An Analysis and Comparison of WMS-III and WMS-IV Verbal Paired Associates: Practical Implications for Neuropsychologists

Author

Anthony Paul Andrews, Nova Southeastern UniversityFollow

Date of Award

2019

Document Type

Dissertation

Degree Name

Doctor of Psychology (PhD)

Department

College of Psychology

First Advisor

Charles J. Golden

Second Advisor

Barry Nierenberg

Third Advisor

Barry Schneider

Keywords

assessment, memory, neuropsychology, Wechsler, WMS-III, WMS-IV

Abstract

The current study examined the performance of clinical outpatients on the Verbal Paired Associates (VPA) subtest from current and prior versions of the Wechsler Memory Scale (WMS). It was predicted that VPA from the WMS-III (VPA3) would be similar in agreement to the WMS-IV (VPA4) and that VPA4 would show a stronger relationship than VPA3 with intellectual abilities, sustained attention, and executive function abilities as assessed by the WAIS-IV; CPT-2; and the TMT, Category Test, and Stroop, respectively. Thirty-six adults were administered both the WMS-III and the WMS-IV, along with the other measures as part of a larger neuropsychological battery. Data were collected from an archival database of individuals clinically referred to an outpatient, university-based neuropsychology clinic.

Results of the current study showed that agreement for scaled scores was lower than expected for both VPA I and VPA II across versions of the WMS. Despite the relatively modest levels of strict agreement, current results did find that 89% of scaled score pairs fell within three scaled score points (i.e., one standard deviation) for both VPA3 and VPA4, and this was reflected in the magnitude of correlation coefficient, which was .76 in the predicted direction.

Results of the study also showed that VPA I and II across versions of the WMS-III and WMS-IV predict neuropsychological functioning similarly. Using the methodology described by Lee and Preacher (2013), direct comparisons found no significant differences between VPA3 or VPA4 in their ability to predict intellectual, attentional, or executive functioning abilities at the .01 level of prediction. The small sample size of the current study, the conservative alpha cut-off used, and the inherent cognitive heterogeneity inherent in a convenience sample of outpatient adults may have contributed to a lack of statistical power to detect real differences and masked otherwise significant differences between the tests.

Implications of the current study may be relevant for research and/or clinical applications. First, it is concerning that the degree of scaled score agreement is low across versions of VPA despite overall similar mean scores, and this low rate of agreement within each participant suggests that the tests themselves may be measuring memory in important, though poorly understood, ways. Overall, participants in the current study obtained identical scores across versions of the test less than half the time. One recommendation is to take caution when comparing results of VPA3 to VPA4 in serial assessments, such as with patients diagnosed with a neurodegenerative disease or litigants assessed by multiple practitioners to inform a trier of fact about a specific cognitive issue at stake in a matter being litigated. This will be become more relevant when WMS-5 is released in the next few years and most serial assessment will involve comparisons to VPA4.

Current results also support the movement towards more transparent and thorough comparison of normative data for clinical populations, along with a specific investigation into the rationale behind future changes to VPA, namely the psychometric approach vs. the theory-driven, neuroanatomically informed approach. For example, while adding semantically related, “easy” word pairs to VPA4 may have achieved a goal of normalizing mean scores on VPA4, it is unclear whether this psychometric victory represents meaningful changes in terms of how memory works in clinical populations, many of whom have been demonstrated to process information in a way dissimilar to non-clinical populations. The current research indicates the need for neuropsychologists to practice mindfully when using VPA to assess memory. Examples include adding alternative auditory memory measures to the test battery and conducting additional research with specific clinical populations to understand how performance on the test relates to in vivo memory functioning.

NSUWorks Citation

Andrews, A. P. (2019). An Analysis and Comparison of WMS-III and WMS-IV Verbal Paired Associates: Practical Implications for Neuropsychologists. .
Available at: https://nsuworks.nova.edu/cps_stuetd/133