper is ready to be considered for publication.

The paper is now framed within sustainable food consumption domain and discusses how consumers' perception and behavior related to environmental sustainability is associated with local produce purchase decision making.

 

Comments:

The literature review should be a part of the introduction to summarize the main background for the paper.

Response: I have added more relevant papers in the introduction, including studies on sustainable food consumption. Literature review section has also been merged into the introduction.

 

3.1. Data: How were the participants chosen? It is not clear to me what group of people took part of the study. Were the participants invited? Was it advertised? – these information are important to estimate the limitations of the study and if all groups in a society are covered.

Response: The participants were chosen through a registered panel by filtering people who conduct grocery shopping regularly. Elaboration on this has been added in the data section.

3.2 should be structures differently and separate between the data sampling and steps of analysis. It would be easier for the reader to follow the study if it is clearly stated, 1. What are the questions in the survey and possible outcomes (data collection), and then, 2. how the author used the regression model.

Response: I have re-organized those sections accordingly.

What is the difference between equation 1 and 3? They seem identical.

Response: Thank you for pointing it out. The estimation models have been updated based on the comments by another reviewer. I have two separate equations now.

L245 – 245 there is a reference to a much later table. Usually, it the tables should be presented in the order, they are mentioned in the text.

Response: The placement has been adjusted.

L271 As not everyone will know the currency and how it relates to internationally currency, it would be good to give an estimate in US$ as well, maybe in brackets behind original currency.

Response: I have added the converted nominal to US$ inside brackets.

L267 – 274 – The ethnic distribution is related to the state distribution but not the other categories. Would there be data that can be related to the survey data to show if the study group is representative for the whole population?

Response: I have added national statistics on gender, citizenship, income, and median age for comparison.

It is a bit confusing to have the groups Singaporean and Chinese – could the author please clarify if this refers to the where they were born, the passport they hold or how they feel themselves? Are there other groups that took part in the study but could not be matched to one of the two?

Response: Singaporean refers to Singapore citizens and permanent residents, while Chinese refers to the ethnicity, as elaborated in the text. Details have been added in the table.

Figure 1 to 4 show a very similar picture and distribution with always 80 – 90% of people agreeing (3-5) more or less with the statement. That makes me wonder, if the way the questions are asked are bias. There should be some questions included that would increase the participants to think the other way around, so that they do not automatically click on the agree buttons.

Response: The result is consistent with the national survey mentioned below, which has also been referred in the paper. The estimation analysis uses the variable converted into percentile to address the skewed data.

Also, was there an analysis done, if the same group of people always disagreed? Is there a correlation with the answers to income, age group, etc.?

Response: Checking the data, there were only a few people disagreed to more than 10 out of 14 questions. As for correlation, I have added another analysis whose results is reported in Table 5.

Willingness to pay: I am a little confused with Table 5. The % for not willing to pay more should sit a bit more separate, otherwise it looks like the whole table presents the % of people that are not willing to pay.

Response: Table layout and wordings have been edited.

L339-340 This statement needs a reference.

Response: Reference has been included for the statement.

L342 also needs a reference

L347-348 Another statement and the attempt of an explanation, however, this should be backed up by the literature.

Response: Reference has been added to the statement.

These is a trend through the whole paper. If the author makes a statement it should be presented with a reference and discussed further. Please, revise accordingly.

Response: I appreciate your comments. I have included references wherever needed.

Also, the reference list looks short and there could be more recent studies referenced.

Response: I have reviewed more recent studies in introduction to provide background to our study

There are quite big tables presented, that could be moved to a supplementary. The result section could be condensed to the main and important results and outcomes. I do not see a proper analysis on the findings, only the presentation of the survey.

Response: I have adjusted the estimations which resulted in less tables that are most relevant.

 

The beginning of the discussion repeats parts of the introduction. The first 2 paragraph add nothing new and could be delated.

Response: Thank you for pointing out. I have rewritten the section.

Other studies are mentioned in the discussion but not referenced.

Response: I have double checked the studies mentioned and added missing references.

Overall, there is not really a discussion present in the paper.

Response: Focusing around the issue of sustainable food consumption, I have reviewed relevant papers based on which I discussed how this study is linked with previous findings.

Reviewer 2 Report

The paper addresses the behaviour of Singaporean consumers with respect to local food, and their related willingness-to-pay. What makes the paper exceptionally interesting is the context in which the development of a local food system is a top-down process, due to the lack of historical background, and local food production, due to the scarcity of capacities including available land is made practically possible by recent modern technologies only. In general, I believe that this paper has a clear potential to be a very good paper, when the suggested corrections are addressed.

 

Major comments

1. The major shortcoming of the paper is that it currently lacks a clear focus, and though they are related, the topics are too diversified. I would recommend a one idea – one paper approach which might make the paper more focused and more concise.

1. a. Addressing the local and organic character is mixing, while organic issues are practically not introduced in the motivation or the background (nor they are properly discussed later). Perhaps the focus should be directed on local issues only, and discuss organic matters in another paper.

1. b. There are different estimations with and without reasons for purchase. Perhaps the use of an ordered probit/logit model should be considered to analyse consumers who do not, or rarely, or often buy local food, and have one single analysis. (This single framework would allow different model arrangements, e.g. with and without reasons for purchase.) But relying on an existing probit model and focusing on the first or the second approach only (estimation of the general purchasing behaviour vs. inclusion of reasons) could also work, depending on the rethinking of the aims of the paper. Again, what I would recommend is to choose one approach, one framework and stick with that all along the paper.

2. Additionally, having only four categories for first, and eleven ones later is confusing. It is recommended to focus only on the WTP for (non-organic) eggs, vegetables, fish or seafood, and alternative proteins in general, or perhaps the same four categories and their organic counterparts, where applies, when the decision is to keep a local-organic comparison. But I advise against introducing other or more categories than before.

3. While the paper is properly motivated, and the aims are clear, the methods section would benefit from some clarifications at some points (detailed below).

4. More than once I felt that a specific issue is not properly introduced, and some questions remained, though I found answers at a later part of the text. I would recommend to discuss all the relevant questions when an issue is first mentioned. E.g. justification of the choice of the four (eleven) produce; representativeness of the sample, ethnical diversity of Singapore, etc.

5. Comparing the results to that of existing specific literature is missing. E.g. men having a higher WTP for local food might be surprising in the light of current evidence in other contexts. A broader discussion of the role of vegetarianism also appears to be worthwhile.

Minor comments

1. Lines 78-83. Introducing the intention-behaviour gap is recommended here.

2. Line 137. The date of the Feldmann-Hamm paper is not correct. (2015, instead of 2014.)

3. Lines 145-152 and 153-160. The distinction between the two paragraphs is not clear. Unification is recommended.

4. When the WTP for local food is discussed (lines 167-179), the very important paper of Printezis et al (2019) must be acknowledged.

5. Section 3.1. The method of data collection needs to be described more in detail. How were the respondents recruited? Did the sample collection stop after 400 responses (why 400?), what is the response rate, etc. Also, I would recommend to give a hint at this point about the expected representativeness.

6. Table 1. The scale of the answers should be presented in the table, or in the related text (I am aware that the scale is given later in Fig. 1 – explanation at the first mention is recommended). Additionally, it is not clear how the specific items were used to create categories for the analysis.

7. Table 3. Presenting the distribution of frequencies for categorical variables rather than the mean and standard deviation seems more appropriate.

8. Section 4.3. Specific items within categories are compared to one another (e.g. ‘The most acknowledged impact of climate change by the respondents is that of rising sea levels’). However, the statements are not based on statistical analysis, thus the significance of differences in the given sample cannot be taken as granted.

9. The survey questions might be placed in an appendix, instead of inserting some parts in the main text. This would also help making the paper more concise.

10. Line 333-335. It is difficult to judge this statement without knowing how the measure ‘I make an effort to fight climate change’ was calculated (see also minor comment #6). More specifically, the role of vegetarianism is not clear in shaping the final result. See also the next comment.

11. Line 338-340. The conclusion is not justified by the analysis. In this paper the locality itself was directly addressed, thus if non-Chinese people are more willing-to-buy local alternative protein or fish, it is not because they generally BUY more of these items. It is because they PREFER local sources more (perhaps as being vegetarians/pescaterians they are more aware of the potential sources and the role of Singapore in producing such items, or simply, they prefer local produce due its superiority in terms of taste, freshness, etc.).

12. Lines 341-345. The current explanation is not convincing. Is there any evidence for the local food being bought in a bulk is cheaper (and especially cheaper than bulk prices of supermarkets)? Usually bulk price is more associated with supermarkets. Perhaps there is a correlation between household size and cooking for children under 12.

13. Line 345-346. There might be an additional explanation. Is the ratio of vegetarian people evenly distributed among people with different income, or there are more of them among the rich?

14. Table 6. Kindly unify the number of digits shown (this applies to the numbers shown in brackets in Table 4, also to Tables 7-8).

15. The role of vegetarianism should be considered with respect to Table 7, too.

16. Table 8. Changing the arrangement (rows and columns) is not recommended. Kindly use the same logic all along the paper.

17. Kindly discuss the limitations of this analysis.

I really liked this paper, and I wish all the best for the revisions.

 

References

Printezis, I., Grebitus, C., & Hirsch, S. (2019). The price is right!? A meta-regression analysis on willingness to pay for local food. PloS one, 14(5), e0215847.

Author Response

I greatly appreciate all your constructive comments and suggestions. I hope the revisions I have made is satisfactory.

Major comments

1. The major shortcoming of the paper is that it currently lacks a clear focus, and though they are related, the topics are too diversified. I would recommend a one idea – one paper approach which might make the paper more focused and more concise.
   
   Response: Thank you very much for your suggestion. This paper now focuses on the topic of sustainable food consumption and discusses how consumers' environmental perceptions and behvaior affect demand for local produce in Singapore. I have decided to remove the discussion of organic food and vegetarianism and focus on the "local" aspect.
   
   
   
2. a. Addressing the local and organic character is mixing, while organic issues are practically not introduced in the motivation or the background (nor they are properly discussed later). Perhaps the focus should be directed on local issues only, and discuss organic matters in another paper.
   
   Response: Thank you. I have dropped the discussion on organic matters - hence excluded the organic food and alternative protein items in the analysis.
   
   
3. b. There are different estimations with and without reasons for purchase. Perhaps the use of an ordered probit/logit model should be considered to analyse consumers who do not, or rarely, or often buy local food, and have one single analysis. (This single framework would allow different model arrangements, e.g. with and without reasons for purchase.) But relying on an existing probit model and focusing on the first or the second approach only (estimation of the general purchasing behaviour vs. inclusion of reasons) could also work, depending on the rethinking of the aims of the paper. Again, what I would recommend is to choose one approach, one framework and stick with that all along the paper.
   
   Response: Thank you very much for your suggestion. I followed your advice and decided to run a ordered probit by combining the first two estimation on purchase and reasons. 
   
   
4. Additionally, having only four categories for first, and eleven ones later is confusing. It is recommended to focus only on the WTP for (non-organic) eggs, vegetables, fish or seafood, and alternative proteins in general, or perhaps the same four categories and their organic counterparts, where applies, when the decision is to keep a local-organic comparison. But I advise against introducing other or more categories than before.
   
   Response: I have selected only the food items relevant to the discussion point of this paper. As the price point is different for each food item, the analysis still remained for each item.
   
   
   
   
5. While the paper is properly motivated, and the aims are clear, the methods section would benefit from some clarifications at some points (detailed below).
   
   Response: Thank you. I have revised accodringly.
   
   
6. More than once I felt that a specific issue is not properly introduced, and some questions remained, though I found answers at a later part of the text. I would recommend to discuss all the relevant questions when an issue is first mentioned. E.g. justification of the choice of the four (eleven) produce; representativeness of the sample, ethnical diversity of Singapore, etc.
   
   Response: I have included explanation whenever it is first discussed.
   
   
7. Comparing the results to that of existing specific literature is missing. E.g. men having a higher WTP for local food might be surprising in the light of current evidence in other contexts. A broader discussion of the role of vegetarianism also appears to be worthwhile.
   
   Response: I have added more recent relevant papers to provide the background and connected with the findings of this study.

Minor comments

1. Lines 78-83. Introducing the intention-behaviour gap is recommended here.
   Response: Thank you for the advice. I have included it and discussed in the results section as well.
2. Line 137. The date of the Feldmann-Hamm paper is not correct. (2015, instead of 2014.)
   Response: Thank you for pointing it out. Corrected.
3. Lines 145-152 and 153-160. The distinction between the two paragraphs is not clear. Unification is recommended.
   Response: Revised to clarify that first is preference and second is WTP.
4. When the WTP for local food is discussed (lines 167-179), the very important paper of Printezis et al (2019) must be acknowledged.
   Response: Thank you very much for your suggestion. This is a very important paper indeed.
5. Section 3.1. The method of data collection needs to be described more in detail. How were the respondents recruited? Did the sample collection stop after 400 responses (why 400?), what is the response rate, etc. Also, I would recommend to give a hint at this point about the expected representativeness.
   The respondents were selected from panels registered with a vendor company. I have added the description in data section. The sample size is essentially because of the budget limitation.
6. Table 1. The scale of the answers should be presented in the table, or in the related text (I am aware that the scale is given later in Fig. 1 – explanation at the first mention is recommended). Additionally, it is not clear how the specific items were used to create categories for the analysis.
   

   Response: The analysis now uses individual question without grouping them as each item in the same group do not necessarily produce similar results.

7. Table 3. Presenting the distribution of frequencies for categorical variables rather than the mean and standard deviation seems more appropriate.
   Response:  Amended the table.
   
   
8. Section 4.3. Specific items within categories are compared to one another (e.g. ‘The most acknowledged impact of climate change by the respondents is that of rising sea levels’). However, the statements are not based on statistical analysis, thus the significance of differences in the given sample cannot be taken as granted.
   Response: Yes, I understand. This is just the part of descriptive analysis and I hope it is not confusing.
   
   
9. The survey questions might be placed in an appendix, instead of inserting some parts in the main text. This would also help making the paper more concise.
   Response: The most important questions are now placed in data section.
   
   
10. Line 333-335. It is difficult to judge this statement without knowing how the measure ‘I make an effort to fight climate change’ was calculated (see also minor comment #6). More specifically, the role of vegetarianism is not clear in shaping the final result. See also the next comment.
    Response: The analysis now uses each item without grouping them.
    
    
11. Line 338-340. The conclusion is not justified by the analysis. In this paper the locality itself was directly addressed, thus if non-Chinese people are more willing-to-buy local alternative protein or fish, it is not because they generally BUY more of these items. It is because they PREFER local sources more (perhaps as being vegetarians/pescaterians they are more aware of the potential sources and the role of Singapore in producing such items, or simply, they prefer local produce due its superiority in terms of taste, freshness, etc.).
    Response: Yes, I understand. Thank you for your comments. I have revised the section to address your point.
    
    
12. Lines 341-345. The current explanation is not convincing. Is there any evidence for the local food being bought in a bulk is cheaper (and especially cheaper than bulk prices of supermarkets)? Usually bulk price is more associated with supermarkets. Perhaps there is a correlation between household size and cooking for children under 12.
    Response: I have revised the estimation model based on your helpful advice and I hope it is now clearer.
    
    
13. Line 345-346. There might be an additional explanation. Is the ratio of vegetarian people evenly distributed among people with different income, or there are more of them among the rich?
    Response: Such data is not available unfortunately. 
    
    
14. Table 6. Kindly unify the number of digits shown (this applies to the numbers shown in brackets in Table 4, also to Tables 7-8).
    Response: Amended.
    
    
15. The role of vegetarianism should be considered with respect to Table 7, too.
    Response: Thank you. As this paper has narrower focus, I have excluded the vegetarianism discussion.
    
    
16. Table 8. Changing the arrangement (rows and columns) is not recommended. Kindly use the same logic all along the paper.
    Response: Understood and amended.
    
    
17. Kindly discuss the limitations of this analysis.
    Response: I have added in the discussion section.

I really liked this paper, and I wish all the best for the revisions.

Thank you very much. I truly appreciate all lf your helpful comments and advice.

Round 2

Reviewer 1 Report

The paper improved greatly through revision and I could see that the authors addressed all my main concerns. There are only some issues, I would like to see addressed and then the paper will be ready for publication.

1. Please check carefully the text for typos and spelling mistakes (some addressed below). It would be worth using a professional language revision service.

2. Table 7 and 8: As these tables are really big, I would suggest moving it to the supplementary and just writing about the main outcomes and interesting results in the text. Maybe add instead a graph or figure that summarise the results of the table comprehensively.

 

Minor comments:

L70 typo: “to be an important the factor that influences”

L73-74 This sentence is grammatically wrong. Please rephrase.

L197 There might be a formatting issue as there is a S shown before the dollar sign.

L190ff about the local produce in Singapore: This section should not be separated from the rest of the introduction but be included before the end. The introduction should end with a paragraph describing the aim of the study/paper.  

L264 there is a dot before “not willing to pay”

Table 5: maybe highlight the lowest and highest numbers to show what group agrees most in one question.

 

Author Response

Thank you very much for your feedback.

Please find my responses to your comment in the attached document.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear Author,

First, I deeply appreciate your efforts related to the revision; your paper is considerably concise and easy-to-follow now. I also found the extensions of the literature review interesting and relevant. However, from a methodological point of view, what I would recommend is to continue simplification – I am sorry if my related suggestion was not clear in the first review round. I believe that instead of having binary dependent variables, the use of ordinal ones would be more meaningful (this is also the way how you interpreted your findings in the concluding section). Thus you could have 3+7 models altogether (instead of the current 3×3 and 7×3), which could form the columns of Table 7 and 8. It would allow you to directly estimate the importance of a variable in case of an increasing commitment towards consumption (never -> sometimes -> often), and in case of increasing WTP - and add the marginal effects in a separate table only for those variables that prove to be significant, if you think that important. It might happen that some minor details would remain hidden in case of this arrangement; however, major conclusions would easier to be drawn, and if you think so, the detailed models could go in appendices.

Moreover, it would be important to discuss your findings not only in the light of existing statistics (the related part makes your results very convincing), but in the light of previous literature. I agree with you that there is a strong need for context (country-)specific analyses, but it is important to see the factors that are more generalizable than others.

Additionally, I have identified a couple of minor issues, detailed below, that might be worthy of your attention.

Furthermore, though I am not a native English speaker, I have a feeling that the short time that is generally allowed by the journal for revisions prevented the new parts of the text from going through proof-reading. Typos and grammatical mistakes have been identified (e.g. Line 70; Line 74.75; Line 103-104, and several more). Kindly check the language carefully.

 

Minor comments

Data collection: How the initial 7,891 people, whom were invited to participate were identified?

Line 253. A reference remained here for four items, while the revision addresses only three.

Line 262-263. I understand now your point on insisting on the seven different specific products, and I accept your justification. I just recommend to list them here, to give a reader an overview about them. (Simply place here current lines 301-304.)

Table 3. The table can be simplified by merging columns 2 and 3 (or simply eliminating the ‘Variable description’ column). As far as I can see, there is a difference only in the elaboration of the high vs non-high income; the specifics can be added in notes.

Line 330. The word “Singapore” was left out (“Singapore citizens”)

Line 331. “With regards to ethnic composition” – kindly add that these data refer to the sample.

The use of first and third person (‘I’ vs. “we”) is mixed.

Author Response

Thank you very much for all of your helpful comments.

Please find my responses to your comments in the attached document.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

Dear Author,

Congratulations again for your efforts, I think that the merit of your paper is constantly increasing. I have only minor comments now that do not concern methodology or the structure of the paper.

I agree with your solution as to distinguish between three different levels of WTP: none, up to 5% and more than 5% as it is explicitly stated in lines 277-728. However, I would recommend to use this categorization consistently throughout the paper (see lines 310-315 and 428-431; and also Table 6), I think higher level details are not that important for an international readership.

Table 3, variable gender is not specified, thus the readers don’t know which number considers males and females.

In the previous versions of the manuscript you gave only 4 digits in the Tables that give the estimation results. I would recommend to go back to this strategy (or lower the number of digits even more). Furthermore, I can see that the other reviewer suggests simplifications with respect to these tables, and I agree. Perhaps you can show only the significant results, and write n.s.¹ elsewhere (you can publish the whole table – including standard robust errors – in an appendix, if you feel like), thus the table could be simplified greatly. Additionally, please fix the first row of the tables (the titles) as to they should be seen on new pages, for convenience.

Note: ¹: n.s.: not significant

 

The discussion of the results in the light of existing studies still needs to be developed; at present this is the only major remaining shortcoming of the paper. There is a need for specific references in Section 5, too: who found something similar (or dissimilar) things in other contexts?

I think that after addressing these issues, the paper will be ready for publication. Good luck.

Author Response

Please refer to the uploaded file for my responses.

Author Response File: Author Response.pdf