Reference metadata describe statistical concepts and methodologies used for the collection and generation of data. They provide information on data quality and, since they are strongly content-oriented, assist users in interpreting the data. Reference metadata, unlike structural metadata, can be decoupled from the data.
The data provide information on persons with tertiary education in the fields of Science, Technology, Engineering and Mathematics (STEM) in the labour market. They are secondary datasets derived from the primary data collected through the EU Labour Force Survey primary data (EU-LFS).
The STEM statistical domain is linked to the wider Human Resources in Science and Technology (HRST) domain, which is also a secondary domain based on EU-LFS data.
These indicators have been developed for analytical and monitoring purposes in the fields of employment, education, skills and research policy. It provides information on the labour-market situation of STEM graduates, a population group of particular relevance for innovation, competitiveness, and the green and digital transitions. The indicator is also relevant in the context of the European Research Area Policy Agenda 2025–2027 and Horizon Europe priorities, notably those related to research careers, talent, gender equality and the effective use of highly qualified human capital. They complements broader labour-market indicators by helping to assess the extent to which STEM qualifications are translated into labour-force participation and employment, and may support the identification of skills shortages, mismatches and underutilisation of human resources.
3.2. Classification system
STEM data are built in accordance with relevant international classification systems:
Persons with tertiary education in STEM fields are classified by fields of education and training by International Standard Classification of Education (ISCED) - ISCED-F 2013;
The composition of the workforce based on its labour status according to the Resolution of the 13th International Conference of Labour Statisticians (ICLS), convened in 1982 by the International Labour Organisation (ILO);
As a general rule the EU-LFS covers all economic sectors.
3.4. Statistical concepts and definitions
STEM statistics provide information on persons with tertiary education in STEM fields at a given point in time. They describe the educational profile of individuals in a given year. The target population consists of persons aged 15–74 with tertiary education (ISCED 2011 levels 5–8) whose field of education falls within the STEM disciplines defined below.
Field of Education (STEM Definition):
According to ISCED-F 2013, includes:
05: Natural sciences, mathematics and statistics
06: Information and Communication Technologies (ICT)
07: Engineering, manufacturing and construction
Students graduating at the ISCED2011 6th level should, however, already be counted as a part of the STEM since enrolment in education at the ISCED2011 level 6 normally requires a degree at the ISCED2011 level 5.
The EU-LFS provides population estimates for the main labour market characteristics, such as employment, unemployment, people outside the labour force, hours of work, occupation, economic activity and other labour related variables, as well as important socio-demographic characteristics, such as sex, age, education.
The definitions of employment and unemployment are the same as used in the primary data source EU-LFS (see EU-LFS metadata under Related metadata section) and follow the definitions and recommendations of the International Labour Organisation. The definitions are clearly stated in the Article 2 of the aforementioned Commission Implementing Regulation 2019/2240.
Breakdowns are provided by sex, age (15–34, 35–54, 55–74, 20–64 and 15–74), labour status, sector of economic activity and occupation, across STEM fields of study.
The age group 15–74 is used to provide a broad view of the labour-market situation, consistent with standard EU-LFS practice for employment, unemployment and inactivity statistics. The age group 20–64 is used for employment analysis in line with the standard EU policy framework for monitoring employment rates among the core working-age population. In the case of STEM graduates, this age group is also particularly suitable for analysing employment outcomes, as it excludes most persons still in initial education.
Labour status distinguishes between persons who are employed, unemployed, and outside the labour force.
To ensure relevance, robustness and sufficient reliability of the results, the variables on sector of economic activity and occupation are presented in aggregated form. The following aggregations are used for economic activity (NACE Rev. 2):
Nace Rev.2 Aggregation structure
Section
Title
A
Agriculture, forestry and fishing
B-D-E
Mining and quarrying; electricity, gas, steam and air conditioning supply; water supply, sewerage, waste management and remediation activities
C
Manufacturing
F
Construction
G-I
Wholesale and retail trade, transport, accommodation and food service activities
J
Information and communication
K_L
Financial and insurance activities; real estate activities
M
Professional, scientific and technical activities
N
Administrative and support service activities
O
Public administration and defence; compulsory social security
P_Q
Education; human health and social work activities
R-U
Arts, entertainment and recreation; other service activities; activities of household and extra-territorial organizations and bodies
The following aggregations are used for occupation (ISCO-08):
ISCO08 Aggregation structure
Major Group
Title
OC1
Managers
OC2
Professionals
OC3
Technicians and associate professionals
OC4_5
Clerical support workers, service and sales workers
OC6-8
Skilled manual workers
OC9
Elementary occupations
STEM statistics versus HRST statistics
The concept of ‘Human Resources in Science and Technology’ (HRST) relates mainly to the education of persons irrespective of their actual professional occupation (‘people who have successfully completed tertiary education or who are employed in science and technology occupations where such education level is normally required’). In contrast, the concept of STEM relates to the actual level and fields of education of persons.
3.5. Statistical unit
The unit of measure for the HRST_STEM data collection is the number of persons in thousand. Shares are expressed as a percentage of persons with tertiary education in STEM fields.
Individual person. The observation units of measurement for which results are obtained from the EU-LFS survey are persons in private households.
3.6. Statistical population
The STEM data covers all persons usually residing in private households in the territory of the reporting countries (Member States of the European Union). Persons living in collective or institutional households do not belong to the target population and are excluded from the EU-LFS.
The data measures the number of persons who graduated in STEM fields disaggregated by labour status, occupation, age, sex and by NACE Rev.2 activity.
The conditions of the above educational requirements are considered according to the internationally harmonized standards of ISCED2011 levels 5-8 and ISCED-F 2013. Labour status is classified according to the definitions used in the EU Labour Force Survey (EU-LFS). Data stem from the EU-LFS.
Eurostat excludes anyone below the age of 15 or over the age of 74 from the STEM population so the STEM statistics are based on the age-group 15-74 years. For the user it may be advisable to use the age-group 20-64 which is available in the STEM tables.
3.7. Reference area
European Union, EU Member States and EFTA countries.
3.8. Coverage - Time
Data are available from 2021 onwards.
3.9. Base period
Not applicable.
STEM statistics are expressed as absolute values, measured as the number of persons in thousands, and as relative values, measured as percentages. The data are disseminated in several datasets with the following units of measure:
Labour status by sex: number of persons in thousands and percentages;
Labour status by age: number of persons in thousands and percentages;
Economic activity (NACE Rev. 2): number of employed persons in thousands and percentages;
Occupation (ISCO08): number of employed persons in thousands and percentages.
Absolute values refer to persons with tertiary education in STEM fields. Percentage indicators are calculated with reference to the target population defined by the breakdown concerned. In each case, the numerator is the number of persons with tertiary education in STEM fields in the category observed, while the denominator is the corresponding total population in that category.
In the dataset on economic activity, percentages are calculated using as denominator the total employed population aged 15–74 in the relevant NACE Rev. 2 category. In the dataset on occupation, percentages are calculated using as denominator the total employed population aged 15–74 in the relevant occupational group.
In the occupation table, persons employed in armed forces occupations are excluded from the target population.
For details concerning the computation and interpretation of units and measures used in STEM tables see Annex 1 under Annexes section.
The EU-LFS annual average dataset is used. For more information see EU-LFS metadata under related metadata section.
6.1. Institutional Mandate - legal acts and other agreements
As part of the science and technology domain and as a secondary domain derived from EU-LFS data, STEM statistics are covered by Regulation (EU) 2019/1700, , the Integrated European Social Statistics Framework Regulation (IESS FR), and by Commission Implementing Regulation (EU) 2019/2240 for the labour force domain, applicable from 1 January 2021.
STEM statistics follow the rules and guidelines of the primary data sources (EU-LFS). For more information see EU-LFS metadata under related metadata section.
8.1. Release calendar
Data are available for release T+5 months after the end of the reference year T.
8.2. Release calendar access
Not applicable.
8.3. Release policy - user access
Equal access is provided to all users through Eurostat’s dissemination channels. For methodological information on the primary source see EU-LFS metadata under related metadata section.
Annual dissemination.
10.1. Dissemination format - News release
None.
10.2. Dissemination format - Publications
STEM statistics may be used in Eurostat publications and in Statistics Explained articles available on Eurostat’s website.
EU-LFS anonymised microdata are available for research purposes. Please refer to access to microdata.
10.5. Dissemination format - other
Not applicable.
10.6. Documentation on methodology
Metadata for the domain Employment and Unemployment which is the main source for EU-LFS based statistics (See EU-LFS metadata under related metadata section).
10.7. Quality management - documentation
The main quality issues relate to the primary data sources (see EU-LFS metadata under Related metadata section).
11.1. Quality assurance
The quality assurance of STEM statistics is linked to the quality of the EU-LFS data. Major milestones in the improvement of EU-LFS quality has been the adoption of Regulation (EU) 2019/1700.
11.2. Quality management - assessment
Quality assurance for STEM statistics is dependent on the primary source (EU-LFS) which goes through the validation procedures before dissemination. However, with respect to the highest quality assurance the raw data used for compilation of STEM statistics are submitted to validation tests and quality checking. The outcomes of the calculations are as well controlled and compared inside domain and among domains. (see EU-LFS and Education metadata under Related metadata section).
EU-LFS and derived STEM statistics have overall high quality. National LFS surveys are considered as reliable sources applying high standards with regard to the methodology. However, the EU-LFS, like any survey, is based upon a sample of the population. The results are therefore subject to the usual types of errors associated with random sampling. Based on the sample size and design in the various Member States, Eurostat implements basic guidelines intended to avoid publication of figures that are unreliable or to give warning for low reliability of the figures.
For a detailed description of the methods and concepts used, as well as for other documents related to the EU-LFS, please consult the EU-LFS (Statistics Explained) webpage.
12.1. Relevance - User Needs
The STEM database is based on data coming from the EU-LFS on highly skilled human resources, described as essential for the development and diffusion of knowledge. They constitute the crucial link between technological progress and economic growth, social development and environmental well-being. Countries and international organisations highlighted the political and economic importance of a need for internationally comparable, harmonized and high-quality data on human resources.
12.2. Relevance - User Satisfaction
Not applicable. A user satisfaction survey may be organised once the STEM domain is well established in the STI statistics.
12.3. Completeness
An analysis of completeness of STEM variables has been conducted for the purpose of populating the new tables in Eurostat’s online database. The results show good completeness.
13.1. Accuracy - overall
As STEM statistics are derived from a sample survey, they are affected by sampling variability and possible non-sampling errors. Their overall accuracy therefore depends on the quality of the primary source data; see the EU-LFS metadata under the related metadata section.
13.2. Sampling error
STEM data comes from the EU-LFS survey. These data are, as with any sample survey based upon a sample of the population, subject to sampling errors. For the EU-LFS survey, experience shows that at national level the survey information provides sufficiently accurate estimates for the levels and structures of the various aggregates into which the labour force is divided, provided that analyses of this type are confined to levels of a certain size.
Reliability of the results is assured by the size of the samples and the sampling methods used, in addition to careful and thorough planning of the various survey operations and rigorous administration of all phases of the survey. Based upon the sample size and design in the various Member States, Eurostat implements basic guidelines intended to avoid publication of figures which are statistically unreliable.
See further details in the metadata for the primary sources under Related metadata section.
13.3. Non-sampling error
See EU-LFS metadata under related metadata section.
14.1. Timeliness
Data are available for release T+5 months after the end of the reference year T.
14.2. Punctuality
It depends on timeliness and availability of primary data, as STEM statistics are normally updated when data for all Member States are available.
15.1. Comparability - geographical
Geographical comparability is considered to be good, as the data are derived from harmonised EU-LFS concepts, definitions and classifications applied across countries.
However, when interpreting differences between countries, users should take into account differences in national labour markets and education systems.
15.2. Comparability - over time
Metadata for the domain Employment and Unemployment which is the main source for EU-LFS based statistics (See EU-LFS metadata under related metadata section).
15.3. Coherence - cross domain
STEM statistics are coherent with Eurostat Labour Force Survey statistics, as they are derived from the same source. Processing errors are considered negligible. Users should, however, take into account the specific STEM field definition when making comparisons with other domains.
15.4. Coherence - internal
With EU-LFS.
Not available.
17.1. Data revision - policy
To further specify the general Eurostat revision policy, the following revision policy has been established for the STEM secondary statistical domain: The EU-LFS team updates the STEM data each time there is an update with the EU-LFS data.
No data revisions are made at this secondary level but instead made at the primary level. Exceptionally, if an error fully attributable to processing at the secondary level is detected, the STEM data would be revised although the primary data (see the ‘Related metadata’ section) remain unchanged.
17.2. Data revision - practice
When new data are provided and validated:
The already disseminated data are updated following the update to these EU-LFS data. The complete time series are updated once a year when new annual Labour Force Survey data are available for all Member States (normally in mid-April of year T+1). However, there may be frequent (relatively minor) updates to the source data (Labour Force Survey data) over the year, in which case the already disseminated data are updated accordingly.
All reported errors (once validated) result in corrections of the disseminated data. Reported errors that are deemed to be significant are corrected in the disseminated data as soon as the correct data have been validated. Corrections for other errors are carried out in connection with the annual scheduled data dissemination. This is consistent with Eurostat error management policy.
18.1. Source data
Data are extracted from the EU-LFS, with data re-aggregated into STEM stocks. The EU-LFS is a survey of households. The basic data are transmitted to Eurostat by the Member States and other countries participating.
18.2. Frequency of data collection
Yearly.
18.3. Data collection
See EU-LFS metadata under related metadata section.
18.4. Data validation
STEM data are validated using comparisons with previous years and the situation in other countries. If anomalies are detected, they are reported back to the primary data sources within Eurostat (EU-LFS). If needed countries are contacted for an explanation. If the result is considered not reliable enough, the data are not disseminated. Breaks in series are identified and flagged in a similar procedure.
For more information see metadata for the primary sources under Related metadata section.
18.5. Data compilation
See EU-LFS metadata under related metadata section.
18.6. Adjustment
See EU-LFS metadata under related metadata section.
The data provide information on persons with tertiary education in the fields of Science, Technology, Engineering and Mathematics (STEM) in the labour market. They are secondary datasets derived from the primary data collected through the EU Labour Force Survey primary data (EU-LFS).
The STEM statistical domain is linked to the wider Human Resources in Science and Technology (HRST) domain, which is also a secondary domain based on EU-LFS data.
These indicators have been developed for analytical and monitoring purposes in the fields of employment, education, skills and research policy. It provides information on the labour-market situation of STEM graduates, a population group of particular relevance for innovation, competitiveness, and the green and digital transitions. The indicator is also relevant in the context of the European Research Area Policy Agenda 2025–2027 and Horizon Europe priorities, notably those related to research careers, talent, gender equality and the effective use of highly qualified human capital. They complements broader labour-market indicators by helping to assess the extent to which STEM qualifications are translated into labour-force participation and employment, and may support the identification of skills shortages, mismatches and underutilisation of human resources.
14 April 2026
STEM statistics provide information on persons with tertiary education in STEM fields at a given point in time. They describe the educational profile of individuals in a given year. The target population consists of persons aged 15–74 with tertiary education (ISCED 2011 levels 5–8) whose field of education falls within the STEM disciplines defined below.
Field of Education (STEM Definition):
According to ISCED-F 2013, includes:
05: Natural sciences, mathematics and statistics
06: Information and Communication Technologies (ICT)
07: Engineering, manufacturing and construction
Students graduating at the ISCED2011 6th level should, however, already be counted as a part of the STEM since enrolment in education at the ISCED2011 level 6 normally requires a degree at the ISCED2011 level 5.
The EU-LFS provides population estimates for the main labour market characteristics, such as employment, unemployment, people outside the labour force, hours of work, occupation, economic activity and other labour related variables, as well as important socio-demographic characteristics, such as sex, age, education.
The definitions of employment and unemployment are the same as used in the primary data source EU-LFS (see EU-LFS metadata under Related metadata section) and follow the definitions and recommendations of the International Labour Organisation. The definitions are clearly stated in the Article 2 of the aforementioned Commission Implementing Regulation 2019/2240.
Breakdowns are provided by sex, age (15–34, 35–54, 55–74, 20–64 and 15–74), labour status, sector of economic activity and occupation, across STEM fields of study.
The age group 15–74 is used to provide a broad view of the labour-market situation, consistent with standard EU-LFS practice for employment, unemployment and inactivity statistics. The age group 20–64 is used for employment analysis in line with the standard EU policy framework for monitoring employment rates among the core working-age population. In the case of STEM graduates, this age group is also particularly suitable for analysing employment outcomes, as it excludes most persons still in initial education.
Labour status distinguishes between persons who are employed, unemployed, and outside the labour force.
To ensure relevance, robustness and sufficient reliability of the results, the variables on sector of economic activity and occupation are presented in aggregated form. The following aggregations are used for economic activity (NACE Rev. 2):
Nace Rev.2 Aggregation structure
Section
Title
A
Agriculture, forestry and fishing
B-D-E
Mining and quarrying; electricity, gas, steam and air conditioning supply; water supply, sewerage, waste management and remediation activities
C
Manufacturing
F
Construction
G-I
Wholesale and retail trade, transport, accommodation and food service activities
J
Information and communication
K_L
Financial and insurance activities; real estate activities
M
Professional, scientific and technical activities
N
Administrative and support service activities
O
Public administration and defence; compulsory social security
P_Q
Education; human health and social work activities
R-U
Arts, entertainment and recreation; other service activities; activities of household and extra-territorial organizations and bodies
The following aggregations are used for occupation (ISCO-08):
ISCO08 Aggregation structure
Major Group
Title
OC1
Managers
OC2
Professionals
OC3
Technicians and associate professionals
OC4_5
Clerical support workers, service and sales workers
OC6-8
Skilled manual workers
OC9
Elementary occupations
STEM statistics versus HRST statistics
The concept of ‘Human Resources in Science and Technology’ (HRST) relates mainly to the education of persons irrespective of their actual professional occupation (‘people who have successfully completed tertiary education or who are employed in science and technology occupations where such education level is normally required’). In contrast, the concept of STEM relates to the actual level and fields of education of persons.
The unit of measure for the HRST_STEM data collection is the number of persons in thousand. Shares are expressed as a percentage of persons with tertiary education in STEM fields.
Individual person. The observation units of measurement for which results are obtained from the EU-LFS survey are persons in private households.
The STEM data covers all persons usually residing in private households in the territory of the reporting countries (Member States of the European Union). Persons living in collective or institutional households do not belong to the target population and are excluded from the EU-LFS.
The data measures the number of persons who graduated in STEM fields disaggregated by labour status, occupation, age, sex and by NACE Rev.2 activity.
The conditions of the above educational requirements are considered according to the internationally harmonized standards of ISCED2011 levels 5-8 and ISCED-F 2013. Labour status is classified according to the definitions used in the EU Labour Force Survey (EU-LFS). Data stem from the EU-LFS.
Eurostat excludes anyone below the age of 15 or over the age of 74 from the STEM population so the STEM statistics are based on the age-group 15-74 years. For the user it may be advisable to use the age-group 20-64 which is available in the STEM tables.
European Union, EU Member States and EFTA countries.
The EU-LFS annual average dataset is used. For more information see EU-LFS metadata under related metadata section.
As STEM statistics are derived from a sample survey, they are affected by sampling variability and possible non-sampling errors. Their overall accuracy therefore depends on the quality of the primary source data; see the EU-LFS metadata under the related metadata section.
STEM statistics are expressed as absolute values, measured as the number of persons in thousands, and as relative values, measured as percentages. The data are disseminated in several datasets with the following units of measure:
Labour status by sex: number of persons in thousands and percentages;
Labour status by age: number of persons in thousands and percentages;
Economic activity (NACE Rev. 2): number of employed persons in thousands and percentages;
Occupation (ISCO08): number of employed persons in thousands and percentages.
Absolute values refer to persons with tertiary education in STEM fields. Percentage indicators are calculated with reference to the target population defined by the breakdown concerned. In each case, the numerator is the number of persons with tertiary education in STEM fields in the category observed, while the denominator is the corresponding total population in that category.
In the dataset on economic activity, percentages are calculated using as denominator the total employed population aged 15–74 in the relevant NACE Rev. 2 category. In the dataset on occupation, percentages are calculated using as denominator the total employed population aged 15–74 in the relevant occupational group.
In the occupation table, persons employed in armed forces occupations are excluded from the target population.
For details concerning the computation and interpretation of units and measures used in STEM tables see Annex 1 under Annexes section.
See EU-LFS metadata under related metadata section.
Data are extracted from the EU-LFS, with data re-aggregated into STEM stocks. The EU-LFS is a survey of households. The basic data are transmitted to Eurostat by the Member States and other countries participating.
Annual dissemination.
Data are available for release T+5 months after the end of the reference year T.
Geographical comparability is considered to be good, as the data are derived from harmonised EU-LFS concepts, definitions and classifications applied across countries.
However, when interpreting differences between countries, users should take into account differences in national labour markets and education systems.
Metadata for the domain Employment and Unemployment which is the main source for EU-LFS based statistics (See EU-LFS metadata under related metadata section).
