The accuracy of AHP method don't have any mathematical proofs. In my opinion, "the size of n" is not important in that case. Moreover, if you have more alternatives (bigger n) you have to make more comparisons. For n we have to make n(n-1)/2 comparisons.
have a nice day WS
The question on n deals with the issue of sample size in AHP. According to the majority of published applied AHP, the n has been less than 50; in many cases it's less than 20. The quality of AHP respondents (mostly expert on the researched issue) justifies the small sample size in AHP.
follow the link to see a paper by Prof Saaty, on how many respondents should be used , where he concludes
"To engage judges to help with a decision should not be a random matter. One needs to
know the area of expertise needed to make that decision and select a judge or judges
that have both knowledge and practical experience with the matter. In this case one
expert judge may suffice unless political expediency requires that several judges from
different constituencies are necessary. In that case one might select several judges if
they are available."
Article How Many Judges Should There Be in a Group ?
- Badges
- Science topic