ABSTRACT
In the simplest context, the issue of conditional inference arises in relation to the exponential model when three items are put on test at the same time but the time at which the first failure occurs goes unrecorded. In the most general context, the issue of conditional inference for the one-parameter exponential model appears to arise when data are gathered following a Type II progressively censored scheme that contains at least one observation censored on the left. The conditional method based on ancillary statistics allows a dimensionality reduction in the process of deriving inferences about unknown parameters. The method is applied to the one-parameter exponential model when the data are gathered following a Type II progressively censored scheme that includes at least one observation censored on the left.
