Statistical processing

Source data

The survey covers companies in the private, non-financial urban industries with 5-9 full-time employees. The statistics are based on questionnaire information from a sample of companies. The sample is stratified by industry groups and size. The brutto sample contains 1,649 companies, of which 1,548 answer the questions. Data is collected via digital reporting. The questionnaires can be found here:

Spørgeskema mikro virksomheder_2021 Questionnaire micro enterprises 2021.pdf

Spørgeskema mikro virksomheder_2022 Questionnaire micro enterprises 2022.pdf

Spørgeskema mikro virksomheder_2023 Questionnaire micro enterprises 2023.pdf

Questionnaire small entreprises draft - SVA_2024.pdf

Frequency of data collection

Yearly.

Data collection

Digital reporting via the form. The form can be seen here.

Data validation

A number of checks and validation mechanisms as well as reporting aides for the reporter’s understanding of the questionnaire are built into the digital form. This includes among other things checks on sum totals of quantitative fields (e.g. that reported percentages add to 100 per cent) as well as filter and routing mechanisms (e.g. so that information cannot be reported in fields, where the enterprise should not provide any information).

Once data is received by Statistics Denmark a number of additional checks are performed. These are performed in part at macro level, where e.g. the results for a given variable for an activity group or size class are compared to the ones from the previous survey period. Further validation is carried out at micro level, i.e. at the level of the individual enterprise. This includes e.g. comparison of the reported e-commerce figures with those reported previously. In both macro and micro level validation background information from Statistics Denmark’s Statistical Business Register is used. Finally, in addition to this, checks involve identifying outliers, i.e. reported data with extreme values, for certain variables. In some cases validation results in recontacting the reporting enterprise for a clarification of the reported data.

Data compilation

The published results are raised to the level of the population.

A stratified random sampling is used on the basis of the activity of the enterprise and the number of employees. By grossing up a weighting and calibration using regression techniques is applied to the weight of the individual enterprise. Imputation is not used, neither in the case of partially lacking information from the enterprise (item non-response), nor in the case of completely lacking reporting from an enterprise (unit non-response). Both, instead, is handling through reweighting as part of the grossing up procedure.

Adjustment

No further corrections are undertaken than those already described under validation and treatment.