Statistical processing

Source data

The energy census for the manufacturing industry, is a primary source, which means that data is formed in this census. All companies that, according to the Danish Business Register, are located in industry (DB07-branch 10-33) and parts of raw material extraction (DB07-branch 08-09), and that have at least 20 employees, are selected for the statistics. All selected companies are contacted for reporting.

Frequency of data collection

Information on the consumption and use of energy is collected every two years.

Data collection

The statistics is conducted as a web-based questionnaire. Information on the data collection is available (in Danish) at Reporting.

Data validation

The two primary sources of inaccuracy in the statistics are:

• Reports containing errors, in particular on the energy unit. These errors are found be comparing with the costs.

• Typing mistakes when data are entered. Such errors is mainly detected by comparing with data for previous years.

At the time of typing, all data go through plausibility checks. These consist of comparisons of consumption and the related costs. By this many unit-errors are avoided.

After that follows a more detailed editing process where data are checked thoroughly by using the general tools and methods in Statistics Denmark, in particular comparing with earlier data for the working place.

For the few enterprises having quotas for CO2, The Energy Agency collects particular information which since 2018 are available in the data validation.

All discovered discrepancies are checked and corrected by contacting the involved companies.

Data compilation

Error-checked data is first compiled to publication level. An adjustment is then made for omissions and for units where the company uses rented premises and therefore does not have information on energy consumption itself, e.g. because it is included in the rent. The statistics is a census, but imputation is performed for non-responses or missing information. The imputation method is based on the responses received, and uses a stratified approach based on industries and company size. The response rate is normally 97-99 percent, and therefore imputation is made for 1-3 percent. A targeted telephone reminder procedure ensures that there are no missing responses from large and critical units, and therefore the imputation has a minimal impact on the overall results. The energy consumption are for all types of energy compiled in GJ (Gigajoule), using current values for energy contents based on data from the Danish Energy Agency.

Adjustment

There are no further corrections.