The effectiveness of three strategy training methods was compared with 15 learning disabled (LD) children of two age groups (9 to 10 and 11 to 12 years old) and 15 nonLD children in the same age group. The methods (practice, in which the Ss were to find the best way to remember; specific strategy method, in which the Ss used a specific cumulative rehearsal strategy; and a metamemory method, in which Ss were trained in a self instructional, self monitoring strategy) were compared for effectiveness in increasing efficiency of verbal rehearsal strategies during actual training (immediate effects), on a delayed test of the task in which training occurred (maintenance effects), and on a delayed test on a nontrained task (generalization effects). Findings revealed that both LD groups employed rehearsal strategies, but not as efficiently as same aged controls. There appeared to be a lag in the development of strategies in LD Ss. Training methods did not differ in effectiveness, but training resulted in more efficient strategy use during actual training trials and in better performance on the nontrained task for all Ss. Metamemory (the individual's knowledge of and awareness of memory) knowledge on task and strategy variables did not differ between LD and nonLD Ss, but older LD Ss did not have as much confidence in their memory abilities as older normal and younger LD Ss. Younger LD Ss were not as aware of the need for strategy implementation or of other tasks when rehearsal might be an efficient strategy, indicating some delay in development of metamemory understanding. (Author/CL)