The reported data is scaled to the total population of professional units with main activity in construction. No numbers are imputed. The total employment in each construction industry and each type of construction work is seasonally adjusted. The cross between construction industry and type of construction work is not seasonally adjusted.
The statistic is bases on a questionnaire among construction companies. The sample is approximately 1.500 professional units. The professional units are chosen with the aid of registry information. In the sample all professional units with at least 39 registry employees have been chosen. The rest of the sample is based on random stratified sampling. There are 4 size groups, defined by the intervals 5-9 employees, 10-19 employees, 20-39 employees and at least 39 employees. The limit of 5 employees, under which no professional units are selected, has been set to reduce the burden placed on small companies.
Data is collected each quarter.
The employment numbers are gathered by online reporting from the professional units in the sample. The survey is mandatory.
The reported numbers are controlled for errors. they are compared to reported numbers from earlier periods. If the reported employment numbers are substantially different from what is expected, the company is contacted to confirm the numbers. It is likely that not all errors are discovered. Therefore the statistic has some uncertainty connected to wrong reports. However, the magnitude of this problem is believed to be small.
The reported data is scaled to the total population of professional units with main activity in construction. The population, to which we scale, include professional units with less than 5 employees, event though these are not in the sample. For scaling a method called ratio estimation is used. The method uses employee information from the Danish Business Register from the preceding 4 quarters.
The total employment in each construction industry and each type of construction work is seasonally adjusted. The cross between construction industry and type of construction work is not seasonally adjusted. The season adjustment corrects for holidays(Easter) and the number of days in the quarter. The seasonally adjusted population totals are constructed as the sum of the seasonal adjustment for the individual construction industries. The seasonally adjusted employment numbers for the different types of work are normalized to sum to the population total.