This paper reviews evaluation reports on the use of integrated learning systems (ILSs) in elementary and middle schools throughout the United States in light of the following questions: Do these systems work? Do students make more progress using an ILS than learning in "traditional" ways? Could the cost of software and equipment be better spent on smaller class size, higher teacher salaries, more print materials, inservice training in classroom management, or on training for using other innovative instructional programs? The following studies on the implementation of ILS are summarized: (1) the WICAT ILS in individual schools and Chapter I and compensatory education programs; (2) the use of Computer Curriculum Corporation (CCC) in Chapter I pull-out programs in public and parochial schools, including a 4-year study of CCC use in the Los Angeles (California) schools; (3) Jostens Learning Corporation's Basic Skills implementation; (4) a field experiment using computers in fifth through eighth grade mathematics in 31 schools; and (5) integrated systems piloted in New York City using CCC, WICAT, ESC, Prescription Learning, Wasatch, Ideal, Plato, CNS, and others. The analyses show that results differ as much based on different conditions of the study as on the different software packages used; students generally do somewhat better than expected using ILSs and sometimes the results are substantially superior; and evaluation data are not sufficient for use by a school district as a basis for acquisition and investment decisions. Tables showing comparisons of the various studies are included. (ALF)