Study 1
Heller and colleagues (2015) conducted a randomized controlled trial at the Cook County (Ill.) Juvenile Temporary Detention Center (JTDC) to determine if a cognitive intervention influenced both the probability of readmission to detention and the number of readmissions for participating youths across an 18-month follow-up period. The sample included 2,693 male residents of the Cook County JTDC from 2009–2011. Data was collected via intake forms that provided basic demographics and addresses; admissions logs, which the admissions staff used to record who enters the facility each day; and the JTDC’s housing roster, which captured the residential unit in which a youth was located on each day to measure receipt of treatment. The outcome of interest, recidivism, was measured by re-entry into the JTDC facility. Youths’ data was also linked to other arrest databases to assess this measure.
Male youths entering the facility were randomly assigned into one of four CBT residential units or to one of four “status quo” units, which did not receive any treatment. However, as there were systemic reasons that some study youths could not comply with random assignment, the JTDC staff members were allowed to disregard random assignment when necessary to ensure the safety and well-being of staff and youth residents. For example, re-admitted youths who had already been assigned to a CBT unit were generally placed in this unit again regardless of random assignment. Additionally, those who had a history of conflict with others or were physically, mentally, or emotionally immature would not be placed in the treatment group.
The average age for the treatment group (n=1,371) was 16 years. The treatment sample was 84 percent Black,13 percent white, 3 percent Hispanic, and 1 percent other race/ethnicity (participants could choose more than one race/ethnicity). Based on administrative arrest records, 30 percent of the treatment sample entered the detention center during the study directly, without an associated arrest; 18 percent entered the detention center because of an arrest for a violent crime; 9 percent for a property crime arrest; 7 percent for a drug crime arrest; and the majority (35 percent) for some other crime arrest (the remaining 1 percent of the data was missing for this item). The average age for the control group (n=1,322) was 16 years.
The control sample was 83 percent Black, 14 percent white, 3 percent Hispanic, and 1 percent other race/ethnicity (participants could choose more than one race/ethnicity). Thirty percent of the control group entered the detention center without an associated arrest, 17 percent entered the detention center because of an arrest for a violent crime, 10 percent were arrested for a property crime, and 8 percent were arrested for a drug crime.
“Spell numbers” counted how many juvenile detention admissions the youths had, up to and including the focal spell. The average number for both groups was 3 spells, and no significant differences were found between groups on any baseline characteristics. An intent-to-treat analysis was conducted using regression with a linear probability model, and an instrumental variable analysis was conducted using random assignment as an instrument for participation. No subgroup findings were conducted.