José Ramírez, PhD, Global Quality Engineering, Amgen
When we measure a quantity – say, a concentration – the result is usually a single number like 0.25 mg/mL. Other times, however, we only know that our measurement is below or above a limit, or that it falls in a given interval. We call this type of data censored because we do not know its actual value. In this presentation we will discuss the different types of censoring and give examples of how they arise. We will demonstrate how to perform distribution analyses for censored data using the Life Distribution platform, drawing analogies to the Distribution platform, and will show how to calculate lower- and upper-tolerance bounds using distribution percentiles.