Practices when changes are made to statistical data
The situations in which changes may be made to already published data can be divided into two groups: data revisions and errors.
Data revisions improve the quality of statistics
Revisions of published statistical data are part of the normal production of statistics and mean improvement of quality. The principle is that statistical data are based on the best available data and information concerning the statistical phenomenon. On the other hand, the revisions are communicated as transparently as possible in advance. Advance communication ensures that the users can prepare for the data revisions.
The reason why statistical data become revised is often caused by the data becoming supplemented. Then the new, revised statistical figure is based on a wider information basis and describes the phenomenon more accurately than before.
Revisions of statistical data may also be caused by the calculation method used, such as annual benchmarking or updating of weight structures. Changes of base years and used classifications may also cause revisions to data.
Seasonally adjusted time series of short term statistics become revised because of the calculation method used. Additional information on a new time series observation is exploited in model-based methods and this is reflected as changes in previous releases. Revisions of the latest figures of seasonally adjusted time series are discussed in connection with the releases and quality reports of statistics.
A summary table of the revisions that have taken place is also published in connection with key short term statistics and some annual statistics. The table shows how the data for the statistical reference periods have changed between the first and the most recent release.
The reasons for revisions in each set of statistics are given in the quality report (under Data revision - practice). The description also explains the usual magnitude of the revisions, how long the data will be subject to revisions after their initial release, and whether certain data are subject to larger revisions than others.
If data are revised substantially more than usual, the reasons for it are separately explained in the release.
Users are informed about errors
Examples of errors in statistical releases include erroneous figures in release texts, figures or tables, or use of incorrect concepts. Unlike revisions, error situations are unexpected deviations in the normal production of statistics.
Errors are corrected and the users are informed about them as quickly as possible. Statistics Finland informs about significant errors to the same extent and through the same channels as in the releasing of the original data.
Notations about corrections and their points of time are added to the corrected releases.
Typographical and other form errors in statistics are corrected as fast and flexibly as possible and no separate notifications are made of them on the web pages.
Information about changes made to statistical databases is posted on the Changes in the StatFin database page. Only the latest data are updated in databases.
If the release time of statistical data changes from the announced point of time, information about the change is posted on Statistics Finland's release calendar.