Perceived cognitive load: meaning and application

You have shortened an application form from twelve fields to eight, yet users still describe it as a hassle. That experienced mental work is called perceived cognitive load. This article explains what the concept means, how to measure it and why you should not confuse it with task complexity, performance or a bodily signal.

Perceived cognitive load is the mental work that a task, choice or message costs in a person's experience.

It describes an experience, not a property of the task. Two people may find the same task differently demanding, for example through prior knowledge, effort or presentation. It is usually measured with self-report, such as the Paas scale or the NASA-TLX. Physiological measures such as pupil size or EEG can add information, but measure something else and are no substitute.

Neurofactor uses perceived cognitive load as a working concept: the degree to which people experience a task, choice or message as mentally demanding. The term has no single originator or validated instrument of its own; it combines three lines of research.

The first is cognitive load theory. Sweller (1988) described how a common problem-solving strategy demands so much processing capacity that little remains for learning. The theory later distinguished intrinsic load (complexity of the information relative to the prior knowledge of the person processing it), extraneous load (load caused by the way information is presented) and germane load (processing that contributes to learning). In their revision, Sweller, Van Merriënboer and Paas (2019) no longer treat germane load as a source adding to total load, but as a redistribution of working memory resources towards learning-relevant activities.

The second line is working memory research. In the multicomponent model of Baddeley (2012), several components work together, each with limited capacity. So new, interrelated information can quickly feel demanding.

The third line is ergonomic research on mental workload. Longo, Wickens, Hancock and Hancock (2022) describe mental workload as something that arises from the interaction between task and person. The word perceived stresses the experience of that interaction, not the task on paper.

Cognitive load theory comes from education. Applying it to forms, choices and messages is a translation, not a direct outcome of that research.

Objective task complexity covers features such as steps, simultaneous elements and time pressure. Experienced load does not follow from these one to one, for at least four reasons.

• Prior knowledge. What is one unit for an experienced user is several separate elements for a beginner (Sweller et al., 2019).
• Effort. Yeh and Wickens (1988) show that investing more effort can improve performance while experienced workload rises.
• Presentation. In two experiments by Song and Schwarz (2008), participants judged an exercise routine and a recipe as more time-consuming and harder when the instructions used a hard-to-read font. The task itself was the same.
• Context. Hart (2006) notes that NASA-TLX ratings can be influenced by what someone did just before.

So a shorter task is not automatically lighter, and low experienced load does not prove the task is simple.

Mental workload research uses three kinds of measures: self-report, performance and physiological signals. They can rise together, diverge or be insensitive to a change; Longo et al. (2022) call such dissociations a central challenge.

Self-report is closest to the concept. The NASA Task Load Index (NASA-TLX), developed by Hart and Staveland (1988), asks for ratings on six dimensions: mental demand, physical demand, temporal demand, own performance, effort and frustration. An unweighted variant is often used alongside the original weighted version; Hart (2006) finds no clear answer on which is more sensitive. The scale by Paas (1992) is shorter: a single rating of invested mental effort on a nine-point scale. Sweller et al. (2019) see this simplicity as an advantage, but studies differ in labels and scale points, which hampers comparison.

Physiological indicators measure something else: a bodily response possibly linked to effort. In cognitive control tasks, pupil dilation typically increases with task demands, but luminance, baseline pupil size, age and surprise need to be controlled (Van der Wel & Van Steenbergen, 2018). EEG has also been studied as a load indicator (Sweller et al., 2019); What is EEG? explains what it records. Physiological signals are also influenced by factors unrelated to the task (Longo et al., 2022).

So report self-report and physiological data separately: a larger pupil response does not prove that someone found the task demanding.

The study by Iyengar and Lepper (2000) shaped the idea of choice overload: more options might reduce motivation and satisfaction. However, a meta-analysis of 50 experiments found an average effect of virtually zero (D = 0.02; 95% CI −0.09 to 0.12), with considerable variation between studies (Scheibehenne, Greifeneder & Todd, 2010). A later meta-analysis found that overload occurred mainly with complex choice sets, difficult decision tasks, uncertain preferences and an effort-minimising goal (Chernev, Böckenholt & Goodman, 2015). The number of options alone predicts little.

Decision fatigue also calls for caution. Vohs et al. (2008) reported that making many choices reduced subsequent self-control, assuming a limited mental resource (ego depletion). A preregistered replication in 23 laboratories with 2,141 participants found an ego-depletion effect of d = 0.04, with a confidence interval that includes zero (Hagger et al., 2016). That replication tested a different paradigm but undermines the assumed mechanism. Do not treat decision fatigue as an established explanation.

Processing fluency is the subjective experience of ease or difficulty in processing information (Alter & Oppenheimer, 2009). Easily processed information is more often judged true and likeable. People can also use the processing difficulty of a message as a cue for the effort of the action itself (Song & Schwarz, 2008). As this rests on a few experiments with students, it is an indication, not a general law. Test it with your own message.

The concept helps with forms, product use, choice moments and messages.

1. Describe the task objectively. Count steps, fields, terms and decisions.
2. Define for whom and in which situation. Prior knowledge and context shape the experience.
3. Choose a suitable self-report measure. Use one effort question per step, or the NASA-TLX to see which dimension drives the load. Fix the choice in advance.
4. Combine with behaviour and explanation. Drop-off points, errors and open answers show where the load lies. Report any physiological measures separately.
5. Compare variants. A difference between two versions says more than a single score.
6. Link it to the decision. Before collecting data, record which difference matters for your decision.

Experienced load can be a pain point in itself or amplify another one; see What is a target group’s main pain point?. An association such as ‘complicated’ or ‘hassle’ can also act as a deal-breaking association.

This example is fictitious. A business software provider sees that many visitors start the trial account form but do not complete it.

Decision question. Should the form be shorter, or is the experienced effort elsewhere?

Available information. Drop-off points per step and reports about unclear questions.

Suitable approach. The team counts fields and decisions per step. New and experienced software buyers then complete two versions: the current one and one with the same fields plus an explanation of the subscription choice. After each step comes one effort question on a nine-point scale, with an open question at the end.

Possible interpretation. If experienced load peaks at the subscription choice rather than with the number of fields, further shortening is probably not the solution. No results are known for this example.

Next step. Redesign the choice step, compare again and check whether new and experienced buyers differ.

• Equating load with length. Fewer fields do not automatically reduce experienced load.
• Mixing up measures. A physiological signal is not self-report.
• Reading a score as an absolute measure. There is no general threshold for load that is too high; Hart (2006) also sees no such limit usable across tasks yet.
• Treating popular effects as fact. Choice overload and decision fatigue are context-dependent or disputed.
• Seeing all mental work as a problem. In learning, part of the load is useful; the aim is to reduce avoidable load, not all thinking.

Self-report usually happens afterwards and depends on how people interpret the scale. Group averages say little about individuals.

Perceived cognitive load shows how much mental work people experience, regardless of how complex a task looks on paper. It is most useful when measured with suitable self-report, compared across variants and kept separate from performance and physiological signals. Treat broad claims about choice stress or decision fatigue cautiously: the evidence is mixed.