B-175: Computerized Vs. Non-computerized Performance Equivalence of a Word List Task

Abstract

Objective: Traditional neuropsychological testing has relied on pen-and-paper test delivery methods. However, there is a significant modern trend toward computerized testing, urging researchers to validate these methods. This study aims to investigate computerized and non-computerized test performance on a cross-culturally adapted word list task. Method/Data Selection: Participants (N = 207) were recruited from undergraduate classes at the University of North Carolina Wilmington, as well as through convenience sampling. ANCOVAs were conducted across 19 dependent variables quantifying task performance with age and education as covariates, and gender included as an additional factor. Two one-sided t-tests were conducted to establish equivalence within clinically meaningful bounds drawn from the Rey Auditory Verbal Learning Task. Benjamini-Hochberg corrections were applied to adjust for multiple comparisons with a false discovery rate of 0.05. Results/Data Synthesis: ANCOVAs revealed no significant difference in performance between computerized and non-computerized following the correction for multiple comparisons. Education acted as a significant predictor for performance across five variables, F = 4.53 – 11.66, η2p = .025 – .059, p < .001 – .035; while gender was a significant predictor for one variable, F(1, 185) = 5.73, η2p = .030, p = .018. Age was not indicated as a significant predictor, and the overall model significantly explained variance across two variables, F = 2.84-3.38, p = .011-.026. The observed group effect was within equivalence bounds for 18 of the variables (p < .05; Figure 1). Conclusions: Overall, the current study supports the general trend in the literature, suggesting equivalence between computerized and non-computerized formats based on bounds derived from a similar task. The present task is not yet proposed for clinical application, and additional validation analyses are underway.