Abstract:
|
The 2014 Survey of Juveniles Charged in Adult Criminal Courts (SJCACC) will use a hybrid estimation strategy combining data or estimates obtained from a (partial) census, a probability sample, and in one or more instances, model-based or model-assisted estimation. A census of cases will be obtained from about 25 states which could provide statewide, electronic, case-level data files, representing roughly 50-60% of the population. The remaining states will be represented through a stratified, probability proportional to size sample of primary sampling units (PSUs) where case level data are obtained or collected locally. In one or more instances, the state's data availability circumstances allow for either a model-based or model-assisted estimation strategy as an alternative to the census or sample approach. Examples include calibrating statewide case-level data from a prior year to control totals available for the study reference year, and adjusting case-level data for a subset (probabilistic or otherwise) of counties to control totals for the entire state. The purpose of this paper is to present and discuss the overall approach for this hybrid estimation strategy.
|