Harris & Boyd's test for partitioning the reference values.

To decide upon the need for partitioning, Students t test or Mann-Whitney^ U test are usually applied to compare the reference samples under study. But these tests may be too sensitive for this purpose: if the sizes of the two subgroups compared are large enough, small differences between subgroups without clinical importance may become statistically significant. For this reason, Harris & Boyd developed a special statistical test in order to find out the need for partitioning under a diagnostic point of view (2—3).