Road Accident Data Collection Form Design Research Project
5 Draft form design
In order to develop a revised form, we used insights gathered from the literature review and consultations. We also took the opportunity to engage with our expert advisors to provide input into an internal project team workshop.
The session was structured such that the insights gathered in the literature review and consultation tasks were presented to our expert advisors. The advisors, Jeremy Broughton, Richard Cuerden and Caroline Wallbank contributed to this in terms of providing their own expert opinions as well as challenging and raising questions to further enhance the development of the form.
General use of the form
The interviews carried out in the consultation revealed that none of the Police Officers interviewed complete a 'STATS19 form' at the scene of a collision; instead, notes are made in notebooks, telephoned in or recorded on a PDA. The interviewees suggested that this process was normal practice for the collection and inputting of STATS19 data. Those that used notes completed a form on a computer (often a template that was typed in) at the Police station and sent it on or completed an online recording system.
Therefore it was not deemed appropriate to design a paper-based form, since the findings from the interviews suggested it is unlikely that this would be used. It was proposed that a simple electronic version of a form could be designed that could minimise errors and that an officer could complete for each collision based on notes made at the scene.
The draft form was designed in Microsoft Excel, making use of the validation drop down menus, which were highlighted in the literature review as improving the accuracy of the data entered. Users can only select items on each option list, and an error is displayed if a different response is entered. Drop down lists were chosen despite the finding from the literature review that they can be frustrating for users who are entering familiar information. The decision to use the drop downs was made based on feedback from the consultation, i.e. that interview respondents felt that dropdown menus would address the problems with the accuracy of the existing data returns. Further information about the data field or definitions from STATS20 was entered so that it is displayed when a cell is selected.
Cells other than those where data are required have been greyed out and locked so that they are not selectable, and moving between cells using the TAB, ENTER or arrow keys scrolls through only those selectable cells. This means that data cannot be entered in other cells. Within each Excel file one sheet was created for each of collision circumstances, contributory factors, vehicles (up to 3) and casualties (up to 4). Provision for higher numbers of vehicles or casualties was considered after the vignette study.
The use of an electronic form enables some logic checks to be carried out at the data entry stage, which should reduce the number of forms that are returned to be checked at a later stage, for example, to check that only pedestrians have the pedestrian location, movement and direction data completed.
One respondent of the consultation suggested putting the most frequently used values at the top of the list for each field. This was considered as part of the workshop, and compared with the order as listed in STATS20 or alphabetical order. Our experts thought that putting the most common options first would not be suitable since the other options may be overlooked. For example, for vehicle manoeuvre 'going ahead other' is the most frequent value (Transport Scotland, 2014, Table 14), listed as the last option in STATS20, but our experts felt that all other options should be considered before this option is used.
Details of revised form design
Figure 4 shows the draft revised accident circumstances form produced following the workshop.
Figure 4: Draft accident circumstances form
The changes to the accident circumstances form were:
- Road type field. Users reported confusion for motorways and that for some cases multiple options appeared to apply
- adding 'or motorway main carriageway' to dual carriageway
- reordering to put single carriageway and dual carriageway at the end of the list to help suggest that the 'special cases - roundabout, slip-road, one-way street' should be used if they are present rather than single or dual carriageway
- 2nd Road Class. It was reported that this is not always completed for collisions involving a vehicle using a private drive or entrance
- add 'or private drive' to unclassified
- It was suggested changing the order so that the conditions with and without high winds were adjacent.
The following fields are calculated automatically:
- The accident severity is automatically calculated based on the casualty from the accident with the highest severity level.
- The number of vehicles and number of casualties are filled in automatically from the vehicle and casualty data.
- The day of week is calculated automatically from the date entered.
The following data validation and logic checks were incorporated:
- The recording of the Ordnance Survey Grid References appeared problematic. Data validation were applied with the following criteria:
- Eastings between 0 and 500,000
- Northings between 500,000 and 1,300,000
- For junction accidents the following checks are carried out:
- If junction detail is not equal to 'not at a junction' then red text is shown to indicate to complete the junction control, 2nd road class fields
- If 2nd road class is M, A(M), A or B then red text is shown to complete the 2nd road number field
- If junction is not completed and any of junction control, 2nd road class and 2nd road number are, then red text shows to fill in junction detail.
Figure 5 shows the draft revised vehicle details part of the STATS19 form.
Figure 5: Draft vehicle details form
The changes to the vehicle details form were:
- Vehicle type. Some STATS19 users reported confusion over 'other' vehicle types, for example, ambulance and motor caravan. A review of data suggests that 'vans' are often coded as 'other' also. Some interviewees suggested entering the motorcycle type was not always easy.
- The text box that displays when a user click on the vehicle type field shows some of the text from STATS20 (Department for Transport, 2011) giving definitions of vehicles, including examples of vehicles included under van, HGV and other vehicles.
The following checks were included within the form:
- Pedal cycles and motorcycles not allowed to be left hand drive vehicles
- Pedal cycles and motorcycles not allowed to have overturned
Vehicle manoeuvre, compass points and junction location of vehicle were also reported as confusing or inconsistent, but no form design changes were identified which could address these.
Figure 6 shows the draft revised casualty details part of the STATS19 form.
Figure 6: Draft casualty details form
The draft form includes a box shown when casualty severity is selected to indicate the injuries that are classed as serious (from STATS20). This should help users to identify seriously injured, with lesser injuries classed as slight.
The following data validation or logic checks were also included:
- If casualty class is pedestrian then red text alerts user to complete pedestrian location, movement and direction.
- If the casualty is a driver then the spreadsheet checks the age entered against the driver age of the vehicle corresponding to that driver. Red text alerts if these are different.
Users reported confusion over the term 'participant'. This refers to a vehicle/driver, casualty or uninjured pedestrian.
There was also confusion between similar descriptions of factors relevant for vehicles/drivers and pedestrians which have different codes:
- Failed to look properly (405 for drivers, 802 for pedestrians)
- Failed to judge other person's/vehicle path or speed (406 for drivers, code 803)
- Impaired by alcohol (501 for drivers, 806 for pedestrians)
- Impaired by drugs (502 for drivers, 807 for pedestrians)
- Careless, reckless or in a hurry (code 602 for drivers, 808 for pedestrians)
The form was designed for users to select the type of participant (vehicle, casualty or uninjured pedestrian) first, and the relevant reference number.
The next field is the factor type, and is designed to only show those types relevant for the participant selected:
- For vehicle/driver, all factor groups apart from 'pedestrian only' are shown
- For injured and uninjured pedestrians, only the 'pedestrian only' and 'special codes' are shown
Once the factor type has been selected, the drop down list of options only shows the factors in the group selected. This should ensure that any vehicle specific code is assigned to a vehicle and that the correct code is used when the descriptions are similar.
Figure 7 shows the draft contributory factors part of the STATS19 form showing the factors reduced to those in the 'injudicious action' group.
Figure 7: Draft contributory factors form
The revised form was developed using insights from the literature review and the consultation as well as expert opinion provided during the expert advisor workshop. The first decision about the redesign of the form was to ensure that it was in a usable format for the Police and other users. Based on the information about current approaches to collection of accident data received as part of the consultation, a paper-based form was not considered to be the most ideal means of data collection, therefore the revised form was developed using Microsoft Excel.
The revised Excel form was designed to reduce or eliminate some of the accuracy issues highlighted in earlier tasks, while also incorporating any relevant best practice guidelines identified in the literature to enhance its layout and design.
The revised form included tabs or worksheets in Excel for each of accident circumstances, vehicle details, casualty details and contributory factors, using drop down lists wherever possible to minimise errors. Some logic checking between fields was introduced and modifications to the text for some field names and labels were also included.