SURVEY DESIGN USING INDIVIDUAL NUMERICAL SCALES IN THE FRAMEWORK OF ANALYTIC HIERARCHY PROCESSES

This paper discusses the adequacy of a generalization of Saaty’s 1-9 scale proposed by Liang at all (2008) in the attempt to identify individual scales. Several surveys in completely different areas were conducted on different topics. Comparisons among the consistency index-as a measure of a “good answer” and the previously mentioned scale reveal a non monotonic correspondence among those two criterions. Also, the individual scale considered – which is in itself a generalization of other similar scales for measuring individual responses – is not uniquely determined for a single respondent and is very often contradictory. Yet, the potential benefits in determining individual scales of measurement are enormous and maybe the most important one is getting rid of the myth of the good appliance of the “law of large numbers” in social sciences.