Memory Representations, Tree Structures, and Parameter Polysemy: Comment on Cooper, Greve, and Henson (2017)

Publication date: Available online 2 June 2017
Source:Cortex
Author(s): David Kellen, Henrik Singmann
Cooper, Greve, and Henson (2017) discussed the use of different approaches for measuring item and source memory, and how choices among these can affect the comparison between different groups (e.g., younger versus older adults). The authors argue that the tree structure adopted in the specification of item- and source-memory retrieval in multinomial processing tree models implies a theoretical commitment to the way memories are represented. According to the authors, this commitment can affect the conclusions that are taken from the results and produce different model fits in experimental designs involving confidence-rating judgments. Reported model fits suggest that an alternative tree structure provides a superior account of the data. The present comment argues that the particular tree structure used does not enforce any commitment to a particular structure, as long as the trees are well defined in the sense that parameters are monosemic across the entire structure of the model. The different fit results reported by CGH are due to their reliance on tree structures that do not ensure that all parameters are monosemic.



http://ift.tt/2rRGI6g