When CIA director Leon Panetta declared on a Sunday talk show in late June that “we’re looking at maybe 50 to 100” al Qaeda fighters in Afghanistan, “maybe less,” some commentators took this as a political turning point. British journalist and author Stephen Grey commented via his Twitter account, for example, that the statement “could change the whole war debate.”

No doubt the 50 to 100 figure will be repeated by officials and pundits for some time to come; given the paucity of available information, it will factor heavily in debates over America’s strategic interests in Afghanistan. This is unfortunate, as every available indicator suggests that Panetta’s figure is unreliable. Worse, it may be evidence of a lack of rigor within the U.S. intelligence community.

Data points gleaned from coalition operations seem to undercut Panetta’s claim. Consider the case of Kunar Province.

The same day that Panetta said there were only 50 to 100 al Qaeda fighters in Afghanistan, the U.S. military announced an operation in Kunar targeting “al Qaeda and Taliban leadership in the area.” While the names of targeted al Qaeda leaders were not disclosed, two major known al Qaeda leaders in Kunar are Abu Ikhlas al Masri, al Qaeda’s operations chief for Kunar, and Qari Zia Rahman, who is considered the top regional commander in Kunar and Nuristan.

A report published last year by the Institute for the Study of War provides more detailed information about the insurgency in Kunar. It notes a 2008 estimate by provincial officials, which coincides with information provided to us by an American intelligence source, that there were “at least 2,000 insurgents in the mountains of Kunar.” Though the number fluctuates as insurgents crisscross the border with Pakistan, it is significant that about half of these fighters were “believed to be foreigners, including Pakistanis, Arabs, Chechens, and Uzbeks.” Most foreign fighters are assessed to be loyal to al Qaeda.

The number of insurgents in Kunar has probably risen since 2008. More fighters are reported to have moved there since the United States began abandoning its Kunar outposts last fall. Even if 90 percent of the foreign fighters in Kunar belong to the Pakistani Taliban and only the remaining 10 percent are Arabs, Uzbeks, Chechens, and so on, that would suggest that Kunar alone has a significant al Qaeda problem—with about as many al Qaeda-affiliated fighters as Panetta estimated for the entire country.

But the number of al Qaeda fighters is almost certainly higher still. One American intelligence officer who recently returned from Afghanistan told us: “About half of the insurgents in Kunar are foreign fighters, and given that most of the foreign fighters are loyal to al Qaeda, that would easily put the group’s numbers at approximately 1,000 in Kunar alone.” The military’s intelligence agencies do not concur with the figures Panetta cited.

Another difficulty in measuring al Qaeda’s presence in Afghanistan is defining “al Qaeda.” The distinctions between many militant leaders or groups and al Qaeda have broken down over time. Qari Zia Rahman, for example, straddles al Qaeda and the Taliban. Hakimullah Mehsud’s Tehrik-i-Taliban Pakistan is closely allied with al Qaeda; it sponsored the failed Times Square bombing and aided an al Qaeda operative in the suicide attack that killed seven CIA officials in Khost province. And Sirajuddin Haqqani, head of a powerful network in Pakistan, is believed by some intelligence officials to be a member of al Qaeda’s shura majlis, or leadership council.

It is possible that the unreliable estimates of al Qaeda numbers are politically motivated—designed either to demonstrate progress or to hasten U.S. disengagement from Afghanistan. Another possibility, and perhaps cause for greater concern, is that they reflect poor analytical methods.

It is unclear what methodology CIA analysts used to arrive at the 50-100 figure. (This opacity can work to the advantage of anyone trumpeting the figure, as those without the proper security clearances cannot get a sense of how the sausage is made.) One strong possibility, however, is that the analysts ran through recent intelligence and embraced the fixed numbers they came across without subjecting them to rigorous scrutiny.

The analyst we spoke with about Kunar noted that forming an estimate for even a single province requires specific, in-depth knowledge of that place. It would take detailed data for “every province in Afghanistan to derive a real approximation of the enemy” in the country as a whole. That would require a great deal of research, effort, and analysis—which Panetta’s estimate seems to lack.

Of course, unreliable estimates do not start and stop with Leon Panetta. Rahm Emanuel claimed in late June that about “a half of al Qaeda has been eliminated in this last 18 months.” Similarly, George W. Bush claimed in 2004 that three-quarters of al Qaeda had been killed or captured. Such declarations can only confuse the public debate—or produce confused policy that ultimately costs lives.

Daveed Gartenstein-Ross is the director of the Center for the Study of Terrorist Radicalization at the Foundation for Defense of Democracies (FDD) and coeditor of The Afghanistan-Pakistan Theater: Militant Islam, Security, and Stability (FDD Press, 2010). Bill Roggio is the editor of The Long War Journal and a senior fellow at FDD.