Standardizing scores makes one case’s value depend on other cases’ values.

Evidence

The statistical mean and standard deviation usually change as the underlying data change. In fact, with standardization the score for any one case can easily change if the scores for the other observations are altered. A case might not experience any real-world change, yet its coding shifts because values for other cases are changing. From a semantic as well as descriptive perspective this seems odd: Why should the score for a case depend on how other cases happen to be coded, or on changes in their values? Of course, if the concept is relative and relational to begin with, e.g., inequality or positional goods, then how other cases are coded does and should matter. — Gary Goertz, Social Science Concepts and Measurement, 863

Categorization

• XREF: Connects to the reading on relative vs. absolute measures of inequality and positional goods (e.g. Hirsch’s social limits); the parenthetical concession is the pivot worth tracking.