The evidence from studies of component selection in paired-associate learning is reviewed and then considered in relation to Martin's (1968) hypothesis of encoding variability and Rudy's (1974) model of variations in the associative process. Component-selection tasks are also compared to verbal concept-formation tasks. (Editor)

This study attempted to answer two questions: should variable or constant grapheme-phoneme correspondence be introduced initially and should paired-associate or conceptual training be used in initial training. (Author/RK)