UNDERSTAND THE CONCEPT
A right-censored observation means an item survived to a known age, but its eventual failure time was not observed. Excluding such items can make a population look worse than the evidence supports. Record the observation end and why tracking stopped.
Use methods that handle censoring rather than treating observation end as a failure. Retirement due to poor condition can also be informative and may need more careful treatment than an unrelated study cutoff. Keep the data collection policy with the analysis.
Questions to take back to your team
- Is the end-of-observation reason recorded?
- Are replacements confused with failures?
- Does the chosen analysis method support the censoring pattern?
Further reading
Explore the primary references below. Requirements and applicability depend on your organization and jurisdiction.
