Can Differentiation Adequately Account for the Influence of Word Type on Episodic Recognition Memory? Ii Can Differentiation Adequately Account for the Influence of Word Type on Episodic Recognition Memory?

iii Certification I certify that this is my original work and any text that is not my own has been quoted and attributed appropriately in the references. I also declare that I am the student whose name appears below and that this text has not been previously submitted for assessment. I certify that I have made and retained a copy of this thesis. _________________________________ Kimberley McFarlane iv Acknowledgements First and foremost, I would like to thank my supervisor, Professor Michael Humphreys, for taking the time to share his expert knowledge of this interesting field with me. His guidance has been invaluable this year. I would also like to thank Doctor Angela Maguire for her assistance, support, and encouragement. In addition, it would be remiss of me not to thank Doctor Amy Criss of Syracuse University for allowing me to use and modify her simulation code for the purposes of this paper. In relation to this, thanks are also extended to Doctor Scott Bolland for providing verbal feedback on the code modifications that I made. Finally, I would like to thank Kimberley Harper for sharing the workload in testing participants. v Abstract In episodic recognition memory, differentiation is the assumption that a study item's pre-existing memory trace is updated when additional study for that item is provided. The differentiation models commonly suppose that episodic memory encoding conforms to this process. Although these models have received considerable support within the literature, results inconsistent with their predictions have also been found. The present paper examined conflicting findings that resulted from study list strength manipulations with rhyming word stimuli and semantically related stimuli. As part of the investigation into this discrepancy, 79 university students participated in a computer-based recognition memory task. In this task, word categories of varying length (short vs. long) and word type (rhyming vs. taxonomic) were presented either five times or once within a mixed study list. Following study, an old-new response paradigm was used to examine recognition memory performance. Results from both the rhyming and taxonomic category stimuli were largely consistent with the previous findings in the literature, indicating that word type does appear to influence recognition memory, even within a mixed study list. These findings are interpreted primarily in terms of word type similarity predictions made by one of the differentiation models. Other possible explanations are also discussed.