Load measures need to be read jointly with task performance

What it is

A single workload score — whether from a subjective scale, a secondary-task measure, or a physiological indicator — is not interpretable on its own. A cognitive load measurement only becomes meaningful when read alongside primary task performance. This practice is called joint interpretation of workload and performance.

Why it happens

Load and performance are not monotonically linked; the same load reading means something different depending on the performance level it's paired with. High load with maintained performance signals compensatory effort regulation — resources are near their limit but haven't collapsed. High load with degraded performance signals genuine resource overload. Low load with low performance may simply reflect insufficient engagement or motivation, unrelated to capacity. Low load with high performance is the ideal operating region. These four combinations cannot be told apart by looking at load alone — collapsing compensation and overload into the same conclusion is exactly the mistake a single load score invites.

Studying it

A rigorous design collects a load indicator and at least one performance indicator (reaction time, error rate, completion time) as parallel dependent variables, and analyzes the interaction between them rather than treating load as the sole outcome. A common pitfall is reporting a rise in load score as proof that "the experience got worse" without checking whether performance actually suffered — a load increase with stable performance supports "the user is working harder to keep pace," not "the user's capacity fell short," and these two conclusions must be kept distinct in the writeup.

Where it stops holding

Joint interpretation depends on the performance measure having enough range to show a decline. If a task is easy enough that performance sits near ceiling for nearly everyone, performance won't drop even when load is already high — an unchanged performance score in that condition cannot be used to infer that the load is acceptable.

Applying it

• When reporting or assessing a product's cognitive load, never present a single load score alone; show it alongside the matching completion time or error rate.
• Use the four combinations of load level × performance level as the interpretive frame for classifying observations, rather than discussing the load number in isolation.
• When you find "load is high but performance hasn't degraded," flag it as a compensatory-load case that needs attention rather than treating it as acceptable — this pattern often precedes fatigue and rising error rates over sustained use.

Related

• Same group: A9.02.1 Subjective scales and secondary-task methods measure different things
• Adjacent: A9.08 Performance-based measurement and the secondary-task method · A9.05 Load and interface complexity
• Search terms: workload-performance relationship · compensatory effort · joint interpretation · ceiling effect