A psychometric comparison of automated neuropsychological assessment metrics with laptop and tablet administration

Objective Determine whether Automated Neuropsychological Assessment Metrics administered with laptop and tablet platforms produces comparable scores. The study quantified the direction and magnitude of differences using a clinically relevant analytical approach. Method Randomized, within-subjects, counterbalanced, repeated-measures study with 302 active-duty service members. Participants completed the 10-subtest battery in two consecutive sessions (Tablet–Laptop, Laptop–Tablet, Tablet–Tablet, and Laptop–Laptop). Analyses included between-group comparisons, within-subject change, and Reliable Change Index base rates. Results Laptop scores were significantly higher than tablet scores on 9 of 10 subtests, with effect sizes from d = 0.11 to 1.15. Within-subjects analyses indicated a consistent pattern of performance declines in the Laptop–Tablet Group and improvements in the Tablet–Laptop Group. Reliable Change Index base rates showed that 73% of participants in the Laptop–Tablet Group reliably declined on ≥2 subtests versus 10% in the Tablet–Laptop Group. After correcting tablet controller timing artifacts, effects on fast reaction time (RT) tasks were substantially attenuated, while effects on learning-, memory-, and accuracy-based subtests remained significant with large effect sizes. The Laptop–Tablet Group decline rate decreased from 73% to 57% but remained 4 times the same-platform reference rate. Conclusion Tablet and laptop scores may not be interchangeable. Hardware timing corrections attenuated effects on fast RT subtests but did not reduce effects on accuracy-based measures, indicating factors beyond RT measurement error contribute to cross-platform differences. Interpreting tablet-based administration with laptop-based normative data or baselines could misclassify examinees. Until validated tablet-specific interpretive criteria are established, cross-platform comparisons should be avoided.