In 2007 Don Bolles, drummer for the punk rock band the Germs, was arrested for possession of the banned depressant GHB, commonly known as the "date rape drug," after police in Newport Beach, California, pulled over his van because of a broken taillight. The charges were dropped after a lab test showed that the bottle of "yellowish goop" in his toiletry bag contained exactly what the label said: an all-natural, peppermint-infused cleanser produced by Dr. Bronner's Magic Soap.
In 2012 Robert and Adlynn Harte were awakened by sheriff's deputies pounding on the door of their house in Leawood, Kansas, demanding entry so they could search the place for the marijuana plants they were sure they'd find there. The deputies forced Robert Harte to the floor at gunpoint and then confined him, his wife, their 7-year-old daughter, and their 13-year-old son to the living room sofa while they spent two and a half hours looking in vain for evidence of criminal activity. It turned out that the "saturated plant material" deputies found in the couple's kitchen garbage was not marijuana but loose tea.
In 2013 Alexander Bernstein spent a month in a Pennsylvania jail after state troopers pulled over the rental car in which he was riding, found "two brick-size packages" of white powder in the trunk, and charged him with possessing and conspiring to distribute cocaine. His bail, originally set at $500,000, was reduced to $25,000 four weeks later, which is how he finally managed to get out of jail. Five days after his release, a lab test showed the white powder was what he had claimed all along: homemade soap.
These are just a few examples of the periodically publicized brushes with the law caused by unreliable but widely used field tests for illegal drugs. Many people suffer consequences worse than the embarrassment, legal costs, and temporary loss of freedom experienced by Bolles (who spent three days in jail), the Hartes (who had to mount an expensive legal battle just to find out why their house was searched and so far have been rebuffed in their attempt to win compensation), and Bernstein (who last fall sued the state troopers responsible for his arrest and detention). In last week's New York Times Magazine, ProPublica reporters Ryan Gabrielson and Topher Sanders estimate that thousands of Americans have been not just falsely accused but falsely convicted of drug crimes based on erroneous field test results.
These tests are generally not admissible in court, but that does not matter much, since the vast majority of drug cases—90 percent or more—are resolved by plea deals. Gabrielson and Sanders found that "guilty pleas are allowed based on field tests alone in dozens of jurisdictions, among them Atlanta, Boston, Dallas, Jacksonville, Las Vegas, Los Angeles, Newark, Philadelphia, Phoenix, Salt Lake City, San Diego, Seattle and Tampa." Focusing on Harris County, Texas, Gabrielson and Sanders found that 212 people were convicted of possessing a substance that a field test mistakenly identified as an illegal drug between January 2004 and June 2015.
One of those erroneously incriminated defendants, a Louisiana woman named Amy Albritton, was arrested in 2010 after Houston police pulled over her car and found a white crumb they mistakenly identified as crack cocaine. Although she was innocent, the prospect of up to two years in prison, combined with everyone's pseudoscientific certainty that she was guilty, persuaded her to accept a plea deal that promised a short jail sentence. She ultimately spent just three weeks in jail, but that was the least of her punishment. She lost her job as the manager of an apartment complex in Monroe, Louisiana, and her new status as a felon made it impossible to find steady work that paid nearly as well, which in turn made it difficult to care for a son with cerebral palsy. Her whole life was upended by a crumb on the floor of her car that a lab test later found was not any sort of illegal drug. Albritton had no idea she had been exonerated until Gabrielson and Sanders tracked her down and got in touch with her.
Without going through the records of every county in the country (or at least a representative sample), it is hard to say how many people have had experiences similar to Albritton's. But Gabrielson and Sanders take a stab at it, considering the data from Harris County, the ubiquity of field tests, the widespread willingness to accept plea deals based on those tests, and the large number of arrests for drug possession (1.3 million in 2014, the most recent year for which FBI data are available). "If Albritton's case is one of hundreds in Houston," they write, "there is every reason to suspect that it is just one among thousands of wrongful drug convictions that were based on field tests across the United States." Later they say Albritton is "one among unknown tens of thousands of Americans whose lives have been torn apart by a very flawed test."
How flawed? That question also is hard to answer with precision. Gabrielson and Sanders note that a Las Vegas study found one out of three samples identified as cocaine was in fact something else. In Florida, they write, "21 percent of evidence that the police listed as methamphetamine after identifying it was not methamphetamine, and half of those false positives were not any kind of illegal drug at all." In 2013 Travis County, Texas, District Attorney Rosemary Lehmberg decided to stop accepting guilty pleas in drug cases without a lab test after her office discovered a dozen false positives from field tests during a six-month period.
One source of false positives is nonspecific tests. The Scott reagent test that was used to incriminate Albritton involves dropping a suspected drug sample into a vial of cobalt thiocyanate, which is supposed to turn blue in the presence of cocaine. But as Gabrielson and Sanders note, "cobalt thiocyanate also turns blue when it is exposed to more than 80 other compounds, including methadone, certain acne medications and several common household cleaners."
The Hartes, the Kansas couple whose house was raided in 2012, drew police attention because Robert visited a hydroponics store with his son to buy supplies for a school project. They were falsely implicated by a KN reagent test that twice indicated Adlynn's tea was marijuana. Those results supposedly confirmed suspicions that the Hartes were growing marijuana and extracting THC from it, even though the plant matter in the trash neither looked nor smelled like cannabis. Johnson County Sheriff's Deputy Mark Burns, who found the supposed contraband, admitted in a deposition that the material was "hard to identify" but said "he never considered that the substance could be tea leaves because he had never in his life seen loose tea leaves."
The test that supposedly clarified matters—a test that was crucial both to obtaining a search warrant and to a federal judge's rejection of the Hartes' subsequent lawsuit—seems to be incapable of distinguishing between cannabis and many common organic substances. In 2008 and 2009, Claflin University biotechnologist Omar Bagasra used the NarcoPouch KN reagent test to screen 42 legal plant products. Thirty-three of them tested positive for marijuana, including spearmint, peppermint, basil, oregano, patchouli, vanilla, cinnamon leaf, lemon grass, bergamot, lavender, ginseng, anise, gingko, eucalyptus, rose, cloves, ginger, frankincense, vine flower, chicory flower, olive flower, cypress, and St. John's wort. Several of those are common ingredients in herbal tea. Another test, the NIK Duquenois-Levine Reagent System, did much better but still misidentified cypress, patchouli, and eucalyptus as marijuana while generating inconclusive results for oregano, spearmint, and lavender.
Even exposure to air seems to trigger some tests. The Marshall Project's Alysia Santo reports that a lieutenant with the Hillsborough County, Florida, sheriff's office did some experiments with a NARK II field test and "found that just opening the test bag to the air produced the same shade of purple as exposure to methamphetamine." At a 2009 press conference highlighting Bagasra's findings, marijuana policy activist Adam Eidinger opened a NarcoPouch KN reagent test bag, waved it around, and resealed it, which caused a color change. "This is just air," he emphasized.
Another source of false positives is human error. Police may perform a test incorrectly, misunderstand the instructions about what different colors mean, or misperceive the colors in the poor lighting conditions that are common during roadside stops. Weather conditions also can affect results.
Despite the national publicity attracted by cases in which people are falsely incriminated by field tests, cops often claim to be unaware of the problem. Burns, the Kansas deputy who did not know what loose tea was, had worked at the Johnson County Sheriff's Office for 15 years and had eight years of experience with drug cases. But he testified that "he had no knowledge that anything other than marijuana could test positive on a marijuana field test kit and that he was 'not aware' of the possible occurrence of false positive test results."
Evidently Burns did not read his own affidavit in support of the search warrant for the Hartes' house, which plainly states that "this test is presumptive but not conclusive for the presence of marijuana." Burns' boss, Sheriff Frank Denning, also did not take that warning to heart. Denning, who has more than four decades of experience in law enforcement, said "the only false positive results of which he is aware are the results at issue in this case."
Charles McClelland, Houston's police chief until last February, and Ray Hunt, head of the local police union, likewise told Gabrielson and Sanders they did not realize there was a problem with false positives from drug field tests. Gabrielson and Sanders' exposé is a wake-up call for all of the cops and prosecutors who continue to put unjustified faith in tests that lie.
This article originally appeared at Forbes.com.