Rowland, Christopher A., authorDeLosh, Edward, advisorRhodes, Matthew, committee memberDik, Bryan, committee memberRobinson, Daniel, committee member2015-08-282015-08-282015http://hdl.handle.net/10217/167240Despite a substantial literature describing the memory benefit resulting from testing (i.e., memory retrieval), relatively few investigations have attempted to detail how retrieval acts as a memory modifier. One core issue concerns the extent to which testing and studying effect fundamentally similar or different processes or components of memories. The present paper introduces two computational models, both based in REM theory (Shiffrin & Steyvers, 1997) and designed to provide a plausible basis for describing the testing effect at a more mechanistic level than existing theories. The two models are derived from the same set of core assumptions about the functioning of the memory system, and differ only in their specifications of the components of memories that are modified as a result of retrieval. The “Item Model” (IM) assumes that retrieval serves primarily to strengthen the target item content representation of information that is retrieved. In contrast, the “Context Model” (CM) assumes that retrieval serves to embed additional contextual information into the target memory trace, facilitating the subsequent ability of the memory system to locate such items. This manuscript provides coverage of relevant areas in the literature that have bearing on the IM and CM, details the implementation of the models and their larger framework, and reports on 4 experiments designed to test contrasting predictions of the IM and CM. Experiment 1 observed a testing effect using a mixed list, but not a pure list design, implying that testing may serve to enhance the search process by strengthening context information in memory. Experiments 2-4 were designed to examine the effects of reinstating contextual information during final testing on the testing effect. Experiments 2 and 3 found that reinstating either perceptual contextual elements (Exp. 2), or semantic context cues (Exp. 3) at the time of final test did not significantly impact the magnitude of the testing effect. However, Experiment 4 found that reinstating the initial learning mental/temporal context at the time of final test mitigated the magnitude of the testing effect. Potential nuanced interactions between testing and context in memory are discussed.born digitaldoctoral dissertationsengCopyright and other restrictions may apply. User is responsible for compliance with all applicable laws. For information about copyright law, please see https://libguides.colostate.edu/copyright.contextretrievaltesting effectmemorycognitiontestingText