The report describes the historical development of halfway houses, their current operations, and a framework for reconciling theory and operations. The study maintains that few methodologically sound evaluations of halfway houses have been completed because of the use of insensitive outcome measures and vague program goals and objectives. A review of existing evaluations suggests some conclusions about halfway houses. Houses are as effective in preventing criminal behavior in the community as alternatives that involve community release. The placement of a halfway house in a community neither increases crime nor decreases property values. Houses assist their clients in locating employment but not necessarily maintaining it. Houses can provide for the basic needs of their clients as well as other forms of release. At full capacity, houses cost no more, and probably less, than incarceration, although they cost more than parole and outright release. The available capacity of halfway houses is only partially used at present, thus driving up actual per diem costs. Evaluations of halfway houses tend not to result in changes in house operations. Appended client flow narrative; evaluations surveyed by house affiliation, design type, and evaluating agency; bibliography of evaluations surveyed; and 4 figures