Christopher R.
Bilder, Ph.D.
University of Nebraska-Lincoln
Department of
Statistics
340 Hardin Hall North,
East Campus
Lincoln, NE 68583-0963
Phone: (402) 472-2903,
Fax: (402) 472-5179
Website:
www.chrisbilder.com
E-mail:
and
Instead of testing each individual, observations are made on groups of
individuals pooled together with group size s > 1. In
medical-screening experiments involving infectious diseases, grouping might
involve pooling blood samples from several individuals to test for seropositivity. In most applications, the group response is binary,
classified as either positive or negative.
Images from a past presentation:
Individual testing
Problems: Cost and time to test all 16 people!
Group testing
If the group tests negative, then all 16 individuals
are declared negative. Thus, it took just one test to determine 16
people are negative; it would have taken 16 tests with individual testing to
come to this same conclusion. If the group tests positive, then at
least one of the 16 individuals is a positive. Various retesting
strategies can be implemented to determine who is positive. Note that
the American Red Cross screens blood donations in groups of size 16.