Sampling Bias
Updated 2026-08-15
INTRODUCTION
English translation pending.
CORE DEFINITION
Sampling bias occurs when the selection procedure makes some members of a population more likely to enter the sample than others on dimensions that matter to the question. Sources include non-random selection, self-selection, non-response, and coverage gaps in the sampling frame. Survivorship bias, which reads only the cases that endured, is the best-known instance. The qualification is that bias is a property of the procedure rather than of sample size: a large biased sample estimates the wrong quantity ever more precisely. Random selection and weighting address it; no arithmetic on the data removes it afterward.
SCAFFOLDING EFFECT
Reduce cognitive load
- Interrogate the frame: ask which members of the population could never have been selected. - Hunt the missing: examine who refused, dropped out, or failed to survive being counted. - Separate size from bias: check that a bigger sample reduces error rather than refining a wrong estimate.
Anchor fast decisions
Inference from sample to population assumes the two share the same distribution on the variable of interest. Selection procedures that correlate with that variable break the assumption, so the sample mean converges to a different quantity than the population mean. Increasing sample size narrows the interval around that wrong value, which increases confidence without increasing accuracy. Non-response acts the same way, because those who decline differ systematically from those who answer. The distortion survives all downstream analysis, so it must be repaired at the design stage.
MINIMUM ACTION
In progress 0/1Practice this model in one real situation:
account_treeGenealogyexpand_more
menu_bookReferencesexpand_more
Source support: Explicit
- en.wikipedia.orghttps://en.wikipedia.org/wiki/Sampling_biasverified
PRIVATE NOTES · Only visible to you
SAVED Q&A
ENTRY Q&A · Private saving available
Ask with a clear boundary
thinkingmodels answers from published entry context only.
Your question is sent to thinkingmodels. The answer uses public entry context only.
RELATED MODELS