AI, machine learning and deep learning · Introductory

WHAT IS SELECTION BIAS?

Written by SignalsLab

Early edition: these lessons are being reviewed.

Content

Summary

Selection bias is an error arising from non-random participant selection, leading to inaccurate study conclusions. It manifests through sampling bias, time interval issues, data subset selection, and attrition. It is distinct from the bias-variance trade-off, which relates to model oversimplification and under-fitting.

Intuition

When researchers determine what should be studied, a flawed selection process means the resulting conclusions may not be accurate. This happens because the sample does not represent the population correctly, whether due to systematic exclusion, early termination at extreme values, arbitrary data rejection, or participant dropout. Understanding these specific errors helps distinguish data collection flaws from model complexity issues.

Derivation

  1. Selection bias arises when participant selection is not random.
  2. This non-random selection leads to inaccurate study conclusions.
  3. The bias manifests in four specific ways: sampling, time interval, data, and attrition.
  4. Sampling bias occurs when some population members are less likely to be included.
  5. Time interval bias occurs when a trial is terminated early at an extreme value.
  6. Data bias happens when specific subsets are chosen to support a conclusion.
  7. Attrition bias is caused by the loss of participants who did not complete the trial.

Worked example

Problem

A researcher terminates a clinical trial early because the initial results show an extreme value. Is this selection bias or bias-variance trade-off?

  1. Identify the cause: The trial was terminated early at an extreme value.
  2. Check definition: Time interval bias occurs when a trial is terminated early at an extreme value.
  3. Distinguish: This is not model oversimplification (bias-variance trade-off).
  4. Conclusion: This is selection bias (specifically time interval bias).

Answer This is selection bias (time interval bias).

Common mistakes

  • Confusing selection bias with the bias-variance trade-off, which relates to model oversimplification and under-fitting.

Animation

Animation coming soon

The animated version of this lesson is being prepared for the web. In the Signals AI app, lessons are drawn step by step.

Code

This is a text lesson: it has no animation code yet.

ccBYNC© 2026 SignalsLab. Lesson text shared as is under CC BY-NC 4.0.