Testing effects of free recall on organization in whole/part and part/whole transfer

Testing of to-be-learned material facilitates subsequent learning of new material. We investigated this forward effect of testing in two experiments using the whole/part and part/whole transfer paradigms with categorized word lists. Learning was assessed for recall of individual words, higher order categories, and category clustering. In each experiment participants learned two lists in which the number of tests on the first list was varied. The first list contained either twice as many items as the second list (whole/part paradigm) or half as many items as the second list (part/whole paradigm). In the experimental condition, the part list contained half the items of the whole list. In the control condition, the two lists were unique. In the whole/part paradigm, learning of the part list was poorer in the experimental than in the control condition. Although testing during whole list learning facilitated learning of the part list, it did not moderate the negative transfer effect. In the part/whole paradigm, learning of the whole list was better in the experimental than in the control condition, and this positive transfer effect was strengthened by repeated testing of the part list. The findings are discussed in the context of discrimination and encoding explanations of the forward effect of testing.