Impact evaluations are one type of outcome evaluation available to a re-entry initiative. These assessments consider both short- and long-term effects of a given program and enable program administrators to assess the types of released prisoners that are most likely to benefit from a re-entry program and why, as well as to quantify the specific benefits achieved by the program. Impact evaluations require specific data collection systems that track each client as he or she progresses through the program, as well as data about program activities and services offered.

An experimental or random sample evaluation is the ideal design for measuring a program's impact. It compares the outcomes of a "treatment" group--persons designated to receive particular services that are designed to achieve specific outcomes--and a "control" group, who receive no services. This design, however, can be impractical because it requires program administrators to randomly select some individuals to be in a treatment group and receive the program services while placing other individuals in a control group and depriving them of services. Experimental designs can also be costly, requiring significant staff resources to track treatment and control groups over time. It is thus not surprising that experimental evaluation designs are the exception rather than the rule in criminal justice research. State and local policymakers and practitioners are nonetheless encouraged to make use of random sample selection whenever possible.

When random sample selection is not feasible, there are a variety of research designs that may be substituted without seriously compromising the utility of the research findings. One alternative is the quasi-experimental design, for which a comparison group, comprised of persons who are similar to program participants (but not randomly selected from the same pool), is compared to a group of program participants. For example, a group of male prisoners from one state who were convicted of drug offenses and received inpatient drug treatment might be compared with a similar cohort from another state who did not receive treatment. Another impact evaluation design is a before and after comparison, which compares the situation of participants prior to their participation in the program with their situation after participation.