This is a longitudinal, comparative evaluation of the therapeutic effectiveness of hospital hostels for the long term mentally ill. The empirical work was in two parts. The first consisted of a pilot operation in which the study's methods and instruments were tested at a single hospital hostel. The second consisted of the use of those methods and instruments, in ways which were informed by the pilot experience, at three other hospital hostels, over a period of twelve months. The methods used included both quantitative and qualitative techniques, some of which (viz., four questionnaires) were wholly devised by the author. The primary aim of the study was to identify the probable key factor(s) involved in the effectiveness, or otherwise, of hospital hostels, with a view to, if necessary, recommending their adjustment in the interest of the optimisation of the effectiveness of the hospital hostel. The study's hypothesis, that residents of a hospital hostel with a lower level of restrictiveness than others would tend to achieve better social and clinical outcomes than residents of those other hostels, was confirmed. However, the most crucial factor for residents' outcomes was not the level of restrictivenes per se, but the appropriateness of the match, or 'fit', between it and their overall level of disablement. The appellation long stay tended to distort in a negative direction the staff's perceptions and judgements of their hostel's resident group's overall level of disablement, leading them to adopt a more restrictive approach than was necessary, desirable or appropriate. Inter alia, it is recommended that the long stay label be proscribed, at least in respect of hospital hostels, and that all such facilities be known in future as rehabilitation hostels.