Reference intervals remain the most common means of determining whethe
r an observed value suggests the presence of disease. When creating re
ference intervals, one must look at the requirements for the preanalyt
ic, analytic, and postanalytic phases. For the preanalytic phase, inve
stigators use exclusion and partition criteria to create the sample po
pulation. During the analytic phase, the investigator should use contr
ol charts to ensure that the system is in good control. During the pos
tanalytic phase, the investigator uses the nonparametric approach, whi
ch involves establishing the 2.5 and 97.5 percentiles reference limits
. This is the fourth and final article in a continuing education serie
s on statistics. Other articles in this series dealt with control char
ts, the use of linear regression to compare methods, and linearity ass
essment in the laboratory. The exam follows this article.