A serial dilution test estimates the concentration of a microbe in a broth by inoculating several tubes with portions of the broth. The test may include both presumptive and confirmation steps. For the confirmation step considered here a portion of the contents of each tube indicating a change in the presumptive step is streaked on a plate. Several colonies are selected and examined to determine if the tube contained the target microbe. A statistical model accounts for the possibility of not selecting the target microbe from a tube that contains it and a similar appearing microbe. Simulations show that this model sometimes gives estimates similar to those currently used, but at other times can make large corrections. Also, a commonness measure helps check the validity of the assumptions of this statistical model.