Sixteen male prisoners participated in the study. Participants ranged in age from 30 to 70 (M = 43). Participants’ demographic information was collected but not approved for dissemination. Exclusion criteria included individuals under the age of 18, documented mental neurocognitive or neurodevelopmental disorders, and individuals who had less than one year left on their sentences. As inmates serving sentences, all participants were convicted of one or more crimes. In order of prevalence, the most common convictions included theft, possession or trafficking of narcotics, robbery, receiving stolen goods, assault, battery, murder, property destruction, counterfeiting public seals or documents, and use of counterfeited seals or documents.

All individuals were incarcerated in a maximum-security detention treatment center throughout the study. The study did not modify the prison’s existing daily schedule. All inmates in the prison were allowed access to various activities from 8:00 am to 8:00 pm, including job skills, leisure, and meals. Leisure and job skills varied depending on the availability of instructors, and they included activities such as theater, yoga, and cooking, among others. All prisoners were locked in a cell between 8:00 pm and 8:00 am. Before data collection, all procedures were approved by the Dipartimento dell’amministrazione penitenziaria, Roma. Participants were recruited in person after approval. Informed consent was obtained from all participants prior to data collection.

The project was designed to target the most common and detrimental problem behaviors exhibited by the inmates. Potential targets were identified by asking the prison staff what were the most problematic in terms of frequency of occurrence, injuries to inmates, and disruption to the daily routine of the prison. The final list of problem behaviors tracked for all participants included property destruction, self-injury, and aggression toward others. Property destruction was defined as throwing, disassembling, or breaking any item. Common examples of property destruction included breaking lights and throwing furniture. Self-injury was defined as any attempt to harm their own person by any means. The most common examples of self-injury included cutting and burning with a lighter. Aggression toward others included any attempt to harm another person and most commonly involved striking, kicking, and attempting to choke other prisoners or staff members.

The prison staff collected data by recording whether or not each participant engaged in any target behaviors daily. The data were then aggregated per day and presented as the number of participants in each group who engaged in aggression, self-injury, and property destruction. Participants also collected data during the intervention but only the prison staff member’s data were graphed.

A second independent observer collected data on 75% of the total days. The days evaluated were randomly selected across all conditions. The independent data collector shadowed the prison staff as the participants followed their daily schedule and recorded data without interacting with the prison staff or prisoners. Interobserver agreement (IOA) was completed by comparing the data sheets of the prison staff members and the secondary observer. An agreement was reached if the prison staff and the independent observer recorded the same number of problem behaviors across all participants for a given day. If there were any discrepancies between the count, then a disagreement for that day was recorded. IOA scores were calculated by dividing the number of days with agreements by the total number of agreements plus disagreements multiplied by 100. Total IOA was 100%.

We used a concurrent multiple baseline design to evaluate the effectiveness of the intervention. The participants were distributed randomly and evenly into four groups. Each group first experienced a baseline condition and was then sequentially exposed to the DRO and self-monitoring intervention.

During baseline, there were no programmed changes to the daily routine in the prison while prison staff collected data on the targeted behaviors. The four groups of four participants remained in baseline until they were sequentially introduced to the intervention.

The differential reinforcement of other (DRO) behavior intervention and self-monitoring consisted of earning a token each day if the participants engaged in any behavior aside from those targeted (i.e., property destruction, aggression, or self-injury). A DRO procedure was selected because the prison warden approved the procedure, the intervention was relatively easy to implement, and past studies demonstrated the effectiveness of DRO procedures in reducing challenging behaviors with other populations (Wilder et al., 2023).

At the beginning of the intervention, each prisoner was presented with a contract describing the intervention procedures. Participants were given a token board that they kept in their cell, on which they were instructed to track their behavior. Tokens were blue stickers with a smiley face and were delivered daily by the prison staff after the prisoners returned to their cells for the evening. After the prisoners returned to their cells, the staff asked them to report whether they had engaged in any targeted responses during the day. The staff praised accurate responses and corrected prisoners who misreported. The tokens were then cashed in weekly on Saturdays.

Each week on Saturday, the participants had the opportunity to exchange seven tokens for extra edibles (e.g., pastries). If a participant did not earn seven tokens, the participant did not receive an edible pastry, and the rule about needing to earn a token each day was restated. If participants earned a token every day during the intervention, they could place one additional phone call to a family member and a small party.

The maintenance phase, which occurred when the intervention was discontinued, was identical to the baseline. It was conducted for all participants from day 41 to day 43.

A social validity assessment was delivered before data collection and at the end of the project. The lead author administered the social validity survey questions to each participant. The first author read the questions to each participant individually and recorded their answers. The questions were presented on a Likert scale from 1 to 5, in which the participants ranked how much they agreed with the statement (1 = not at all to 5 = completely agreed). The questions included and the results of the evaluations are presented in Table 1.