Statistical processing
Contact info
Statistics Denmark, Short Term StatisticsSimon B. Halifax
+45 3917 3088
Get as PDF
Monthly business survey data is collected from a sample of approximately 800 manufacturing enterprises, 900 enterprises in construction, 1400 enterprises in retail and 2900 enterprises in services. Data is reported either online at http://www.VIRK.dk or in paper questionnaire, and consists of qualitative assessment of changes in net sales, employment, inventories, orders, etc.. The collected data are enumerated from the sample to the total enumeration population (similar to Frame or Survey Population) weighted by employment data, and distributed at industry and enterprise size groupings. If seasonal variation patterns are detected for the indicators that constitute the composite confidence indicator, data is adjusted for this.
Source data
The statistics is compiled from data collected via questionnaires. The sample is based on information in the Central Business Register. The survey is carried out with a panel based stratified sample Because of drop-out the sample is updated two times a year.
The sample is stratified into lines of industries as is seen in that statbank tables.
Frequency of data collection
Data is collected each month during the first 2-3 weeks of survey month
Data collection
Data is collected either by paper questionnaire send to the respondent including postage pre-paid envelope, or by online reporting. In the latter case, respondents receive an e-mail request to log in to the reporting system VIRK, and fill the online questionnaire. Respondents are urged to report online, and if so the respondent the subsequent months only receives an email request about reporting.
Notification is send the day before deadline.
Response rate is about 77-80 pct. on average. Deadline for reporting to the current month's questionnaire is usually the second Friday of the month. Here, about two third of the answers have arrived. The rest, one third, arrives within two weeks and are included in the statistics.
Data validation
To fill the questionnaire one has to tick one of three options for each question. This does not call for data validation. Non-response errors are corrected for in the enumeration process. By online reporting the system warns the respondent if a question is not ticked. However in both cases, online reporting or reporting by mail, it is accepted that some questions are left un-ticked.
In the processing, basic data is checked if bigger deviations as compared to last month occurs. Usually the reason is either unidirectional change in assessments by several big companies or missing reporting from some few big companies.
Data compilation
The enterprises' responses are weighted according to their average number of full-time employees. Within each stratum answers are distributed into percentages answering increase, unchanged and decrease adding up to 100 pct. The group of non-responding enterprises within the stratum is assigned same percentage distribution as the group of responding enterprises (which is considered a imputing method).
Hereafter, the responses are enumerated to the number of full-time employees of the total population, within each industry-grouping.
Confidence indicator for a variable is calculated and denominated balance. Is is the difference between percentages of enterprises (weighted by number of employees) that have answered increase and decrease (or too small and too large). The balance figures do not measure the size of the increase or decrease, but only indicates that an overweight of enterprises expect either increase or decrease.
The Composite confidence indicator is the arithmetic average of balance figures.
Especially regarding the questions on Production limiting factors, e.g. lack of workforce, the indicators reflects the share of companies that have ticked of the specific questions. The share of companies is weighted according to the average number of full-time employees within the specific companies. Hereafter the result is enumerated to cover the full population within each industry grouping.
Adjustment
Data is seasonally adjusted, if seasonal pattern is detected, and only for the indicators that constitutes composite confidence indicators.