What study design best isolates the causal impact of a road safety program on crashes?

• A. Observational before-after
• B. Case study
• C. Cross-sectional survey
• D. Randomized controlled trial

Answer: (D)

To prove a causal impact, you need a design that controls for confounding, and randomization does that by balancing factors across groups. In a randomized controlled trial, units such as sites, drivers, or communities are randomly assigned to receive the road safety program or to serve as a comparison group without the program. This random assignment makes the groups similar on both measured and unmeasured factors that could affect crash rates. After implementing the program, you compare crash outcomes between the two groups; because the groups started off alike, differences in crashes are most plausibly due to the program itself, with a clear estimate and statistical precision.

Other designs struggle to establish causality. An observational before-after study can show change over time but may be biased by trends, seasonal effects, or other simultaneous changes. A case study provides rich detail but no proper comparator. A cross-sectional survey captures exposure and crashes at one point in time, so it cannot determine which came first or rule out confounding.

So, randomized controlled trials offer the strongest evidence for the program’s causal effect on crashes.