I have gone through number of papers based on MCDM techniques. Varied number of decision experts or Decision makers have been used including starting 3 to 18.
Is there any minimum number prescribed to be used in MCDM analysis
or any formula based on number of factors ?
Looking for feedback to clear my confusion!
The lowest number of experts that I have found is 5 respondents, the maximum number is around 50 or more little bit. You can conclude that this is based on your preference. However, as much as you can collect that would be better. At the same time, this will make your paper beneficial to the readers, a high contribution, and not be rejected by reviewers. I would recommend over 30 respondents.
Thanks Prof Mazen.
I have seen dozens paper using different MCDM techniques for various applications using number of experts ranging from 5-10. Hardly MCDM based papers crossing experts more than 15.
I am wondering if there is any formulae like sample size in quantitative analysis.
Dear Mohammad
In MCDM there could be say 20 criteria and many of them related to 9 different issues like Health, Finance, Economics, Engineering, Environment, Sustainability, Transportation, Job generation, City mobility, etc.
Therefore, it would be necessary to count with 9 different experts. The problem is not how many experts we need, but how many experts in each specialty we need.
Suppose that the project is the selection of two types of means to cross a river, by a tunnel or by a bridge, and we need to choose one of them.
It is seen that each expert must compare in turn with the other eight experts.
In a criterion like cost, the finance guy must confront and discuss with the engineering guy.
Is it realist for the finance point of view to say that financing is 3 times more important that engineering? On what grounds? What authority has him to assert this, if he does not know the engineering aspects, and vice versa? How can two people negotiate if they are speaking different languages, and each of the ignoring the importance of the other?
This is one of the fallacies of the pair-wise comparison procedure.
Assume that there are 20 criteria as mentioned, with the same characteristics and only 4 experts, how can they manage?
I ask anybody if this is RATIONAL, or enters in the category of playing to lottery
Ten years ago, doing my master's degree, I used AHP a lot, and researched a lot about it and MCDM, but I found nothing that determined the minimum number of experts or respondents.
I used common sense and decided that if it was something to be decided collectively, 03 people by criterius, would be an interesting number, as there would never be a tie in premises, whether negative or positive, and in the end the calculation would decide whether my preposition was correct or not.
I hope I helped with your decision.
Dear Antonio
Of course nothing was said about the number of experts needed in AHP, I tbhink I was the first to point it out. AHP assumed or perhaps its developers did not see this point. However, as I said before, if you have say 9 criteria related to health, environment, economics, engineering, government, financing, land use, etc , as most real-life projects, it is obvious that the DM, perhaps expert in one or two fields, cannot take decisions on aspects and fields that he/she does not dominate. I don't think it is too difficult to underrstand this.
You at least proposes three experts for criteria, but it does not solve the problem, first in so doing you would need, in my eample 27 experts, and also baceause there is not guarantee that for criterion health for instance, you may have a physician, an engineer and one accountant. If for say criterion government yoy have an accountant, a doctor and a finance guy, how the accountant may help if he does not know anything regarding for instance effects of NOx on human health?
I notice, along the years, reviewing more than 150 papers published and my own experience, that in general, people use AHP because it is the most employed method, it is easy, and fundamentally, the user does not need to think. He accepts blindly absurds like pair-wise comparison and assignning weights, how ? Easily, using his inthition, and happily accepts that the weights, that are not wreights, but trade-offs, can be used in evaluation alternatives.
It is strange that the DM did not think about the alquimia neded and hidden in the method, by which what is in his/her mind can be translated to the real weight? Where is the common sense?
If he/she did , well, it is easier to accept it.
I could continue with other beauties of this metbod, but many people prefer to ignore them.
I do not know you, neither what is your field, but I wonder if in your extense use of AHP, you realized that you cannot use it in the criterioa are not independent. Pls. think about this
Coming back to you proposition, could you explain how would you check if it is correct or not, since you do not have any yradstick to compare it to?
Thank you for your answer
- Badges
- Science topic
- Similar topics
- Mathematics