Statistical methods used in the guidelines research identified information items related to sentencing decisions. (The project started as an effort to build on the pioneering work done on parole decisionmaking.) A 2-year feasibility study had concluded that simple methods could provide a workable decisionmaking aid for judges at sentencing and that judges were willing to cooperate in developing and using sentencing guidelines. Four counties were then chosen as pilot implementation sites for the newly developed methods of establishing sentencing policy. The volume analyzes each State's criminal code and the local courts' sentencing practices in each site. Project staff designed and pretested a coding manual and data collection instruments, and gathered and analyzed a construction sample of sentencing decisions to develop guideline models. Judges commented on early results; a validation sample tested the predictive efficiency of the models. A historical analysis of the construction and validation samples gave judges an explicit picture of previous sentencing decisions to use in determining future sentencing policy. The report acknowledges that, despite the uniformity of approach at each site, differences in sampling, data collection, analysis, modeling procedures, and research strategies occurred. Tabular data are given; study instruments and sentencing grids are appended. For the remaining volumes, see NCJ 58874 and 82360.