Course contentsShow
← Data Science
Lesson 804 of 2,14521. Survival AnalysisFree lesson

Right Censoring

Understanding right censoring when subjects exit before the event occurs or the study ends.

Right Censoring

You'll learn what right censoring is and why it's crucial for handling incomplete survival data.

What Is Right Censoring?

Right censoring occurs when you know a subject's survival time up to a certain point, but you don't observe the actual event of interest. This happens in two common scenarios:

  1. The study ends before the event occurs (you run out of time)
  2. The subject exits early (they move away, drop out, or experience a competing event)

The key insight: you know the subject survived at least until their last observation, but you don't know when (or if) the event eventually happened.

A Simple Analogy

Imagine tracking how long light bulbs last in 100 homes. You start your study in January and plan to end in December. By December:

  • 60 bulbs have burned out (you recorded exact failure times)
  • 40 bulbs are still working (right-censored observations)

For those 40 bulbs, you know they lasted at least 12 months, but you don't know their true lifespan. Throwing them out would bias your results toward shorter lifespans. Right censoring lets you include this partial information: "still alive at 12 months."

Why This Matters

Without accounting for censoring, you'd either:

  • Ignore subjects who didn't experience the event (wasting data and creating bias)
  • Incorrectly treat their last observation as the event time (wrong!)

Survival analysis methods are specifically designed to handle right-censored data correctly, extracting maximum information from both complete and incomplete observations.

Key Takeaway: Right censoring captures subjects whose event time is unknown but known to exceed their last observation time—essential for analyzing real-world time-to-event data without bias.