This project focuses on pre-intervention risk factors and specific mechanisms of action which operate during the delivery of an empirically validated prevention intervention program (Functional Family Therapy: FFT) for high risk, indicated youth. Direct measurement of such mechanisms and the relationships between them, can enhance our understanding of the active ingredients of change. Of particular interest are processes that operate on a moment-by-moment basis and which predict if not mediate program outcomes. The factors studied consist of multi-source, multi-systemic measures in four domains: 1) Pre-intervention characteristics of the indicated youth and parent figure(s), family functioning, and demographic variables; 2) Program characteristics, i.e., the competence with which program content is delivered and its' relationship to interventionist experience; 3) Youth and parent immediate response to intervention content and process during the critical first phase of program delivery; and 4) Two levels of Program Outcome: Dropout versus program completion (program """"""""dosage"""""""") and long-term outcome (one-year post intervention measures of youth targeted behaviors in drug and related behavioral realms). Program retention receives particular attention as a NIDA priority (NIDA, NIH, 1999 Publication #99-4180, p.3) which represents the """"""""...variable with the most consistent relationship to positive outcome"""""""" (Williams & Chang et al., 2000). Analyses address two Specific Aims:
Aims 1 a-d identify the relationships among the pre- intervention characteristics of youth, families, and interventionists, and their association with pre-intervention official criminal records and outcomes. These relationships will be examined using structural equations models with a sample of 480 families.
Aims 2 a-d examine mediating effects of intervention process on outcomes. Measures of intervention content and process include independent observer rating and coding of actual program delivery, allowing for the identification of immediate (microsequential) and longer transactional patterns. Analyses for Aim 2 use multi-group (engaged versus non-engaged) growth curve modeling preocedure with a sample of 240 families.
