Czech Statistical Office

Society Development Statistics Department

Na padesátém 3268/81, 100 82, Praha 10, Czech Republic

The Adult Education Survey (AES) covers adults’ participation in education and training (formal - FED, non-formal - NFE and informal learning - INF). The 2022 AES focuses on people aged 18-69. The reference period for the participation in education and training is the twelve months prior to the interview.

Information available from the AES is grouped around the following topics:

• Participation in formal education, non-formal education and training and informal learning
• Volume of instruction hours
• Characteristics of the learning activities
• Reasons for participating
• Obstacles to participation
• Access to information on learning possibilities and guidance
• Employer financing and costs of learning
• Self-reported language skills

For further information see the 2022 AES legislation (http://ec.europa.eu/eurostat/web/education-and-training/legislation) and the 2022 AES implementation manual (http://ec.europa.eu/eurostat/web/education-and-training/methodology).

- Classification of Learning Activities (CLA, 2016 edition)
- International Standard Classification of Education 2011 (ISCED 2011)
- Classification of Occupations 2008 (ISCO 08)
- Classification of economic activities Rev. 2 (NACE Rev. 2)

AES covers all economic sectors.

Definitions as well as the list of variables covered are available in the 2022 AES implementation manual (http://ec.europa.eu/eurostat/web/education-and-training/methodology).

For specific divergence from AES concepts see table 15.1 “Deviations from 2022 AES concepts and definitions” in annex “CZ - QR tables 2022 AES (excel)”.

Individuals, non-formal learning activities.

Individuals aged 18-69 living in private households.

All parts of the country were included in the survey.

2007 pilot testing (01-03/2008)
2011 AES (07-12/2011)
2016 AES (07-12/2016)

Not applicable.

Number, EUR.

Data collection for 2022 AES: 11.07.2022-30.12.2022

At European level:

Basic legal act: Regulation (EU) 2019/1700
Implementing act: Commission Implementing Regulation (EU) 2021/861

At national level:

No complementary national legislation

Not applicable.

Interviewers keep the list of addresses of the households which they should visit separately from the questionnaires. Questionnaires themselves are therefore anonymous from the beginning.

Strict internal rules are in place at the Czech statistical office to protect data. They are set mainly in the following documents: the CZSO Statistical Confidentiality Policy and the CZSO Security Policy. These policies lay down principles of confidential statistical data protection as well as security of operated important and critical information systems in accordance with the relevant legal regulations, especially:

• Act No. 89/1995 Sb., on the State Statistical Service, as amended,
• Act No. 2016/679 - Protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation)

Anonymised data are available to people in central office (project manager of the survey and people from coordination of households surveys department).

Data are stored in oracle database, only people with authorised logins can access the data.

Release calendar is publicly available on this webpage: https://www.czso.cz/csu/czso/catalogue-of-products

Basic data from AES will be published in Statistical Yearbook in the end of November 2023.

Publication is planned to be published in the end of year 2023. The day of publication will be specified.

On webpages of the Czech Statistical Office: https://www.czso.cz/csu/czso/catalogue-of-products

Users can find information about data future availability in release calendar. There are press releases when publications are published. Users can find the released data in public database, in online publications and on websites. Data are available for all users at the same time. In case of publishing data from AES survey, there are no deviations from general policies.

Every 6 years.

Not available yet.

Not available yet.

Not available yet.

Not applicable.

Anonymized microdata can be disseminated. Only for researchers and for a fee.

Pursuant to the provision of the Section 17 of the State Statistical Service Act, confidential statistical data may be provided for statistical purposes or for the purposes of scientific research. The relevancy of the data for applicant's purposes is examined every time. Data may be provided only in the form, which does not allow direct identification of the reporting unit the data pertain to. For that purpose the data are anonymized in the necessary extent in order to minimize the risk of disclosure via indirect identification of the statistical unit concerned.

Not available.

Not applicable.

See the interviewers instructions for documentation on methodology (in Czech only).

Not applicable.

https://www.czso.cz/documents/10180/23183847/politika_kvality_cz.pdf/9fc1fdda-0303-4049-b734-92837a215d7f?version=1.0 (in Czech only)

There was a course of quality management principles in the Czech Statistical Office in 2021 for managers.

The last time when internal audit on household surveys took place was in 2016.

Quality of the survey is considered to be good. No major issues were encountered during administration of the survey.

The survey theme - education - is usually well received among the general public, and respondents are rather willing to participate in the survey.

The most problematic part of the survey is the length of the questionnaire for people who participated in some education. The amount of time spent on the details of learning activities is excessive if the respondent participated in 2 or more learning activities (especially if more than 1 person is interviewed within the household). It was too detailed and therefore too long interview. Managers of interviewer are afraid that quality of the data may be destroyed by quantum of the questions. There is a big pressure on interviewers, because some respondents want to refuse and some really refuse the interview due to its length.

The reference period of 12 months is in many cases too long for respondents to remember all learning activities they took part in (even with "pre-reminders") so the number of non-formal learning activities or the participation rate in non-formal education may be underrepresented.

Costs and hours in education were sometimes difficult to remember, so they were estimated. Respondents do not always have information about the provider of education. Education of parents (especially by older respondents), level of foreign language skill, guidance in education, long lists of options or categories were in some cases problematic. Concept of Informal Learning was for respondents difficult. However interviewers were trained, they got lots of examples, for respondents themselves it was difficult to find the line between intentional and unintentional learning.

Sensitive variables – income, dropped education, questions about health (the last mentioned were for respondents out of topic in AES, therefore they (and also interviewers) were annoyed, why they were asked such questions in Adult Education Survey).

Proxy interviews remain to be an issue, however compared to the previous AES data collection, their number was significantly higher (12 % for AES 2016, 25 % for AES 2022).

Participation rate in non-formal education – benchmark to LFS data (more frequent) to the goals of The European Pillar of Social Rights Action Plan in cooperation with Ministry of Education, Youth and Sports based on requirement of European Commission.

Overall demand on data from AES (Ministry of Education, Youth and Sports; Department of Adult Education and Personnel Management Charles University; Working Group Life-long learning).

User satisfaction is consulted during bilateral meetings. The meetings are held when the users propose to organize such meetings.

All requested variables were included.

Not applicable.

As far as sampling errors and coverage are concerned, accuracy of 2022 AES is in many aspects related to accuracy of the LFS as the AES was a follow up of two LFS waves.

Young people, employed persons, and people living in highly populated areas are overall more difficult to reach. Socially excluded households are more likely to give up participation in LFS. Weighting procedures were used to adjust for unit non-response.

Regarding non-sampling errors, there may be some underreporting of participation in non-formal education and training activities or their number due to:

• proxy answers (25 % of all interviews) which were allowed for hardly to reach respondents and
• the reference period of 12 months which is in many cases too long for respondents to remember all learning activities they took part in (even with "pre-reminders")

Mistakes can happen during data collection when the interviewer records answers incorrectly or the respondent does not understand the question. Item non-response rate is very small for all variables, so this should not lead to big errors. Bigger problem is unit nonresponse because usually young people with high income tend to refuse to be surveyed. For younger people proxy answers are used very often.

Errors are minimised by trainings for interviewers who are then able to explain meanings of the questions by other words or clarify ambiguities. The CZSO has also different modes of collection to be able to access data from as many people as possible.

Coefficient of variation does not exceed 10 % for all but 2 values.

The sampling error reflects the fact that only a particular sample was surveyed rather than the entire population. It is estimated by the standard error and can be expressed by the square root of the estimate of the sampling variance. The estimation of the sampling variance should ideally take into account the sampling design (e.g. the stratification).

More information on methodology for calculating precision estimates is detailed in the sections below.

The calculation of the coefficients of variation, standard errors and confidence intervals was executed using R software (survey package) with respect to design effect.

One of the main indicators: "Participation in non-formal education and training (age 25–69) in the last 12 months" (individuals who ticked at least once 'yes' in questions 66 to 69 of the 2022 model questionnaire)

• Number of respondents (absolute value for ‘Yes’ answers): 3592
• Estimated proportion (in %): 40.5
• Standard error (in percentage points): 0.50

For details see table 13.2.1 “Sampling errors - indicators for 2022 AES key statistics” in annex “CZ - QR tables 2022 AES (excel)”.

Non-sampling errors are covered by items 13.3.1 - 13.3.5 below.

Some over-coverage exists in the CEUs (census enumeration units) register in terms of recorded residential status (usually about 10% of the samples are coded as “administrative waste” – non residential (recreational or commercial space, temporarily empty flats – no persons with usual residence, demolished buildings not yet deleted from the register). Small CEUs (in terms of number of residential dwellings) are currently dropped from the sampling frame due to practical sampling reasons (approximately 1.26% of the dwellings).

See table 13.3.1.1 “Over-coverage - rate” in annex “CZ - QR tables 2022 AES (excel)”.

Not applicable.

No major measurement errors were encountered.

No official pilot testing of the AES questionnaire on a sub-sample of population was carried out, however the questionnaire was pre-tested several times among social survey experts, selected interviewers and small sample of public.

Questionnaire was translated and the questions formulated by project manager of the survey. The questionnaire was sent to cognitive testing which was conducted by cognitive laboratory. This laboratory consists of several interviewers from different regions who are employees of the CZSO. Their job is to test questionnaires both cognitively and technically (testing of the programme used for recording answers during the field work). After the formulations of the questions were approved, questions were inserted in the software used by the interviewers and tested again (this time technically) by people in cognitive laboratory. The electronic questionnaire design included many logical checks and controls at both data entry and coding process levels and was tested during a significant period of time. Paper questionnaire was clearly organised and was post-coded into the same electronic questionnaire design after the interview. This means that potential measurement error made during PAPI would have been revealed and re-checked with the respondent.

Mistakes can happen during data collection when the interviewer records answers incorrectly or the respondent does not understand the question. Errors are minimised by trainings for interviewers. The Czech Statistical Office has its own stable pool of interviewers that receive regular training in data collection methods. Almost all interviewers took part in one-day face to face training focused solely on the AES methodology. Others were trained by online training. All interviewers had constant access to detailed interviewer instructions in both paper and electronic form (with explanation notes embedded also directly in the electronic questionnaire), as well as the possibility to contact AES (or other surveys) methodologists at any time. All interviewers have their local supervisors, who took part in the AES interviewers training, as well as attended their own separate AES methodology training. All interviewers get methodological manual where each question is described in detail. This manual also describes how the interviewers should proceed during the visits, which kind of propagation they should use and how to communicate with the household.

Description of measures taken to reduce unit non-response: respondents were usually notified in advance about the survey, its importance and its aim by interviewers during the previous visits for the LFS. A leaflet describing the survey and asking respondents to participate was provided for all potential respondents. Respondents participating in the survey also received small gifts. Although CAPI was strongly preferred as the mode of data collection, CATI or self-administered interview using PDF questionnaire were suggested as possibilities for respondents, that would not otherwise participate in the survey. Proxy interviews were allowed for difficult-to-reach respondents. Non-contact was mainly not an issue for the AES as it was conducted as a follow-up of the LFS and contact with the household was already well established.

Main characteristics of non-respondents: Young people with high-income, people from cities.

See table 13.3.3.1 “Unit non-response - rate” in annex “CZ - QR tables 2022 AES (excel)”.

See table 13.3.3.2 “Item non-response rate” in annex “CZ - QR tables 2022 AES (excel)”.

The occurrence of processing errors was minimised by using the electronic questionnaire (see section 13.3.2 for more information) and standard CZSO social surveys infrastructure. Apart from plethora of our own logical checks, the electronic questionnaire also contained the majority of logical checks suggested and used for data validation by Eurostat.

Not applicable.

The reference period for the 2022 AES were the 12 months prior to the interview.

First national results (published in the statistical yearbook) in the end of November 2023.

AES publication is planned in the end of 2023.

All data were delivered in time.

Survey preparation started in April 2021 by devising the first version of the national version of the 2022 AES questionnaire. Electronic and paper versions of the questionnaire (and their linking up to the LFS questionnaire) were prepared mainly during the last quarter of 2021 and the first quarter of 2022. Both versions were tested and adjusted repeatedly during their production.

The interviewer instructions were prepared during the first half of 2022. Also AES leaflets for respondents containing information about the aim of the survey, the importance to participate etc. were prepared, to be handed by the interviewers during the relevant LFS visits in the 1st quarter of 2022. Methodological training of AES interviewers' supervisors took part in March 2022. Paper questionnaires and interviewers instructions were finished and printed in May 2022. One-day face to face interviewers training took part on 11 separate occasions during June 2022.

The Fieldwork period started on 11st July 2022 and lasted until the end of 2022 with some possible overlap (for data transmission) to the first two weeks of January 2023. Reminders and follow-ups were administered during the fieldwork period using standard CZSO processes (repeated visits in different day times, CATI etc., and for AES the possibility for self-administered interviewing in case of otherwise unreachable respondents).

Processing (data quality control, editing, post-coding, validation, transformation into the Eurostat standard data form) took majority of the first half of 2023 with the first transmission of dataset to Eurostat on 30 June 2023. After consultation with Eurostat experts, another (final) version was sent on 10 July 2023.

See table 14.2 “Project phases - dates” in annex “CZ - QR tables 2022 AES (excel)”.

Not applicable.

No geographical divergence from the concepts and definitions for the 2022 AES set in the legislation and in the implementation manual. Since the sample is withdrawn from the register of buildings, there are no problems with comparability between regions.

See table 15.1 “Deviations from 2022 AES concepts and definitions” in annex “CZ - QR tables 2022 AES (excel)”.

No additional variables related to COVID-19 were collected. COVID could be mentioned in the open category "other reason, state" - reason for not participating in education.

Not applicable.

Data are comparable over time unless stated otherwise. Data which are not comparable are flagged.

See table 15.2 “Comparability - over time” in annex “CZ - QR tables 2022 AES (excel)”.

Not applicable.

Some differences of the population remain between the AES and LFS.

See table 15.3 “Coherence - cross-domain” in annex “CZ - QR tables 2022 AES (excel)”.

Not applicable.

Not applicable.

All statistics are coherent within the dataset.

The average time for answering the questionnaire was 15 minutes.

Socio-demographic characteristics are collected within LFS survey, therefore time spent on these questions is not included when calculating average time of AES survey. Time spent on persuading people to join the survey is also not included.

Measures taken to reduce the cost and burden of the survey: the AES survey is embedded in the Labour Force survey, which helps to reduce time spent in the household and also travelling costs. Thanks to two stages of stratification during sampling, driving distances between households are shorter, therefore travelling costs are also lower. It also saves time of the interviewers.

Not applicable.

Not applicable.

Not applicable.

Data source used for building the AES sample frame was the Labour Force Survey. AES was implemented as a follow-up of the LFS. All households with successfully completed LFS questionnaire on the 3th and 5th (last) LFS wave of the 3rd quarter of 2022 and with at least one member in the AES requested age category (18–69 years) were asked to participate in the AES in the second half of 2022. All individuals within the household (in the requested age category), if willing, participated in the survey.

The Labour Force Survey sample is drawn from the Census Enumeration Units and Building Register. The register contains information about dwellings in buildings. The total sample unit for the Labour Force Survey is the dwelling. Dwellings are selected using stratified two-stage probability sample design. At the first sampling stage, small geographical areas (CEUs - census enumeration units) are sampled as primary sampling units with probability proportional to their size in each NUTS4. In the second stage, 6 dwellings are sampled in each sampled CEU.

Census Enumeration Districts (CEUs) constitute the first-stage sampling units. CEUs are small geographical areas covering the whole territory of the country. They are used as enumeration districts during the census, but their use is more general. Continuously updated geographical register is maintained by the CSU, where these units form the basic geographical layer, on which subsequent aggregations are based. This register is the base for an integrated hierarchical geographical information system and is the base for databases of regional indicators and statistical data.

For each CEU, a list of all buildings is maintained in the register. This list is updated from administrative data of the construction authorities (new buildings’, flats’ or commercial premises’ acceptation protocols, demolitions’ protocols). For each building, the number of dwelling units is recorded.

The sampling of CEUs is stratified by region (NUTS4) and municipality size with following ten categories:

• below 200 inhabitants
• 200 – 499 inhabitants
• 500 – 999 inhabitants
• 1 000 – 1 999 inhabitants
• 2 000 – 4 999 inhabitants
• 5 000 – 9 999 inhabitants
• 10 000 – 19 999 inhabitants
• 20 000 – 49 999 inhabitants
• 50 000 – 99 999 inhabitants
• 1000 000 and more inhabitants

Weighting procedures:

The CALMAR software was used for calibration. The truncated linear method provides trimming. Only calibration was used as the method for correcting non-response.

Factors used:

• Gender
• Age groups (18-24, 25-34, 35-54, 55-69)
• Educational attainment (ISCED 0–2, 3–4, 5–8) was used to control the totals
• NUTS - each NUTS3

See also table 18.1 “Source data” in annex “CZ - QR tables 2022 AES (excel)”.

Every 6 years.

See table 18.1 “Source data” in annex “CZ - QR tables 2022 AES (excel)”.

The AES 2022 questionnaire contained a large number of logical checks and controls as well as pop-up notifications for when improbable data values (outliers, logical inconsistencies etc.) were entered at the data collection level. All statistical classifications used in the survey were embedded directly in the questionnaire in form of "dictionary apps" aimed at helping interviewers with classifications of the given phenomena. A special dictionary containing more than 1300 most commonly occurring non-formal education activities was developed specifically for classifying non-formal education and training activities into ISCED-F 2013. Data and main indicators values were monitored continuously during the whole data collection.

Final results are compared to the previous wave, validation programs are used to check microdata according to Eurostat standards.

First stage of editing data is about imputation of item non-response, i.e. imputation of missing data. Thanks to the good job of the interviewers, the imputation rate is very small. The only-one rate is for the variable households income.

See table 18.5.1 “Imputation - rate” in annex “CZ - QR tables 2022 AES (excel)”.

Not applicable.

Not applicable.

None.