“…Gruber et al (1965) attribute the relative weakness of the effect to the tendency of control 5s to use mediators also, thereby attenuating any difference between their learning and that of instructed 5s. Such a view seems reasonable, given that almost all 5s, regardless of instructions, report using mediators (e.g., Schwartz et al, 1970). It implies, however, that when 5s enter the experiment, they already have learned to use a media-tional strategy.…”