Koriat, Sheffer, and Ma’ayan (2002) documented a phenomenon that they termed the underconfidence-with-practice (UWP) effect: When participants are presented with the same list of paired-associates for several study-test cycles, their JOLs exhibit relatively good calibration on the first study-test cycle, with a tendency for overconfidence. However, a shift toward marked underconfidence occurs from the second study-test cycle onward. The UWP effect was found to be very robust. Based on the memory for past test (MPT hypothesis), we examined the influence of the test experience and learning experience by inserting two judgments in learning and tests stages in two experiments: the second-order judgments (SOJs) and retrospective confidence judgments (RCJs). Second-order judgments (SOJs) pertain to an individual’s confidence in the JOLs themselves; the second judgment called retrospective confidence judgments (RCJs) ,which was used to access the correctness of past retrieval and often are assumed to arise from the retrievability of the to-be-judge item from memory. Participants were instructed that they would be learning 48 word pairs and making JOLs. JOLs were explained as judgment of learning based on what they thought were their chances for recalling the second word when given the first word during a memory test that would happen in a minutes. They were asked to use a scale from 0 to 100%. On trial 1 in Experiment 1, Participants completed the tasks in the order described below： a) Study. Each item was displayed individually for 3s to study. b) Judgments of learning. The participant judged the likelihood of correctly retrieving that item on a subsequent test of memory. c) Second-order judgments. A given JOL was followed immediately by an SOJ in Experiment1. d) Paired-associate recall. The cue was presented in an order of studying, and the participant was instructed to respond with the word that was originally paired with that cue. Experiment 2 was different from Experiments 1 in two aspects: First, SOJs was not used after JOL; Second, A given RCJ followed immediately the test. Participants were asked to rate confidence in the accuracy of each answer. They also used the scale from 0 to 100%. The results showed that the SOJs and RCJs did eliminate the UWP effect in two experiments and proved that the test experience and learning experience influenced the UWP effect indirectly. Such results indicated that UWP effect is affected by many factors, therefore, a single theory hypothesis can not explain all things of the effect.