IN MID-NOVEMBER, just weeks before the deadliest terrorist attack on U.S. soil since 9/11, the Joint Regional Intelligence Center and the sheriffs’ departments of San Bernardino and Riverside counties held the First Annual Inland Terrorism Liaison Officer Conference in Fontana, California. The two-day event — for law enforcement, public officials, and select members of the private sector — included sessions like “Policing Violent Extremism” and “Preventing Lone Wolf Attacks.”

In fact, this part of California’s Inland Empire has become home to a cottage industry of counterterrorism training in recent years aimed at teaching people how to spot would-be terrorists before they attack. By all accounts, those trainings failed to help anyone spot Syed Rizwan Farook and Tashfeen Malik, the married couple who shot and killed 14 people and injured 22 others at a meeting of San Bernardino County Health Department employees on December 2.

Many of the trainings, which focus on helping attendees identify “behavioral indicators” of potential terrorists, were held at the Ben Clark Training Center in Riverside, California, less than 25 miles from where the attacks took place.

These behavioral indicators have become central to the U.S. counterterrorism prevention strategy, yet critics say they don’t work. “Quite simply, they rely on generalized correlations found in selectively chosen terrorists without using control groups to see how often the correlated behaviors identified occur in the non-terrorist population,” Michael German, a former FBI agent who is currently a fellow at the Brennan Center for Justice at New York University School of Law, told The Intercept.

The trainings are based on flawed theories that just don’t stand up to empirical scrutiny, according to German. “The FBI, [National Counter-Terrorism Center], and [Department of Homeland Security] promote these theories despite the fact they have been refuted in numerous academic studies over the past 20 years,” he said.

Yet the behavior indicator training business appears to be booming in California, where the training sessions are sponsored by an alphabet soup of counterterrorism organizations that have sprung up in recent years, including the Joint Regional Intelligence Center; the Los Angeles chapter of InfraGard, a partnership between the FBI and private sector; and the state fusion center.

The Joint Regional Intelligence Center, in turn, has produced dozens of Official Use Only intelligence bulletins focusing on behavior indicators. One intelligence bulletin, from March 2015, identified potential indicators of radicalization including “history of mental instability/illness”; “employment/financial problems”; and “marital/family problems.”

Southern California’s enthusiasm for terrorist spotting dates back to 2002, when it was home to the first Terrorism Liaison Officer program, the controversial initiative that enlists and credentials community members and private sector industry representatives to report any potentially suspicious behavior. The program was first launched out of the Los Angeles chapter of InfraGard — which covers seven nearby counties including Riverside and San Bernardino where the attacks occurred and the perpetrators lived; the program has since been rolled out nationwide.

The Los Angeles chapter of InfraGard has also been a major beneficiary of federally funded grant money for counterterrorism training. In 2013 the Los Angeles County Board of Supervisors doubled the funding of its multimillion-dollar sole source contract with InfraGard to $2,530,000 and extended it through 2018.

One of the companies hired by InfraGard to conduct counterterrorism training is CT Watch, headed by Roque “Rocky” Wicker, who also holds an executive leadership position with the Los Angeles InfraGard chapter. Employees of CT Watch have taught seminars, such as “Threat of ISIS and radicalization in the homeland.”

“The indicators work,” Wicker told The Intercept in an interview. “Behavior indicators work. You just need to train the right people.”

If Wicker is right, then the dozens of trainings held in California over the past year failed to train the right people to spot San Bernardino shooter Farook, who was a state employee and would have interacted with other state officials on a daily basis, or his friend Enrique Marquez, who was indicted Wednesday on charges related to his role in plotting with Farook to carry out attacks in 2011 and 2012.

In the months leading up to the attacks, law enforcement sources say terrorism trainings had increased in response to threats specific to California. For example, a recent issue of Dabiq, the Islamic State’s magazine, had listed potential central California targets.

Recent training sessions held in Riverside include “How to assess the threat posed by a potential lone wolf attacker,” “The Stealth Jihad in the United States,” and “Behavior threat assessment: preventing the Active Shooter,” which took place on October 22, just weeks before the San Bernardino shootings. The last one was designed “to equip law enforcement and security stakeholders with the skills and tools necessary to identify potentially violent individuals, assess the risk they pose of engaging in violence, proactively manage the risk and prevent violent attacks — including active shooter events.”

There were also three separate “Tactical Response to School & Community Violence” active shooter trainings in November, the most recent of which was held on December 2, the same day as the terrorist attack. Participants in the training were among the first to arrive at the scene of the San Bernardino shootings.

As for why behavioral indicator spotting failed to identify the San Bernardino shooters, one problem is simply that that the indicators are overly broad. A local law enforcement official involved in the vetting of suspicious activity reports in the Riverside area told The Intercept that prior to the attack, they’d received hundreds of suspicious activity reports, “most of which turn out to be bullshit. We run them down of course, but mostly, it’s a lot of nothing.”

Research: Sheelagh McNeill