Journal Article

Immediate List Recall as a Measure of Short-Term Episodic Memory: Insights from the Serial Position Effect and Item Response Theory

Brandon E. Gavett and Julie E. Horwitz

in Archives of Clinical Neuropsychology

Published on behalf of The National Academy of Neuropsychology

Volume 27, issue 2, pages 125-135
Published in print March 2012 | ISSN: 0887-6177
Published online December 2011 | e-ISSN: 1873-5843 | DOI: http://dx.doi.org/10.1093/arclin/acr104
Immediate List Recall as a Measure of Short-Term Episodic Memory: Insights from the Serial Position Effect and Item Response Theory

More Like This

Show all results sharing these subjects:

  • Neuroscience
  • Neuropsychology

GO

Show Summary Details

Preview

The serial position effect shows that two interrelated cognitive processes underlie immediate recall of a supraspan word list. The current study used item response theory (IRT) methods to determine whether the serial position effect poses a threat to the construct validity of immediate list recall as a measure of verbal episodic memory. Archival data were obtained from a national sample of 4,212 volunteers aged 28–84 in the Midlife Development in the United States study. Telephone assessment yielded item-level data for a single immediate recall trial of the Rey Auditory Verbal Learning Test (RAVLT). Two parameter logistic IRT procedures were used to estimate item parameters and the Q1 statistic was used to evaluate item fit. A two-dimensional model better fit the data than a unidimensional model, supporting the notion that list recall is influenced by two underlying cognitive processes. IRT analyses revealed that 4 of the 15 RAVLT items (1, 12, 14, and 15) were misfit (p < .05). Item characteristic curves for items 14 and 15 decreased monotonically, implying an inverse relationship between the ability level and the probability of recall. Elimination of the four misfit items provided better fit to the data and met necessary IRT assumptions. Performance on a supraspan list learning test is influenced by multiple cognitive abilities; failure to account for the serial position of words decreases the construct validity of the test as a measure of episodic memory and may provide misleading results. IRT methods can ameliorate these problems and improve construct validity.

Keywords: Learning and memory; Assessment; Statistical methods

Journal Article.  6072 words.  Illustrated.

Subjects: Neuroscience ; Neuropsychology

Full text: subscription required

How to subscribe Recommend to my Librarian

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content.