Three conclusions were drawn about within-class grouping after reviewing the work of several researchers on the following topics: teacher and student behavior, academic achievement, perceptions about reading groups, social class membership and reading group assignment, and the role of group cohesiveness. The conclusions were that: (1) although the evidence is far from conclusive, justification can be found both for grouping or for not grouping (whole class instruction); (2) when instructional reading level conditions are met, no reliable and/or significant differences are found for grouping over whole-class instruction; and (3) the low reading group has numerous psychological and social strikes against it, but there is no strong evidence that these decrease or limit the reading achievement of the group members. A few educators argue that grouping is not strictly a school matter but that it contributes to an undesirable caste system within American society. (Twenty references are attached.) (MG)