Studying the epidemiology of AIDS can be difficult because it requires delving into the sexual activities of individuals, and people are generally reluctant to reveal details of their private lives. For example, asking a man whether he has had a sexual encounter with anyone other than his wife often elicits a look of skepticism and an unreliable answer. Researchers can rarely be certain that they are getting the truth.
In 1987, Joel E. Cohen, a mathematical biologist from Rockefeller University, suggested a statistical tool that could add reliability to the answer. The technique was invented in 1985 by Stanley Warner of Ontario's York University. It has been used periodically since then. Its basic premise is that people will reveal the truth if they believe their answers are secret.
The method requires a question and coin flip. The questioner asks a man, for example, "Have you ever had a sexual encounter with anyone other than your wife?" The man then flips the coin in privacy. He answers no if he has not had an encounter and if the coin comes up tails. Otherwise he must answer yes.
Using this method, a man can answer no because the flip resulted in tails and can answer yes, also because of the coin. A yes does not necessarily incriminate him, and a no does not necessarily mean he is lying.
Totaling the results, the researcher simply doubles the number of no answers to get an almost correct number of nos. The theory is straightforward: given enough flips, the coin should come up tails half the time; this means that only half the total number of men who actually had encounters said no. This is why doubling the number of no’s gives a fairly accurate picture of the true number of nos. Subtracting the number of no’s from the total number of men questioned reveals the number of yes answers.
