The program is overseen by the Sheriff's Office to supervise people recently arrested but released before trial to ensure they appear in court and reduce any new criminal activity.
RELATED: Thief steals garbage bag full of items from SF Walgreens with security filming in plain sight
The contract to fund the program's extension passed the board overwhelmingly in mid-June after the Sheriff's Office said weeks earlier the program has been successful since 2016 in reducing the rate alleged criminals reoffended before trial.
The program is required by law to complete is a Public Safety Assessment, or PSA, which is a measurement tool that helps the court determine whether someone arrested should remain in custody or be released while their charges are pending.
A new independent assessment that aimed to validate the PSA found 55 percent of alleged criminals released in San Francisco reoffended before trial. It's important to note this study was not a direct assessment of the San Francisco Pretrial Diversion Program.
"This is a scandal," said San Francisco Supervisor Catherin Stefani. "They are masking what's actually happening."
Supervisor Stefani was the only board member to vote against approving the $18 million contract.
"The data didn't add up," she said. "It wasn't matching what we were seeing out on the streets."
VIDEO: What is SF doing to fight organized retail theft? SFPD and DA Chesa Boudin respond
On June 3, Crispin Hollings, the chief financial officer for the San Francisco Sheriff's Office presented to the Government Audit and Oversight Committee a graph illustrating more than 90 percent of the city's most violent alleged offenders didn't commit new crimes before trial from 2016 to 2020.
The California Policy Lab, a research arm of UC Berkeley, conducted a study analyzing how effective the Public Safety Assessment is in San Francisco.
The San Francisco Sheriff's Office reported in committee, less than 10 percent of the city's most violent offenders released before trial committed new crimes from 2016 to 2019. Roughly 20 percent or less of other non-violent offenders reoffended during the same time period, according to data from the Sheriff's Office.
But, the validation study found the rate alleged criminals reoffended during those years was actually 55 percent. The study also concluded 74 percent of the city's most violent offenders committed new crimes before trial.
"What did you think after you saw the report?" ABC7's Stephanie Sierra asked.
"I was angry, actually. I was very angry," said Stefani. "We had the highest rates of any other jurisdiction in that study."
By the time the assessment was released on July 1, it was already two weeks after the Board of Supervisors voted to approve the contract.
RELATED: 'Traumatizing': Burglar slams car into SF business to steal reported $20K in lottery scratchers
So, why was the data different?
The I-Team spoke with California Policy Lab's lead researcher Dr. Johanna Lacoe.
"Why do you think the results came out so different?" Sierra asked.
"We have access to data where we can see the very beginning of a case through its resolution," Lacoe said. "That allows us to see the bigger picture."
Lacoe explained the Sheriff's Office figures included smaller samples of data on a quarterly basis, and didn't look deeper at the city's pretrial cases from start to finish.
Yet, Hollings even admitted to board members quarterly reports aren't the most accurate.
"If people are in the program for longer than a quarter of the year, the quarterly reporting doesn't necessarily absolutely reflect what's going on for a longer period of time," Hollings said during a committee meeting in June. "I think the time of a person's full tenure is a good way to do that."
The California Policy Lab added there is trade offs in quarterly reporting. For example, there is a faster turnaround in the analysis of the data versus the PSA validation study that resulted in months of work.
SURVEY: San Franciscans concerned for public safety, 70% say quality of life has gotten worse
According to Lacoe's team, the quarterly rates measure a new filed charge and are used in regular reporting due to the fact they are the fastest turnaround metrics. For example, in the PSA validation report that analyzed data using any new arrest from 2016 to 2019, it's easier for researchers to assess a full outcome over longer periods of time. Whereas, the quarterly rates provide more of a recent view of pretrial cases. Subsequently, a portion of the PSA validation study included people who were released but were also not being supervised by the SF Pretrial Diversion Program.
It's important to note, the PSA validation study also includes arrests that occur throughout the state, based on data requested from the California Dept. of Justice.
The ABC7 I-Team reached out to David Mauroff, the CEO of the city's Pretrial Diversion Program, for further comment on the study's findings.
We received the following statement from his office:
"Although there is overlap between the population included in this study and those referred to SF Pretrial's release programs, the study captures a broader sample and individuals who were released through a number of different methods. Additionally, SF Pretrial supports programs that were not included in the sample based on the study's purview, which pertained to SF Pretrial's release programs only, and did not include its numerous diversion and in-custody programs.
As an organization that values social justice and equity, we are committed to using research and empirics to advance the state of pretrial justice in San Francisco. The validation study provides one opportunity to evaluate San Francisco's progress to date and how local justice partners can understand possible improvements."
But, the Sheriff's Office confirmed to the I-Team nearly 70 percent of the cases involved in this study were released to San Francisco's Pretrial Diversion Program, indicating much of the problem stems locally.
And as Supervisor Stefani fears, it's only getting worse.
"People are reoffending at a rate that seems undeniable and this report shows us that's true," she said.
The report also shows in nearly 30% of cases, judges released individuals against the recommendation of the public safety assessment. (edited)