Improving Drinking Water Quality in South Korea: A Choice Experiment with Hypothetical Bias Treatments
John H. Gibson
Round 1
Reviewer 1 Report
You do not define hypothetical bias. Most people in the water business, including me, do not know what it is. Since I have little idea what the paper is about, it is hard to make sense of it. There is a lot of jargon most readers won't know.
It looks like the analysis was extensive and there is a clearly defined problem. That part was very well done.
Author Response
Referee #1
Your Comment:
You do not define hypothetical bias. Most people in the water business, including me, do not know what it is. Since I have little idea what the paper is about, it is hard to make sense of it. There is a lot of jargon most readers won't know.
It looks like the analysis was extensive and there is a clearly defined problem. That part was very well done.
Authors’ Response:
The reviewer is right that hypothetical bias has not been defined in the paper and not everyone needs to know what it is. We thank the referee for this comment as it has allowed us to correct this omission. We have introduced a sentence in the Abstract:
Hypothetical bias is the difference between what people state in a survey they would willing to pay and what they would actually pay in a real situation.
And one in the Introduction (lines 61-62):
If people overstate for example their willingness to pay for the project then, basing the political decision purely on stated values would lead to wrong decisions
And we have defined hypothetical bias again in the section called Hypothetical Bias on page 5 together with two citations:
It is often the case that stated preference studies demonstrate significant differences between stated versus real values. The difference between the two is called hypothetical bias (Cummings and Taylor 1999, Penn and Hu 2018).
We are convinced that this was an important change and that the paper is easier to read now.
Reviewer 2 Report
The paper has a logical flow and strong structure.
The proposed methods and the obtained results are interesting and could help the population in a sustainable way.
My only concern is the fact that the study was done 5 years ago.
Author Response
Referee #2
Your Comment:
The paper has a logical flow and strong structure.
The proposed methods and the obtained results are interesting and could help the population in a sustainable way.
My only concern is the fact that the study was done 5 years ago.
Authors’ Response:
The project is in the process of implementation now in South Korea and hence the relevance of the study for the specific project has been confirmed; however, the methods applied in the study bear more generality. The mechanisms used to correct for Hypothetical Bias (Cheap Talk and Honesty Priming) can be implemented in further studies. Moreover, the thorough way in which the cost benefit analysis was performed, involving several sensitivity analyses, can serve as an example for future studies.
The basic analysis would not change much if the study would need to be repeated today. There might be some changes in the cost benefit analysis such as, for example, the interest rate used may be lower. However, in our sensitivity analysis the only problem for the feasibility of the project would arise if the interest rate would rise; this is not expected under the current circumstances. At the same time, maybe the benefits are expected to decrease as the WTP for improved water quality might decrease due to the crisis created by Covid-19. However, it is also expected that the demand for improved water filtering will increase during the present circumstances. This, together with the fact that the costs from a water bill represent just a minor part of the monthly household costs in South Korea (<1%), means that we do not expect that the WTP would change much.
We are confident that despite of the fact that the study was conducted 5 years ago, the methods applied are significant today and relevant for any future studies.
We have included a footnote on page 3 (Footnote 6) in the Survey Design and Data Collection section in order to reflect this:
"Even though the survey was conducted five years ago, the methods applied are relevant today and for future studies. Moreover, the present project has served as a basis for the implementation of the water treatments in S. Korea which is happening at the moment (2020)."
Reviewer 3 Report
Unfortunately, this article is not even close to the topic of the journal.
I recommend submitting it to another profile journal. The subject of the research has nothing original, it is rather a theoretical study that is not of interest to the readers of Waters journal!
Some suggestions for authors: the scientific merit of the paper decreases when the first person is used; also, the correct writing of the units of measurement according to IUPAC must be required (i.e., g/L not g/l). However, there are great places throughout this manuscript that need to improve in organization and presentation. For instance, the authors wrote a large body of Introduction but failed to tell us what the main objectives, the importance, and/ or the scientific hypotheses this research are. In my opinion, there are many sentences or some paragraphs that are not closely related to the main objectives and should be deleted. The Abstract must be rewritten!
Another shortcoming of this manuscript is the length of the writing. The section of Abstract, Introduction, Discussion, etc. should be greatly clarified.
Based on the above comments, I am afraid I cannot recommend the publication of this manuscript in Waters!
Author Response
Referee #3
Your Comment:
Unfortunately, this article is not even close to the topic of the journal.
I recommend submitting it to another profile journal. The subject of the research has nothing original, it is rather a theoretical study that is not of interest to the readers of Waters journal!
Some suggestions for authors: the scientific merit of the paper decreases when the first person is used; also, the correct writing of the units of measurement according to IUPAC must be required (i.e., g/L not g/l). However, there are great places throughout this manuscript that need to improve in organization and presentation. For instance, the authors wrote a large body of Introduction but failed to tell us what the main objectives, the importance, and/ or the scientific hypotheses this research are. In my opinion, there are many sentences or some paragraphs that are not closely related to the main objectives and should be deleted. The Abstract must be rewritten!
Another shortcoming of this manuscript is the length of the writing. The section of Abstract, Introduction, Discussion, etc. should be greatly clarified.
Based on the above comments, I am afraid I cannot recommend the publication of this manuscript in Waters!
Authors’ Response:
Many thanks for your comments, we appreciate you candor. The abstract has been rewritten and informs the reader of the objective of the study. The introduction has been edited to ensure a tighter narrative which provides the read of the importance of the study as well as the scientific underpinnings of the study. We have streamlined the manuscript in order to provide a more clear and concise narrative. We have removed the use of the first person throughout and updated measure notation according to IUPAC. Moreover, we have had the manuscript reviewed by a professional copy-editor.
Reviewer 4 Report
Dear Authors,
the manuscript needs many corrections. First of all, the exact adaptation of the text to the requirements of the journal, the style of citation and references, and the unification of the style of tables.
I also propose to separate the conclusions section as a separate chapter.
Yours faithfully,
Rev
Comments for author File: 
Comments.pdf
Author Response
Referee #4
Your Comments:
the manuscript needs many corrections. First of all, the exact adaptation of the text to the requirements of the journal, the style of citation and references, and the unification of the style of tables. I also propose to separate the conclusions section as a separate chapter.
Authors’ Response:
Many thanks for highlighting these points. We have now updated the citation and references in the paper and made the table style unified. We have also made the conclusions its own section.
Reviewer 5 Report
Suggestions to the author:
Line 47:
time annual sales of bottle water increased by 96% between 2009 and 2014, and sales of in-line filters
increased grew by 49% during the same period of time (Database of the Korean Statistical Information
Roger’s comments: The word “increased” appeared twice in the same sentence.
Line 49:
Moreover, this dramatic increase in bottled water (is it sales or disposal?) exacerbates the negative effects of…
Line 52:
systems in S. Korea’s Guem River Basin and in the waterworks providing it for the purpose of providing with drinking water
Line 53 to 54:
The two treatments are: granular activated carbon (GAC), and ozone plus GAC treatment.
GAC Carbon is usually added to the process of filtration, and ozone treatment is added to coupled with the system of chlorine disinfection as an additional method to remove fine particles and to create chemical reactions
Line 57:
present study also discusses the most appropriate environmental solutions for improving long-run long-term
Line73:
something that seems to enhance its efficacy. Consumers were reminded that if they spend more one on a
Comments to the author:
Is cost the biggest limiting factor based on the analysis? Or cultural factor? This is not clear to me after reading the manuscript.
Was a multivariate analysis was considered by the author? The reason I asked this question is this could have reduced the number of variables considered in the analysis and perhaps minimize the confounding effects arising from so many variables.
I was a little bit surprised by the fact that water color as an attribute was not an issue for the test subjects since people tend to generally react to visual or tactile attributes of water.
Author Response
Referee #5
We thank the referee for reading the manuscript so carefully and for giving us these valuable comments. We are going to address them point by point below:
Referee #5 Suggestions to the author:
Line 47:
time annual sales of bottle water increased by 96% between 2009 and 2014, and sales of in-line filters
increased grew by 49% during the same period of time (Database of the Korean Statistical Information
Roger’s comments: The word “increased” appeared twice in the same sentence.
Good point. This has been corrected!
Line 49:
Moreover, this dramatic increase in bottled water (is it sales or disposal?) exacerbates the negative effects of…
In order to clarify this the sentence was changed to:
Moreover, this dramatic increase in sales of bottled water leads to more disposal of water bottles and exacerbates the negative effects of perception of undrinkable tap water via increased marine litter.
It was important to clarify this. Thank you!
Line 52:
systems in S. Korea’s Guem River Basin and in the waterworks providing it for the purpose of providing with drinking water
This has been corrected, thank you.
Line 53 to 54:
The two treatments are: granular activated carbon (GAC), and ozone plus GAC treatment.
GAC Carbon is usually added to the process of filtration, and ozone treatment is added tocoupled with the system of chlorine disinfection as an additional method to remove fine particles and to create chemical reactions
This has been corrected, thank you.
Line 57:
present study also discusses the most appropriate environmental solutions for improving long-run long-term
This has been corrected, thank you.
Line73:
something that seems to enhance its efficacy. Consumers were reminded that if they spend more one on a
This has been corrected, thank you.
Referee #5 Comments to the author:
Comment 1:
Is cost the biggest limiting factor based on the analysis? Or cultural factor? This is not clear to me after reading the manuscript.
Authors’ Response:
From the CBA results presented in Table 13 it appears that the main factor that could make the project unfeasible are the benefits. If the benefits appear only during 10 years or the benefits decline to zero after the first year, the project becomes unfeasible. The benefits need to appear for 10 years or more in order for the project to be feasible.[1]The reviewer is right that if the citizens would not see value in implementing the water treatments then this would be the main limiting factor. If the households would continue to buy bottled water even after the filtering of tap water out of ‘cultural habit’, then the project could become unfeasible. There are studies that show that it is rather the perceived poor quality of water than the actual quality that hinders S. Korean people to drink water from the tap (Um et. al. 2002). Hence, if S. Koreans would still perceive the water to be polluted even after the installment of the two systems, then they would not drink it and the benefits of the project would be limited. However, our survey shows, that S. Koreans do see value in implementing the two water filter treatments and are willing to pay for it. Moreover, as mentioned in the paper, a study performed by Jo et al. (2015) showed that when directly asked, if they are going to drink water from the tap if it’s quality will be improved, a vast majority (approx.85%) of S. Korean consumers said they would. Additionally, when excluding approx. 15% of the population from our sample, the percentage that might not change their behavior out of cultural habit, the project still remains feasible (as the results of the sensitivity analysis in Table 13 show). Hence, based on the results of our survey (performed with on a representative sample of households in S. Korea), and on our results of the sensitivity analysis, we have to conclude that the ‘cultural factor’ is not expected to be the biggest limiting factor in this case. But this is an important factor to mention and we have included footnote 21 on page 19 to mention this.
"It is important to note that in the present analysis the surveyed households appear to be willing to pay in order to improve the tap water quality and hence it is assumed that they will start drinking water from the tap more frequently once the treatments are implemented. Moreover, when asked explicitly in a different study if they will drink tap water when its quality will be improved, a vast majority of consumers (84.3) answered that they would [45]. Additionally, when performing a sensitivity analysis excluding 15.5% of the sample from the benefits, the consumers that might not change their behavior out of cultural habit, the project remains feasible (Table 13, row 11). Hence, the cultural factor does not appear to be a big limiting factor in the present analysis."
Referee #5 Comment 2:
Was a multivariate analysis was considered by the author? The reason I asked this question is this could have reduced the number of variables considered in the analysis and perhaps minimize the confounding effects arising from so many variables.
Authors’ Response:
The reviewer is correct that initially many confounding variables have been considered for the present analysis. The choice of attributes was informed by previous studies. As explained in the literature review we chose these attributes as they have been shown to be significant in previous studies. We do take into consideration however, in the Latent Class Non-Attendance Models if these attributes ‘have not been attended’. Subsection LCM-ANA starting on page 10 explains exactly how. Hence, if an attribute is irrelevant to the consumer, this is reflected in their lower, or zero willingness-to-pay. With respect to other explanatory variables, as can be seen in Table A3 in Appendix 3 the original analysis contained no less than 28 individual specific/socio-economic factors. As mentioned on page 9 we only chose in the end the factors that were significant in the analysis.
"The ASCs of the socio-economic factors are chosen when their coefficients are significant at least in one option at the 95% significance level."
From the multitude of socio-economic factors, only 3-4 variables were significant for each model and were chosen for the further analysis. For RPL1 ‘Elderly’, ‘Bill’ and ‘Environ’ and for RPL2 ‘Elderly’, ‘Bill’, ‘Environ’ and ‘Fulltime’ (see Table 2 on page 10). The rest of the variables considered are the 4 attributes (safety, taste & odour, colour and price), dummies related to hypothetical bias treatments (Dboth, Dcheap, Dhonest) and their interaction with the price as we need to observe how the hypothetical bias treatments are impacting on the price. Hence, we did consider how to minimize the confounding effect that could arise from so many variables initially considered. This is an important aspect and we thank the referee for pointing it out.
We have included the following comment in footnote 16 on page 9 to emphasize this matter:
"As we use Attribute-Non Attendance (ANA) and chose only the socio-economic factors that are significant at 95%, we don’t need to perform a multivariate analysis to reduce the number of variables considered in the analysis."
The language in the methodology section of the paper has been tightened to reflect the discussion above, to make clearer the modelling choices made, and the rationale for these.
Referee #5 Comment 3:
I was a little bit surprised by the fact that water color as an attribute was not an issue for the test subjects since people tend to generally react to visual or tactile attributes of water.
Answer:
The reviewer is again right that the fact that water color was not an issue for the test subjects is in general surprising as people tend to react to this attribute; however, considering the way in which the attribute levels for this attribute were defined in the present study, makes it maybe less surprising. As explained on page 4 the color of drinking water is linked to the concept of True Color Unit (TCU)[2]. The current standard for the color of drinking water in S. Korea is five TCU. Tap Water Public Relations Association, S. Korea (2013) reported GAC can reduce the color of drinking water to less than 4 TCU and the GAC + Ozone can usually remove the color of drinking water to less than 3 TCU (Choi, 2007). However, Bean (1962) reported that the 3 TCU level of drinking water color is the human detection limit. And here is the reason why water color appears not to be an issue for the S. Korean sample. As it can be seen from Part A of the Survey description (page 26) the TCU changes are very small and beyond the human detection limit. Hence, as commented on page 11 (second paragraph) this is the reason why we consider that the color attribute was ignored in most classes and the willingness-to-pay to improve it was close to zero.
"This result is expected because people cannot presumably detect the differences between 5 and 3 TCU, and this was also suggested by the RPL results."
[1] We are not worried too much about the project life as previous projects had a business life above 10 years.
[2] One TCU corresponds to the amount of colour exhibited under the specified test conditions by a standard solution containing one milligram of platinum per litre.
Round 2
Reviewer 1 Report
Well done.
Reviewer 4 Report
Dear Authors, Please adapt the text to the editorial requirements (removal of footnotes) and use the format presented on the MDPI website.Best regards, REV
Reviewer 5 Report
Suggestions have been provided.
