Familiarity is thought to emerge when there is a match between the features currently presented and the features stored in memory, and it could be affected by both restudying and testing. It is still not clear whether all types of features can lead to familiarity. Though previous studies identified some types of features that could produce familiarity in the absence of recall, it is inconclusive as to whether semantic information gave rise to the familiarity feeling in recognition. No study has yet examined whether semantic features of Chinese can produce familiarity in the absence of recall. Furthermore, another open question remains as to whether the effects of testing and restudying on recollection and familiarity are dissociative.

The RWCR paradigm provides a quantitative way to explore above questions. RWCR refers to the phenomenon that even when participants are unable to recall a studied item when cued at test, they can still reliably discriminate between cues that resemble studied items and cues that do not. In this research, we explored the RWCR effect in Chinese characters learning, and used RWCR to examine the effectiveness of testing and restudying in recognition.

In Experiment 1, the final recognition test was taken immediately after the learning phase. A total of 60 college students were engaged in a 2 (Study Status: studied vs. nonstudied) × 2 (Study Strategy Type: testing vs. repeated study) mixed design experiment, to study different influences of testing and restudying on effectiveness of recollection and familiarity. In Experiment 2, 41 college students participated in a mixed design experiment. The procedure was identical to that of experiment 1, except that the final recognition test was taken one week after learning.

The results showed that (1) when the final test was taken immediately after the study phase, both restudying and testing lead to better recollection performance than the studying-once encoding did, and there was an advantage of restudy encoding over test encoding. (2) Restudy encoding led to better familiarity performance relative to testing and study-once encoding, and there was no difference between test encoding and studying-once encoding. (3) When the final test was delayed for one week, there was no difference in the recollection performance after either restudy encoding or test encoding. (4) Familiarity performance in restudy encoding condition declined faster than that in the test encoding condition. (5) The semantics of Chinese could elicit the RWCR effect.

The results demonstrated that the initial testing increased recollection relative to studying words once, whereas it did not affect familiarity. In addition, recognition tasks showed better familiarity on restudied words than on tested words, and test could enhance long-term retention of the tested material. Further, the semantics of word would also elicit the RWCR effect in logographically scripted language (i.e., Chinese).