Statistical standardization serves rescaling, not semantic validity, so researchers should avoid it.

Evidence

In summary, statistical standardization is usually motivated by getting a common scale, not by increasing the semantic validity of the resulting variable. In addition, it relies heavily on the empirical distribution of data. Both these features lead to a recommendation that one should not statistically standardize variables. — Gary Goertz, Social Science Concepts and Measurement, 870

Categorization

• XREF: Connects to Paul Milgrom’s point that formal machinery often silently imports assumptions about the world it measures; scaling choices carry substantive commitments.
• QUESTION: If statistical standardization is routinely wrong, why does it remain near-universal in quantitative social science practice?