Accuracy and reliability

Overall accuracy

No quantitative measures of uncertainty have been calculated besides the sample uncertainty, which is 0.5 pct. for the total employment. For the construction industry totals the sample uncertainty is under 3.5 pct., while the totals for the types of construction work are under 5 pct. For the cross between industry and type of work, the sample uncertainty can be as high as 50 pct., but generally lies in the range 5-15 pct. The uncertainty of the quarterly changes, however, is far smaller, since the professional units in the sample are not replaced each quarter. The over coverage is estimated to be under 1 pct., un-weighted. The number has been estimated from the number of respondents that point attention to the fact that they have been placed in the wrong industry. The uncertainty that results from non-response, wrong reported numbers and misunderstandings has been sought minimized by control and re-contact with the companies. The variable that the companies are asked about is well defined, so misunderstandings are estimated to be few. The response rate is high.

Sampling error

Sample uncertainties for the scaled employment numbers have been calculated. For the total employment the sample uncertainty is approximately 0.5 pct. For the construction industry totals the sample uncertainty is under 3.5 pct, while the totals for the types of construction work are under 5 pct. For the cross between industry and type of work, the sample uncertainty can be as high as 50 pct, but generally lies in the range 5-15 pct.

Non-sampling error

The statistic has a number of sources of systematic errors. The systematic errors are estimated to be small. The over coverage is estimated to be under 1 pct., un-weighted. The number has been estimated from the number of respondents that point attention to the fact that they have been placed in the wrong industry. The uncertainty that results from non-response, wrong reported numbers and misunderstandings has been sought minimized by control and re-contact with the companies. The variable that the companies are asked about is well defined, so misunderstandings are estimated to be few. The response rate is high.

Quality management

Statistics Denmark follows the recommendations on the organisation and management of quality set out in the Code of Practice for European Statistics and the accompanying implementation model, the Quality Assurance Framework. Read more about these on the page Quality in official statistics.

Quality management is supported by common principles, guidelines and managerial responsibility for quality in statistical products and processes. The aim is to ensure a systematic and consistent approach to quality across the entire statistical production.

Quality assurance

Quality assurance at Statistics Denmark is integrated into the statistical production processes and is carried out in accordance with the principles of the Code of Practice for European Statistics. Read more on the page Quality in official statistics.

Quality assurance includes the continuous use of standardised procedures, documentation in line with international standards and internal controls related to data collection, data processing and dissemination. Experience gained from production is continuously used to identify needs for adjustments and improvements.

Quality assessment

The quality of the statistic is estimated as being high. For the total employment the sample uncertainty is approximately 0.5 pct. For the construction industry totals the sample uncertainty is under 3.5 pct, while the totals for the types of construction work are under 5 pct. For the cross between industry and type of work, the sample uncertainty can be as high as 50 pct, but generally lies in the range 5-15 pct. The over coverage is estimated to be under 1 pct., un-weighted. The number has been estimated from the number of respondents that point attention to the fact that they have been placed in the wrong industry. The uncertainty that results from non-response, wrong reported numbers and misunderstandings has been sought minimized by control and re-contact with the companies. The variable that the companies are asked about is well defined, so misunderstandings are estimated to be few. The response rate is high. By publication, around 90 pct. of the professional units have answered.

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

Only final numbers are reported. In case of changes in the statistical methods, revisions of earlier numbers are performed. The last time this happened was in the 3. quarter of 2013.