Detention Overrides Can Become a Rule-Eating Exception

Detention risk assessments are hardly a sexy subject. But in the juvenile justice system, they are in fact a moral declaration of fairness.

It is a reflection of a system thinking about what factors contribute to a risky enough situation that detaining a youth before trial is necessary. It says, “We will detain or commit youth, but only if the data tells us it is necessary.“

That is, of course, unless decision makers choose to “override” the risk assessment instrument (RAI) score, supplanting the data-driven conclusion for their own instinct in a given case. The vast, vast majority of such decisions fall in the “override-up” bucket, meaning the youth received a more severe restriction than his RAI score would suggest.

A recent panel discussion, held at the Annie E. Casey Foundation’s Juvenile Detention Alternative Initiative (JDAI) conference, revealed how easy it is for overrides to become a rule-eating exception if nobody has an eye on things. In one jurisdiction, nine out of ten screened youths were detained even though less than one in ten scored in a risk category that warrants detention.

Nate Balis, who leads Casey’s Juvenile Justice Strategy Group, warned attendees at the conference against the “danger of complacency” when it came to de-incarceration efforts.

The current political climate favors the movement for moral and fiscal reasons, but “the narrative can change in a heartbeat,” Balis said. “We might squander a chance by setting the bar too low.”

Youth Services Insider was provided an example of what Balis fears at a panel discussion later that morning entitled “Overrides: The Dirty Little Secret of Objective Admission Decisions.”

It was a discussion about risk assessment instruments and detention decisions. The former must influence the latter, according to the playbook for detention reform developed by Casey through JDAI.

But in two states whose leaders shared their history at the panel, probation and court staff frequently ignored low- and moderate-risk scores, opting to “override” the score and detain a youth anyway.

“I’ve assisted several sites and I’m still surprised at how often I’m stopped in my tracks,” said Kelly Dedel, a consultant for the foundation. The RAI “looks good on the surface, but then you dig in and see that the risk score is not driving decisions because of override factors.”

Much of the decline in detention and incarceration has been caused by the fact that there are simply fewer juveniles being arrested (see pages six and seven of this report), thus creating a much smaller pool of kids who even could be locked up. That shift is attributable to some combination of overall behavior and law enforcement priorities and strategies.

Actual reform of the detention process comes from intentional efforts to control the flow of juveniles into either pre-trial detention facilities or post-adjudication placements. And risk assessment instruments are really the fundamental driver of that effort.

Most RAI use available data to produce a “score” that indicates the likelihood that a juvenile would endanger people if released, or that he would fail to appear in court. They also gauge the juvenile’s needs (mental health problems or special education) and potential deficits (drug use, being a runaway).

JDAI, she said, “places a premium on these decisions being based on research and data.”

A reasonable number of override decisions is not necessarily a bad thing, and in fact might suggest the presence of thoughtful managers. After a discussion about risk instruments at last year’s JDAI conference, we asked veteran juvenile justice expert David Steinhart what a rational override percentage would be for an RAI. In other words, is there a sweet spot where you’d know that a RAI was being relied upon, but not blindly?

His estimate: 15 to 20 percent.

In one county Dedel recently studied, she said the RAI was administered to 4,726 youths. Just seven percent of them scored as “high risk,” the classification that normally leads to detention.

The actual percentage of those 4,726 youths who were detained? Ninety percent.

“It’s baffling,” said Dedel. “They’ve spent a lot of time building them, picking known risk factors for it, and they’re not using them.”

In 2014, the state RAI for Missouri identified 5,392 youths for either release or an alternative-to-detention program. But 16 percent of those youths were detained anyway.

But that isn’t what worries Jay Rodieck, Missouri’s juvenile coordinator for the Office of State Courts Administrator. It’s the fact that in 18 of his counties, the “override up” rate is over 30 percent. The highest override-up rate for 2014 was 85 percent, he said at the panel discussion.

In Nebraska, 56 percent of the detained youths in 2014 landed there as a result of an override up. Nearly 250 of them had actually scored for pre-trial release without restriction of any kind.

So what causes these kinds of high override rates? In Dedel’s experience, there are a few big culprits:

  • Mistrust: Probation and/or court staff do not support the RAI, or do not have confidence in the system’s alternatives to detention.
  • Safety: Specific threats related to the victim of a crime.
  • Policy: Automatic exceptions that exist either in practice or protocol that mandate detention regardless of the RAI score.

That last one is the most pervasive, according to Dedel. It also might be the most fixable.

One of Missouri’s big outliers was St. Louis City, where the override rate was 76 percent in 2013. By 2014, the county had knocked it down to 57 percent, and the rate thus far for 2015 is 31 percent.

This was accomplished mostly by removing some of the policy overrides in St. Louis, explained Cathy Horejes, chief deputy juvenile officer for the 22nd Judicial Circuit. The biggest exception was a mandate that if a parent or guardian would not pick up the youth, they went to detention.

Sometimes there just wasn’t an adult available to get the youth, said Horejes. Other times, “parents were cool with detention so they wouldn’t come get them.”

Neither reason works anymore in St. Louis. “We started taking the kids home,” said Horejes.

Youth Services Insider has no clue how many of the JDAI sites, or any other systems relying on RAI, are bothering to check for aggregate or derivative override rates.

Kudos to Missouri and Nebraska for identifying a pattern and trying to get a handle on things, because any jurisdiction with these kinds of override rates is really just perpetuating a mirage of reform. It’s like building a drawbridge to a castle, then putting a regular bridge next to it that anyone could use.

Sure, overall detention numbers might be going down because there are fewer kids getting arrested. But has the proportion of juveniles being detained changed because of anything done on purpose?

Not really, if a system is passing on options below detention when its own RAI recommends that. If and when juvenile arrests go back up, said system will see its detention rolls go right back up.

Print Friendly, PDF & Email

John Kelly, Editor in Chief, The Chronicle of Social Change
About John Kelly, Editor in Chief, The Chronicle of Social Change 1204 Articles
John Kelly is editor-in-chief of The Chronicle of Social Change. Reach him at