摘要: Testing often enhances memory, but memory can be harmed by interference from similar or competing items. This thesis examines two situations in which it has been proposed that testing harmful to because the test itself increases susceptibility interference. Experiments 1-8 investigate effect of generating errors during new learning. Participants learned definitions for unfamiliar English words, translations foreign vocabulary, either a response and being given corrective feedback, reading word its definition, selecting choice followed feedback. In final all feedback led significantly better correct definition than making incorrect choices, suggesting benefits generation are not restricted correctly generated Even when information is novel, errorful may play powerful role potentiating encoding Metacognitive judgments learning revealed participants were strikingly unaware this benefit, judging less effective method choosing, fact was better. Predictions reflected participants’ subjective experience A second series experiments (Experiments 9-10) claim reactivating consolidated destabilizes it, more susceptible English-Swahili pairs (List 1) on Day 1 with 3. When List reactivated form reminder immediately before Finnish words 2) 2, testing, far impairing enhanced revealing effect. Furthermore, 2 disrupted there no test, immunized against