LGRAS (02) 4 - THE POSSIBLE QUARTERLY RELEASE OF ROAD ACCIDENT STATISTICS
1.1 This paper informs Members of the Group about DTLR's likely forthcoming quarterly release of road accident statistics for Great Britain, and seeks their views on the possible quarterly release of road accident statistics for Scotland.
2 The likely quarterly release of figures for Great Britain
2.1 DTLR is currently investigating the feasibility of producing a new series of quarterly road accident statistics for Great Britain. A decision will be taken in the summer, after checks have been carried out. It is anticipated that the first quarterly release would be in November, and would cover the first and second quarters of 2002.
2.2 DTLR's intention is to make figures for a quarter available 18 weeks after it ends. DTLR would base the quarterly figures on the data that were available 14 weeks after the end of the quarter. It would estimate any missing components of the GB totals (this would be necessary if there were any months in the quarter for which DTLR had not received sufficiently complete returns for one or more Police Force areas). DTLR has been testing the estimation process over the past year, and will compare the trial estimates with the final figures for 2001, once they are available. Checks so far suggest that the quarterly figures should produce reliable indications of key trends.
2.3 DTLR would initially release the quarterly totals on an experimental basis. The figures would be made available to interested parties, on paper and on the DTLR Web site. There are National Statistics rules regarding the announcement in advance of the release dates of forthcoming publications. The dates of release of the quarterly figures would be preannounced, but it is not expected that the quarterly figures would be accompanied by a news release while the series is provided on an experimental basis.
2.4 Because of the "volatility" of quarterly figures (particularly if they include estimated components for some Police Force areas for some months), DTLR would release only a few overall summary totals for GB as a whole. It would not release the numbers who were killed - the quarterly figures' breakdown by severity would be restricted to "Fatal & Serious" and "Slight" casualties. Similarly, DTLR envisages only releasing quarterly figures separately for four main types of road user: "pedestrian", "pedal cyclists", "two wheeled motor vehicle users" and "car users", together with the "all road users" total. Road class or type of road would be available for "built up" and "non-built up" roads, and for "major roads" and "minor roads". Categories would be shown separately, rather than cross-classified, so that (for example) there would not be a figure for pedestrians on built up roads. The figures would be produced for GB as a whole: DTLR would not release quarterly figures for any areas within GB, such as Scotland. DTLR has produced an example set of tables which show these few "key" figures for a number of periods, including the latest quarter, the same quarter of the previous year, the latest 12 months, the 12 months before that, the 1994-98 average (the baseline for the casualty reduction targets for 2010). There would also be two tables with some longer-term series.
2.5 DTLR would also release a table which would show how up-to-date each of the Police Force areas is in its contribution to the quarterly figures for Great Britain as a whole. The table would show the latest complete 12 months figures for each of the 51 Police Force areas in GB, and state which months they cover, so it will be clear which Police Force areas are behind with their returns. An example of this table, which was distributed at the most recent meeting of the Standing Committee on Road Accident Statistics, showed that Scotland had none of the eight worst performers (for which two or more months' figures had to be estimated to produce the quarterly figures). There were also nine Police Force areas (two of them were Scottish) for which one month's figures had to be estimated. We appreciate the efforts of Scottish data suppliers, who are making their returns more promptly than in some other parts of GB.
2.6 Members of the Group are invited to note the position on the likely forthcoming release by DTLR of quarterly totals for GB as a whole. Members of the Group representing data suppliers are also asked to note the importance of making complete road accident statistics returns within the agreed timescales, so that any problems can be resolved, and the figures reconciled, in good time for the data to be used to produce the quarterly totals.
3. The possible quarterly release of figures for Scotland
3.1 If DTLR does start to release quarterly figures for GB as a whole, what (if anything) should the Scottish Executive do?
3.2 First, is there any "demand" for the quarterly release of road accident statistics for Scotland? There are several reasons why it might be felt that there is no need for the Scottish Executive to release quarterly figures for Scotland. For example:
- there appears to have been no demand for such figures in the past five years - the Transport Statistics branch cannot recall any enquirers seeking quarterly figures.
- until about mid-1996, DTLR used to release quarterly figures for GB as a whole which were not accompanied by a separate Scottish release of quarterly figures. To the best of our knowledge, the lack of quarterly Scottish figures did not cause any concern in Scotland, nor did the discontinuation of the GB quarterly figures. (NB: the DTLR publication had a table, like that mentioned in paragraph 2.5, which showed the latest complete twelve months' totals for each Police Force, listed in what appears to be codenumber order within the "most recent" month, without any sub-total for Scotland. So the DTLR publication did not provide the latest quarter's figures for Scotland.)
- the Scottish Executive is not required to release quarterly figures for Scotland whenever quarterly figures are released for GB/UK - for example quarterly releases of traffic estimates and rail passenger numbers for GB as a whole are not accompanied by a release of the corresponding Scottish figures.
- to release figures for Scotland each quarter could lead to undue attention being given to short-term fluctuations, which are unimportant in the context of the long-term trends.
- the production and checking of quarterly releases, with the associated commentary and briefing, and the distribution of such material, would take time which might be better used (say) improving the road accident statistics data in other ways, or in developing other types of transport statistics.
However, if there is a clear "demand" from users of the road accident statistics for the release of figures each quarter, the Scottish Executive would wish to meet it. Therefore, Members of the Group are asked whether the Executive should release quarterly figures for Scotland. For example, do they believe that such figures should be made available? If so, would they use them? If so, for what purposes? If not, why not? If they feel that it would be better not to release figures each quarter, why is this?
3.3 If there were sufficient "demand" for the Executive to release quarterly figures for Scotland, it would have to decide how they should be produced and made available. As there are only eight Scottish Police Forces, it is not worthwhile trying to develop a Scottish version of DTLR's method to estimate the figures for any months for which a complete return has not been received from one or more Police Force areas. Because a few areas account for large proportions of the total numbers of accidents and casualties, there could be a considerable margin of uncertainty around any Scottish totals which included estimates resulting from backlogs in one or more of the larger areas. Therefore, if it were to release quarterly figures, the Scottish Executive might release Scottish totals which were based on:
- the figures for the latest quarter, in the case of those Police Force areas for which apparently complete returns had been received for that quarter, and
- the figures for the latest three months which happened to be available, for any Police Force areas for which there was a backlog (such areas would be identified as such in the release).
The "change in the past year" would then be calculated using the previous year's figures for the corresponding months for each area, and the "change from the 1994-98 average" would be calculated using the "latest four quarters / 12 months available" figures for each area. Such an approach should provide a good enough indication of whether the numbers were rising or falling, compared with the same "quarter" a year earlier, and how the latest "year" compared with the baseline for the casualty reduction targets for 2010. If it appears that there is sufficient "demand" for the quarterly release of Scottish figures to warrant their production, Members of the Group are asked whether they feel that this would be an appropriate method of producing them.
3.4 If there were sufficient "demand" for the Executive to release quarterly figures, it would have to decide what should be made available, and how, to whom and when. For example, if such figures would be produced purely to inform a few interested parties, such as those who had obtained the GB quarterly figures and wondered about the position in Scotland, it would seem appropriate for:
- the detail of the figures to be limited to a few overall totals for the main groups of road user and road type (i.e. along similar lines to the approach which would be taken by DTLR for any quarterly release of figures for GB as a whole);
- the Scottish figures to be made available on the Executive's Web site;
- paper copies to be sent out using a limited version of the road accident statistics publications distribution list - e.g. only to members of LGRAS and any other people who asked to be sent the figures;
- the Scottish figures to be released without an accompanying news release (even if DTLR issued a news release on the GB figures), to reduce the likelihood of undue attention being given to one quarter's figures;
- the timing not to be determined by the timing of the release of the GB figures - the Scottish figures could appear earlier, as there would be little need to synchronise the release of the SE and DTLR figures if the SE figures did not have accompanying news releases. Under National Statistics rules, the release dates of forthcoming Scottish figures would have to be announced in advance.
If it appears that there is sufficient "demand" for the quarterly release of Scottish figures to warrant their production, Members of the Group are asked for their views on the level of detail that should be provided, the methods to be used to release them, and the timing of release - for example, would they like to see changes to the approach which is outlined above?