Which term describes a situation where fatal crash counts may be influenced by factors outside the program, leading to misattributed improvements?

• A. Type I error
• B. Confounding factors
• C. Regression to the mean
• D. Sampling error

Answer: (B)

External influences that are linked to both implementing a program and the outcome being measured can distort evaluation results. When fatal crash counts appear to improve, changes outside the program—such as variations in enforcement, weather, or other concurrent road safety efforts—can influence crashes and be mistaken for the program’s effect. This creates a confounding effect, where an outside factor drives the observed change, making it look like the program caused the improvement even if it didn’t. Confounding factors is the best term for this situation because it identifies the outside variable that can misattribute the cause of the observed outcome.

For context, Type I error is about wrongly concluding there is an effect when none exists; regression to the mean is the tendency of extreme values to move toward the average on subsequent measurements; and sampling error is random variation due to not perfectly representing the population.