Capturing additional information about the organization of entries in the lexicon from verbal fluency productions

Abstract

Troyer and colleagues [Troyer, A. K., Moscovitch, M., & Winocur, G. (1997). Clustering and switching as two components of verbal fluency: evidence from younger and older healthy adults. Neuropsychology, 11(1), 138–146] developed a seminal method to measure clustering and switching behaviors during verbal fluency (VF) productions. We sought to expand the reach of their system by modifying the scoring rules. Compared to the Troyer system, our modifications yield comparable estimates of interrater reliability and similar patterns of correlation with demographic characteristics for both clustering and switching in healthy adults. However, two objective measures of word relatedness (interword interval timing and latent semantic analysis) confirm that our revisions capture additional information about the organization of entries in the lexical network.

Keywords:

• Verbal fluency
• Semantic network
• Lexical retrieval
• Language
• Neuropsychology
• Cognitive psychology.

Funding

This research was supported by National Institute of Mental Health (NIMH) [grant number MH60504] [grant number MH077852] (D.J.S.); the Therapeutic Cognitive Neuroscience Fund (B.G.); and the Benjamin and Adith Miller Family Endowment on Aging, Alzheimer’s, and Autism Research (B.G.).

Conflict of interest

Under an agreement with Psychological Assessment Resources, Inc., D. J. Schretlen and T. D. Vannorsdall are entitled to a share of royalty on sales of the Calibrated Ideational Fluency Assessment test used in the study described in this article. The terms of this arrangement are being managed by the Johns Hopkins University in accordance with its conflict of interest policies.

Notes

¹ By defining clusters as multiword strings, our scoring differs from that of Troyer et al. (1997), which defines both strings of related words and single words that are unrelated to their neighbors as clusters. The Troyer system yields numerous single-word clusters. Theoretically, we accept the idea that a “cluster” could consist of only a single word. However, because one cannot measure interword intervals for fewer than two words, a single-word “cluster” cannot be corroborated by the measurement of interword intervals. Consequently, we restrict the scoring of clusters to strings of two or more successive words whose relationship is defined by one of our scoring rules.

² In our scoring system, clusters can be contiguous or separated by single, nonclustered words. Because we define a switch as any transition between clusters or nonclustered words, the number of switches is not perfectly correlated with the number of clusters. This overcomes a conceptual limitation of the Troyer et al. (1997) system (Mayr, 2002).