A military comparison group is considered the most closely comparable to the military participants. It includes men who served in the military at the same time, in the same types of military occupations and paygrades, and in similar settings as the participants, but who did not participate in Operation CROSSROADS. Whatever the unmeasured nonradiation risk factors the participants may have carried (e.g., behavioral, such as cigarette smoking; environmental, such as living on a ship; or demographic, such as education and income), the comparison group is likely to be equivalent, reducing the opportunity for biased study results.

Comparison with the general U.S. male population is a useful adjunct to a military comparison group. Scientists and other readers of mortality studies are used to seeing U.S. rates and can weigh interpretation accordingly. A very large drawback to using U.S. population rates exists, however. The "healthy worker effect" or, in this case, the "healthy soldier effect" makes interpretation of these comparisons difficult; we discuss this in Chapter 3.