A single-item assessment for remaining mental resources: development and validation of the Gas Tank Questionnaire (GTQ). (3rd September 2018)

• Record Type: Journal Article
• Authors: Monfort, Samuel S.; Graybeal, John J.; Harwood, Amanda E.; McKnight, Patrick E.; Shaw, Tyler H.
• Is Part Of: Theoretical issues in ergonomics science. Volume 19:Number 5(2018)
• Page Start: 530
• Page End: 552
• Publication Date: 2018-09-03
• Subjects: Mental workload -- attentional resource theory -- ego depletion -- single-item measures -- NASA-TLX -- Gas Tank Questionnaire; Human engineering -- Periodicals; Human engineering -- Research -- Periodicals; 620.82
• DOI: 10.1080/1463922X.2017.1397228
• Languages: English
• ISSNs: 1463-922X

Abstract

ABSTRACT: Measuring workload accurately is critical for human factors research and applications. However, if the instruments we use to measure workload are burdensome for participants to complete, we may find ourselves in the ironic position of increasing workload simply by measuring it. We present a pair of studies that suggest the NASA-TLX, a widely-used workload metric, increases workload for participants who complete it, and consequently may hurt performance and upwardly bias the estimates it produces. We found that repeated administration of the NASA-TLX increased workload by 18%, and more than doubled the performance decrement on a simple vigilance task. Single-item measures may be a simple way to avoid burdening participants unnecessarily while still obtaining valuable insight into participant cognitive state. In line with the attentional resource theory, we validate a measure that asks participants to 'think about [their] brain as an engine, ' and to indicate inside a rectangle 'how much gas [they] have left right now.' The Gas Tank Questionnaire performs well when compared against other subjective measures and when validated against physiological measures. Just as human factors professionals are careful to design systems to minimise unnecessary cognitive load, researchers should take similar care when selecting measures for studies.