Canberra Manual

Why a separate manual

In the early 1990s, as the manual puts it, there was no generally accepted statistical framework in the OECD area for analysing human resources in science and technology, few countries had any formal definition, and there was very little internationally comparable data (§6). Education and labour force statistics contained relevant information, but they had rarely been collected with science and technology policy in mind (§7).

The manual set out to provide a framework for data on the stocks and flows of HRST (§8), built on existing classifications and data sources, while also guiding future surveys (§§9–10). Policy questions behind it included skill shortages and surpluses, ageing of the workforce, international mobility of specialists (“brain drain” and “brain gain”) and the participation of women (§26, §35, §§40–46).

The definition

“HRST are people who fulfil one or other of the following conditions: a) successfully completed education at the third level in an S&T field of study; b) not formally qualified as above, but employed in a S&T occupation where the above qualifications are normally required.”

Canberra Manual (1995), §49

The two conditions correspond to two views of the labour market. Qualifications describe supply: who could work in science and technology. Occupations describe demand: who is actually needed and employed there (§17, §48).

The two criteria behave differently. Education is a one-off event: once you have a degree, you are HRST for life, whatever you do. Occupation is renewable: someone who is HRST only because of their job stops being HRST when they change jobs, retire or become unemployed (§54). Education is also easier to measure; occupation data often rely on self-description and are less reliable (§55).

“Science” is meant broadly. The manual notes that English speakers tend to think of the natural sciences and engineering, while the German Wissenschaft covers more fields. It therefore starts with a wide definition and then marks out the fields of greatest interest (§19, §50).

Core, reserve and HRST by occupation only

Combining the two tests gives three groups (§§57–58, Figure 3.1):

HRST core

People with a tertiary qualification in a science and technology field who also work in a science and technology occupation. Example from the manual: a university professor with a PhD in economics, or a dentist with her own practice.

HRST reserve

People with the qualification who do not work in such an occupation: an unemployed marine biologist, an engineer at home raising children, a professional sportswoman with a medical degree.

HRST by occupation only

People in science and technology jobs without the formal qualification: a programmer who did not finish secondary school, or a sales manager who learned on the job.

A secretary in an R&D department or a clerk in a university library is not HRST, unless they have a tertiary qualification: the work is in a scientific setting, but at too low a skill level (§58).

Because qualification counts for life, the definition is deliberately wide. “Taxi drivers with PhDs are included, and so are retired people, whatever their age” (§59). The reasoning is that basic knowledge can be “polished up” and used again. As a practical limit, the manual recommends giving priority to data on HRST aged 70 or less (§60).

University level and technician level

HRST is split by skill level (§62). University-level HRST hold a first or postgraduate university degree, or work in an occupation that normally requires one. Technician-level HRST hold a tertiary award below a first university degree, or work in an occupation that normally requires one. Someone qualified at one level and working at the other counts at university level (§§64–65).

In the 1976 version of the International Standard Classification of Education (ISCED) used by the manual, university level meant ISCED categories 6 and 7, technician level ISCED 5 (§§67–68). A separate category for PhD holders was considered and dropped on practical grounds: doctoral education varied too much between countries, and matching occupations were hard to identify (§20, §70).

The choice matters for the numbers. Citing OECD education indicators, the manual notes that including technician-level staff almost doubles the number of HRST in the OECD area (§23).

Which fields count: core, extended, complete

For international comparisons, the 21 ISCED fields of study are grouped into seven broad fields (§71). Not all of them are always collected. The manual proposes three types of coverage, chosen for each combination of level and field (§§72–76):

So a PhD in meteorology or a university degree in economics falls into the core coverage; a PhD in English literature or a technician award in electrical engineering into the extended coverage; a non-university award in singing or languages only into complete coverage (§77). Countries can adjust these choices, which is why the manual asks for data to be presented in detailed form (§76).

Which jobs count

Occupations are classified with the International Standard Classification of Occupations (ISCO), in the manual's case ISCO-88 (§§78–79). Two major groups are central: Professionals (major group 2), whose work mostly requires university-level skills, and Technicians and associate professionals (major group 3), whose work mostly requires technician-level skills. Everyone working in these two groups is HRST (§§81–85). Legislators, senior officials and managers (major group 1) and the armed forces (major group 0) are considered of “some relevance” (§81).

Stocks and flows

A stock is the number of people meeting the definition at a given moment, for example the number of physics PhDs employed in a country on a given date. A flow is movement into or out of the stock during a period, usually a year (§§106–110). The accounting is simple: stock at the start, plus inflows, minus outflows, equals stock at the end (§114).

Inflows are new graduates, people hired into science and technology jobs without the formal qualification, and qualified immigrants. Outflows are emigrants, deaths and people without a qualification who leave their science and technology job (§116, Figure 4.1). Changes within the stock, such as moving to another sector or earning a higher degree, are “internal flows” (§§111–113). The manual pays particular attention to the education “pipeline” and to foreign students (§27).

HRST versus R&D personnel

HRST and the R&D personnel of the Frascati Manual overlap but are not nested (§105):

• HRST is much broader: it includes qualified people who do not do R&D, those who are unemployed, retired or outside the labour force.
• R&D personnel is broader in one respect: it includes everyone working on R&D regardless of qualification, including supporting staff such as clerks and craft workers, who do not meet the HRST skill threshold.

The manual draws a similar comparison with UNESCO's “scientific and technical personnel”, which is based on the activity people perform and therefore includes auxiliary staff but leaves out the reserve (§102).

How Eurostat measures HRST today

Eurostat publishes HRST statistics based on the Labour Force Survey and the Canberra definitions, updated to current classifications:

• HRSTE (by education): people who have completed tertiary education, ISCED 2011 levels 5 to 8;
• HRSTO (by occupation): people employed as professionals or as technicians and associate professionals in the ISCO classification;
• HRSTC (core): people who meet both criteria.

The figures, according to Eurostat (data extracted January 2026):

• In 2024 the HRST core in the EU amounted to 55.2 million people, a quarter (25.1%) of the total labour force, up from 18.9% in 2014.
• In Poland the HRST core made up 27.7% of the labour force in 2024, above the EU average, and its share rose by 7.8 percentage points between 2014 and 2024, one of the largest increases in the EU.

As with any indicator built on qualifications, a rising share reflects both changes in the economy and the expansion of higher education: anyone who graduates enters HRST by definition.

Edition and status

The manual was published by the OECD in 1995, prepared with the European Commission (DG XII and Eurostat), UNESCO and the International Labour Organisation. It describes itself as the fifth manual in the Frascati family and the first produced jointly by the OECD and the European Commission/Eurostat, and says it “has yet to be tried out in practice”, with a revised version expected after testing (§33).

Some parts are visibly dated: the classifications (ISCED 1976, ISCO-88), the data sources and the references to the European Union of twelve countries. The core of the manual, the two-test definition and the split into core, reserve and occupation-only, is what Eurostat still uses. We have not found a Polish edition, so the Polish version of this page uses our own translations.