Abstract Operators controlling a power grid were subjects in a mental workload experiment in which they performed their normal tasks under simulated quiet and busy conditions. Workload was assessed using the NASA Task-Load Index (TLX). It was found that there were significant differences between the two conditions, but there were no order effects. The value of the weighting procedure was assessed by correlating unweighted and weighted workload ratings. Although the correlations were high, they were not perfect, indicating some value for weighting the dimensions of the TLX. Large individual differences were observed in perceived workload.