On the evening of October 5, 2013, state troopers ran a sobriety checkpoint on Route 1A in Beverly, Massachusetts. A trooper smelled liquor coming from Lindsay Hallinan's car; a second watched her fail a series of field sobriety tests, then had her breathe into a Draeger Alcotest 9510. The machine printed a blood alcohol reading of 0.23 percent, nearly three times the legal limit. Hallinan's lawyer told her the number made her case unwinnable, and she pleaded guilty that November.
The printout did not say that nobody in Massachusetts had a written method for confirming the machine behind it was telling the truth. The state's Office of Alcohol Testing (OAT), a unit of the State Police crime lab, had put the Alcotest 9510 into service statewide in June 2011 with no formal certification protocol; the steps for calibrating one were passed on, a judge later found, by word of mouth around the lab. After a ten-day hearing in a case consolidating more than five hundred challenges to the device, the judge ruled in February 2017 that the Alcotest 9510 itself could produce reliable readings, but that OAT's unwritten method, used from June 2011 through September 2014, had not — and declared every result from that window presumptively unreliable.
A worse finding followed. Every test produced a certification worksheet, pass or fail. In August 2017, a records request turned up hundreds the lab had never handed over, despite a standing court order to produce every one. Of nearly two thousand worksheets OAT eventually released, only eleven recorded a failure; a further 432 failed worksheets had been withheld outright. The lab's own supervising scientist considered a failed calibration exempt from disclosure, and told her staff to pull any failed worksheet before it went out. A six-week state investigation found OAT's discovery practices dysfunctional, built on serious errors of judgment, and enabled by an institutional culture its own investigators later described in words the state's highest court would go on to quote.
On April 26, 2023, the Massachusetts Supreme Judicial Court ruled on Hallinan's own appeal and said what OAT's conduct had cost beyond her case. The Commonwealth had already agreed to notify roughly twenty-seven thousand people that their case rested on an Alcotest 9510 reading the lab could not vouch for. The court set the period of presumptive unreliability at June 1, 2011, through April 18, 2019, once Massachusetts met a judge's remedial measures, among them an accreditation application found substantially likely to succeed, and held that every one of those defendants is entitled to a conclusive presumption that OAT's misconduct happened and came before their plea. Each still has to show, as Hallinan did, that the misconduct affected their own decision to plead guilty — but none of them has to prove OAT's conduct was egregious to begin with.
A breath test is a number a machine manufactures, and that number is only as good as the record proving the instrument was working when it produced it. That record is what a calibration worksheet is for. Withholding the failing ones did not make the Alcotest dishonest. It meant that, for years, nobody outside the lab could tell a reading from a properly working instrument apart from one the lab itself already knew had failed.
The worksheet behind a failed calibration was not a technicality. It was the only thing standing between a number and a fact.