• Entry
  • Reader's guide
  • Entries A-Z
  • Subject index

Calibration refers to the degree of correspondence between probabilities and observed relative frequencies. Suppose that when a patient is admitted to an intensive care unit (ICU), a physician assesses the probability that the patient will survive until hospital discharge. If data on these probabilities and the resulting outcomes (survival or death) are collected for a large number of patients, the data can be organized by the numerical probabilities. For example, the relative frequency of survival among all the patients for whom a probability of 70% was assessed can be determined. If this relative frequency is 70%, and the relative frequencies for other probability values also match those probabilities, the probabilities are said to be perfectly calibrated. If the probabilities and their associated relative frequencies differ, ...

  • Loading...
locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles