Accuracy and reliability

Overall accuracy

The overall uncertainty relates to both sampling uncertainty or random error as well as non-participation in the statistics and partial responses, where answers are missing to some of the form's questions. Both affect the uncertainty of the study's results only to a limited extent.

Sampling error

The results are based on responses from 1,500 companies out of a total population of 14,326 companies with 5-9 full-time employees. The overall response rate in 2021 was 91 per cent.

Sample uncertainty calculations are made every year. Examples of variables from 2021, with estimated proportion of companies and associated lower and upper limits at 95 per cent. statistical confidence interval:

• The proportion of micro-enterprises that have a website is estimated at 84 per cent. The standard error is 0.89 per cent. This means that the actual value is with 95 per cent. probability between 82 per cent. and 86 per cent
• The proportion of micro-enterprises that use cloud solutions is estimated at 34 per cent. The standard error is 1.15 per cent. This means that the actual value is with 95 per cent. probability between 32 per cent. and 36 per cent

The uncertainty calculations are available upon contacting the statistics' contact person.

Non-sampling error

Other uncertainty relates to the systematic uncertainty, i.e. the uncertainty that is not related to the sample (the random error). The main cause of other uncertainty may be that the population cannot be completely captured, or covers too widely. This happens, for example, when the form is sent to companies with fewer than 5 employees. This typically results in the company not answering the form. The questionnaire can be experienced as a great burden for some companies, who may end up not answering all questions with equal accuracy. Uncertain memory can lead to questions being answered inaccurately, just as technically complex questions can be too difficult to answer accurately. In case of non-response, the weights for the companies that have responded are adjusted so that it is assumed that the companies that have not responded are similar to the companies that have responded. Random sampling as close to the time of datacollection as possible, guidance on technical questions on the form and a limited number of questions are examples of measures aiming to reduce these uncertainties.

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

The relevance of the statistics is considered to be significant as the content is largely developed in collaboration with users. The statistics mainly contain yes-no questions, which are relatively easy to answer. This improves the quality of the statistics by means of more accurate and reliable answers. The statistics are timely because the results are published approx. one month after the end of the data collection and five months after the start of the data collection. The availability of the statistics is ensured by sharing metadata, i.e. uncertainty calculations, method descriptions and the precise formulation of the questions. Statistics Denmark has limited opportunities to influence the quality of the responses. It is not possible for us to check the correctness of the individual answers, e.g. regarding selected technical solutions, such as artificial intelligence, are To give an indication of the data quality, uncertainty calculations have been prepared, see more in section 5.2. Throughout, the response rate in the survey is high and the extent of partially deficient responses limited. The survey method, including random sampling and calculation of weights, is continuously assessed.

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

Not relevant for this statistics.