# 107

Dear David Jorge-García , Vicente Estruch-Guitart , Pablo Aragones-Beltran

I have read your paper:

How does the type of MCDM method affect the results of the prioritization and assessment of ecosystem services? A case study in the Ebro River Delta (Spain)

My comments

1- In the abstract you state “Regarding how stakeholders structure criteria, networked processes are, theoretically, more appropriate than the hierarchical techniques such as AHP, since they capture all the relationships among criteria”

This is true; however, it appears to be a contradiction in that you say later that AHP, ANP and DEMATEL are used. First you say that AHP cannot be used, and after that you enunciates that it can be used in a triad with ANP and DEMATEL. Consequently, I don’t see, as you express, that AHP may influence prioritization.

2- In page 3 “sociocultural factors such as DMs’ perceptions or biases due to their personal and professional backgrounds”

Exactly, very good point. It also depicts the drawbacks of the AHP and ANP methods

3- “On the contrary, ANP can draw a network incorporating feedback and interdependent affinities within and between clusters”

Not in my opinion. ANP does not incorporate feedback. Saaty introduced that concept but never demonstrated its existence, therefore, it is only an assumption.

4- “For this purpose, the compatibility of the priorities and the correlation between the rankings obtained will be analyzed”

So, now you are oblivious to the AHP limitation of working only with independent criteria?

5- “AHP is unsuitable for complex processes where most criteria are interconnected.”

True, but if you know that, why do you use it?

6- “ANP determine the consistency of the DMs’ judgements.”

True, reaching transitivity within a 10%, FORCED by a formula that corrects what the DM estimated or preferred. And normally it is not a minor correction because the Fundamental Table is logarithmic

What connection does this have with real life, that is, why is it assumed that what is in the DM mind is applicable to real life?

For instance, why he/she declares that real-life is transitive? Is there any theorem that justifies the automatic transfer of these mind estimates to real-life?

Of course not, and AHP users take it as if it were true, even if not proof is given. Very scientific indeed!

7- In Fig. 1 what is the input or contribution of AHP? Prioritization is not done also by ANP?

8- In Section 2.2 you are again using AHP, even when you recognized that it cannot be employed with interdependent criteria, as the same Saaty explicitly said. Could you explain it? In addition, it is very easy to prove this.

9- “In this step, each DM has to build its 12x12 matrix determining the existing influences among the ESs.”

False, in network theory, the arrows connecting nodes only denote precedence, not influence. Another assumption without any foundation

10- “In this step, each DM has to build its 12x12 matrix determining the existing influences among the ESs.”

That is, subjectively, and based on what?

11- Page 9 “Firstly, AHP simplifies reality when it comes to giving more priority to some services than others”

AHP like any other MCDM method builds a model that is a simplification of reality; as a consequence, we can’t expect accurate results from any of them. But there are degrees of simplification; the problem is that AHP simplifies so much that it does not even consider the data in a real problem; it fabricates that data using assumptions and intuitions, which depend on the DM feelings and mood.

Other rational methods do, like PROMETHEE and TOPSIS, that start working with real values, and use reasoning, not feelings.

You say that AHP results are worse than those of ANP and DANP, and I don’t have a doubt about it, but curiously, you use AHP when your previously said that it can’t be used due to its intolerance to criteria relationships. I understand that, by respect to the reader, an explanation should have been given.

12- “Compatibility among the priorities obtained by the three methods.”

This is expected, since AHP works with intuitions, while DEMATEL analyzes and investigates the potential existence of causal relationships. The first is a lottery, the second is reasoning

You are then comparing oranges and bananas. Same with ANP, since there is nothing in the network that indicates the influence between two criteria, the arrow only indicates precedence. And this precedence may also be due to reasons like structural relationships, when for instance C5 cannot start if C3 is not completed, which, by the way, is the basis of the Critical Path Method.

In my opinion the scales have nothing to do; DEMATEL uses a lineal scale, when Saaty derived his by assuming that it replicates the psychological laws of Weber and Fechter, that link responses (in a logarithmic scale), to the variation of the stimulus. As far as I know stimulus is not related to intuition. Stimulus is´” something that causes growth or activity “(Cambridge Dictionary). Weights are a relative measure of the importance of each criterion

I hope these comments may help

Nolberto Munier