Selection Bias

The people or items actually studied can end up a poor match for the whole group they're meant to represent. Results get misleading whenever certain kinds of people or items are left out, whether on purpose or by accident.

This bias is a systematic distortion that occurs when the sampled subset differs non-randomly from the target population, altering the observed associations. It compromises external validity and can induce spurious correlations wherever inclusion depends on exposure or outcome.

A news website runs a poll asking readers "Do you support this new policy?" and reports 80% in favor. But only people who already visit that particular website — who likely share its editorial slant — bothered to answer, so the result tells us nothing reliable about what the general public actually thinks.

A retrospective cohort study of occupational chemical exposure and lung-cancer mortality recruits workers still employed at the plant as of the study start date, inadvertently excluding workers who became ill and left employment earlier. That exclusion truncates the highest-exposure, highest-mortality group right out of the data, attenuating the estimated hazard ratio toward the null. Reweighting using payroll records that capture the departed workers shifts the estimate substantially higher, revealing a dose-response relationship the original selection had masked entirely.

People with certain traits are more likely to end up included, so the results lean toward those traits. Leaving others out shifts the averages and hides the real relationships in the data.

Unequal inclusion probabilities from enrollment rules or censoring create an asymmetry in the observed distribution, constraining what the data can actually show. Sampling frames and eligibility criteria weight the observations unevenly, biasing the resulting estimate.

Including as wide a range of people as possible, while reducing rules that exclude groups, is the direct fix. Checking whether the sample actually resembles the whole population of interest keeps the result honest.

Randomized sampling frames, weighting adjustments, and sensitivity analyses correct for the unequal inclusion probability directly. Inverse-probability weighting or explicit selection models mitigate the bias introduced by the known selection mechanism.

Nonrandom participation; Attrition over time; Measurement-dependent inclusion

An adversarial actor can deliberately engineer selection bias by designing eligibility criteria or recruitment channels that systematically exclude populations whose outcomes would contradict a desired conclusion — for example, a pharmaceutical sponsor enrolling only healthy, young trial participants to inflate efficacy estimates and suppress adverse-event rates. In observational contexts, a motivated analyst can restrict the analytic sample post-hoc to a subgroup where the target association is strongest, then report findings as if they apply universally. Survey or polling operations can weaponize voluntary response mechanisms to over-represent sympathetic demographics, manufacturing consent for a preferred narrative.

Pre-register sampling frames, eligibility criteria, and exclusion rules before data collection to prevent post-hoc sample manipulation and enable independent audit. Apply inverse-probability weighting or selection models calibrated against external benchmarks (e.g., census data) to correct for known asymmetric inclusion probabilities. Conduct sensitivity analyses — including worst-case bounds and E-value calculations — to quantify how much unmeasured selection would be required to nullify observed associations, making covert bias harder to conceal.