Measuring the “unmeasurable”

Abstract

From time to time, I hear people say something like “You just can’t measure that” or “You really can’t put a numerical score on that.” The point is often made in connection with highly abstract factors and beliefs held by people. What I think these people are really saying is that the factor is too complex to be represented by a single number; too many facets and subordinate qualities are associated with the factor to make a single numerical score meaningful. For example, how would you put a number on the depth of the relationship between a married couple? Yet, I personally believe that you can measure anything. If you’re willing to put the time, effort, and investment in the measurement process, you can come up with a numerical score that can truly represent any factor, no matter how abstract. Wouldn’t human intelligence be considered a highly complex and abstract factor with many facets and subordinate qualities? Yet, IQ scores are assigned to most people and these numerical scores seem generally accepted by almost everyone. The Software Engineering Institute of Carnegie-Mellon University assigns software organizations a score from 1 to 5 to represent their maturity in software engineering (a score with enormous implications for the organization’s business development capability) and yet this “measure” goes virtually unchallenged. Yes, I’m pretty certain that you can measure anything, so let me describe a few approaches for assigning a numerical score to represent abstract factors using measurement techniques that are considered accurate and meaningful.