Elements of an evaluative study include the problem statement, the type of evaluation (i.e., impact evaluation), and data-gathering techniques. If the evaluation compares two sets of data or two sample groups, it must determine the validity of the basic data. Comparisons fall into two types: experimental and quasi-experimental. Experimental comparisons include random assignment and control groups, while quasi-experimental comparisons may include less sophisticated techniques. The agency administrator must explain to the researcher exactly what information is needed for the evaluation and its specific purpose. The discussion participants include a research associate at the Pretrial Services Resource Center and a director of agency pretrial services.