This study investigated the effects of an inductive reasoning training program on children's performance on intelligence test tasks, the range of transfer, the long-term effects of training over 4 months, and the effectiveness of group training. Participating were 47 third-grade children of average ability. The 23 children randomly assigned to the training condition received a 3-week course in inductive reasoning in ten 30-minute sessions. The training program consisted of six forms of inductive reasoning tasks: (1) generalization; (2) discrimination; (3) cross-classification; (4) recognizing relations; (5) discriminating relations; and (6) system formation. Abstract material was used in 15 percent of the tasks. The remaining 85 percent consisted of either concrete material, such as blocks or picture and figure problems from the children's everyday life. Children were trained in groups of three or four, by six researchers who were not involved in administering the pre- or posttests. Control group classes completed the regular school curriculum. The Raven Coloured Progressive Matrices Test and arithmetic tasks were used as a pretest, immediate posttest, and posttest four months later. Results indicated that there was a significant, positive training effect on children's performance on inductive reasoning tasks. Far-far transfer was also observed, because the children were able to solve arithmetic problems involving relations between numbers and their common attributes which related to inductive reasoning, in which they had received no training. (Contains 43 references.) (Author/KB)