Censoring (statistics)

Censoring (statistics) is a concept in statistics and epidemiology that refers to the situation in which the value of a measurement or observation is only partially known. This can occur for a variety of reasons, such as when a value falls below or above a certain detection limit or when a study ends before all subjects have experienced the event of interest.

Types of Censoring

There are several types of censoring in statistics, including:

• Right censoring: This occurs when a subject's survival time is known to exceed a certain time, but it is unknown by how much. This is the most common type of censoring in survival analysis.

• Left censoring: This occurs when a subject's survival time is known to be less than a certain time, but it is unknown by how much.

• Interval censoring: This occurs when a subject's survival time is known to fall within a certain interval, but the exact time is unknown.

• Random censoring: This occurs when the censoring time is a random variable that is potentially independent of the survival time.

Applications in Medical Research

Censoring is a common issue in medical research, particularly in survival analysis where the outcome variable of interest is 'time until an event occurs'. For example, in a study of survival times of cancer patients, some patients may still be alive at the end of the study. These patients' survival times are right-censored.

Handling Censoring in Statistical Analysis

Censoring can introduce bias into the analysis if not properly accounted for. Several statistical methods have been developed to handle censoring, including the Kaplan-Meier estimator and the Cox proportional hazards model. These methods provide ways to estimate survival functions and compare survival rates between groups while taking censoring into account.

See Also

• Survival analysis
• Kaplan-Meier estimator
• Cox proportional hazards model
• Likelihood function
• Maximum likelihood estimation