Ordinal categories built on an underlying ratio variable reveal unequal spacing between groups

Evidence

It is worth contrasting figure 5.2 with what the previous chapter did with the ordinal World Bank groupings of poverty–wealth. The core difference is that the ordinal categorization of the World Bank has an underlying ratio variable of GDP per capita. Because of this underlying data, we can see that the distance covered by groups is definitely not equal. For the PTS we do not have such underlying ratio data. As a result we have to use other means to try to do the mapping. However, the basic semantics, logic, and methodology are not really that different. — Gary Goertz, Social Science Concepts and Measurement, 797

Categorization

• XREF: Connects to measurement debates over whether ordinal scales from survey data can legitimately be treated as interval — the underlying ratio variable here justifies unequal distances. — connects to: A racial binary can be vertically ordered across multiple axes of subordination
• SEED: Angle: ‘The hidden scale behind the index’ — how latent continuous data lets you rescue ordinal categories from the equal-spacing assumption.