ROAD ACCIDENTS SCOTLAND 2000
Annex B The Collection of Road Accident Statistics
The Road Accident statistics are compiled from returns made by police forces. For each injury road accident known to have occurred in their areas, the police authorities complete a statistical return (which is called a "Stats 19" return), which provides details of the accident circumstances, separate information for each vehicle which was involved in the accident, and separate information for each person who was injured in the accident. Annex C shows the kinds of information that were collected in the statistical returns in 2000, using examples of forms that could have been used for this purpose. (In practice, each police force uses its own system, often involving the information being keyed into the computer by the officer, or by the clerical staff whom the officer telephones to report the accident, rather than the use of paper forms.) Such information was collected with effect from 1999, following the implementation of the changes to the Road Accident Statistics system that were recommended in the 1997 Quinquennial Review.
The statistical returns cover all accidents in which a vehicle is involved that occur on roads (including footways) and result in death or personal injury, if they become known to the police. It should be noted that the vehicle need not be moving, and need not be in collision - for example, the returns include accidents involving people alighting from buses. Road accidents in which no-one is injured ("damage only" accidents) are not included in this definition, and the Scottish Executive does not receive statistics of such accidents. This publication therefore cannot give any figures for "damage only" accidents.
Full guidance on the completion of the "Stats 19" statistical returns, including detailed notes and definitions of the coverage of the returns and of the information to be provided in each field, is given in a document produced by the Department of Transport Local Government and the Regions (DTLR), called "Instructions for the Completion of Road Accident Reports" (which is also referred to as the "Stats 20").
The returns for accidents in Scotland are submitted every month by the police authorities, either directly or with the assistance of a local Council, to the Scottish Executive. All the returns should first be subject to the validity and consistency checks specified in a document called "Procedures for Submitting Road Accident Data to The Scottish Office" (as the Scottish Executive was formerly known). This document is also referred to as the Scottish Edition of "Stats 21". The Scottish Executive also applies these checks, and clears any errors that it finds with the Police authorities. The returns are added to the Scottish Executive Transport Statistics branchs database, which contains statistical information about all injury road accidents in Scotland since 1979.
The Transport Statistics branchs records for accidents which occurred on Motorways and A roads are copied to the Road Network Management and Maintenance Division of the Scottish Executive, which maintains a database of information about trunk roads. From all the Motorway and A road accidents, the ones which occurred on trunk roads are identified using their road numbers and their grid co-ordinates, and the information about them is put onto the Road Network Management and Maintenance Division database. The Transport Statistics branch is subsequently informed which of these accidents occurred on trunk roads, and its database is updated accordingly.
Similar returns are made throughout Great Britain. The Scottish Executive sends a copy of the Scottish data to DTLR, which holds a database of accident records which covers the whole of Great Britain.
Changes to the "Stats 19" road accident statistics system, which were implemented in 1999
Government statistical surveys are reviewed regularly to ensure that they continue to provide essential information for government, and to reduce, where possible, the amount of "form filling" and other costs of data provision faced by the data suppliers. The collection of road accident statistics data is reviewed every five years by the Standing Committee on Road Accidents Statistics (SCRAS), whose members represent central government, local authorities and police forces.
In 1997, SCRAS consulted many organisations about their ideas for improving the road accident statistics, and subsequently prepared proposals which responded to changes in the needs for data and which could be adopted without increasing costs significantly. The report of the 1997 Quinquennial Review was published by the then Department of the Environment, Transport and the Regions as DETR Statistics Bulletin number (98) 14.
The changes which were implemented in 1999 are described in "Road Accidents Scotland 1999", pages 146-151. Pages 152-153 of that publication gives the frequency of use of each code of the new and modified variables. Two particular points should be noted about the changes made in 1999.
Vehicle type: coding of motor caravans
The vehicle type code formerly used for 'Minibus/motor caravan' (code 10) was changed:
- Minibus: the "code 10" category now covers only minibuses;
- Motor caravans are not identified as a separate category - they are now included with 'Other motor vehicles' (code 14).
As a result, the figures for the categories described in the tables as "minibus" and "other" are on different bases for (a) 1998 and earlier years and (b) 1999 and later years. The scale of the discontinuity is not known, because motor caravans have not been identified separately in the statistical returns. However, it is likely that the change in the way in which motor caravans are recorded in "Stats 19" has contributed to the fall in the "minibus" figures between 1998 and 1999, and the rise in the "other" figures.
Driver and casualty postcodes, and estimated distances between homes and the locations of accidents
Postcodes were added to the "Stats 19" returns in 1999. It was accepted that their collection would have to be phased in, as they became readily available from police administrative systems, and so the "not known" code was used more often than should be the case in future. There are also codes for non-UK residents and for parked and unattended vehicles.
The "straight line" (or "as the crow flies") distance between the location of the accident and the home of a driver, rider or casualty was estimated using the postcode of the person's home. The grid co-ordinates of the "centre" of the postcode were obtained from the General Register Office for Scotland's "postcode directory" file. These were taken as an approximation to the grid co-ordinates of the person's home, and used in conjunction with the grid co-ordinates of the location of the accident (as reported by the police) to estimate the distance. A similar approach was used in the small proportion of cases where there was only the start of a postcode (eg the police might record "EH10" if they knew that someone lived in Edinburgh 10, but they could not provide the full postcode) or where only the postal district or postcode sector could be matched with the postcode directory. A distance could not be estimated if the postcode were blank, coded "not known" or "non-UK resident", did not contain a valid postal district, or were for a place outwith Scotland.

