SELF-REPORT MEASURES OF COGNITIVE WORKLOAD
A long-standing approach in research studies is to ask individuals directly how much workload they have experienced while performing a task. The best known exemplar of this approach is the NASA-TLX, developed by Hart & Staveland, 1988.
Although its name suggests that it is a measure of task load, in fact it is an attempt to quantify the level of workload perceived by the individual (as described above). Thus, it is a subjective, self-report measure.
As a research tool, the NASA-TLX has proved useful and has demonstrated that individuals are often able to distinguish among tasks with high and low mental workload. However it has at least two limitations.
Its first limitation is that it is a snapshot rather than a continuous measurement of the workload experienced by the individual. The typical administration of the NASA-TLX comes at the end of a task, and the measurement necessarily suffers from recency effects. If the individual experienced high workload at the start of the task and then his workload diminished, he will probably not characterize the task as causing high workload because he has most recently experienced only low workload. Similarly, if he experienced low workload for most of the task and then had high workload at the end, his estimate of workload from the task will likely be inflated. Thus, important information is lost.
The second limitation is that the workload estimate is based on an opinion given by the test subject himself; it is not an objective measurement. If we are to accept this estimate of workload, we must assume that the subject is both truthful and accurate.