Accuracy and reliability
Contact info
Labour and Income, Social StatisticsDaniel F. Gustafsson
+45 39 17 35 89
Get as PDF
As a result of the smaller sample size and the lack of persons overlapping over between months the quality is lower in the monthly results compared to the quarterly results. This means that only selected key numbers on employment and unemployment are published.
Overall accuracy
The quality of the quarterly LFS is higher compared to the monthly LFS. This is due to the larger sample size and the overlap of respondents between quarters that do not exist between months.
Sampling error
Each week around 2,600 people aged 15-74 are drawn from the population, but it doesn't matter which 2,600 are drawn. Depending on who ends up in the sample, the result can be smaller or larger than if the entire population were in the sample. This uncertainty is called sampling uncertainty and must always be taken into account when interpreting samples, especially when analyzing smaller subgroups.
In order to provide a good description of the uncertainty associated with different either larger or smaller groups in a sample-based survey, intervals of confidence are often used rather than specifying the uncertainty in the form of the standard deviation or variance. In the Labor Force Survey, we have chosen to use intervals of confidence at a 95 significance level. This means: if the survey was repeated 100 times, in 95 out of 100 cases the estimate would be bounded by this interval, while only in 5 cases the estimate would range above or beneath these limits.
The interval of confidence for e.g. the employed persons normally is +/- 22,000, while for the unemployed it is +/- 13,000.
Coefficient of variation for published variables (June 2024): LFS-unemployment rate: 5.5 +- 0.59 Employment rate: 69.5 +- 0.60 Labour force participation rate: 73.6 +- 0.60
Non-sampling error
The non-response in the Danish LFS is relatively large, typically at a level around 50 pct. This is handled by an advanced weighting scheme. The weighting scheme in the monthly LFS is not as complex as the quarterly weighting scheme. The LFS operates with a weighting scheme where the incoming survey results are weighted before being published, so they state results for the whole population aged 15-74. There is drawn on registers as auxiliary information on, e.g. age, gender, region, educational level and status and socio-economic status. In addition, the register-based unemployment is also a part of the weighting scheme.
The monthly weighting scheme is a simplification of the quarterly weighting scheme, since the amount of auxiliary information as well as the number of subgroups is reduced. It has been necessary to simplify the model due to the monthly sample being much smaller than the quarterly sample. This applies e.g. for register-based information and age.
Key figures for sample and answers (June 2024): Sample size: 12957 Number of responses: 5259 Response rate: 41.4 Non response: 58.6
Rate of non-response (June 2024): Refusals: 10.7 No telephone contact: 46.1 Other: 1.8
Quality management
Statistics Denmark follows the recommendations on organisation and management of quality given in the Code of Practice for European Statistics (CoP) and the implementation guidelines given in the Quality Assurance Framework of the European Statistical System (QAF). A Working Group on Quality and a central quality assurance function have been established to continuously carry through control of products and processes.
Quality assurance
Statistics Denmark follows the principles in the Code of Practice for European Statistics (CoP) and uses the Quality Assurance Framework of the European Statistical System (QAF) for the implementation of the principles. This involves continuous decentralized and central control of products and processes based on documentation following international standards. The central quality assurance function reports to the Working Group on Quality. Reports include suggestions for improvement that are assessed, decided and subsequently implemented.
Quality assessment
As a result of the lesser data and the lack of overlap between months, the quality of the monthly LFS is lower compared to the quarterly LFS.
Every quarter a sample of 34,320 persons is drawn based on the CPR-register. Around 2 pct. of this sample can not be contacted either because they are dead or migrated. From the remaining group the typical response rate is between 55 pct.
As with all other surveys that are based on random samples, there is some uncertainty associated with the survey. The uncertainty is due to the selection of the sample and the structure of the non response. Absence is when you do not get to conduct an interview with a selected person. The non response increases the uncertainty of the study because there is not the same probability of obtaining an interview with all selected persons. In other words, there are certain types of groups that are more often not interviewed, and this affects the representativeness of the survey. This applies, for example, to unemployed persons, people with short degrees and ethnic minorities and young people aged 15-24. However, this is handled to a large extent through the enumeration and use of register-based auxiliary information. These are used for calibration and weighting, whereby people who are typically underrepresented in sample surveys will be given a higher weight.
Data revision - policy
Statistics Denmark revises published figures in accordance with the Revision Policy for Statistics Denmark. The common procedures and principles of the Revision Policy are for some statistics supplemented by a specific revision practice.
Data revision practice
There will be ongoing revisions of the monthly results. Firstly, the data is benchmarked against the quarterly series. A new quarterly observation will thus move the monthly observations, also back in time. Secondly, new information about the seasonal pattern will be reflected in changes in the seasonally adjusted series. Both seasonally adjusted and non-seasonally adjusted data may be revised current plus three years back in time. For example, the publication of figures for September 2022 will potentially contain changes to the figures from and including January 2019.