Censored observations that can be considered to have the same probabilities of failure at later times as those individuals remaining under observation.