For develop a measurement model of construction waste management, I use AHP for prioritisation the factors. Because I have to collect the data from experts, I'm looking for the right way to calculate the sample size.
There is no general rule for the sample size. It can be ranged from 5-9000.
The number depends on the aim of the research. If the survey is aimed at weighing or criteria the sample should represent the target population interested to those factors.
Attached please find a publication that might be helpful,
Grateful for your kind reply. I have to use expert and as I found so far, AHP doesn't need large sample size. However, I try to prove it by some methods. unfortunately I couldn't find any methods for it.
Thank you M. Omidvari ! Would you please provide the reference of " 7 number is good, 7 is gold number and its suitable for more than 90% cases. " I appreciate your help!
AHP require the expert who is representative to the problem in hand. The number of expert may be any thing from 3 to a higher number. If you are taking a focus group of experts and develop a consensus using Delphi method or any other method then you will get only one pairwise comparison matrix.
Small sample size can adversely affect several aspects of any research, including the data analysis and concomitant interpretation of results. The major advantage of AHP over other MCDM methods is that it does not require a statistically significant (large) sample size to achieve sound and statistically robust results (Dias and Ioannou 1996; Doloi 2008). Some researchers argue that AHP is a subjective method for research focusing on a specific issue, hence, it is not necessary to employ a large sample (Lam and Zhao 1998). Others argue that because AHP is based on expert judgments, judgments from even a single qualified expert are usually representative (Golden et al. 1989; Abudayyeh et al. 2007; Tavares et al. 2008). Moreover, it may be unhelpful to use AHP in a study with a large sample size because ‘cold-called’ experts are likely to provide arbitrary answers, which could significantly affect the consistency of the judgments (Cheng and Li 2002). Much of the popularity of AHP in CM could be attributed to its ability to handle small sample sizes.