Professional Detailing2 min read

Right-Censored Durability Data: A Test Ending Is Not a Coating Failure

Record right-censored durability observations correctly when a coating still meets the criterion at the end of a study.

Updated · NANOLAB Knowledge

When a durability study ends while a sample still meets its defined criterion, the failure time is not known. The observation is right-censored: the sample lasted at least until the last valid observation, but its eventual endpoint was not observed.

Label the event correctly

In an illustrative twelve-month study, a sample still meeting the criterion at month twelve has an observed duration of at least twelve months under that procedure. Assigning it a failure time of twelve months would be wrong. Claiming an indefinite lifetime would also exceed the evidence.

If another sample stops meeting the criterion between scheduled visits at months eight and ten, the exact event time is not necessarily known either. The observation schedule limits the precision of the result.

Preserve the information needed for analysis

Record start date, observation dates, last acceptable observation, event definition and whether an event was actually observed. Include withdrawals and their reasons. A simple spreadsheet containing only a final “months lasted” column loses important distinctions.

For formal service-life analysis, use methods appropriate to censored data and obtain statistical support where needed. Ordinary averages of observed failures can misrepresent a dataset that includes many continuing samples.

Keep claims tied to the test

A statement such as “still met the defined criterion after twelve months in this study” is more precise than “lasts twelve months” when the endpoint has not been reached. It also leaves room for the fact that another environment or maintenance routine may produce a different result.

For NANOLAB evaluation, define the relevant criterion before starting: appearance, a specified water-behaviour result or another measured property. A change in one criterion is not automatically disappearance of every coating function. Clear event definitions and censoring records make long-running trials more useful as evidence accumulates.

Related guidance

Technical references: NIST: Censored reliability data.

For product-specific application instructions, consult NANOLAB technical documentation or ask about your surface and intended use.