The testing effect in immediate recognition: tests of the episodic context account

Abstract

ABSTRACTThree experiments explored when testing produces immediate advantages over restudying in old/new recognition tests. According to the episodic context account, the study context is reinstated with every accurate retrieval during relearning, enabling accurate discrimination of the originally studied word from similar foils. Context plays a little role when the foils are extremely dissimilar, at least on an immediate test. We predicted that recognition in a plurality discrimination test would be better for tested than restudied items. Experiment 1 used one study-test cycle and produced no testing effects. Experiment 2 used two study-test cycles and produced a testing effect for plurality discrimination but not for new words. Experiment 3a–e systematically varied the number of study and test opportunities. We replicated Experiment 2 in Experiment 3a, but using fewer study or test cycles did not produce a testing advantage. The results are interpreted as supportive of the episodic context account.