Evaluating Cities,Multidimensionality,and Diminishing Returns

Abstract

In order to evaluate multiple alternatives such as cities across multiple criteria such as climate and housing, one needs to be sensitive to the importance of working with a common measurement unit that does not lose important information. One also needs to consider diminishing returns, since goingfrom 0 to 1 in almost any scoring system should generally mean more than going from 100 to 101. Perhaps the best way to deal with those two problems is to convert the original raw scores into part/whole percentages. Doing so is more valid and simpler than working with ranks, 1-5 scales, logarithms, or multiplying across scores. Doing so produces more meaningful results when evaluating cities like Pittsburgh and San Francisco.