The aim of the research is to examine haw pupils behave in small group discussions, what pupils gain and which pupils benefit most. A distinction is made between 'grouping' and 'groupwork', and the origins, claims and related research evidence for the latter are examined. The actual use of groupwork in selected Midlands schools is reported. This is based partly on the findings of an S.S.R.C. research programme (ORACLE), on which the researcher was employed. Reasons for the discrepancy between theory and practice are suggested. Empirical investigations involved developing a series of four Tasks, which were administered to 24 groups. The 96 discussion sessions were recorded and analysed, using a specially designed category system. Results showed the importance of 'questioning' and 'challenging' to the progress of the discussions. Analysis also revealed the high level performance of low-achieving boys' groups, the equal participation of members of mixed ability groups, the consensus orientation of girls' groups, and also the severe problems experienced in mixed sex groups.