An introduction to Bayesian credible intervals for sampling error in DNA profiles

Experts presenting statistical DNA evidence are often challenged about their ability to use databases of only a few hundred people to claim probabilities in the billions. This challenge is countered with the inclusion of a sampling error calculation in the case report. Sampling error is present in every DNA calculation because the statistics rely on a sample (database) of individuals to estimate population frequencies of alleles and this error must be quantified. This paper explains in simple terms how Bayesian estimation theory is used to produce Bayesian credible intervals for sampling error calculations in DNA profiles.