Statistical processing

Source data

The Labour Force Survey is quarterly based on a stratified sample. The sample was reduced in the 1st quarter of 2016. The reduction will be implemented successively and the sample size will be reduced from 40,532 individuals to 34.320 persons aged 15 to 74 years in the 1st quarter of 2017 when the reduction is fully implemented. Until the year 2020 the LFS has been collected at the individual level for 15-74-year-olds. From 2021 the population has changed to also include the age group from 75-89 years. On a quarterly basis the sample has thus increased from 34,320 people to 36,020 people.

In 2022q1, a new stratification was introduced in the LFS. Register data on employment and register unemployment are utilized to a greater extent, in order to obtain a greater number of responses in some of the groups that suffer from low response rates. For starters, the population is divided into four groups: 1) in stable employment, 2) in registered unemployment, 3) neither in stable employment nor registered unemployment and 4) persons aged 75 to 89. In addition, groups 1 and 3 are divided into the age groups 15-29, 30-64 and 65-74. This results in eight different strata, which are used in the sample selection from 2022q1 onwards. As the LFS consists of four panels, each appearing for six quarters, there will be a gradual implementation of the new stratification until 2023q2.

Different administrative resources are used to select the sample. Administrative sources are also used to obtain various background information on the people interviewed, for example on educational level or workplace.

These registers (among others) are being used for the Labour Force Survey: · Central Population Register (CPR) · Population Register · The Register of Labour Market Statistics (RAM) · Register based-labour force statistics (RAS) · Education classification (DISCED) · Employees . Register of income . Business statistics register ·

Frequency of data collection

Every respondent is interviewed about one specific reference week and the interviews are conducted daily all year round. Respondents are surveyed four times. First two quarters in a row and after a break for two quarters the respondents are interviewed again for two quarters in a row. This implies that half the sample is renewed each quarter.

Data collection

The interviews are conducted by online interview or telephone.

Every respondent is interviewed about one specific reference week (Monday to Sunday) and the interviews are conducted every day all year. The survey is conducted quarterly and is based on a sample of the population. The interviews are conducted by online interview or telephone. From 2017, all the invitation letters to the survey will be sent via E-box, with the exception of those who are not enrolled in E-box. These people will receive a normal letter.

Sample size: The Labour Forces Survey is based on a quarterly sample. The sample is divided into 13 sub-samples of equal size, one for each week in the interview quarter, and people are interviewed with reference to one of the reference weeks.

Q2 2024 In connection with this publication, the seasonal adjustment models have been reviewed and updated. This affects the main series for primarily employment and LFS unemployment. The biggest change is in the LFS unemployment rate, where the seasonal outlier that was added in connection with the publication for Q4 2023, see News from Statistics Denmark February 22, 2024 - no. 46 , has been replaced with a level change. This has been done as it is evident in the subsequent quarters' data that there is a permanent shift to a higher level of LFS unemployment. The permanently higher level is a result of a change in the data collection method, where respondents now have four weeks to answer the questionnaire as opposed to the previous two weeks. This has resulted in a change in the composition of the responses. However, the changed composition is not fully corrected in the weight calibration, which is why a data break in the LFS unemployment rate of up to one percentage point can be observed. The break is due to a higher response rate among some groups that have previously been very unlikely to respond. In the coming years, the weight calibration will be monitored with a special focus on the data break. The new level of LFS unemployment is higher than the number of registered unemployed. The methodological and data reasons for this are described in Ledighedsbegreber.pdf (In Danish only)

Q1 2024 Over the past two quarters, the LFS unemployment rate has increased significantly, which to a large extent can be attributed to a change in the data collection methodology implemented in Q4 2023.

Q4 2023 The unemployment figures for 2023q4 are classified as a seasonal outlier, as an increase in the length of the data collection period has caused a shift in the seasonal pattern. This causes the seasonally adjusted figures to exhibit a high level, whereas the seasonally adjusted figures do not. Read more about the changes in the document on 4th quarter 2023: [Beskrivelse af særlige forhold ved udgivelse af AKU for fjerde kvartal 2023.pdf] (In Danish only)(/Site/Dst/SingleFiles/kvaldeklbilag.aspx?filename=1b8e8a9e-f35f-4ce2-933d-ca9ae4820be7)

New regulation in the LFS from 2021 From January 1st 2021 the LFS is adapted to a new EU framework regulation Regulation 2019/1700. The new regulation applies to all EU countries and is introduced simultaneously throughout the EU by the 1st of January 2021. The new regulation entails changes in definitions of some variables and in the order in which the questions are asked. In addition a number of new variables are included at the same time as other variables are removed. One of the more significant changes concerns temporary / long-term absence, which means that persons who, for certain reasons of absence are completely absent from work and where the absence has lasted or will last for three months or longer and are not receiving a salary or unemployment benefit beyond the three months are not counted as employed.

Questionnaire 2021 Until the year 2020 the LFS has been collected at the individual level for 15-74-year-olds. From 2021 the population has changed to also include the age group from 75-89 years. On a quarterly basis the sample has thus increased from 34,320 people to 36,020 people.

The new weighting scheme has also been changed to also include the age group 75-89 years. For both the age group 65-74 year olds and 75-89 year olds it applies that so far they are only reported to (and published by) Eurostat and are not included in the national dissemination via http://www.Statbank.dk.

Before 2016, individuals who had research protection were not interviewed, but this protection has been removed from the first quarter of 2016. This effectively means an expansion of the number of people that actually can be interviewed. This is compensated by reducing the number of people who are drawn out to the first panel from 10,133 persons to 8,580. The total quarterly sample will in the long term be reduced from 40,532 to 34,320 people in Q2 2017 when the reduction is fully implemented. A series of tests on this show, however, that it does not appear to have influenced the figures for the labor market participation.

Different administrative resources are used to select the sample. Administrative sources are also used to obtain various background information on the people interviewed, for example on educational level or workplace.

These registers (among others) are being used for the survey: - Central Population Register (CPR) - Population Register - The Register of Labour Market Statistics (RAM) - Register based-labour force statistics (RAS) - Education classification (BUE)

The Labour Force Survey is based on telephone interviews and is conducted every day, every week, all year. The survey is a rotating panel survey including four waves each quarter. Due to the design respondents participate in the survey several times. During one and a half years respondents participate four times. First in two quarters in a row, then an interval of two quarters and then participations in two quarters again. The purpose of the design is to have a theoretical overlap of 50 percent in order to be able to measure both quarterly and yearly changes of employment and unemployment.

The survey has a multimode design and is conducted via telephone interviews and web interviews. The invitation to the survey is sent via e-box with a direct link to the web questionnaire. The respondent has four weeks to respond to the survey. During the first week, data is only collected through web interviews, after which the telephone interviewing starts. Around 2/3 of the interviews are collected via web, the rest from telephone interviews.

• Questionnaire (in Danish only).pdf
• Dokumentation and Concepts
• International definitions.
• Codification

Data validation

In the Labour Force Survey 'depending interviewing' is used , which means that previous answers are used to reduce the response burden and to limit the amount of time used to manually seek out errors. The interviewed persons validate previous answers.

All interviews are auto-encoded by computer and then manually encoded and corrected. The respondents will be asked about occupation and industry, and it is these data, which is used from previous rounds of interviews. Some of these codes of occupation and industry will be encoded automatically. The remaining fields are manually encoded.

All data are aggregated and the weighted and not weighted data are compared with data from the previous quarter and to the same quarter the year before.

Data compilation

Each quarter a sample is selected from the Population Register. The sample is divided into 13 sub-samples of equal size, one for each week in the interview quarter, and people are interviewed with reference to one of the reference weeks. However, around 15 per cent of the sample size cannot be contacted either because they have passed away, or have emigrated. Of the remaining group the response rate is usually between 63 to 68 per cent. The population used to for enumeration of sample data, usually has its reference two quarters prior to the stated quarter.

Stratified sampling and weighting: In order to measure unemployment adequately, former unemployed people are selected with a higher probability than others. The sample e.g. of earlier registered unemployed is disproportionately increased, due to the coherence between people registered as unemployed in an earlier quarter and in the present one. The purpose is to ensure a sufficient number of observations of unemployed people to be able to make proper analysis of them. This stratification is taken into account in the weighting of the results. Furthermore, in weighting the following distributions are taken into account: gender, age, registered unemployment, income, socio-economic status, education, immigration, region and mobility.

Before 2016 , individuals who had research protection were not interviewed, but this protection has been removed from the first quarter of 2016. This effectively means an expansion of the sample for which compensated by reducing the number of people who are drawn out to the first panel from 10,133 persons to 8,580. The total quarterly sample will in the long term be reduced from 40,532 to 34,320 people. A series of tests on this show, however, that it does not appear to have influenced the figures for the labor market participation.

Adjustment

The main figures of the LFS are seasonally adjusted: Employed, unemployed, persons outside the labour market.

Since employment and unemployment are now only somewhat affected by the special circumstances caused by COVID-19, the whole COVID-19 period can now be treated more comprehensively in the seasonal adjustment. Therefore, the time series for seasonally adjusted employment, unemployment and persons outside the labor force have been reviewed, in order to improve on the treatment of outliers in the seasonal adjustment. The review of the seasonal models affects the seasonally adjusted figures in the Statbank from 2019q1 onwards.

The seasonally adjusted series in Statbank Denmark are revised three whole calendar years plus the current year. Older data are basically final.