Food Purchasing Decisions and Environmental Ideology: An Exploratory Survey of UK Shoppers
Round 1
Reviewer 1 Report
Barker et al. Sustainability Food Purchasing Decisions and
Environmental Ideology: An Exploratory Survey of UK Shoppers.
General Comments
This is a very interesting paper both in terms of its empirical content and also in its thorough approach to the sustainability literature. The Introduction (and the Discussion) clearly outline the previous work in this area so that the reader can follow the arguments that underlie the analyses of the survey data.
The survey findings show that the two subscales of the NEP had quite different associations with normative purchasing attributes; the ecocentric values scale being positively related to the perceived importance of health, animal welfare and environmentally related criteria and the DSP scale (anthropocentric values) was associated with greater importance ratings of quality, taste, safety, price and convenience criteria. The only predictor of ecolabel recognition was the DSP values scale. The findings confirm the view that the overall NEP scale is best considered as two, quite different subscales.
Specific comments
Abstract: this is short but sufficient.
Introduction: This is thorough and well referenced and clearly argues for the three main aims of the study.
Methods: these are comprehensive and well described
Recruitment: it is clear that this was a convenience survey, the limitations of which are well described in the Discussion.
Survey measures: These are adequately outlined though more justification of the need to include the Social Desirability Scale may be useful, either here or preferably in the Introduction.
Grocery Shopping influences: well described.
Eco-label recognition: Well described.
Statistical analyses: these are comprehensive and thoroughly described.
The location of the study: initially the location is stated to be “a northern English City” but as the paper progresses it is clear that it is Sheffield. I suggest that the location is stated as Sheffield throughout the paper.
Results: Again, the findings are simply stated and comprehensible. However, the Tables should be checked for clarity, for example, Table 2 refers to “an Eco label Prercognizoni score” – what is this? Could you briefly define it in a legend? Table 3, check the column headings to ensure "Odds Ratio" is in each column
Discussion: This flows a logical pathway through the main findings and is easy to follow. The paper clearly shows the associations of the two dimensions of the NEP instrument and the limited associations of ecolabels. The limitations of the study are made clear and several useful new research directions are indicated, for example, the need to confirm and explore the “Light Green” and “Dark Green” consumer typologies and the opportunities to extend this approach to other environmentally reliable behaviours (such as meat eating). The challenge for communicators may be to find ways to resolve the conflict between the two opposing NEP dimensions in consumers’ minds!
Comments for author File: 
Comments.pdf
Author Response
Environmental Ideology: An Exploratory Survey of UK Shoppers.
General Comments
This is a very interesting paper both in terms of its empirical content and also in its thorough approach to the sustainability literature. The Introduction (and the Discussion) clearly outline the previous work in this area so that the reader can follow the arguments that underlie the analyses of the survey data.
The survey findings show that the two subscales of the NEP had quite different associations with normative purchasing attributes; the ecocentric values scale being positively related to the perceived importance of health, animal welfare and environmentally related criteria and the DSP scale (anthropocentric values) was associated with greater importance ratings of quality, taste, safety, price and convenience criteria. The only predictor of ecolabel recognition was the DSP values scale. The findings confirm the view that the overall NEP scale is best considered as two, quite different subscales.
Specific comments
Abstract: this is short but sufficient.
We have extended the abstract to include the implications of our findings. Please see line 24
Introduction: This is thorough and well referenced and clearly argues for the three main aims of the study.
Methods: these are comprehensive and well described
Recruitment: it is clear that this was a convenience survey, the limitations of which are well described in the Discussion.
Survey measures: These are adequately outlined though more justification of the need to include the Social Desirability Scale may be useful, either here or preferably in the Introduction.
We have added a justification for including SDS in the methods. Please see line 220.
Grocery Shopping influences: well described.
Eco-label recognition: Well described.
Statistical analyses: these are comprehensive and thoroughly described.
The location of the study: initially the location is stated to be “a northern English City” but as the paper progresses it is clear that it is Sheffield. I suggest that the location is stated as Sheffield throughout the paper.
This has been done
Results: Again, the findings are simply stated and comprehensible. However, the Tables should be checked for clarity, for example, Table 2 refers to “an Eco label Prercognizoni score” – what is this? Could you briefly define it in a legend? Table 3, check the column headings to ensure "Odds Ratio" is in each column
The Word Processing package corrected the original to Italian (as none of the authors speak Italian we are at a loss to explain this).
Discussion: This flows a logical pathway through the main findings and is easy to follow. The paper clearly shows the associations of the two dimensions of the NEP instrument and the limited associations of ecolabels. The limitations of the study are made clear and several useful new research directions are indicated, for example, the need to confirm and explore the “Light Green” and “Dark Green” consumer typologies and the opportunities to extend this approach to other environmentally reliable behaviours (such as meat eating). The challenge for communicators may be to find ways to resolve the conflict between the two opposing NEP dimensions in consumers’ minds!
Reviewer 2 Report
This is a very good research paper and I recommend publication but with a few queries and suggestions for amendments to the paper as follows’
The data comes from a survey in 2014, why the long gap between the survey date and this paper? Maybe because the survey was part of a Masters’ thesis. I have no problem with this but perhaps make it clear in the survey section, rather than at the end of the paper. Why no follow up project to what is a good piece of research? The analysis is maybe a bit too statistically dry and could be made more readable by summarising the key implications of each statistical factor. Table 6 contains a lot of useful data and maybe merits more than a paragraph. I am not sure that Figure 1 is correct in inferring that all routes lead to environmental shopping, it surely needs three or more pathways e.g. anthropocentric-neutral-environmentally concerned. Figure 2 not needed Figure 3 shows that environmental factors and locality, seasonality, and organic are not key factors, but has there been a change since 2014? Appendix 1 not needed as the questions are by and large repeated in the Tables. Replace with available from authors
Author Response
please see attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
This paper is well written, and the topic is very timely as it relates to both environmental values/worldviews and food shopping. The empirical work presented slightly lets down the paper to be honest, as there are a number of problems which would benefit from being corrected in my opinion before the paper is ready for publication. I have made a number of suggestions regarding how I feel the paper could be improved, and I hope that the authors find them useful and constructive. I have divided the comments below by section:
INTRODUCTION
============
The language and writing throughout is very good, and the text is easy to follow. The introduction is concise and too the point (which is good in my opinion), but I wanted to point out the following observations:
1) I recommend that the authors consider reviewing the literature (perhaps looking in nutrition journals like Appetite) to examine what work has been done in the UK looking at ECF using dietary data (e.g. the NDNS)
2) On line 137, you say that ECF purchasing behaviour is difficult to assess. This is not strictly true. There are ways to assess such things using purchasing data, such as the kantar dataset, or other such large surveys of purchasing.
3) lines 118-123 seem irrelevant to me, and I would suggest removing this paragraph.
4) line 146-147. Suggestion that recognising eco-labels is a facilitating behaviour for environmentally driven food purchase. I am not convinced by this argument, and even though references are provided, I think it would be useful to elaborate the argument a bit more. In my opinion, eco-label recognition is just that - recognition. A person can recognise a brand without necessarily wanting to purchase that brand.
METHODS
=======
Whilst the methods are well described, I found that there were some areas where it would benefit from having some more detail provided, and there were some issues that I found to be problematic in terms of the suggested approach:
5) The questionnaire also contains measures of weight and height (assume to have been included to estimate BMI). Why was this data not reported upon?
6) The Social Desirability Scale seems to be a problematic measure on two levels: (a) the scale reliability seems low, but the measure seems to be scaled using a yes-no scoring, but it is not actually clear how the scoring is done. The ‘not sure’ category should NOT be seen as half-way between yes and no. Is the score on this scale simply a total of the number of ‘yes’ answers? (b) The measure itself does not appear to be particularly meaningful in the context of the current study, and unless a much better justification for inclusion is included, I would recommend removing it from the paper completely. I understand that one might want to control for something like social desirability, but I am not convinced at all that this measure is a good way to assess it.
7) Line 191-192. You say that respondents were asked to rate importance from 1-5, but the questionnaire provided indicated that the response scale was a 7-point one.
8) Lines 194-203. The eco-label recognition ‘score’ is particularly problematic in my opinion. There seems to be an assumption that this is a kind of knowledge test that can be objectively scored. If this was the case, then I would have expected a lot more detail of the scoring itself, and even then it would take a lot to convince me that this was a good idea. I am not convinced that it is necessary to have this assumption though. Why not simply take the data at face value (i.e. a multidimensional indicator of what attributes people associate these different logos with). I actually think the results would be far more interesting presented this way. The companies associated with the logo’s might also be interested to know what the logo’s are associated with. I would recommend consulting with a statistician to explore different ways to reconsider this data as a multidimensional portfolio-type measure. It might be that something like MDS or MCA could be used for this.
9) Section 2.5 (statistical analysis). If the intention is to suggest that the results of the EFA are particularly important, then I would expect the results of this analysis to be provided in more detail (e.g. with the loadings in a table, etc..). However, I don’t actually think the analysis necessarily merits this, as it tends towards overstatement in my opinion. In the abstract you claim that the results provide “strong evidence” that the scale should be considered as two subscales, but not a lot of evidence is provided for this. The fact that the single factor and the 5-factor model were found to have ‘adequate fit’ in the CFA is not in itself strong evidence that NEP and DSP should be considered as two distinct subscales. However, assuming it does suggest this, you do not actually use the scale as two distinct subscales, as you are relying on the factor scores from the EFA instead. This strikes me as a questionable decision. Why not simply use the two subscales (based on the evidence from the CFA that you should do so) directly. You may well expect there to be a correlation between the two, but if you are claiming these to be two distinct constructs, then this would potentially be a simpler way to do this.
10) Line 230-235. The justification for the quadrant approach is not convincing. I would strongly recommend re-considering this approach, and perhaps including the interaction between NEP and DSP as a separate term in a model (e.g. by adding DSP as a moderator between NEP and an outcome variable).
11) Throughout the text you mis-use the term ‘ranking’ when you should say ‘rating’. You do not have any ranking questions, and although strictly speaking, summary data can perhaps be described as having a rank, I would strongly recommend avoiding this, as the paper often implies that you have measured data using ranking questions – which you do not.
12) Line 246-257. You suggest that the summary is of participants rating things as ‘important’ or ‘very important’, but the questionnaire does not have an ‘important’ label, and simply suggests a 7-point scale. If ‘7’ is ‘very important’, what is ‘important’ (is it 6?). It is not clear why cluster analysis has been used here. I would suggest either considering PCA here, or perhaps even using cluster analysis to partition respondents into different response categories. I recommend talking to a statistician to get advice on the most appropriate selection of exploratory techniques to use.
13) Section 3.4. Related to above – It may be worth considering reducing the ‘importance’ variables to a summary of some kind (perhaps using PCA), and then including this summary variable in a regression model. The approach used here (and including sections 3.5 and 3.6) is not making best use of the data in my opinion. I would recommend removing the quadrant analysis completely, and processing the data in a different way (see suggested approach above).
DISCUSSION
==========
14) Line 307. Ranking should be clanged to ‘rating’. This needs checking throughout this paragraph, and indeed throughout the whole paper.
15) Section 4.3. You measured weight and height, but did not explain why. Was there some literature that led you to believe that people with a high BMI would think or act in a different manner?
16) You use the term ‘predict’ and other such terms in a way that is perhaps tending towards over-statement. Your findings are associations, and should be treated as such.
17) Section 4.5. You should reconsider this in relation to the comments made earlier about the two sub-scales. My view is that the data would be better dealt with by genuinely using the measure as two sub-scales, but this requires the analysis to be re-done (but would actually be simpler). It is important not to over-interpret findings. This is one small study, and the findings should be reported as they are, as they add to the overall scientific record.
18) Line 417. You say that the correlation is low, but would you not actually expect it to be low given the fact that you are using factor scores (even though oblimin rotation is used), as if I understand it correctly, the technique itself will try to minimise the between-factor correlation.
Overall, I think the paper has the potential to be a useful contribution to the scientific record, but I think it would benefit from being presented in a slightly simpler manner, being careful to avoid over-statement. I hope that the comments provided in this review are useful in helping the authors to improve the paper.
Author Response
please see attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
I am satisfied that the authors have taken on board the comments I made in the first review.
