The American-led military coalition in Afghanistan backed off Tuesday from its claim that Taliban attacks dropped off in 2012, tacitly acknowledging a hole in its widely repeated argument that violence is easing and that the insurgency is in steep decline.

In response to Associated Press inquiries about its latest series of statistics on security in Afghanistan, the coalition command in Kabul said it had erred in reporting a 7 percent decline in attacks. In fact there was no decline at all, officials said.

Defense Secretary Leon Panetta, who is among the senior officials who had publicly repeated the assertion of an encouraging drop-off in Taliban attacks last year, was disturbed to learn of the error, said his spokesman, George Little.

"This particular set of metrics doesn't tell the full story of progress against the Taliban, of course, but it's unhelpful to have inaccurate information in our systems," Little said.

A coalition spokesman, Jamie Graybeal, attributed the miscounting to clerical errors and said the problem does not change officials' basic assessment of the war, which they say is on a positive track as American and allied forces withdraw.

The 7 percent figure had been included in a report posted on the website of the coalition, the International Security Assistance Force, on Jan. 22 as part of its monthly update on trends in security and violence. It was removed from the website recently without explanation. After the AP asked last week about the missing report, coalition officials said they were correcting the data and would re-publish the report. As of Tuesday afternoon it had not reappeared.

It was not clear whether or how the Pentagon might correct a separate report _ its semi-annual report to Congress on security progress in Afghanistan, which used some of the same Taliban-attack statistics. The report was sent to Congress in December.

"We'll look at any adjustments that need to be made" to that report, Little said.

U.S. and allied officials have often cited declining violence as a sign that the Taliban have been degraded and that Afghan forces are in position to take the lead security role across the country when the last U.S. combat troops leave Dec. 31, 2014.

In mid-December, Panetta said "violence is down" for 2012 and Afghan forces "have gotten much better at providing security" in areas where they have taken the lead. He said the Taliban could be expected to continue to attack, "but overall they are losing."

Little said Panetta was briefed only "very recently" on the erroneous data.

U.S. and alliance officials try to measure progress against the Taliban from a variety of angles. Those include, for example, indications that the Taliban have lost much of their influence in population centers.

"The fact that 80 percent of the violence has been taking place in areas where less than 20 percent of the Afghan population lives remains unchanged," Little said.

The Taliban have lost a good deal of territory since a 2010 surge of U.S. forces in the southern provinces of Helmand and Kandahar, and they failed to recover it during the past two fighting seasons. Even so, they are resilient, and they are expected to severely test Afghan forces as the U.S. and its coalition partners step further into the background this year and complete their combat mission next year.

Many people, including coalition officials, have cautioned against the heavy reliance on statistics in assessing war progress. Yet the figures often are highlighted when they fit the narrative being promoted by leaders in Washington and other allied capitals.

"It is disturbing that, after 10 years of war, no reliable count of trends in violence exist even in terms of deaths, the most visible form of violence and one that is only a small portion of the actual causes and patterns of violence in the war," Anthony Cordesman of the Center for Strategic and International Studies wrote in February 2012.

Graybeal did not fully explain erroneous reporting of 2012 Taliban attacks by the International Security Assistance Force. It was not clear, for example, at what point the data errors began or who discovered them.

"During a quality control check, ISAF recently became aware that some data was incorrectly entered into the database that is used for tracking security-related incidents across Afghanistan," Graybeal said earlier.

He said an audit determined that portions of the data from unilateral Afghan military operations were "not properly reflected" in the trends ISAF had reported in its monthly updates.

"After including this unilateral ANSF (Afghan National Security Force) data into our database, we have determined that there was no change in the total number of EIAs (enemy initiated attacks) from 2011 to 2012," Graybeal said.

"This was a record-keeping error that we recognized and have now corrected," he added.

While ISAF routinely reports trends in Taliban attacks, it does not reveal exact numbers of attacks. Judging from its illustrative charts, however, it appears that there were more than 28,000 Taliban "enemy initiated" attacks in 2011.

The coalition defines enemy initiated attacks as those by small arms, mortars, rockets and improvised explosive devices, or IEDs. But it does not include IEDs that are found and cleared before they explode.

Trends in Taliban attacks are one yardstick used by ISAF to measure war progress. Others include the state of security in populated areas, the number of coalition and Afghan casualties, the level of economic activity, the degree to which civilians can move about freely and the performance of Afghan security forces.

Graybeal said that even though the number of 2012 Taliban attacks was unchanged from 2011, "our assessment of the fundamentals of campaign progress has not changed. The enemy is increasingly separated from the population, and the ANSF are currently in the lead for the vast majority of partnered operations."

___

Follow Robert Burns on Twitter at http://www.twitter.com/robertburnsAP