I have conducted a survey on hospital waste management and now I want to evaluate the responses.
I used AHP to receive the judgments of the respondents. However, I don't know how to add the different individual results of the respondents so as to get one combined judgment. What is the standard method of adding individual results?
Dear,
This is a commom problem, and there is not a simple solution. If you want just one answer, you must use the geometric mean, because the Saaty scale is a rational, and not linear.
But what I suggest you is calculate as many results as answers you have, and after that, try to explore those results, without means, because in this way, you are being compensative.
Totally I'm agree with Enrique. On the other hand, you can use the SPSS software to process these surveys, obtaining some statistical metrics that can be very useful to you study.
ho matter how many respondents have the problem, you can solve it with this software.
i'm familiar with spss and also R. i was trying to suggest that analyzing a 100 responses individually might not be very useful. that's why i wanted to find a way to get a combined judgment.
Dear colleagues,
As I said, the simple mean is not correct to combine the answers. It must be at least, a geometric mean.
But if I were you, I would calculate the 100 answer individually, and after then, combine the results. It is methodologically more correct.
Your question is not vivid. If your judgments are in linguistic scales, you can better use grey theory (or fuzzy set theory) for first converting into grey values, take the average and then convert back to crisp ratings. Please read my publications.
Hi colleague,
Looks like u r in China... Well not relate to the question but it is my hometown.
Back to the question, I agree with Henrique that it isnt correct to combine the answers by simple means. However, concerns typical AHP usage, mostly they are using geometric means as there isn't any other established approach.
In my previous research I derived all answer individually and employing information entropy to combine them (I am working on this method, and the paper is under developing for publication). However I need to notice you that it doesn't influence the final result a lot actually regarding my small number of decision makers, the rank hasn't been changed. Still methodologically speaking, I don't agree with simple means.
Hope these can help you.
For typical AHP, you can use software Expert Choice (it is use geometric means).
The software can save and restore the respondents' judgment; also to combine it. It is also easy to use and interpret.
Good luck.
- Badges
- Science topic