Related Experiment Video
Updated: Sep 25, 2025
Tactile Vibrating Toolkit and Driving Simulation Platform for Driving-Related Research
Published on: December 18, 2020
Cognitive Load Assessment Scales in Simulation: Validity Evidence for a Novel Measure of Cognitive Load Types
Sara K Greer1, Donna B Jeffe, Arushi Manga
1From the Division of Pediatric Critical Care Medicine (S.K.G., A.M., A.R.E.), Department of Pediatrics, and Department of Medicine (D.B.J.), Washington University in St Louis School of Medicine, St Louis, MO; and Anesthesiology and Clinical Affairs (D.J.M.), Khalifa University, Abu Dhabi, United Arab Emirates.
Introduction:
Cognitive load (CL) theory provides a framework to inform simulation instructional design. Reliable measures of CL types (intrinsic [IL], extraneous [EL], and germane load [GL]) in simulation are lacking. We developed the novel Cognitive Load Assessment Scales in Simulation (CLAS-Sim) and report validity evidence using Kane's framework.
Methods:
This quasi-experimental study tested the effect of a segmented/pause-and-debrief or standard/end-of-case-debrief intervention on pediatric residents' performance and self-rated CL in 2 complex- and simple-case simulations. After each simulation, participants completed 22 items measuring CL types. Three validity inferences were examined: scoring (instrument development and principal component analysis); generalization (internal consistency reliability of CL-component items across cases); and extrapolation [CLAS-Sim correlations with the single-item Paas scale, which measures overall CL; differences in primary task performance (high vs low); and discriminant validity of IL under different instructional-design conditions].
Results:
Seventy-four residents completed both simulations and postcase CLAS-Sim measures. The principal component analysis yielded 3 components: 4-item IL, 4-item EL, and 3-item GL scales (Cronbach's α, 0.68-0.77). The Paas scores correlated with CLAS-Sim IL and total CL scores in both cases ( rs range, 0.39-0.70; P ≤ 0.001). High complex-case performers reported lower IL and total CL (analyses of variance, each P < 0.001). In multivariate analyses of variance, CLAS-Sim IL, GL, and total CL varied across both cases by arm (each P ≤ 0.018); the segmented-debrief arm reported lower IL than the standard-debrief arm in both cases (each P ≤ 0.01).
Conclusions:
The CLAS-Sim demonstrates preliminary validity evidence for distinguishing 3 CL types but requires further study to evaluate the impact of simulation-design elements on CL and learning.
