Monthly incarceration and disposition data from the restitution sites from the period January 1977 through December 1979 were analyzed with the interrupted stochastic difference equation model. For each of the five sites evaluated, a similar series of analysis steps were followed. First, the dependent series (incarcerations) were plotted and visually inspected with special attention given to overall trends, outliers, possible discontinuities in series behavior coincidental with restitution project startup, and any other obvious series characteristics. Next, the autocorrelation and partial autocorrelation functions for the series were estimated and displayed in correlograms. Examination of these correlograms provided indications as to the nature of possible time dependence in the incarceration series. Following this nonstructural inspection of series characteristics, procedures were followed which led to intervention model specification, estimation, diagnosis, and interpretation. After the determination of appropriate incarceration/disposition specification, all intervention model variables were entered into a stepwise regression procedure. Results of the assessment for each site are presented. Three sites provided clear indications of a downward trend in incarceration trends associated with the establishment of their restitution projects, one site showed an increase in both incarceration level and trend coincidental with project startup, and one site displayed an immediate increase in the number of youths incarcerated combined with a strong decrease in the incarceration trend subsequent to the beginning of the restitution project. Tabular data and 30 references are included.