Clustering at a scale’s endpoints signals the scale is too short.

Evidence

An important part of concept design is whether one can find actual cases with the maximum or minimum values on the scale of the basiclevel concept. The advice to the concept-builder is to think seriously about this issue. One might want to stretch out the scale so that there are very few (if any) cases at either extreme. The underlying idea is that if there are large clumps of cases at either extreme then in reality the end points are not the ideal types and one could go further in either direction. — Gary Goertz, Social Science Concepts, 556

Categorization

• QUESTION: Does this ceiling-effect diagnostic generalize beyond democracy scales to other ordinal measures, e.g. poverty or development indices?