Statistical Classification

added in DDI-Lifecycle 3.3

Statistical Classification is a specialized class of code lists that are managed and shared between statistical agencies. Statistical agencies frequently collect data on similar topics; industry, occupation, education, communication, etc. To meet the needs of statistical agencies the structure needs to manage the development of classifications over time, noting changes and comparison as well as who owns the classification. Comparisons are created between a series or versions of the same statistical classifications as well as between classifications covering the same type of information.

The DDI Lifecycle set of classes dealing with Statistical Classification has a strong relationship with the GSIM model reflecting the ability to organize and describe hierarchical classifications with levels that have meaning in themselves.

A Statistical Classification describes levels in terms of their meaning, relationship to other levels, and the Classification Items contained within the level. DDI focuses on the ability to maintain these objects over time and to reuse contents (such as a classification item) in more than one Statistical Classification (primarily over time). The primary chain of information is the Statistical Classification composed of an organization of one or more levels defined by a concept, containing one or more Classification Items. Statistical Classifications can be grouped in a Classification Series (example North American Industry Classification System (NAICS) versions 2003, 2014, and 2017). Classification Series can be organized into a Classification Family. An example would be Industry which may include Classification Series for NAICS, SIC, and the International Standard Industrial Classification.

Comparison and linkages within and between classifications and series are captured by Classification Maps and often expresses in a Classification Index which supports searching and cross-classification exploration.

DDI has expanded Representations to support the direct use of Statistical Classifications as Value Representations or Response Domain. The structure is similar to that used for expressing Code Lists, Geographic Structure Codes, or Geographic Location Codes. It allows for the use of the full structures, specific levels, ranges of codes, or specified codes. This allows the user to retain the context of codes within their original classification structure.

Figure 1. Overview of Statistical Classification