Skip to content
Back to search results

HISCO Occupational Coding

Introduced by
Marco H. D. van Leeuwen, Ineke Maas, and Andrew Miles
Year
2002
Type
measurement-classification
Family
occupational-measurement
Font size

What does it do?

HISCO, the Historical International Standard Classification of Occupations, is a coding scheme that translates the messy free-text occupational descriptions found in historical records into a single standardized, hierarchical classification.

Built by van Leeuwen, Maas, and Miles on the framework of the modern ILO ISCO-68, it lets a 'cordwainer' in eighteenth-century England, a 'cordonnier' in France, and a 'Schuhmacher' in Germany all resolve to the same shoemaker code. Occupational titles are the single richest indicator of social and economic position in pre-modern records, appearing in censuses, marriage registers, tax lists, and directories. But their idiosyncratic spelling, multilingual diversity, and archaic vocabulary make raw comparison impossible. HISCO provides the controlled vocabulary and coding rules that turn these strings into analysable data, forming the indispensable first layer beneath class schemes, mobility studies, and occupational structure reconstruction.

The intuition behind the method

A historical record might describe a man as a 'whitesmith', 'tinman', 'ferblantier', or simply 'worker in metals', and a researcher wants to know whether these are the same job, different jobs, or the same job seen through different clerks' habits. HISCO answers by giving every distinct occupation a fixed numerical address in a tree of related work, organised the way the modern international standard organises today's jobs. Coding is like sorting an enormous multilingual lost-property office: each shoe, hammer, or quill on a title must be matched to the right drawer, accounting for dialect, abbreviation, and the clerk's loose phrasing. Crucially, HISCO records not just what the occupation is but its status (master, journeyman, apprentice) and what is made or handled. Once every title carries this stable code, a researcher can finally count how many shoemakers a town had in 1750 versus 1850, regardless of how each enumerator chose to spell the trade.

History

HISCO grew out of the Historical Sample of the Netherlands and parallel European projects that needed comparable occupational data across countries. Van Leeuwen, Maas, and Miles published the scheme and its index in 2002, deliberately anchoring it to the ILO's ISCO-68 so historical and modern data could speak. It rapidly became the international standard, adopted by the North Atlantic Population Project, IPUMS, and numerous national historical population databases. A 2011 companion volume extended the apparatus into derived schemes, notably HISCLASS, and refined coding guidance. Ongoing work has produced automated and machine-learning coding tools and country-specific extensions, while the core five-digit framework has remained stable for two decades.

When to use it

Use HISCO whenever a historical dataset contains occupational titles and you need them in comparable, analysable form, especially across regions, languages, or long time spans. It is the necessary precondition for HISCLASS social-class coding, for reconstructing occupational structure, and for any mobility analysis built on occupations. It is most valuable in large micro-data projects, civil registers, censuses, and genealogical databases, where consistent coding unlocks comparison. It is less essential for a single small dataset analysed qualitatively, though even there it aids comparison with the wider literature. Avoid treating HISCO as a class measure in itself; it classifies occupations, and social standing requires a further scheme such as HISCLASS.

Limitations and considerations

  • Labour-intensive coding requiring linguistic and historical expertise
  • Female, service, and by-employment occupations are systematically under-recorded
  • Identical titles can mask very different real work across contexts
  • Built on ISCO-68 categories that fit some pre-industrial trades awkwardly

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Enables consistent occupational comparison across languages, regions, and centuries
  • Hierarchical structure supports analysis at varying levels of aggregation
  • Status, relation, and product modifiers preserve socially salient detail
  • Serves as the validated foundation for class coding and mobility research

Applications

  • Coding census and civil-register occupations in IPUMS and NAPP projects
  • Preparing occupational data for HISCLASS social-class assignment
  • Comparing occupational composition across European countries over time
  • Building genealogical and population-database occupation fields consistently
  • Underpinning intergenerational mobility studies using marriage-register occupations

How it works

How it works

1

Coding starts by extracting the verbatim occupational strings exactly as written in the source, preserving original spelling, language, and any qualifiers. Faithful transcription matters because abbreviations, status words, and the product worked on all carry information that later coding steps exploit. Each unique string is typically catalogued with its frequency, so that the most common titles can be coded first and applied automatically to identical entries across the dataset.

Formula
O = \{o_1, o_2, \ldots, o_n\}\ \text{(verbatim title strings)}
2

Raw strings are then normalised: spelling variants are reconciled, abbreviations expanded, and the language identified. This step builds a dictionary mapping the bewildering orthographic diversity of historical clerks onto canonical forms. Normalisation must be careful not to erase genuine distinctions; 'master weaver' and 'weaver' differ in status, and a normalisation that collapsed them would destroy information that HISCO is specifically designed to retain in its status modifiers.

Formula
o_i \to \tilde{o}_i\ \text{(spelling, language, abbreviation normalized)}
3

Each normalised title is assigned a five-digit HISCO code locating it within the hierarchical occupational tree inherited from ISCO-68: major group, minor group, unit group, and specific occupation. Coders consult the published HISCO index and decision rules, which resolve ambiguous or compound titles. Well-documented projects code the high-frequency titles by hand and propagate codes to identical strings, reserving manual judgement for rare or ambiguous descriptions.

Formula
\tilde{o}_i \mapsto c_i \in \{\text{HISCO 5-digit codes}\}
4

Beyond the core occupation, HISCO captures three modifiers: status (employer, employee, master, apprentice), relation (whether the title describes the person or a relative, as with 'shoemaker's widow'), and product (the material or good worked). These modifiers preserve socially salient distinctions that the bare occupation code omits and are essential inputs to downstream class coding, since skill and supervisory position depend heavily on master-versus-journeyman status.

Formula
(c_i,\, s_i,\, r_i)\ \text{status/relation/product modifiers}
5

Finally coding quality is assessed, classically through inter-coder reliability measured by Cohen's kappa, where independent coders code a sample and agreement beyond chance is quantified. Discrepancies are adjudicated to refine the coding dictionary and decision rules. Validation also checks coverage, the share of titles successfully coded, and flags systematic gaps, such as female or service occupations that historical sources and the scheme alike tend to under-specify.

Formula
\kappa = \frac{p_o - p_e}{1 - p_e}\ \text{(inter-coder agreement)}

Common pitfalls

  • Coding the bare occupation while discarding status modifiers needed for class analysis
  • Over-normalising and collapsing genuinely distinct titles into one code
  • Ignoring under-enumeration of women's and secondary occupations in the source
  • Reporting coverage without inter-coder reliability, hiding inconsistent coding

Frequently asked questions

How is HISCO different from modern ISCO?

HISCO adapts the structure of ISCO-68 but adds occupations and distinctions specific to the pre-industrial and early-modern world, such as artisanal trades and status modifiers like master and journeyman. It also incorporates multilingual historical title dictionaries. The shared backbone means HISCO-coded historical data can be compared, with care, against modern ISCO-coded data, supporting very long-run occupational analysis.

Can HISCO coding be automated?

Partly. High-frequency titles are commonly coded once and propagated to identical strings, and machine-learning classifiers now suggest codes for new titles. But ambiguous, compound, or rare descriptions still require expert human judgement, and automated suggestions should be validated against a hand-coded sample. The most reliable projects combine automation for volume with manual coding and inter-coder reliability checks for quality.

Does HISCO measure social class?

No. HISCO classifies occupations, not social standing. Two occupations with similar status can sit in very different HISCO groups, and class depends on skill and supervisory dimensions that the occupation code alone does not express. To obtain class you apply a derived scheme such as HISCLASS, which translates HISCO codes plus status modifiers into ranked social classes.

Related methods

Sources

  1. van Leeuwen, M. H. D., Maas, I., & Miles, A. (2002). HISCO: Historical International Standard Classification of Occupations. Leuven University Press.

    ISBN: 9789058671967
    Copied.
  2. van Leeuwen, M. H. D., & Maas, I. (2011). HISCLASS: A Historical International Social Class Scheme. Leuven University Press.

    ISBN: 9789058678577
    Copied.