Measuring Restriction of Living Environments for Children and Youth

The Restrictiveness of Living Environment Scale (ROLES) has served since 1992 as the primary way of conceptualizing the “restrictiveness” of a child's living situation. It is a key performance dimension used by many mental health, child welfare and juvenile justice agencies in helping children move to and function in less restrictive settings. However, changes in youth-serving systems created a pressing need to revise the ROLES. The ROLES ordered list of environments is neither mutually exclusive nor exhaustive, and the level of restrictiveness for any type of treatment setting varies widely going from program to program or youth to youth. The objective of this study was to revise the ROLES using an approach that would reduce the measure's shortcomings while retaining its simplicity. Specifically, the goals of this study were to 1) create a conceptually-based reliable and valid measure that allows for precise measurement of environmental restrictiveness; and 2) provide an empirical basis for a simplified typology of restrictiveness that is intuitive and flexible.

The Restrictiveness Evaluation Measure for Youth (REM-Y) and the Restrictiveness Evaluation Measure for Self (REM-S) were created using a process of expert review, cognitive interviewing, review by experts in culture, review by alumni of foster care and pilot testing. The REM-Y/S was emailed to over a thousand child organizations and to older youth and completed on 602 youth. Rasch Measurement Model analysis allowed for description of the item and response option calibrations of the REM-Y/S, including the differential item function (DIF) across placement settings and across child and youth characteristics as well as the transformation of ordinal into interval data.

Cluster analysis of the 602 responses produced a four-cluster solution which translates to descriptive general environment types of “low,” “moderate,” “elevated,” and “high.” Low restriction environments have few or no limitations on what youth can do, where they can go in the community and environment, and who they can be with and for how long. Supervised or unsupervised independent living, living with family or friends, and adoptive homes are typical of this cluster. Moderate restriction environments are ones in which there are a few limitations on what the youth can do. Parents' home for young children, regular foster care, dormitories, and Job Corps are typical for this cluster. Elevated restriction environments have more far reaching restrictions, such as time limits for communication and internet access, and interaction with friends. Treatment foster care, group homes, and less intensive residential treatment are typical of this cluster. High restriction environments have the greatest limitations on what the youth can do, where they can go in the community and environment, and who they can be with and for how long. Most residential treatment centers, wilderness camps, drug and alcohol rehabilitation facilities, youth correction/detention centers and emergency shelters are typical of this cluster.

The results also show that there is a degree of overlap among clusters; suggesting that when a more precise measure of restrictiveness is required, use of the REM-Y is warranted rather than simply using the General Environment Types.