Statistics applicable to the inference of a victim's blood type from familial testing.

Abstract Two recently reported cases involving inference of a victim's blood type using parental testing are reviewed. The statistic "frequency of non-excluded couples" was used in these cases to interpret the bloodstain evidence. Two other statistics are introduced: the frequency of non-excluded stains, and the likelihood ratio. Both the frequency of non-excluded couples and the frequency of non-excluded stains present an incomplete assessment of the non-exclusion because each considers only a portion of the data. The likelihood ratio utilizes all the data and provides an accurate assessment of the evidential significance.