Crash programme: Sampling and Weighting

Learn how crash investigation programmes use representative sampling, ethical data collection and weighting to produce reliable road safety insights.

Representative crash data through sampling and weighting

Sampling is used in crash investigation programmes because it is not feasible to conduct detailed investigations for every crash. By selecting a representative sample of crashes and applying appropriate weighting factors, the findings can be extrapolated to the wider crash population, ensuring that results are both cost-effective and statistically representative.

Further information can be found in our report, In-Depth Crash Investigation Programme.

Sampling Strategy

Two main sampling approaches can be used:

Retrospective sampling

Investigations take place after the crash, using pre-existing records, there is a lower cost, but generally provides less detailed and lower-quality data.

On-the-scene sampling

Delivers more comprehensive data, but requires greater resources and higher costs.

The size of the sampling area depends on the chosen methodology:

For retrospective sampling, the sampling area can extend up to 40–50 km from the team headquarters.
For on-the-scene sampling, the investigation team should be able to reach the crash site within 30–40 minutes of notification. This requires defining an operational radius in kilometres.

Clear crash reporting procedures are essential:

Retrospective sampling can rely on reporting channels with the police and hospitals, subject to agreed criteria.
On-the-scene sampling requires multiple notification sources, including police, emergency medical services, and the fire brigade.

Strong partnerships with healthcare providers are particularly important for retrospective investigations:

Collaborate with hospitals receiving crash victims within the sampling area.
Engage first aid and emergency response personnel.

Injury data collection should include:

Detailed information on injury type and severity.
Diagnostic data required for AIS (Abbreviated Injury Scale) coding.
Follow-up information to monitor potential long-term disabilities at 3, 6, and 12 months, where applicable.

Ethics and Confidentiality

Ethical approval should be obtained from an ethics committee affiliated with a university or hospital in the study area. If no local ethics committee is available, approval may be sought from:

Universities or hospitals in other regions.
Institutions located in other countries, where appropriate.

Ethical review is a critical component of the investigation programme, ensuring:

Compliance with ethical and legal requirements.
Protection of participant rights and privacy.
Secure handling and storage of sensitive data.

Establishing a robust ethical framework strengthens the credibility, integrity, and reliability of the investigation process.

Weighting Strategy

A clear weighting strategy is essential to ensure that in-depth crash investigation data is representative and that findings can be reliably extrapolated.

One commonly used approach is to compare the crashes investigated in depth with national crash statistics using key variables that are strongly correlated with other crash characteristics. To support this process, it is important to collect at least a subset of the crash data available at the national level for the study area.

Applying appropriate weighting factors during analysis enables results from the in-depth sample to be scaled to the national context, improving the validity and policy relevance of the conclusions.