Chronometric Normalization for BYOD Digital Assessments: Separating Hardware Latency from Cognitive Fluency
Abstract
The transition to digital, Bring-Your-Own-Device (BYOD) environments has vastly expanded the scale and accessibility of psychological and educational testing. However, unstandardized hardware introduces systematic mechanical latencies that contaminate timed cognitive performance, leading to measurement error and possible diagnostic misclassification. Furthermore, naive device-specific normalization strategies are inadequate, as they inadvertently suppress meaningful sociodemographic variance that inherently correlates with device ownership. To address this methodological gap, we introduce a chronometric normalization framework designed to isolate and remove hardware-driven latency without altering underlying cognitive capacity. The model leverages a brief motor task to estimate individual mechanical delay and calibrates this correction against a device-minimal, untimed cognitive task that is substantially less sensitive to hardware speed. This anchors the calibration so that real between-group differences in ability are retained. We demonstrate the framework’s practical utility using data from COSMOS, a large-scale dyscalculia screening battery administered to 54,538 students across 910 schools using the hardware available at each school. Results indicate that while trackpad users initially exhibited severe, artifactual underperformance on timed fluency tasks, the chronometric correction successfully neutralized this hardware penalty. By realigning score distributions while preserving structural demographic differences, this open-source method ensures that subsequent norming and diagnostic decisions reflect true cognitive architecture rather than the mechanical limitations of the user’s interface.
Related
- Explicit diagonalization of the linearized Calderón problem on the disk: multiplicity normalization and exponential spectral attenuation
- Self-normalization for Spectral Density Integrals
- Automorphic Transcendence Degree, Growth, and Skew Normalization
- Normalization by Evaluation with Substitutions as Operations
- Plane separating continua inscribe rectangles
BibTeX
@misc{fraga2026chronometric,
title = {{Chronometric Normalization for BYOD Digital Assessments: Separating Hardware Latency from Cognitive Fluency}},
author = {Eudald Correig Fraga and Carlos García Sancho and Marc Colomer Canyelles and Brian Butterworth},
howpublished = {PsyArXiv},
year = {2026},
month = sep,
doi = {10.31234/osf.io/d35fc_v1},
url = {https://osf.io/preprints/psyarxiv/d35fc_v1/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Chronometric Normalization for BYOD Digital Assessments: Separating Hardware Latency from Cognitive Fluency
AU - Eudald Correig Fraga
AU - Carlos García Sancho
AU - Marc Colomer Canyelles
AU - Brian Butterworth
PY - 2026
DA - 2026/09/23
PB - PsyArXiv
DO - 10.31234/osf.io/d35fc_v1
UR - https://osf.io/preprints/psyarxiv/d35fc_v1/
AB - The transition to digital, Bring-Your-Own-Device (BYOD) environments has vastly expanded the scale and accessibility of psychological and educational testing. However, unstandardized hardware introduces systematic mechanical latencies that contaminate timed cognitive performance, leading to measurement error and possible diagnostic misclassification. Furthermore, naive device-specific normalization strategies are inadequate, as they inadvertently suppress meaningful sociodemographic variance that inherently correlates with device ownership. To address this methodological gap, we introduce a chronometric normalization framework designed to isolate and remove hardware-driven latency without altering underlying cognitive capacity. The model leverages a brief motor task to estimate individual mechanical delay and calibrates this correction against a device-minimal, untimed cognitive task that is substantially less sensitive to hardware speed. This anchors the calibration so that real between-group differences in ability are retained. We demonstrate the framework’s practical utility using data from COSMOS, a large-scale dyscalculia screening battery administered to 54,538 students across 910 schools using the hardware available at each school. Results indicate that while trackpad users initially exhibited severe, artifactual underperformance on timed fluency tasks, the chronometric correction successfully neutralized this hardware penalty. By realigning score distributions while preserving structural demographic differences, this open-source method ensures that subsequent norming and diagnostic decisions reflect true cognitive architecture rather than the mechanical limitations of the user’s interface.
ER -