Which statement correctly describes the potential bias in before/after evaluations if not accounted for?

• A. It can be biased by regression to the mean if not accounted for
• B. It cannot be biased
• C. It always overestimates benefits
• D. It relies solely on expert opinion

Answer: (A)

When you compare measurements before and after, extreme starting values tend to drift toward the average on the next measurement due to random variation. This regression to the mean can make it look like the intervention caused a change when, in fact, the shift would have happened anyway. If you don’t account for this, your evaluation becomes biased. It isn’t true that there can be no bias, and the change isn’t guaranteed to be a positive one or always reflect expert judgment—the bias here stems from statistical tendency, not from the intervention being misinterpreted. To reduce this bias, use control groups, randomization, or statistical adjustments that separate the intervention effect from natural variation.