Blood samples: probability of discrimination.

The paper shows how to calculate the probability of being able to discriminate between two blood samples using a single test. The distribution of the population with respect to this test criterion is assumed to be known. The paper goes on to describe how to calculate the probability of being able to discriminate between two blood samples using a battery of such tests. The tests are assumed to be independent. The actual calculations are made for a battery of seven such tests. Finally, two conditions which tests of good discriminating power should satisfy are stated and proved.