Suspensions occur when units under test do not fail during the duration of the test. Because this condition is referred to as censoring, suspensions are also called censored units. If, for example, some of the units under test survive the test, the data is said to be right-censored because the suspensions occur on the right side of a timeline representing the test. Likewise, when a random sample of units is drawn, some of the units might fail prior to the start of the test. In this case, the data is said to be left-censored because the suspensions occur on the left side of a timeline representing the test.

If unit failures are caused by different failure modes than the one under test, you must record failure times for these suspensions. When test data is collected in the field, it can contain a mix of all possible censoring types. In such cases, you generally do not know the exact time that the suspended unit failed, which results in interval-censored data. In these cases, you know only the interval in which the suspension occurred.

Other types of censoring include time-censored data and failure-censored data. In time-censored data, censoring is the result of units surviving past the end of the pre-determined test time, which means that the number of recorded failures is random. In failure-censored data, censoring is a result of stopping the test when a predetermined number of failures is reached, which means the time at which the test is stopped is random.

Data for suspensions must not be ignored. In the Weibull Parameters pane for a parametric life data set, Suspensions is a checkbox. When you select it, Failure/Suspension is shown so that you can indicate whether a data point is for a failed unit or a suspension.