SELECTION BIAS:
A bias in the estimated association between an exposure and disease or outcome, that arises from the study procedures used, or factors that influence, study participation or selection of individuals into an analysis1,2. An important element for understanding Selection Bias is that the relationship between the exposure and disease/outcome is different from those who participate compared to those who are theoretically eligible to participate, but do not2.
Selection Bias is often erroneously confused with Sampling Bias; which is a bias that arises due to systematic differences in participant characteristics between the study sample and other popualtions1. Narrowly restricting the eligibility criteria for participants to take part in a study is not necessarily a Selection Bias, if it does not truly distort the relationship between the exposure and disease/outcome in the sample compared to the population from which they came; as this restriction does not affect the internal validity of the study (biases are unfavourable, and affect internal validity). In this case, restriction is an issue of generalizability and therefore not a bias.
Often Selection Bias is defined as: “A systematic difference between those selected into a study and those who are not”. This definition is oversimplified and problematic as it does not address whether the estimated association between the exposure and disease/outcome is distorted or not; which is a key element to understanding Selection Bias and its difference between Sampling Bias or generalizability.
The term ‘Selection Bias’ is sometimes used as a broadly-applied high-level category, or synonym for any selection-related biases in a study which leads to conclusions about an exposure/predictor and disease/outcome relationship that are systematically different from the truth3. For example, Publication Bias, may be considered a type of Selection Bias for systematic review studies. Also see: Sample Selection Bias, Ibrahim-Spitzer Bias, Structural Selection Bias, Sampling Bias, Publication Bias, and Bias (Definition).
References:
1. Porta M, ed. A Dictionary of Epidemiology. Sixth ed. New York, NY: Oxford University Press 2014. (Link to Reference)
2. Lash TL, VanderWeele TJ, Haneuse S, Rothman KJ. Modern Epidemiology. Fourth ed. Philadelphia, PA: Lippincott Williams & Wilkins 2020. (Link to Reference)
3. Celentano DD, Szklo M. Gordis Epidemiology. Sixth ed. Elsevier 2018. (Link to Reference)