A censored observation records only that a unit survived past a certain time (right censoring, which reliability engineers call a suspension), failed before a time (left censoring) or failed between two inspections (interval censoring). Fitting only the failures and ignoring the survivors badly underestimates life. Maximum likelihood and rank adjustment methods use suspensions properly, and a test stopped at a fixed time or after a fixed number of failures is called Type I or Type II censored.
Test stopped at 1,000 hours with 3 fails out of 40. The other 37 are censored, so put them in the fit as suspensions.
Throwing out the units that did not fail is the costliest mistake in life data analysis: it makes the product look far worse than it is.
Also heard as
- censored data
- suspension
- right censoring
- runout
Related terms
Weibull distribution
Flexible life distribution with a shape and a scale parameter, the standard model for fitting time-to-failure data.
MTTF
Mean time to failure: expected life of a non-repairable item, the average time until its first and only failure.
Acceleration factor
Ratio of life under normal use conditions to life under a harsher test condition, used to turn test hours into field time.