Accuracy vs. Probability In A Paternity Test

In paternity testing the word "accurate" refers to the paternity test process in the laboratory, and verifying it is free from mistakes. This happens by following strict guidelines to ensure the integrity of each DNA sample that is processed.

"Probability" of paternity in paternity testing is defined as the calculated likelihood that the alleged father is the biological father of the child being tested.

Example: In paternity testing, when all genetic systems are analyzed and concluded to be a match, it shows with 100% accuracy that the probability of paternity is 99.99%. This is based on the correctness of the procedures dealing with genetic testing, sample handling, and reporting of the laboratory performing the paternity test.

Why will paternity test results never show 100% probability of paternity?

Without having genetic data (or profiles) from every human male in the world, the probability of paternity can only reach 99.99%. This is because science cannot rule out 100% of all untested males.

The probability of paternity is based on Bayes’ Theorem. This model assumes a prior probability. The laboratory assumes (in most cases) there is a "prior probability" (a 50% chance) that the alleged father is the biological father of the child being tested. This assures the calculations are not based on probable information from the participants ("He said" vs. "She said"). Prior probability is a standard used in the DNA testing industry because as a laboratory, they have no prior information regarding the case. This prior probability gives both paternity and non-paternity equal influence.

The probability of paternity in a paternity test is calculated based on the Combined Parentage Index (CPI), which is a likelihood that the shared alleles (or markers) between the child and the tested man are, or are not common among the population, and the commonality is given a measurement of strength (the lower the number the more common it is in the population, and the higher the number the less common it is) and this genetic evidence is presented on the paternity test report.

This CPI in a paternity test is then used to calculate the probability that this man tested is the biological father of the child.