The Pentagon's recent purge of seven Arabic-language specialists -- people working in a field vital to the war on terrorism -- provides dramatic evidence that the country's policy on gays in the military urgently needs review.

All seven language specialists were released because they are gay. The question that needs to be answered is whether known homosexuality is a good reason to sacrifice a service member.

Ten years ago, the last time the nation seriously considered the matter of gays in the military, much of the debate was based on conjecture. Many allied nations either banned gay service members, as the United States did, or had unwritten "don't ask, don't tell" policies, which made careful study of the matter difficult.

Even those hostile to lifting the ban conceded that their position was based on hypotheticals, because no evidence existed to support their position.

But as David Burrelli, a national defense specialist for the Congressional Research Service, noted in congressional testimony: "The very existence of the policy itself prevents empirical research from discovering whether or not open homosexuals would, in fact, prove to be disruptive." In other words, where gays aren't allowed to identify themselves, it is impossible to assess their influence on military effectiveness.

Today we have numerous real-world instances of known gays serving in the military, providing clear evidence that open homosexuality does not impair military effectiveness. For starters, 23 nations allow gays to serve openly in their militaries. The Center for the Study of Sexual Minorities in the Military, a research center at the University of California at Santa Barbara, has published studies of gay men and lesbians serving in the Israeli, British, Canadian and Australian militaries.

Rather than relying on anecdotes or conjecture, our researchers asked experts to assess the effect on military performance of lifting the ban on gays. While isolated social disruptions required a few service members to be transferred, not a single informant observed any impairment of military effectiveness after the bans were lifted.

The ban's defenders argue that foreign militaries are not relevant to the United States because, as the world's sole superpower, our military must be stronger than any other. Certainly the United States has more international obligations than other countries do. But the question is not how similar our missions are to those of other nations but whether the United States is any less capable than other nations of integrating gays into its military. There is no conclusive evidence that U.S. commanders are peculiarly unable to direct those under their command to behave as a professional fighting force.

The second area of empirical evidence comes from the U.S. military itself. Despite the Pentagon's insistence that the presence of known gays would threaten morale and military effectiveness, known gays already are serving in the U.S. military. A recent study by Maj. John W. Bicknell of the Naval Postgraduate School in Monterey, Calif., shows that more than 39 percent of the Navy's sailors "personally know a homosexual service member." In addition, dozens of reports by service members indicate either that they know of gay peers in their ranks or that they themselves are gay and have served in units where their sexuality was widely known. No evidence has tied these units to decreased performance.

When asked to provide evidence to support the need for "don't ask, don't tell," champions of the ban can point only to opinions. Northwestern University Prof. Charles Moskos, the primary architect of the current policy, defended the ban by telling Congress that "empirical data" showed that sex integration degraded military effectiveness. But his "data" turned out to be opinions. "According to a Roper poll taken in Desert Shield," he testified, "45 percent of those who served in mixed-gender units in the Gulf said that there was enough sexual activity to degrade military performance."

Another defender cited "testimonies from [Norman] Schwarzkopf, from [Colin] Powell, as well as thousands of other testimonies" that, he said, all concluded that the presence of openly gay service members would undermine the military. But in similar surveys taken before Canada lifted its ban, 62 percent said they would refuse to fight alongside openly gay troops. According to the study by the Santa Barbara center, none resigned after the ban was lifted.

Opinions certainly have a role to play in public debate, but they do not constitute empirical evidence, and it's unwise to assess exclusionary policies by polling the excluders. Evidence from real militaries that have lifted real bans on gays provides a real test of the effects of gay troops on military effectiveness. By all accounts, gays -- even serving openly -- have now passed that test.

Not only does evidence show that allowing gays to serve openly does not undermine military effectiveness, the purge of intelligence experts and linguists suggests that the ban is hurting national security. If any evidence exists to defend this nine-year-old "interim" policy, it should be brought forward for evaluation. If not, the Pentagon should acknowledge that lifting the ban would not undermine combat effectiveness.

The writer is senior research fellow at the Center for the Study of Sexual Minorities in the Military at the University of California at Santa Barbara.