Location quotients (LQ) are a transformation of raw data stored in contingency tables.

This transformation has been used frequently in studies involving international trade (under the name ‘revealed comparative advantage’) and in multiple other settings where a total can be disaggregated under at least two independent criteria and thus arranged in a cross sectional table: eg. employment by industry and city, etc.

The spread of LQ values is influenced by row, column and table totals, which prevents us from comparing LQ levels safely across entities, over time or between datasets.

This issue dramatically undermines the usefulness of the location quotient transformation.

In this paper I offer viable paths for characterizing these so-called size distortions in location quotients.