Abstract: This research explored the effect of material difficulty and study times on the absolute accuracy of immediate and delayed Judgment of Learning (JOL). The design of this experiment was 5(learning times)×2(material difficulty), and every participant joined all conditions. All items which composed by 24 paired easy items and 24 paired difficult items were assigned separately in the immediate and delayed conditions. The difficulty of all items was evaluated by passing rate. 45 participants who were undergraduates attended this experiment and studied all the items for five times. We adopted the classical paradigm of the judgment of learning: study-judgment-recall. All responses which were grades of judgment of learning and memory performance were recorded by computer, while absolute accuracy of JOL was evaluated as indexes. We collected 37 effective data in the end. The results were as follows:(1) In the immediate-JOL condition, the Underconfidence-With-Practice effect appeared both in easy and difficult items;(2) In the delayed Judgment of Learning condition, there was no Underconfidence-With-Practice effect in any items, but there were no bias in easy items and Overconfidence-With-Practice effect in difficult items;(3) Both Underconfidence-With-Practice and Overconfidence-With-Practice effects disappeared gradually following the repeating times.

Key words: judgment of learning, Underconfidence-With-Practice effect, Overconfidence-With-Practice effect