This review of the literature discusses 15 studies that have been conducted on the effect of computer-based instruction on the academic achievement of at-risk students in elementary schools in the United States. The at-risk population consists of those students who are in the free- and reduced-school lunch programs, those students who are non-white, those students who are in Title I schools, and those students who have scored in the lowest quartile on academic achievement tests. The studies fall into the categories of mathematics, reading, telecommunications, Integrated Learning Systems, and higher order thinking skills. The study on mathematics was inconclusive. The majority of the six studies on reading were significantly positive. The two telecommunications studies were mixed-one positive for the treatment group and one positive for the control group. The four studies on Integrated Learning Systems mostly showed significant gains. The higher order thinking skills study showed positive significance in one category. (Contains 16 references.) (Author/AEF)