Assessment Function Analysis of Human-Automation Team Performance: A reanalysis of data from Yamani and McCarley (2018)

Abstract

Yamani and McCarley (2018) used workload capacity analysis to quantify automation usage strategy in a speeded length-judgment task and showed that operators delayed their responses under difficult task conditions. Contrary to predictions of the proximity compatibility principle, the results showed comparable operator performance between displays that embedded the aid’s decisional cue within the stimuli and those that did not, depending on task difficulty. This study reanalyzes the data of Yamani and McCarley (2018) employing functional principal component analysis of assessment functions (Townsend & Altieri, 2012) that combines both response times and accuracy for an analysis of workload capacity conditional to accuracy. The results indicate two components, early and late RT periods, both increased workload capacity and suggest that the integrated display may speed operator responses only under easy task conditions. Applied implications of the results are discussed.