Premium
THE PROBLEM OF MISCLASSIFICATION OF MORTALITY STATISTICS
Author(s) -
Heathcote C.R.
Publication year - 1988
Publication title -
australian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.434
H-Index - 41
eISSN - 1467-842X
pISSN - 0004-9581
DOI - 10.1111/j.1467-842x.1988.tb00475.x
Subject(s) - statistics , mathematics , section (typography) , multinomial distribution , matrix (chemical analysis) , diagonal , computer science , materials science , geometry , composite material , operating system
summary The problem of misclassification of cause of death in mortality tables is discussed through the introduction of a classification matrix. This is a stochastic matrix whose off‐diagonal entries are the probabilities of a misclassification. An example is given in Section 2 which shows how a classification matrix can be estimated. An unbiased linear estimate of the unknown true numbers of deaths in different categories is obtained in Section 3 and its asymptotic normality is established. It is shown that the usual classical estimate is biased. Section 4 deals with the special case of a classification matrix whose rows define independent multinomial distributions.