National Records of Scotland

Preserving the past, Recording the present, Informing the future

August 2009 Revision of the Historical Series of Divorce Figures

August 2009 Revision of the Historical Series of Divorce Figures

For many years, the Registrar General's Annual Review and the Vital Events Reference Tables' Notes and Definitions have stated that divorces are counted on the basis of the date of the decree. However, until Summer 2009, that was not quite correct. In fact, the divorce statistics were based on the value of a field in the National Records of Scotland (NRS) database which was a 'proxy' for the year in which the decree was granted. In the vast majority of cases, the value of the 'proxy year' was the year in which the decree was granted, but there were exceptions (which are described below). The use of the 'proxy year' was intended to ensure that (as with the statistics of births and deaths) the published figures for previous years would never change. This would simplify matters for users of the statistics because they would not need to worry about the possibility that the numbers that they already have for past years might have been revised slightly.

There will always be divorces, granted in one calendar year, for which data are not supplied until the next calendar year. In order that almost all such cases were counted against the correct year, NRS set the 'proxy year' equal to the old year for all the divorces, which were granted in the previous year, for which it processed their data before around the end of February. This allowed plenty of time for any backlog of data from the old year to be processed, and enabled those cases to be counted against that year in the preliminary figures published in the first half of March. For example, all the statistical records for divorces granted in, say, 2007 which NRS processed up to about the end of February 2008 were given a 'proxy year' of 2007, and were counted against 2007.

Later in the year, when NRS received data for other divorces which were granted in earlier years, it set the 'proxy year' equal to the current year. This meant that those divorces would be counted in the statistics for the new year. For example, the statistical records for divorces granted in 2007 which NRS processed in (say) April 2008 would be given a 'proxy year' value of 2008, and so would be counted against 2008. The published number of divorces in 2007 would, therefore, differ from the actual number of divorces granted in 2007 - but the difference should represent only a very small percentage of the total, for two reasons. First, there should not be many records of divorces granted in a previous year which were processed after around the end of February. Second, the number of cases excluded from the figures for the old year should be balanced approximately by the number of divorces, granted in earlier years, which were received after about the end of February in the old year, and so were counted in the figures for that year. For example, the number of divorces from 2007 which were counted against 2008 because they were processed after around the end of February 2008 would be approximately balanced by the number of divorces from 2006 (or earlier years) which were counted against 2007 because they were processed after around the end of February 2007.

Unfortunately, at the start of 2009, 'proxy year' values were not allocated correctly in many cases, as a result of which the provisional figure for 2008 which NRS published on 12 March 2009 was incorrect (please go to the correction of the provisional figure for the total number of divorces in 2008 page on our website for further information). To avoid any such problem arising again in future, and to ensure that the published figures were the actual numbers of divorces granted in each year (rather than 'proxy' values), NRS decided to publish figures which were on that basis with effect from the Registrar General's Annual Review for 2008. This involved revising all the figures for previous years in the time-series tables in the Annual Review and the Vital Events Reference Tables. The figures were revised back to 1985, because that is the first year for which there are complete data in our statistical database of records of divorces. A consequence of the change in the method of producing the figures is that the figures for 2008 and earlier years which were published in August 2009 could well be revised in future, because now any late returns for divorces in previous years will be counted, in future published tables, against the year in which the decree was granted.

A spreadsheet is available below which provides a comparison, for 1985 to 2008, of the previously published statistics (which were based on the 'proxy year', using the corrected figure for 2008) and revised figures (which are based on the year in which the decree was granted). This shows that the difference between the two bases is usually only slight. As described above, in most years, counting divorces on the basis of the 'proxy year' produces a result which is very similar to counting divorces on the basis of the year of the decree. This is because only a small proportion of the returns relating to divorces in any given year have a different value for the 'proxy year', and they are usually broadly offset by the cases from the previous year for which NRS put the given year as the 'proxy year'. The spreadsheet shows that, between 1985 and 2008, the only years with large differences between their figures on the two bases are 1993 and 1994 - when, it is presumed, a major backlog (or some other problem) affected 500 or so divorces that were granted in 1993 and led to NRS putting 1994 as the value of their 'proxy year'.

The file below has been made available as an Excel spreadsheet and can be viewed in Adobe Acrobat Portable Document Format (PDF).

When opening the spreadsheet file your browser, depending on how it's been configured, will prompt to either open the file or save it to disk.
Maximum file size is 28 Kb.

PDF files require Acrobat Reader. Download Acrobat Reader free Get Acrobat Reader

Divorces in Scotland - previous published figures ('proxy year'-based), and revised figures ('date of decree'-based)
(Excel    PDF)

All statistical publications