Ascertainment bias

Ascertainment bias — often used interchangeably with sampling bias — is a form of selection bias that arises when the way researchers identify, recruit, or detect subjects makes the study sample unrepresentative of the population the study intends to describe. Because certain individuals are systematically more likely to be captured than others, measured associations and disease frequencies are distorted.

Classic examples make the mechanism clear. A study recruiting participants only from a hospital will overrepresent sicker patients and those with multiple conditions (Berkson bias is a well-known hospital-based variant). Screening programs detect more disease in populations that attend screening — who tend to be healthier and more health-conscious than non-attendees. A genetic study that identifies families through an affected child will overestimate how often the condition appears in offspring.

The defining feature is that the error is systematic, not random: no increase in sample size fixes it, because the sampling method itself is skewed. Prevention requires strategies like random sampling from the full target population, careful choice of comparison groups, and blinding of investigators who assess outcomes so knowledge of exposure status does not influence detection.

Bias types are high-yield biostatistics and epidemiology material on the USMLE Step 1 exam. Be prepared to recognize ascertainment or sampling bias in a study vignette and distinguish it from other selection biases, as well as from measurement, recall, and confounding biases.