Caveats in Environmental Justice, Consumption and Ecological Footprints: The Relationship and Policy Implications of Socioeconomic Rank and Sustainable Consumption Patterns
Round 1
Reviewer 1 Report
The research regarding sustainable consumption patterns is still the current topic and very interesting from the theoretical as well as practical point of view.
The purpose and scope of the research are well defined. The structure of the article would be fully correct if the part “Conclusions” was not missing.
The part “Introduction” contains the short literature review with justifying the undertaken research topic, defining the environmental justice and explaining the possible relationships between social class and sustainable consumption.
Methods used in the research were detailed described. The information when the research was conducted should be only added.
The obtained results were clearly presented and well interpreted. In my opinion they would be more valuable if they were compared to results from other surveys in different countries.
The last four paragraphs of “Results and Discussion” sound like conclusions containing recommendations about interventions to reduce existing barriers to sustainable consumption announced by Authors at the beginning of the article. As this part is important for the final research outcomes it should be developed, extended and presented in separate subsection “Conclusions”.
It is worth emphasizing that the article is well organised, has a proper flow and it is friendly for the reader.
Author Response
REVIEWER 1:
Comment 1: Reviewer twice suggests that we include a conclusions section which was missing in the original draft.
Response: We agree with the suggestion. In the newly revised draft we have separated the findings from the final section which includes “Discussion and Conclusions”. While we could easily divide this into two sections, we feel that many of the comments in the discussion section lead to conclusions and that it would be inefficient to have to recapitulate them in a separate section.
Comment 2: The information when the research was conducted should be only added.
Response: We agree, The dates during which the survey was conducted (November, 2018) along with several other details about the methodology are now included in the Methods section.
Comment: Referee suggests that we include surveys from other countries.
Response: Despite a rather thorough search, we were unable to find other empirical studies comparable to the present research. However, we have cited four Israeli studies which involve comparable analyses of the different realities across Israel’s socio-economic spectrum. These are now cited (lines: 189-90 2014-15; 232, 247)
Author Response File: 
Author Response.docx
Reviewer 2 Report
This is certainly an interesting topic for readers of this journal. However, there are substantial questions/issues that need to be addressed.
1) The introduction makes a big deal of how consumption drives economic growth and ecological footprints. But, numerous scholars/studies have argued/shown that it really is production decisions (by corporations, industries, etc.) that are key to ecological costs. While consumption is certainly important, consumers generally do not control what the ecological costs are for the production of goods. Industry/corporations could, in many cases, improve their use of advanced technology and use more sustainable practices to reduce the ecological footprint of products and services. At least some discussion of the relative impact of consumers and producers is warranted. (Try a search on google scholar for the "Treadmill of Production" for starters.)
2) The authors do not provide enough information about how the study was conducted.
How were people selected for the surveys and then for the interviews? In other words, what kind of sampling was utilized for each? This is especially important for determining the degree to which the survey results are representative of Israelis overall. The authors discuss their results as if they are representative, but the reader is given no basis for judging this. What was the overall response rate to the survey? How were the surveys conducted? Were these done online, in person, by mail? More information is needed about the specific measures used.
3) Overall, there needs to be more focus here. The article spends a great deal of time detailing differences across the deciles but it does not seem that much is accomplished beyond that. The authors make reference to hypotheses, but these are never clearly presented.
Author Response
REVIEWER 2:
Comment 1: “At least some discussion of the relative impact of consumers and producers is warranted. (Try a search on google scholar for the "Treadmill of Production" for starters.).”
Response: Thank you for the excellent suggestion. In the discussion section, we have added an entire paragraph which considers this matter and responds to the social critique expressed in the “Treadmill of Production” model and subsequent supporting positions. (See lines: 406-413)
Comment 2: The authors do not provide enough information about how the study was conducted.
Response: We agree. A new paragraph was added (line 130-143) to the methods section which answers the many questions about the way the survey was conducted, including interviewee selection process, response rates, etc.
Comment 3: “The authors make reference to hypotheses, but these are never clearly presented.”
Response: Although the original draft does refer to the underlying hypotheses behind the survey, we agree that this should have been made clearer at the outset. Accordingly, the research hypothesis is now stated clearly in lines 42-45 at the opening of the article.
Author Response File: Author Response.docx
Reviewer 3 Report
General comments
The topic covered an important issue relevant to the journal. However, the way the paper presented the methods and results does not meet the standard level of an academic article. The authors should be aware that any scientific journal needs to present their methods and results in a precise way so that the readers can replicate the results after reading the paper. Although literature review and conclusion section are one of the important components of an academic article, the paper lacked these materials. They should clarify how the current study performed is different from relevant studies and conclusions should be separated out from the results. Although the topic of the paper is to examine the difference in the ecological footprints among the social income groups, the study did not clarify how the values for the ecological footprint (EF) were calculated. The authors did mention that the EF was based on previous studies, but they did not explain how the EF was numerically estimated in these previous studies. Hence, the readers cannot see if the results discussed in the paper are correct. The paper claims that it used statistical analysis such as ANOVA and Person test, but it did not show these results. The results should be summarized in a table.Minor comments
The reference numbers should be placed in square brackets. Please check the author guideline. 2, Ln82. "Of" should be "of." 2 Ln86-87. A reference for the national survey conducted about consumption patterns in Israel should be provided so that the readers can find out which survey the authors are referring to. The reference style of some of the references must be revised to meet the style of the journal. For example, commas were missing after the year and volume number. Please check the author guideline.Author Response
Reviewer 3
Comment 1: methods and results section need to be expanded and improved.
Response: This comment is similar to that of the previous reviewers. As described above, these sections were expanded and improved.
Comment 2: Need to add a “conclusions” and a “literature review” section.
Response: As mentioned for reviewer 1, the paper now contains a discussion and conclusions section. A literature review has also been added (lines 52-113).
Comment 3: Authors need to clarify how the current study is different from other relevant studies.
Response: We agree, and have now written a paragraph (found in lines 103-118) which fully differentiates the present research from earlier studies on this topic.
Comment 4: Authors have not sufficiently clarified how the ecological footprints were calculated.
Response: In the early part of the findings section (lines 158-164) we describe our calculations as being based on a long list of relevant research as cited in footnotes 37-42. For every category of consumer products, there were at least three different previous surveys characterizing their ecological footprint. As part of the preparation for the analysis, we aggregated these numbers and were able to rank the relative size of the environmental footprints. With a few additional emphases, we believe that the calculation methods are now clear and replicable, based on the extensive citations.
Comment 5:
The paper claims that it used statistical analysis such as ANOVA and Person test, but it did not show these results. The results should be summarized in a table.
Response: In response to this comment, we have now added five additional tables (Table 1-5) which details the results of the Person tests and to a lesser extent, the ANOVA analysis.
General editorial comments
A) The reference numbers should be placed in square brackets. Please check the author guideline. 2, Ln82. "
Response: Done
B) Of" should be "of." 2 Ln86-87.
Response: Done (Due to previous reivsions in the manusxript, this line is now numbered 97.)
C) A reference for the national survey conducted about consumption patterns in Israel should be provided so that the readers can find out which survey the authors are referring to.
Response: Done (Reference number: 50)
D) The reference style of some of the references must be revised to meet the style of the journal. For example, commas were missing after the year and volume number. Please check the author guideline.
Done. See endnote 44.
Author Response File: Author Response.docx
Reviewer 4 Report
The paper aims at disentangling a paradox: economically affluent households on average follow more sustainable consumption patterns than households in disadvantaged social strata, while cheaper consumer goods often have a greater ecological footprint than more expensive alternatives. To avoid generalization, the research provides empirical evidence documenting the association between socioeconomic clusters and sustainable consumption and proposes policy measures to reduce existing barriers to sustainable consumption.
While the paper addresses an interesting and original topic that might result of great interest to the readers, significance of content, quality of presentation and scientific soundness need further improvement.
Concerning the paragraph "Materials and methods": i) it is not specified why the four industries selected (and online purchases) are considered the most polluting; ii) it is not specified if the 600 questionnaires were distributed randomly or according to some stratification criterion; iii) a section might be added to illustrate the main features of the dataset built; iv) it is not specified how the independent variable (socioeconomic ranking) was obtained by the five items collected; v) it is not specified how, using cluster analysis, the five dependent variables have been obtained;vi) a brief explanation of the method used to obtain the individual ecological footprint and a graph/table of the distribution among socioeconomic deciles of the ecological footprint should be added; vii) results of ANOVA and of Pearson test should be presented in a table and commented; viii) in the qualitative analysis how respondents have been selected?
Concerning the paragraph "Results and discussion": i) "in the food sector, no overall ecological footprint index, emerged" should be rephrased as the meaning is not clear ii) Figure 1 does not clearly illustrate data: vegans and vegetarians might be represented on secondary axis, or the graph might be splitted in two graphs (one for Kosher and omnivores, one for vegan and vegetarians) iii) the quality of figure 2 should be improved; iv) when correlations are discussed, estimates should be provided at least in appendix; v) Figure 3 is difficult to interpret, it could be substituted with a table; vi) Figure 4 is difficult to interpret, it could be substituted with a table; vii) ANOVA commented at lines 215-222 is not reported a table should be added; viii) similarly, considerations illustrated at lines 223-229 are not supported by data (data might be added in a table); ix) figure 5 is difficult to interpret, should be modified (trends should be eliminated); x) data on "environmental involvement" and price concerns should be added (this is a potentially original part of the paper); xi) the part after line 269 might be renamed "concluding remarks".
Concerning Figures, as data concern the deciles of population, a useful tool of representation of different behaviour is the Lorenz curve. Authors might use a mix of Lorenz curves and Gini coefficients to represent most of the issues discussed.
Due to the wide amount of data commented, a data Appendix might be added to facilitate the comprehension of the results discussed.
Author Response
Reviewer 4:
Comment: 1
Concerning the paragraph "Materials and methods": i) it is not specified why the four industries selected (and online purchases) are considered the most polluting;
Response: We disagree. Reviewer may not have read these lines closely. Footnotes 29-33 on lines 118-120 provide the requested documentation. While there is a debate in the literature over which industry has the highest environmental impacts, there is support by different experts for a dominant role played each of the four industries selected – each singled out as having “the heaviest, global ecological footprint”. At the same time, we have not found any experts who challenge that these four industries, globally account for the greatest ecological impact. In our own analysis we refrained from ranking these four industries, given the fact that there is no consensus with regards to the relative magnitudes of their environmental impacts.
Comment 2: it is not specified if the 600 questionnaires were distributed randomly or according to some stratification criterion;
Response: We agree. In the new paragraphs added to the methods section in lines (144-147) we go into detail with regards to the different criteria which drove the stratification / selection process.
Comment 3: a section might be added to illustrate the main features of dataset built;
Response: This comment has theoretical validity. But the requested dataset is enormous – and even a table summarizing the different components would be unwieldly for this kind of publication. If editors feel that a table with this information is indeed critical, we are happy to provide one. Yet, in our opinion it will lengthen the article considerable making minimal contribution to the actual readability and validity of the article.
Comment 4: it is not specified how the independent variable (socioeconomic ranking) was obtained by the five items collected:
Response: We believe that lines 151-155 include this information. We paste these lines here for your convenience:
“First, the independent variable, identifying individual socioeconomic ranking, was constructed based on the variables of education, residential district (periphery vs. center), individual income, household Income, receipt of welfare payments and the extent of position and seniority in Israel. These components were drawn from the socioeconomic rating of the government’s Central Bureau of Statistics (CBS). “
Comment 5: it is not specified how, using cluster analysis, the five dependent variables have been obtained;
Response: Agreed. We have now added two sentences (lines 159-162) in order to provide more information and answer this question.
Comment 6: a brief explanation of the method used to obtain the individual ecological footprint and a graph/table of the distribution among socioeconomic deciles of the ecological footprint should be added;
Response: Agreed. See our response to comment 4 of Reviewer 3 above.
vii) results of ANOVA and of Pearson test should be presented in a table and commented;
Response: We agree. See newly added tables 1-5.
Comment 7: in the qualitative analysis how respondents have been selected?
Response: We agree. We have added a description of the selection process, to ber found in lines 147-150.
Concerning the paragraph "Results and discussion":
Comment 8: i-vii
"in the food sector, no overall ecological footprint index, emerged" should be rephrased as the meaning is not clearResponse: Agreed. Sentence modified (see line 194-5)
Figure 1 does not clearly illustrate data: vegans and vegetarians might be represented on secondary axis, or the graph might be splitted in two graphs (one for Kosher and omnivores, one for vegan and vegetarians)Response: We have conferred with colleagues and believe that the present figure is clear – and that dividing into two separate graphs would not provide any real benefits in terms of conveying the results regarding the different dietary patterns according to socio-economic status.
the quality of figure 2 should be improved;
Response: We agree. Graph quality has been upgraded.
when correlations are discussed, estimates should be provided at least in appendix;
RESPONSE: We are doubtful that readers will be interested in the minutiae of correlations between the many variables. If editor feels that this is important, we can prepare such an appendix.
Figure 3 is difficult to interpret, it could be substituted with a table;RESPONSE: We showed colleagues a table versus a graph. Because of the many variables, there was a consensus that the information is better conveyed via a graph. In order to emphasize the point of figure 3 – we have added a new figure 4, which shows how there is no real difference between socio-economic groups in paramount importance attributed to price in making consumer decisions.
Figure 4 is difficult to interpret, it could be substituted with a table;
RESPONSE: Our response here is the same as for comment v: we actually think that the present graph is clearer than a table.
ANOVA commented at lines 215-222 is not reported a table should be added; viii) similarly, considerations illustrated at lines 223-229 are not supported by data (data might be added in a table); ix) figure 5 is difficult to interpret, should be modified (trends should be eliminated); x) data on "environmental involvement" and price concerns should be added (this is a potentially original part of the paper); xi) the part after line 269 might be renamed "concluding remarks".
RESPONSE: The newly added tables 1-5 – responds favorably to these suggestions and makes the requested revisions.
Concerning Figures, as data concern the deciles of population, a useful tool of representation of different behaviour is the Lorenz curve. Authors might use a mix of Lorenz curves and Gini coefficients to represent most of the issues discussed.
RESPONSE: A Lorenz curve and its implications for Israeli society are included now in the introductory section (lines: 73-85). We believe that this is sufficient. Going beyond this would add significantly to the length of the paper, without making a meaningful improvement in the amount of the information conveyed.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
This manuscript is improved. The overall organization of the manuscript could be tightened--it seems a little haphazard in parts. Editing is required.
I appreciate the greater discussion of how the surveys were conducted.
It would also help to have some idea of how the panels are generally recruited initially by Midgam Panel--this helps with determining just how representative these panels, from which the respondents were chosen, are. What the authors report as "response rate" may more accurately be labeled a "completion rate." (If a total of 700 were invited and only 600 participated, then, yes, this would be a response rate. However, it isn't clear in the text; we are only told that the additional 100 people started the survey and did not finish.) It sounds like the sample was selected from the panel using a "disproportionate stratified sampling" technique. Members of deciles were not chosen based on their proportion in the overall population, rather the same number were chosen from each decile. Technically, this means that the sample overall is not "nationally representative," rather the sample is representative of each decile. (It may still be fine, given this, to call the study "a representative national survey" in the abstract...)I understand why the tables were added, but can they be combined or perhaps made more succinct? The information is overwhelming; it is also unclear how these results led to the overall conclusions made by the authors. (Also, there are only correlations presented; weren't ANOVA results also supposed to be included? Aren't "r" from Table 1 and "Pearson correlation" from the other tables the same thing?)
I appreciate the inclusion of the "Treadmill of Production." The author's name is Schnaiberg (it is spelled incorrectly in the text and in the references).
Author Response
December 17, 2019
To the Editors:
Here are our responses to the second round of comments from three reviewers – along with the amended article (in which our second round of changes appear in red letters).
Alon Tal and Meital Peleg- Mizrahi
Reviewer 1:
This manuscript is improved. The overall organization of the manuscript could be tightened--it seems a little haphazard in parts. Editing is required.
I appreciate the greater discussion of how the surveys were conducted.
It would also help to have some idea of how the panels are generally recruited initially by Midgam Panel--this helps with determining just how representative these panels, from which the respondents were chosen, are. What the authors report as "response rate" may more accurately be labeled a "completion rate." (If a total of 700 were invited and only 600 participated, then, yes, this would be a response rate. However, it isn't clear in the text; we are only told that the additional 100 people started the survey and did not finish.) It sounds like the sample was selected from the panel using a "disproportionate stratified sampling" technique. Members of deciles were not chosen based on their proportion in the overall population, rather the same number were chosen from each decile. Technically, this means that the sample overall is not "nationally representative," rather the sample is representative of each decile. (It may still be fine, given this, to call the study "a representative national survey" in the abstract...)
Response
We agree completely. On 144-148 of the revised MS, we now offer a description of how the panelists were recruited and why we decided to ensure equal representation of each socio-economic decile, even though this created over representation of some deciles (and under-representation of others). That’s because the socio-economic decile sizes are not identical but based on income ranges. As suggested, in the abstract, we have changed “nationally representative” to the following sentence: “A national survey, which ensured equal representation of of socio-economic deciles, involving 600 respondents….:
Reviewer 1:
I understand why the tables were added, but can they be combined or perhaps made more succinct? The information is overwhelming; it is also unclear how these results led to the overall conclusions made by the authors. (Also, there are only correlations presented; weren't ANOVA results also supposed to be included? Aren't "r" from Table 1 and "Pearson correlation" from the other tables the same thing?
Response:
These five tables were added in response to the comment of a referee in the previous round of reviews. In its present position, it seems a bit ponderous and “clunky”. We are very happy moving this table, which contains considerable raw data, to an Appendix – in this case, Appendix 2. We have changed the tables in response to the correct observation by the reviewer. Now instead of r2, you will find the Pearson correlation, which is a more precise parameter for presenting these unrefined results.
Reviewer 1:
I appreciate the inclusion of the "Treadmill of Production." The author's name is Schnaiberg (it is spelled incorrectly in the text and in the references).
Response:
Error corrected. Thank you for pointing this out.
Reviewer 2:
The paper does not explain well about the objective of the study.Response
A few sentences were added to final paragraph of the introduction (lines 50-55) which now succinctly set forward the study objectives.
Reviewer 2:
The paper lacks to provide enough evidence to support its significance of the study. It fails to discuss what results can be derived from the study and why their expected findings will be valuable for the readers.
Response
We feel that there is already considerable discussion in the closing paragraphs about the study’s significance. Nonetheless, in light of this comment we have expanded these passages on lines 437-441 to better elucidate the reason why the study matters and what decision makers might take away from the findings.
Reviewer 2:
The authors still did not explain well about how the ecological footprint used in the study is obtained. It is not possible for the reader of the paper to replicate your study by just having the following information. The authors should keep in mind that a scientific paper should be somewhat reproducible by following the study methods."These levels of ecological footprint were determined based on previous studies in each consumer 156 category [38, 39, 40, 41, 32, 42, 43, 44, 45, 46, 47], each product and purchase pattern included on the 157 listed contained in the questionnaire was assigned a numerical value (based on the rankings found 158 in the aforementioned studies). This factor was then multiplied by the amount consumed or the 159 frequency of consumption. The"
Response
We recognize now, due to this and other comments of reviewers that in order to ensure the replicability of this study, we need to provide more detailed information to readers about how the ecological footprints were derived. Accordingly, a brand new Appendix 1 is now added which contains a detailed table which shows the basis for the calculations of ecological footprint by consumption group. The table is extremely long, reflecting the complexity of the calculation process, but we believe that including it will enable subsequent researchers to duplicate our methodology and follow our calculations and conclusions with relative ease.
Reviewer 3:
Comment 1:
Of course authors are free to disagree, this is not a substantive issue and authors selected relevant industries among the most polluting. Anyway, most readers will not have time to study all references, and dubts will remain on the exaustiveness of the choice (i.e. there might be other and even more polluting industries especially on the production side. Therefore, it might be useful to restrict the field of analysis on most polluting consumption patterns and, if possible, to explicitly motivate the selection made).
Response
Selection of the different consumption categories was made after conferring with experts – local and international and based on empirical findings which are meticulously documented in the article. We feel the article already assesses what the reviewer calls “the most polluting consumption patterns”. Nonetheless, we understand from Reviewer 3’s comments that this may not be entirely clear. Accordingly, we have now added a few sentences to emphasize why these four industries were selected. See line 122-127.
Comment 2: ok, thank you.
Comment 3: I agree with authors that a full representation of raw data might be pointless, but I was asking for descriptive statistics, at least for the most relevant items. One table illustrating descriptive statistics and one page of comments might facilitate the comprehension of results and discussion.
Review
In light of the numerous graphs which were added as a result of the previous round of reviews, we question whether the article now needs to also include over two pages of tables with data, especially seeing as they are already aggregated and displayed in the graphs displayed in figures 1-6. If editors feel that this information in numeric form is essential, than we would argue that this should be included as an appendix at the end of the article rather than a table in the heart of it. But, again, given the new Table 1 – a ten page representation of raw data – we feel that this request is now less germane.
Comment 4: a qualitative description is provided in the article, but I was asking for a more formal specification. As an example, by reading lines 151-155 it is not clear wheather the authors implemented a regression model or referred to a composite index to compute the independent variabile. More quantitative information should be provided, including also a table (maybe in Appendix) illustrating results of this intermediate step.
Response
We agree that specific mention of the regression analysis that was conducted should be included in the body of the article. Accordingly, we have added a sentence on lines 154 as well as on 159 to explain to readers what are the independent and dependent variables in the regression analysis.
Moreover, the new Appendix 1 provides ten pages of data which is surely the kind of quantitative information that reviewer is asking for us to provide.
Comment 5: same as comment 4. A table should be added illustrating the main output of cluster analysis.
Response
At present, the presentation of the way we conducted the cluster analysis appears on line 170. This seems from the original draft to be highly intuitive -- but now it is even clearer, based on the new Table of the Sub-variables included in Appendix 1. Therefore we believe providing an additional, probably incomprehensible table inside an article which is already a bit busy with tables will do little to add to readers understanding of the cluster analysis process. The formation of the country’s socio-economic deciles is a complicated process undertaken by the Central of Statistics, a national government agency and is generally considered to be a reliable categorization. Accordingly, we would respectfully suggest that yet another table not be added to an article. which is becoming dominated by tables and figures – at the price of brevity and comprehension.
Comment 6: in my opinion, the method used to obtain the ecological footprint still lacks of clarity. If the authors believe that a detailed analysis might go beyond the scope of the paper, they might at least add a short methodological note in Appendix. Furthermore, Tables 1-5 should be formatted according to the editorial standards and moved to appendices.
Response
Again, by including a ten page Appendix (see response to Reviewer1), we believe that we have answered Reviewer 3’s very legitimate concern. With regards to Reviewer 3’s suggestion that Tables 1-5 be transformed into Appendices – we agree. As mentioned above – in response to Reviewer 1, these tables are now in the Appendixes.
Comment 7: ok, thank you.
Comment 8:
ok, thank you.Reviewer ii) as results and discussion are mostly based on graphical representation of descriptive statistics, authors are invited to refine the graphs included in the paper. Provided that it is not a mistake to represent data as in figure 1, I will try to better explain my concerns.
Concentration curves (Lorenz curves) are mostly used to draw income distributions, but can be also used to represent the concentration (i.e. the unequal distribution) of any phenomenon under inquiry. I did not suggest to add a Lorenz curve for Israel's income distribution. I suggested, instead, to use concentration (Lorenz) curves to represent the unequal distribution of the items illustrated in figure 1, and 6 of the revised version.
If authors anyway believe that an histogram might better represent data, I suggest at least to measure on secondary axis vegans and vegetarians using a more appropriate interval (as an example, min = 0, max = 5).
Response:
We indeed do believe that in the present case, histograms are sufficient and that the presentation of the data benefits from the relative simplicity that they offer.
With regards to adding a revised presentation of vegans and vegetarians, we direct your attention to figure 2 which already includes histograms showing the frequency of vegans and vegetarians among the ten different deciles. (The lack of a clear trend – and the prevalence of vegans/vegetarians across poor and wealthy socioeconomic groups alike emerges quite clearly from this figure.)
So we believe that it would be best to leave this graph “as is”.
Reviewer iii) Ok, thank you, now the graph provides a clear representation of the association between frequency of shopping and economic deciles.
Reviewer iv) If the editor agree, I suggest to add an appendix with correlations not included in tables 1-5.
Response:
It is not clear to us which correlations Reviewer 3 is referring to. There are now six new tables in appendices – one ten pages long and we aren’t certain that another one is critical to improving the presentation of the data.
Reviewer v) Figure 4 better represents the dominance of price concerns over environmental considerations, therefore I suggest to delete figure 3 and keep only figure 4. In figure 3, authors are invited to consider an alternative representation based on an histogram instead of a continuous line and to remove the trend.
Response:
We agree and have removed figure 3 as suggested.
Reviewer vi) Figure 5 (figure 4 in the initial draft) has an incomplete legend (two labels are missing). Furthermore, authors are invited to consider an alternative representation based on a concentration curve.
Response
True. How did we miss that? Thank you for the heads up! This is now corrected. (Legend now contains all eight parameters.)
Reviewer vii) Authors are invited to consider an alterative representation based on concentration curves. If they decide to keep the histogram, authors should at least delete the trends (the dotted lines).
Response
As suggested, we have deleted the trends.
Reviewer viii) As already pointed out, I did not ask to add a Lorenz curve of Israel's income distribution.
Response
Yes, but we feel it is important background information for people unfamiliar with income inequality in Israel. So we are leaving it in.
Other comments
Alon
Reviewer - Lines 132-138: the number of distributed and selected questionnaires seems approximated to round numbers. Are 700 and 600 the real number of distributed questionnaires and complete answers collected or are they approximate values? In the second case, words as "around" or "almost" should be added to clarify that this is not an exact measure.
Response:
Yes, the numbers are precisely 700 and 600 as written down in the text, something made possible by the panel sampling methodology. There is no need to add “around” or “almost” as suggested – as in this case the correct adjective would be “exactly”.
Reviewer lines 139-145:
it seems that questionnaires are equally distributed among deciles, i.e. there are 60 questionnaires for each decile. Is it true or this is a round number? In the second case a preposition like "around" or an adverb like "almost" should be added.
Response
Yes it is true that 60 questionnaires were collected from each decile- as the study design dictated. Again, this preclude the need to add words like “almost” and “around”.
Reviewer lines 233-241:
as the average reader might not know what territories are peripheral and what territories are not in Israel, a short note might be added in appendix to illustrate the classification of territories.
Response:
We are not certain that readers need a thorough geographical explanation, especially as the article is starting to exceed reasonable size limits. But nonetheless, we have added a footnote on the methodology page which describes the four “peripheral” geographic zones used to categorize respondents and the reduced opportunities for residents living in them. If editor feels that this footnote unnecessary, we’re happy for you to remove.
Reviewer lines 316-341:
It seems that this part of "results" paragraph illustrates the major findings of the qualitative analysis. As they seem potentially interesting and original, authors might add further comments and a graph.
Response:
We disagree. The nature of the “qualitative analysis” presented in the results paragraphs does not lend itself to graphical representation. We have added additional comments, about the meaning of the findings, as requested by Review 1. We hope that this will be sufficient.
general comment on "results" paragraph: as results are based on surveyed data, authors might discuss major findings using more moderation. As an example, rather than "confirming" hypotheses, statistical evidence based on surveyed data "support" or "suggest the validity" of the hypotheses discussed at least of Israel.
Response:
We agree and now use a more guarded tone – and have changed a bit of the text accordingly, as suggested by the reviewer.
Author Response File: Author Response.docx
Reviewer 3 Report
1. The paper does not explain well about the objective of the study.
2. The paper lacks to provide enough evidence to support its significance of the study. It fails to discuss what results can be derived from the study and why their expected findings will be valuable for the readers.
3. The authors still did not explain well about how the ecological footprint used in the study is obtained. It is not possible for the reader of the paper to replicate your study by just having the following information. The authors should keep in mind that a scientific paper should be somewhat reproducible by following the study methods.
"These levels of ecological footprint were determined based on previous studies in each consumer 156 category [38, 39, 40, 41, 32, 42, 43, 44, 45, 46, 47], each product and purchase pattern included on the 157 listed contained in the questionnaire was assigned a numerical value (based on the rankings found 158 in the aforementioned studies). This factor was then multiplied by the amount consumed or the 159 frequency of consumption. The"
Author Response
Comment 1
the paper does not explain well about the objective of the study.
Response
A few sentences were added to final paragraph of the introduction (lines 50-55) which now succinctly set forward the study objectives.
Comment 2
The paper lacks to provide enough evidence to support its significance of the study. It fails to discuss what results can be derived from the study and why their expected findings will be valuable for the readers.
Response
We feel that there is already considerable discussion in the closing paragraphs about the study’s significance. Nonetheless, in light of this comment we have expanded these passages on lines 437-441 to better elucidate the reason why the study matters and what decision makers might take away from the findings.
Comment 3
The authors still did not explain well about how the ecological footprint used in the study is obtained. It is not possible for the reader of the paper to replicate your study by just having the following information. The authors should keep in mind that a scientific paper should be somewhat reproducible by following the study methods.
"These levels of ecological footprint were determined based on previous studies in each consumer 156 category [38, 39, 40, 41, 32, 42, 43, 44, 45, 46, 47], each product and purchase pattern included on the 157 listed contained in the questionnaire was assigned a numerical value (based on the rankings found 158 in the aforementioned studies). This factor was then multiplied by the amount consumed or the 159 frequency of consumption. The"
Response
We recognize now, due to this and other comments of reviewers that in order to ensure the replicability of this study, we need to provide more detailed information to readers about how the ecological footprints were derived. Accordingly, a brand new Appendix 1 is now added which contains a detailed table which shows the basis for the calculations of ecological footprint by consumption group. The table is extremely long, reflecting the complexity of the calculation process, but we believe that including it will enable subsequent researchers to duplicate our methodology and follow our calculations and conclusions with relative ease.
Reviewer 4 Report
In this brief report I will initially reply to the comments included in the cover letter, then I will add other comments.
Comment 1: Of course authors are free to disagree, this is not a substantive issue and authors selected relevant industries among the most polluting. Anyway, most readers will not have time to study all references, and dubts will remain on the exaustiveness of the choice (i.e. there might be other and even more polluting industries especially on the production side. Therefore, it might be useful to restrict the field of analysis on most polluting consumption patterns and, if possible, to explicitly motivate the selection made).
Comment 2: ok, thank you.
Comment 3: I agree with authors that a full representation of raw data might be pointless, but I was asking for descriptive statistics, at least for the most relevant items. One table illustrating descriptive statistics and one page of comments might facilitate the comprehension of results and discussion.
Comment 4: a qualitative description is provided in the article, but I was asking for a more formal specification. As an example, by reading lines 151-155 it is not clear wheather the authors implemented a regression model or referred to a composite index to compute the independent variabile. More quantitative information should be provided, including also a table (maybe in Appendix) illustrating results of this intermediate step.
Comment 5: same as comment 4. A table should be added illustrating the main output of cluster analysis.
Comment 6: in my opinion, the method used to obtain the ecological footprint still lacks of clarity. If the authors believe that a detailed analysis might go beyond the scope of the paper, they might at least add a short methodological note in Appendix. Furthermore, Tables 1-5 should be formatted according to the editorial standards and moved to appendix.
Comment 7: ok, thank you.
Comment 8:
i) ok, thank you.
ii) as results and discussion are mostly based on graphical representation of descriptive statistics, authors are invited to refine the graphs included in the paper. Provided that it is not a mistake to represent data as in figure 1, I will try to better explain my concerns.
Concentration curves (Lorenz curves) are mostly used to draw income distributions, but can be also used to represent the concentration (i.e. the unequal distribution) of any phenomenon under inquiry. I did not suggest to add a Lorenz curve for Israel's income distribution. I suggested, instead, to use concentration (Lorenz) curves to represent the unequal distribution of the items illustrated in figure 1, 5 and 6 of the revised version.
If authors anyway believe that an histogram might better represent data, I suggest at least to measure on secondary axis vegans and vegetarians using a more appropriate interval (as an example, min = 0, max = 5).
iii) Ok, thank you, now the graph provides a clear representation of the association between frequency of shopping and economic deciles.
iv) If the editor agree, I suggest to add an appendix with correlations not included in tables 1-5.
v) Figure 4 better represents the dominance of price concerns over environmental considerations, therefore I suggest to delete figure 3 and keep only figure 4. In figure 3, authors are invited to consider an alternative representation based on an histogram instead of a continuous line and to remove the trend.
vi) Figure 5 (figure 4 in the initial draft) has an incomplete legend (two labels are missing). Furthermore, authors are invited to consider an alternative representation based on a concentration curve.
vii) Authors are invited to consider an alterative representation based on concentration curves. If they decide to keep the histogram, authors should at least delete the trends (the dotted lines).
viii) As already pointed out, I did not ask to add a Lorenz curve of Israel's income distribution.
Other comments
Lines 132-138: the number of distributed and selected questionnaires seems approximated to round numbers. Are 700 and 600 the real number of distributed questionnaires and complete answers collected or are they approximate values? In the second case, words as "around" or "almost" should be added to clarify that this is not an exact measure.
lines 139-145: it seems that questionnaires are equally distributed among deciles, i.e. there are 60 questionnaires for each decile. Is it true or this is a round number? In the second case a preposition like "around" or an adverb like "almost" should be added.
lines 233-241: as the average reader might not know what territories are peripheral and what territories are not in Israel, a short note might be added in appendix to illustrate the classification of territories.
lines 316-341: It seems that this part of "results" paragraph illustrates the major findings of the qualitative analysis. As they seem potentially interesting and original, authors might add further comments and a graph.
general comment on "results" paragraph: as results are based on surveyed data, authors might discuss major findings using more moderation. As an example, rather than "confirming" hypotheses, statistical evidence based on surveyed data "support" or "suggest the validity" of the hypotheses discussed at least of Israel.
Comments for author File: 
Comments.pdf
Author Response
Comment 1:
Of course authors are free to disagree, this is not a substantive issue and authors selected relevant industries among the most polluting. Anyway, most readers will not have time to study all references, and dubts will remain on the exaustiveness of the choice (i.e. there might be other and even more polluting industries especially on the production side. Therefore, it might be useful to restrict the field of analysis on most polluting consumption patterns and, if possible, to explicitly motivate the selection made).
Response
Selection of the different consumption categories was made after conferring with experts – local and international and based on empirical findings which are meticulously documented in the article. We feel the article already assesses what the reviewer calls “the most polluting consumption patterns”. Nonetheless, we understand from Reviewer 3’s comments that this may not be entirely clear. Accordingly, we have now added a few sentences to emphasize why these four industries were selected. See line 122-127.
Comment 2: ok, thank you.
Comment 3: I agree with authors that a full representation of raw data might be pointless, but I was asking for descriptive statistics, at least for the most relevant items. One table illustrating descriptive statistics and one page of comments might facilitate the comprehension of results and discussion.
Review
In light of the numerous graphs which were added as a result of the previous round of reviews, we question whether the article now needs to also include over two pages of tables with data, especially seeing as they are already aggregated and displayed in the graphs displayed in figures 1-6. If editors feel that this information in numeric form is essential, than we would argue that this should be included as an appendix at the end of the article rather than a table in the heart of it. But, again, given the new Table 1 – a ten page representation of raw data – we feel that this request is now less germane.
Comment 4: a qualitative description is provided in the article, but I was asking for a more formal specification. As an example, by reading lines 151-155 it is not clear wheather the authors implemented a regression model or referred to a composite index to compute the independent variabile. More quantitative information should be provided, including also a table (maybe in Appendix) illustrating results of this intermediate step.
Response
We agree that specific mention of the regression analysis that was conducted should be included in the body of the article. Accordingly, we have added a sentence on lines 154 as well as on 159 to explain to readers what are the independent and dependent variables in the regression analysis.
Moreover, the new Appendix 1 provides ten pages of data which is surely the kind of quantitative information that reviewer is asking for us to provide.
Comment 5: same as comment 4. A table should be added illustrating the main output of cluster analysis.
Response
At present, the presentation of the way we conducted the cluster analysis appears on line 170. This seems from the original draft to be highly intuitive -- but now it is even clearer, based on the new Table of the Sub-variables included in Appendix 1. Therefore we believe providing an additional, probably incomprehensible table inside an article which is already a bit busy with tables will do little to add to readers understanding of the cluster analysis process. The formation of the country’s socio-economic deciles is a complicated process undertaken by the Central of Statistics, a national government agency and is generally considered to be a reliable categorization. Accordingly, we would respectfully suggest that yet another table not be added to an article. which is becoming dominated by tables and figures – at the price of brevity and comprehension.
Comment 6: in my opinion, the method used to obtain the ecological footprint still lacks of clarity. If the authors believe that a detailed analysis might go beyond the scope of the paper, they might at least add a short methodological note in Appendix. Furthermore, Tables 1-5 should be formatted according to the editorial standards and moved to appendices.
Response
Again, by including a ten page Appendix (see response to Reviewer1), we believe that we have answered Reviewer 3’s very legitimate concern. With regards to Reviewer 3’s suggestion that Tables 1-5 be transformed into Appendices – we agree. As mentioned above – in response to Reviewer 1, these tables are now in the Appendixes.
Comment 7: ok, thank you.
Comment 8:
ok, thank you.Reviewer ii) as results and discussion are mostly based on graphical representation of descriptive statistics, authors are invited to refine the graphs included in the paper. Provided that it is not a mistake to represent data as in figure 1, I will try to better explain my concerns.
Concentration curves (Lorenz curves) are mostly used to draw income distributions, but can be also used to represent the concentration (i.e. the unequal distribution) of any phenomenon under inquiry. I did not suggest to add a Lorenz curve for Israel's income distribution. I suggested, instead, to use concentration (Lorenz) curves to represent the unequal distribution of the items illustrated in figure 1, and 6 of the revised version.
If authors anyway believe that an histogram might better represent data, I suggest at least to measure on secondary axis vegans and vegetarians using a more appropriate interval (as an example, min = 0, max = 5).
Response:
We indeed do believe that in the present case, histograms are sufficient and that the presentation of the data benefits from the relative simplicity that they offer.
With regards to adding a revised presentation of vegans and vegetarians, we direct your attention to figure 2 which already includes histograms showing the frequency of vegans and vegetarians among the ten different deciles. (The lack of a clear trend – and the prevalence of vegans/vegetarians across poor and wealthy socioeconomic groups alike emerges quite clearly from this figure.)
So we believe that it would be best to leave this graph “as is”.
Reviewer iii) Ok, thank you, now the graph provides a clear representation of the association between frequency of shopping and economic deciles.
Reviewer iv) If the editor agree, I suggest to add an appendix with correlations not included in tables 1-5.
Response:
It is not clear to us which correlations Reviewer 3 is referring to. There are now six new tables in appendices – one ten pages long and we aren’t certain that another one is critical to improving the presentation of the data.
Reviewer v) Figure 4 better represents the dominance of price concerns over environmental considerations, therefore I suggest to delete figure 3 and keep only figure 4. In figure 3, authors are invited to consider an alternative representation based on an histogram instead of a continuous line and to remove the trend.
Response:
We agree and have removed figure 3 as suggested.
Reviewer vi) Figure 5 (figure 4 in the initial draft) has an incomplete legend (two labels are missing). Furthermore, authors are invited to consider an alternative representation based on a concentration curve.
Response
True. How did we miss that? Thank you for the heads up! This is now corrected. (Legend now contains all eight parameters.)
Reviewer vii) Authors are invited to consider an alterative representation based on concentration curves. If they decide to keep the histogram, authors should at least delete the trends (the dotted lines).
Response
As suggested, we have deleted the trends.
Reviewer viii) As already pointed out, I did not ask to add a Lorenz curve of Israel's income distribution.
Response
Yes, but we feel it is important background information for people unfamiliar with income inequality in Israel. So we are leaving it in.
Other comments
Alon
Reviewer - Lines 132-138: the number of distributed and selected questionnaires seems approximated to round numbers. Are 700 and 600 the real number of distributed questionnaires and complete answers collected or are they approximate values? In the second case, words as "around" or "almost" should be added to clarify that this is not an exact measure.
Response:
Yes, the numbers are precisely 700 and 600 as written down in the text, something made possible by the panel sampling methodology. There is no need to add “around” or “almost” as suggested – as in this case the correct adjective would be “exactly”.
Reviewer lines 139-145:
it seems that questionnaires are equally distributed among deciles, i.e. there are 60 questionnaires for each decile. Is it true or this is a round number? In the second case a preposition like "around" or an adverb like "almost" should be added.
Response
Yes it is true that 60 questionnaires were collected from each decile- as the study design dictated. Again, this preclude the need to add words like “almost” and “around”.
Reviewer lines 233-241:
as the average reader might not know what territories are peripheral and what territories are not in Israel, a short note might be added in appendix to illustrate the classification of territories.
Response:
We are not certain that readers need a thorough geographical explanation, especially as the article is starting to exceed reasonable size limits. But nonetheless, we have added a footnote on the methodology page which describes the four “peripheral” geographic zones used to categorize respondents and the reduced opportunities for residents living in them. If editor feels that this footnote unnecessary, we’re happy for you to remove.
Reviewer lines 316-341:
It seems that this part of "results" paragraph illustrates the major findings of the qualitative analysis. As they seem potentially interesting and original, authors might add further comments and a graph.
Response:
We disagree. The nature of the “qualitative analysis” presented in the results paragraphs does not lend itself to graphical representation. We have added additional comments, about the meaning of the findings, as requested by Review 1. We hope that this will be sufficient.
general comment on "results" paragraph: as results are based on surveyed data, authors might discuss major findings using more moderation. As an example, rather than "confirming" hypotheses, statistical evidence based on surveyed data "support" or "suggest the validity" of the hypotheses discussed at least of Israel.
Response:
We agree and now use a more guarded tone – and have changed a bit of the text accordingly, as suggested by the reviewer.
