Available Classifications
Alternate structures
Alphabetical Indexes
Correspondence Tables
Free downloads
International Family of Classifications
  Best practices

National classifications

  Statistical Commission
  Expert Group
  Technical Subgroup
  Workshops and Training


Contact us

UN Glossary of Classification Terms

The following Glossary has been approved by the Expert Group on International Economic and Social Classifications as a working document. Comments and suggestions may be sent to the Classifications Hotline.

A | B | C | D | E | F | G | H | I | J | K | L | M | N

O | P | Q | R | S | T | U | V | W | X | Y | Z



Refers to a normalised shortened name (e.g. ISIC Rev 3 for the International Standard Classification of All Economic Activities, Revision 3; CPC V1.0 for the Central Product Classification Version 1.0; etc).


Aggregation is the combination of related categories, usually within a common branch of a hierarchy, to provide information at a broader level to that at which detailed observations are taken. Disaggregation is the breakdown of observations, usually within a common branch of a hierarchy, to a more detailed level to that at which detailed observations are taken. With standard hierarchical classifications, statistics for related categories can be grouped or collated (aggregated) to provide a broader picture, or categories can be split (disaggregated) when finer details are required and made possible by the codes given to the primary observations.


Balance of Payments Manual 5 (BPM5)

The manual describes the methodology for measuring the economic transactions of an economy with the rest of the world. The International Monetary Fund is the custodian of BPM5.

Source: BPM5, para 13.

Best Practices about Classifications

Refers to the approach or procedure recognised as most efficient and effective in producing a desired result. Best practice is based on the experience of experts in particular fields and is usually promulgated through the agreement and endorsement of experts and expert groups.

In the development and revision of international and national classifications, best practice would generally involve the application of practices and procedures promulgated by international and national organisations responsible for classification development in their own particular fields. These practices may well include cost benefit analyses weighing the applicability of final classifications against of their terms of reference, the application of agreed classification principles, an agreed methodology for incorporating local requirements (i.e. an evaluation of the requirements of the society/economy where the classification is to be applied) where they differ from existing standards and the selection of suitable recognised classification characteristics to produce good classifications. The result should optimise the incorporation of these principles in a product that is achievable within budgetary and other constraints.

Refer also to Custodian of a classification.


Represents the limit of a known or recognisable quantity, area or scope. Each classification has its own boundary, as do its constituent categories, such as activities, commodities, occupations etc. Whilst it is possible for the boundaries of individual classifications to overlap, there should be no overlap within individual classifications.

Building blocks (elementary items)

Are the most elementary items of a statistical classification, i.e. the most detailed code for a variable. They may be used alone or in combination to describe a category in one or more classifications, or to compare classifications. A prime example is the Harmonized Commodity Description and Coding System (HCDCS or HS), the categories of which are used not only for the construction of country specific tariff and trade classifications, but also as the building blocks of the Standard International Trade Classification (SITC Rev. 3) and the goods component of the Central Product Classification (CPC). The SITC and the CPC regroup individual HS categories to meet differing statistical needs. Another example is the General Industrial Classification of Economic Activities within the European Community (NACE) which can be combined to reconstruct higher levels of ISIC.



Is the generic term for items at any level within a classification, typically tabulation categories, sections, subsections, divisions, subdivisions, groups, subgroups, classes and subclasses. Classification categories are usually identified by codes (alphabetical or numerical) which provide both a unique identifier for each category and denote their place within the hierarchy. They contain elements which are subsets of the classification to which they belong, such as activities, products, types of occupations, types of education, etc.

Refer also to Item.


Is a title/name used in classifications to depict a particular level within a hierarchy (e.g. Section, Division, Group, Class). It usually refers to the one of the lower levels of a classification, often the lowest (e.g. in ISIC Rev. 3 the lowest level - 4 digit - is referred to as the class, while in the CPC the class level is the second lowest level). Its use is not mandatory.

Refer also to Division, Level.


Is a set of discrete, exhaustive and mutually exclusive observations which can be assigned to one or more variables to be measured in the collation and/or presentation of data. The terms 'classification' and 'nomenclature' are often used interchangeably, despite the definition of a 'nomenclature' being narrower than that of a 'classification'.

The structure of a classification can be either hierarchical or flat. Hierarchical classifications range from the broadest level (e.g. division) to the detailed level (e.g. class). Flat classifications (e.g. sex classification) are not hierarchical.

The characteristics of a good classification are as follows:

- the categories are exhaustive and mutually exclusive (i.e. each member of a population can only be allocated to one category without duplication or omission);

- the classification is comparable to other related (national or international) standard classifications.

- the categories are stable i.e. they are not changed too frequently, or without proper review, justification and documentation;

- the categories are well described with a title in a standard format and backed up by explanatory notes, coding indexes, coders and correspondence tables to related classifications (including earlier versions of the same classification);

- the categories are well balanced within the limits set by the principles for the classification (i.e. not too many or too few categories). This is usually established by applying significance criteria (e.g. size limits on variables such as employment, turnover, etc.)

- the categories reflect realities of the field (e.g. the society or economy) to which they relate (e.g. in an industry classification, the categories should reflect the total picture of industrial activities of the country); and

- the classification is backed up by availability of instructions, manuals, coding indexes, handbooks and training.

Refer also to Nomenclature.

Classification structure

Refers to how the categories of a classification are arranged, grouped and sub-divided.

The categories of a classification can be arranged in either a hierarchical or flat structure. In flat classifications the categories are arranged at a single level. Hierarchical classifications have several levels corresponding to different degrees of resolution (detail) in the measurement (specification) of the variable being observed.

Classification unit

Is the basic unit to be classified in the classification (e.g. in an activity classification this would be the establishment or enterprise, in an occupational classification it will be the job).

Refer also to Observation Unit.


The term is used for the act of classifying or assigning a classification code to a unit or observation.

Code, code letters and numbers

Normally consists of one or more alphabetic, numeric or alpha/numeric characters assigned to a descriptor in a classification. Each code is unique to a property within a classification. If the property changes, then the code should also be changed.

Codes can be linked to other codes with common characters, especially in hierarchical classifications. For example, in ISIC Rev. 3, Technical and vocational secondary education has a class code 8022, which is linked to Group 802 Secondary education and to Division 80 Education.

Statistical compilation, storage and retrieval is facilitated by the use of codes with their descriptors.


Refers to the transformation of a textual information about an observation into a code which identifies the correct category (value) for that observation.

Coding index

A detailed, comprehensive list of entries reflecting the information required as the basis for consistent coding.

The code list (of descriptions and codes) is ordered either alphabetically or numerically and could be stored electronically or in hard copies. It acts as a link between responses and the classification, enabling responses to be coded accurately and quickly to the appropriate category of the classification. It may be necessary to compile different coding indexes for coding different observations to the same classification because of the differences in the information provided. The content and structure of a coding index may also depend on the coding methodology being used (i.e. whether coding is done manually, with computer assistance or automatically).

Coding strategy

Refers to the guidelines for systematically, effectively and accurately allocating codes to the categories in a classification.

Coding structure (coding system)

Refers to the systematic numbering/lettering of all the categories in a classification. The use of standardised conventions when creating coding structures helps in using and comparing classifications.

The following numbering systems could be used:

- a strict decimal numbering system - i.e. each category at a particular level in a classification system would have the same type and number of numeric characters;

- a non-strict decimal numbering system - i.e. at a particular level each category in a classification system would not have the same number of numeric characters. This would have been the case in ISCO-88 if a '0' had not been used to indicate that a particular minor group has not been further sub-divided into unit groups . In a mixed character system, e.g. ISIC Rev. 3, where the tabulation categories are identified by letters, the divisions, groups and classes are identified by two, three, and four digit numbers respectively.


Are abstract summaries, general notions, knowledge etc. of a whole set of behaviours, attitudes or characteristics which are seen as having something in common.

Concepts are used to assist in presenting/conveying precise meaning, categorising, interpreting, structuring and making sense of classifications.


Refer to Correspondence table; Correlation.

Content (of a classification category)

Refers to what is included within the boundary of a classification category.

For example, in an industry classification, the content of a detailed classification category (e.g. class) would include a descriptor or title, a summary stating its scope, a list of inclusions and exclusions and a list of primary activities.

Controlled vocabulary

Refers to the vocabulary to be used for specific classifications which have specific meaning as given by the author or agreed by experts. 'Commodity', 'product' and 'service' would be examples of terms in controlled vocabulary used in industry and product classifications.


A general rule, method, or practice established by consensus, usage, or a formal agreement between parties. A convention can apply to both the adoption of particular classifications and to the ways they are interpreted and used.

Formal international Conventions are often signed to establish organisations such as the Customs Co-operative Council (CCC). The functions and purposes of the organisation are included under the terms of the Convention. They may also be designed to commit governments to certain policies or legislation upon ratification, as do e.g. the ILO Labour Conventions.


Refer to Correspondence table and Concordance.

Correspondence table

Is a tool for the linking of classifications. A correspondence table systematically explains where, and to what extent, the categories in one classification may be found in other classifications, or in earlier versions of the same classification. Methodologically, correspondence tables (also referred to as tables) describe the way in which the value sets of classifications are related, by describing how the units classified to the groups defined for a classification would be classified in other classifications.

Tables are important for the development and harmonisation of international classifications. There are many different circumstances under which one may want to establish relations between classifications, and many forms which these relations may take.

Tables can be precise, depending on convention in order to describe the type of link between tables (e.g. historical, hierarchical or whether they overlap).

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.


Specifies the population from which observations for a particular topic can be drawn.

An understanding of coverage is required to facilitate the comparison of data. Coverage issues are often explained through the use of tables showing linkages (e.g. part or full correspondence); and can also be used to explain the ratio of coverage.

The rules and conventions of coverage are largely determined by concept definitions, scope rules, information requirements and, in the case of statistical collections and classifications, collection and counting units and the collection methodology.

Refer also to Scope.

Coverage ratio

Measures the extent to which observations designated as primary to a particular category are undertaken by units primarily involved with the observations related to that category. In industry statistics, the coverage ratio is the output of goods and services characteristic of a particular industry in proportion to the total output of the same goods and services by the economy as a whole).

Cross reference

Is the linking, tracing or comparing of concepts/categories in one classification or between classifications.

This could be done by specifying inclusions/exclusions, footnotes or descriptors in an annotation. Cross referencing draws users' attention to related concepts/categories, inclusions/exclusions etc. in the same or other related classifications.

Refer also to Table.

Custodian of a classification

Refers to an institution or statistical area which has responsibility for development, maintenance, implementation, promulgation and interpretation of classifications. Collaboration among custodians is essential for harmonisation of classifications.

Classifications are often constructed by, or on behalf of, those responsible for policy implementation. In such cases, the administrative agency will normally be the custodian, sometimes in cooperation with the statistical agency. For example, Customs agencies are often the custodians of tariff classifications, even though such classifications are also used for statistical purposes.

Refer also to Best Practices about Classifications.



Is a statement of the precise meaning of something. In classifications this refers to the explanation of the concepts encompassed in category description and often includes specific examples of what is and is not included in particular categories.

Refer also to Explanatory note

Derived classifications

Are based upon reference classifications. Derived classifications may be prepared either by adopting the reference classification structure and categories, providing additional detail beyond that provided by the reference classification, or they may be prepared through rearrangement or aggregation of items from one or more reference classifications. Derived classifications are often tailored for use at the national or multi-national level (e.g. NACE).

Refer also to Related classifications.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York 2-4 November 1998: Preamble: International Family of Economic and Social Classifications.


Is normally a one line statement/heading/index entry of a category in a classification, designed to convey its content.


Refer to Aggregation/Disaggregation.


Is a title/name used in classifications to depict a particular level within a hierarchy (e.g. Section, Division, Group, Class). It usually refers to one of the upper levels of a classification (e.g. in ISIC Rev. 3 the highest level - 1 digit - is referred to as the division, while in the CPC the division level is the second highest level). Its use is not mandatory.

Refer also to Class, Level.


Economic entity

Refers to a legal or social entity, or a group of entities, that engage(s) in economic activities and transactions in its/their own right, such as corporations, non-profit institutions or government units. An economic entity has legal, administrative, or fiduciary arrangements, organisational structures or other parties having the capacity to efficiently allocate resources in order to achieve objectives. Economic entities are often used as a specific classification unit or a statistical unit.

Source: SNA93, para. 4.5.


Refer to Category.

Enterprise unit

An institutional unit or the smallest combination of institutional units that encloses and directly or indirectly controls all necessary functions to carry out its production activities.

Source: ISIC Rev 3.

Explanatory note

Assists users of classifications to distinguish the boundary and scope of each category. Detail varies from classification to classification, but the intention is to explain precise meaning of categories and the underlying concepts. This is often done through the provision specific examples of inclusions and exclusions and cross references to other categories.

Refer also to Definition.


Family (of classifications)

The family of international economic and social classifications is comprised of those classifications that have been internationally approved as guidelines by the United Nations Statistical Commission or other competent inter-governmental board on such matters as economics, demographics, labour, health, education, social welfare, geography, environment and tourism.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: Preamble: International Family of Economic and Social Classifications.


Generally refers to the area or sphere of operation, observation, activity etc. In classifications, field could describe the scope of individual classifications or their constituent groups, categories or items at a given level.

Refer also to Scope.


Format(s) refers to the style or manner of arrangement or procedure used for the presentation of the content and structure of a classification, and may also refer to hard copy or electronic storage, software used, and text font, size and attributes.


A framework is a multi dimensional classification system that seeks to bring in a range of elements. A framework could include a combination of classifications, code lists and/or data items modules, and generally metadata.

The term framework can also be used to describe the skeleton of classification from which a detailed classification is developed. Such a framework encompasses the concepts to be embedded in a classification (e.g. product and activity) and provides the structure for the classification.

At a broader level, the term framework may be used to describe a family of related classifications, such as those produced by the UN.

Source: ABS, Standard Procedure for Creation of Alternate Industry Views and Frameworks, 1999.



In a hierarchical or tree structure classification, categories are grouped ranging from broad to detailed levels for each set. The categories within each set can be grouped (aggregated) or degrouped (disaggregated).

For example, a multi level hierarchical classification would be structured such that the sum of the detail of each level equates to the level above. In this way, observations can be taken at the level of detail of interest of particular purposes. Observations at the lower levels can be summed to provide observations at more aggregated levels (grouped) and, with appropriate manipulations, observations at higher levels can be inferred at lower levels (degrouped).


Refer to the directions or principles used in the development/building, maintenance and application of classifications. Guidelines are not necessarily mandatory, but are provided as an aid to interpretation and use of classifications.

Refer also to Rules.



Classification harmonisation involves the alignment, wherever possible, of the underlying concepts and definitions of both similar and disparate classifications to produce classifications which can related to the maximum extent possible within the constraints of the requirements of individual classifications.

Harmonisation is the process of combining or comparing data for purposes of analysis, either through the use of similar standard definitions and classifications, or through a complex set of explanations on how to achieve comparisons across standards and classifications.

In the harmonisation of classifications, building blocks for common groupings and regroupings of items from different structures of the classifications are identified. The process is facilitated by reducing or eliminating minor differences among the classifications.

Harmonisation of classifications requires continuous co-ordination and exchange of information between the custodians of the relevant classifications on a regular basis. Without such exchange, different interpretations of similar concepts and categories will occur.

In the harmonisation process the classifications could be described as reference, derived or related classifications.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.


Generally this refers to the title or caption of a document or section of a document. In the context of a classification it is the main descriptor of a category, also referred to as Name or Title (e.g. the heading under Division 01 in ISIC Rev. 3 is Agriculture, Hunting and Related Service Activities).

Refer also to Name (of items) and Title (of a category in a classification).


Refers to the classification structure where a classification is arranged in levels of detail from the broadest to the most detailed level. Each level of the classification is defined in terms of the categories at the next lower level of the classification.

Homogeneity (homogeneous)

One of the characteristics of a good classification is reasonably high homogeneity for its categories. Homogeneity is the measure of the degree to which categories consist of components with similar characteristics and is achieved by systematic grouping and stratifying members of the population being classified.

Homogeneity ratios are defined on a mathematical basis to minimise the variance within a classification.



A listing, usually alphabetical, providing pointers to the location within classifications of the observations contained therein. This may be achieved by references to page numbers, paragraph numbers or classification codes. Indexes often contain terminology not expressly used within classifications (synonyms) and may contain cross-references to related observations.

Refer also to Coding index.


Refers to an article or a unit included in enumeration. In classifications, the term item generally applies to a classification category.

Refer also to Category.



Refer to Correspondence table.


Is part of the controlled vocabulary often used to assist in identification or retrieve information about a classification (e.g. keywords used within a software search facility). The term 'Keywords' reflects this function of the words.


Legal entities

Are entities created for purposes of production, mainly corporations and non-profit institutions (NPIs), or government units, including social security funds. They are capable of owning goods and assets, incurring liabilities and engaging in economic activities and transactions with other units in their own right.

Source: SNA93, para. 1.13.


Level denotes the position within the hierarchy of a category or a group of categories.

Refer also to Class, Division.

Life cycle

Refers to the creation, changes and death of a given classification. A classification can be revised due to a number of factors e.g. changes in industries, changes in international standard classifications, etc. Such changes may include the aggregation of disaggregation of items, changes in terminology, additions and/or deletions etc. These changes will result in either a revised version of the existing classification (where changes are essentially in the detail), or a replacement version (where the changes are substantial, involving structural changes, etc).

The life span of a classification is dependent upon a number of factors, most importantly the rate of change of the observations it describes and time series (stability) requirements.


Refer to Boundary and Scope (universe).


Refers to mapping or linking one classification to another. That is each individual group in one classification should be linked with the most appropriate corresponding group(s) in the other. This allows for better management of classifications in a co-ordinated way, and for the transfer from using one classification to using the other.

The first step when establishing linkages should always be to give to the most detailed groups of one classification the code of the most detailed appropriate group in the other. This then allows, when needed, the groups of one classification to be subsequently aggregated to most of the relevant aggregated groups of the other.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.

Local unit

Is defined as an enterprise, or part of an enterprise, which engages in productive activity at or from one location. The definition has only one dimension in that it does not refer to the kind of activity that is carried out. Location may be interpreted according to the purpose - narrowly, such as specific address, or more broadly, such as within province, state, country, etc. Local units are also used as Statistical Units.

Source: SNA93, para. 5.20.



Refers to an institution or statistical area which has or has been given the responsibility for maintaining and/or updating or revising the classification.

Refer also to Custodian.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: Preamble: International Family of Economic and Social Classifications.


Name (of items)

Refers to the descriptions of classification categories, subcategories or their elements e.g. primary activities.

Refer also to Title (of a category in a classification) and Heading.


Applies to a chain of interconnected persons, things, operations etc. In classifications, networking could result in reference, derived or related classifications.

Exchange of information and knowledge across classifications would be facilitated and implemented if national classifications could be presented as part of the web sites of statistical offices and an international cyber platform on the Internet is used for all the major international classification debates.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.


Systematic naming of things or a system of names or terms for things.

In classifications, nomenclature involves a systematic naming of categories or items.

The terms "nomenclature" and "classification" are often used interchangeably, despite the definition of a "classification" being broader than that of a "nomenclature". A nomenclature is essentially a convention for describing observations, whereas a classification structures and codifies the observations as well.

Refer also to Classification.

Normalised heading and codes

This refers to standardisation of headings/titles and their codes in the classifications. For example, the titles and use of n.e.c. (not elsewhere classified) categories and the codes given to them, as well as the code ending in zero should normally be standardised, e.g. by using the title of the aggregate group when naming its n.e.c. group, and by using 9 as the last digit for n.e.c. categories and 0 to be equivalent with coding to the higher level, e.g. because the information needed for a more detailed code is not available.

Not elsewhere classified (n.e.c.), not elsewhere included (n.e.i.) or not elsewhere specified (n.e.s.) residual category

Applies to a subset of a category (e.g. class, group, etc.) which represents those members of the category that do not belong to any of the other, separately identified, categories. (Note that it should not be used as a 'dump' code for observations for which there is insufficient information to assign a detailed code. Such units should be given a code ending in zero to indicate the appropriate aggregate group.) The significance of the observations (e.g. income, employment etc.) for this category should be relatively low compared to those of the other categories in the same more aggregated group of the hierarchical set.



Is defined as to partly cover, cover or extend beyond two or more classification concepts - i.e. not to be completely separate (e.g. "the group xy in classification N is partly 'overlapping' with category yx in classification N, rev. 1").


Partial correlation/correspondence

Occurs where a category of one classification can be coded to two or more categories of other classifications.

Refer also to Correspondence Table and Concordance.


Refer to Subset.


Is the total membership or population or 'universe' of a defined class of people, objects, or events.

There are two types of population viz. target population and survey population. A target population is the population outlined in the survey objects about which information is to be sought and a survey population is the population from which information can be obtained in the survey. The target population is also known as the scope of the survey and the survey population is also known as the coverage of the survey. For administrative records the corresponding populations are: the 'target population' as defined by the relevant legislation and regulations, and the actual 'client population'.


An introductory statement to a classification, outlining the intent, purpose, scope and coverage of a classification.

Primary unit

Is a unit which always can take one and only one value for the variable for which the classification represents a value set. (E.g. the primary unit to be classified by 'occupation' is the 'job'. To classify a 'person' by 'occupation' one first needs to establish a link between the person and an appropriate 'job'.)

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.


A systematic series of actions directed to some end. It is a usually a set of continuous actions or operations undertaken in a defined manner. Process, when applied to classification development and maintenance, is the means by which the concepts and methodology supporting and underlying classifications are incorporated within classifications, thereby promoting classification best practice.


Reference classifications

Are those economic and social classifications that are a product of international agreements approved by the United Nations Statistical Commission or another competent inter-government board, such as that of the International Labour Organisation (ILO), the International Monetary Fund (IMF), the United Nations Educational, Scientific and Cultural Organisation (UNESCO), World Health Organisation (WHO), or the World Customs Organisation (WCO) depending upon the subject matter area. Thus reference classifications have achieved broad acceptance and official agreement and are approved and recommended as guidelines for the preparation of classifications. They may be used as models for the development or revision of other classifications, both with respect to the structure and with respect to the character and definition of the categories.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: Preamble: International Family of Economic and Social Classifications.

Related categories

Are those categories which have some form of elementary relationship. Such related categories can be meaningfully aggregated to give a broad picture or disaggregated when finer details are required. Related categories often have commonality in their codes, due to their common starting point (e.g. an international standard), although relationships can be applied between classifications with different structures and coding systems provided the concepts embedded within the observations under consideration are consistent.

Related classifications

Are classifications which encompass the same or similar observations within different structures and/or to different levels of detail. They often occur as part of a family of classifications, sometimes with a common starting point, such as an international standard classification.

Refer also to Derived classification.

Revised classification

Refers to a classification which replaces the previous classification. A change in a classification does not necessary result in a change of the name of the classification but can be distinguished by use of a version number (e.g. ISIC Rev. 2 was replaced by ISIC Rev. 3).

A revised classification will normally represent a rethinking of conceptual basis, similarity criteria and/or scope, and should be distinguished from an up-dated classification.


Are statements, decisions, judgements, or precedents which provide operational guidelines on the implementation, use, updating and revision of classifications.


Scope (universe)

The scope of a classification is the coverage or sphere of what is to be observed. It is the total membership or population of a defined set of people, objects, or events.

Refer also to Coverage and Population.


Relates to the splitting/separation of topics. Once the scope/coverage of a classification has been defined (e.g. age, language, industry), breakdowns (e.g. beginning from the top) are made, based on similarity criteria, to form groups of elementary blocks or topics and different categories or levels.

Refer also to Aggregation, Disaggregation and Coverage.

Similarity criteria

Refers to the criteria used to define categories in hierarchical classifications (e.g. the grouping of elementary building blocks).

In ISCO-88 the main similarity criteria are the skill level and skill specialisation needed to carry out the tasks and duties of the jobs. Skill level is the main criterion to delineate the most aggregate categories, while skill specialisation is used to delineate the more detailed categories within the aggregate categories.

Source: ISCO-88

Specialisation ratio

Aids in the assessment of the homogeneity of categories within a classification. Specialisation ratios measure the extent to which observations contained within a category are representative of the population of those observations as a whole (e.g. in industry statistics, the specialisation ratio is the output by an industry of goods and services characteristic to that industry in proportion to its total output).

Refer also to Coverage Ratio, Homogeneity.

Source: ISIC Rev. 3, page 35, para 155.

Standard classifications

Are those that follow prescribed rules and are generally recommended and accepted. They aim to ensure that information is classified consistently regardless of the collection, source, point of time etc.

Statistical classification

Refers to a classification constructed for the collection and presentation of numerical facts systematically collected (i.e. statistics). The usefulness of a statistical classification is enhanced if based on or representing a standard classification.

Structural Links

Are correspondence links where opportunities for (direct) correspondence between the categories of different classifications are difficult or not possible to establish, owing to significant structural differences in the defined value sets that do not allow for common correspondence at a similar hierarchical level in the structure. In some circumstances, an approximate or truncated correspondence may be made by aggregating subclasses of one classification to different structural levels of the other classification.

Source: United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November 1998: A Statement of Best Practices.

Structure (tree)

Provides the means for identifying relationships, usually hierarchical, between categories. A hierarchical classification is based on a tree structure where each set of its detailed categories are subsets of categories at the level about the one in which they contained.

Refer also to Hierarchy.


Refers to a breakdown of a division in a hierarchical classification. Two or more subdivisions form a division.


Refers to a secondary or subordinate set whose elements belong to a larger given set. In classifications a group of categories which are defined for a given purpose and can be defined at a given level. In a hierarchical classification, for example, a subset could be those detailed categories below the broad category of the set.

Refer also to Partition.



Statistically this refers to a list of numbers systematically arranged in columns and/or rows. In classification this usually refers to a correspondence/concordance table or a code list.


Refers to the specific meaning of a word used to define/express a concept. The definitions of terms for a particular classification should be precise, accompanied by explanatory definitions which will make them readily understood by non-experts, and agreed upon by the classification experts and main users of the resulting statistics.

Terminology (classification)

Refers to the system of terms commonly used or adapted for use in a classification. Wording or terminology, which may have broader meanings within the wider community, may have specific meaning within the context of given classifications. For example, 'industry' and 'homogeneity ratios' have unique definitions in the context of industry classifications.

Title (of a category in a classification)

Refers to a word or a limited set of words chosen to reflect the content of the category and to differentiate between categories.

In a hierarchical classification the titles of aggregate groups should always be designed to include all the component subgroups, and nothing else. The title of a 'not elsewhere classified' group should reproduce the title of the broader group to which it belongs with n.e.c. added.

Refer also to Heading and Title (of a category in a classification).


Refers to the specific meaning of a word used to define/express a concept. The definitions of terms for a particular classification should be precise, accompanied by explanatory definitions which will make them readily understood by non-experts, and agreed upon by the classification experts and main users of the resulting statistics.


Units (classified)

Refer to entities, respondents to a survey or things used for purpose of calculation or measurement. Their statistics are collected, tabulated and published. They include, among others, businesses, government institutions, individual organisations, institutions, persons, groups, geographical areas and events. They form the population from which data can be collected or upon which observations can be made.


Validity period

Refers to the time period that any document, classification, etc. could be applicable or used. In classification, usually there is an overlapping time period where the old classification could still be used before being superseded by the revised edition.


Is a characteristic of a unit being observed that may assume more than one of a set of values to which a numerical measure or a category from a classification can be assigned (e.g. income, age, weight, etc. and 'occupation', 'industry', 'disease' etc).


A Statement of Best Practices, United Nations, New York (1998): Fourth Meeting of the Expert Group on International Economic and Social Classifications, 2-4 November (1998).

Balance of Payments Manual Fifth Edition (BPM5), International Monetary Fund (1993).

International Standard Industrial Classification of All Economic Activities (ISIC), Rev. 3, Adopted by the twenty-fifth session of the United Nations Statistical Commission, New York, 6-15 February, 1989.

International Standard Classification of Occupation, ISCO-88, International Labour Office, Geneva (1990).

Preamble: International Family of Economic and Social Classifications, United Nations Statistics Division, Fourth Meeting of the Expert Group on International Economic and Social Classifications, New York, 2-4 November (1998).

System of National Accounts, Inter-Secretariat Working Group on National Accounts (1993).

System of National Accounts Glossary, OECD (1999).