Real-world indicators cut continua into unequal, non-linearly-spaced categories.

Evidence

It is very common for policy and other reasons to break a continuum into three to five categories. These categorization schemes provide concrete evidence that the semantics of the continuum are not linear. The “obvious” way to break a continuum into three categories is to divide it into equal categories; this does not seem very common. Typically, categorization implies nonlinear semantic transformations (e.g., see the discussion below on how the US military treats its test scores). Kelley and Simmons (2019) report that about one-third of their population of global indicators have categorization schemes. — Gary Goertz, Social Science Concepts and Measurement, 588

Categorization

• XREF: Kelley and Simmons (2019) on global indicator categorization schemes links to work on measurement in governance data.
• SHARE: The one-third-of-global-indicators statistic is niche measurement-buff material worth sending to someone working on indicator design.