Multi-attribute task battery performance predicts military multitasking better than multitasking costs

Abstract

Successfully conducting military operations requires proficient multitasking. Thus, it is essential to sufficiently assess soldiers’ multitasking abilities before deployment to predict their success in the open field. There are two established approaches to operationalize multitasking: either using a multi-/dual-task paradigm or assessing individuals’ performance in several subtasks building a cohesive multitasking scenario. However, it is unclear if and which approach may be more suitable for predicting military multitasking. To investigate this, we recruited 25 students/officer candidates of the University of the Bundeswehr in Munich to perform multitasking based on a dual-task paradigm and a multitasking scenario in the laboratory, and military multitasking in a shooting gallery. For the dual-task, individuals were asked to solve math equations and memorize radio signals in single- compared to dual-task conditions. For the multitasking scenario, individuals were required to perform the multi-attribute task battery (MATB), which includes four subtasks to simulate a flight scenario. For military multitasking, individuals were supposed to execute a shooting exercise and perform the same math and radio tasks as performed in the laboratory in single-/dual- and triple -task conditions, simultaneously. We expected the dual-task assessment in the laboratory to predict military multitasking better, given that both tasks shared very similar task requirements. In contrast, we found moderate evidence in favor of individuals’ MATB performance serving as the best predictor of military multitasking, using Bayesian hierarchical regression analyses. Thus, multitasking scenario performance may be more suitable for predicting military multitasking. A general multitasking ability might explain this effect.

Examining the Influence of Cognitive Load and Environmental Conditions on Autonomic Nervous System Response in Military Aircrew: A Hypoxia–Normoxia Study

“…The inlet supply pressure was set at 50 psig (344.7 kPa) and regulated by the built-in pressure reducer. The cognitive loads performed by the subjects under the conditions of Factor A consisted of two aviation-relevant computer-simulated task batteries: Automated Desktop Vision Tester (ADVT) [14] and Enhanced Air Force Multi-Attribute Task Battery (EAF-MATB-II) [15,16]. The tasks were alternated five times, equating to 125 min of continuous performance testing.…”

“…sisted of two aviation-relevant computer-simulated task batteries: Automated Desktop Vision Tester (ADVT) [14] and Enhanced Air Force Multi-Attribute Task Battery (EAF-MATB-II) [15,16]. The tasks were alternated five times, equating to 125 min of continuous performance testing.…”

Always in (partner) action? Working in teams may improve simulated flight performance – but only in the apt cognitive control state

“…exponentially modified Gaussian, or skewed normal distributions (seeHilla et al 2023, for more details). The prior parameters of the reference models (solitary performance in flexible control mode condition) for each task were as followed: tracking task performance ~exponentially modified Gaussian distribution with μ ~ Normal(0.08, 0.01), σ ~ Normal(0.01, 0.025), λ ~ Normal(0.01, 0.025), σrandom ~ Normal(0, 0.025), βCondition ~ Normal(-0.006, 0.25), βPerformance Mode ~ Normal(-0.008, 0.25), βInteraction ~ Normal(-0.002, 0.25); system monitoring task performance ~ exponentially modified Gaussian distribution with μ ~ Normal(2.75, 0.05), σ ~ Normal(0.05, 1), λ ~ Normal(2.5, 0.1), σrandom ~ Normal(0, 1), βCondition ~ Normal(0.03, 1), βPerformance Mode ~ Normal(-0.04, 1), βInteraction ~ Normal(-0.02, 1); resource management task performance ~ Gaussian distribution with μ ~ Normal(500, 200), σ ~ Normal(200, 50), σrandom ~ Normal(10, 10), βCondition ~ Normal(120, 50), βPerformance Mode ~ Normal(-160, 50), βInteraction ~ Normal(-80, 50); and communications task performance ~ exponentially modified Gaussian distribution with μ ~ Normal(2.5, 0.5), σ ~ Normal(0.5, 1), λ ~ Normal(0.5, 0.05), σrandom ~ Normal(0, 1), βCondition ~ Normal(0.02, 1), βPerformance Mode ~ Normal(-0.4, 1), βInteraction ~ Normal(-0.1, 1…”