Designing Technology with the Human Mind in Mind: Part 4
Workload, Stress, & Total System Performance
Intro: This is the 4th in a series of articles to discuss what behavioural science knows about human cognition for those in product development. The 1st article explained why product makers should understand human cognition and introduced the dual system model of cognition (aka “Thinking Fast & Slow”). The 2nd article dove deeper into one part of that dual system model, “The filter.” The 3rd article introduced Working Memory. This article dives deeper into into cognitive workload & stress and how they impact performance. Future articles will discuss leadership, motivation, among other topics. All articles will explain why this knowledge is important to making technology that has a great user experience.
What is cognitive workload?
To understand cognitive workload one first must understand the dual system model of cognition, which I described in the first 3 articles. Briefly, all information that is sensed by the human sensory systems passes through the filter, which can choose to (a) act and bypass working memory, (b) “ignore” information, or (c) pass the information on to working memory. The information passed onto working memory can be individual sensations or it can be patterns. Working memory processes these sensations and patterns to create situation awareness, which feeds into decision making, which results in action (or not) being taken on the world. In working memory there are also the processes of sensemaking (creating new patterns) and metacognition (thinking about thinking).
All processing in working memory requires what are called attention resources. Cognitive workload is defined as the number of resources required by working memory to complete a task in relation to the number of resources available. It should be noted that attention resources are theoretical constructs. As of yet, we cannot point to something physiological and say that is an attention resource. Despite this lack of knowledge as to what resources are physiologically, we know that they exist and that they are limited.
One of the best-known examples of research on the limited resources within Working Memory is George Miller’s Magical Number 7+/- 2. Essentially, Miller found that we can hold 7 chunks of information in working memory. A chunk is a pattern that has been passed on from the filter. As described in the first article, the filter groups sensations into patterns and uses selective attention to determine what patterns or sensations are sent on to working memory. This is done to ensure that the amount of information to be processed does not exceed the number of resources available. Because if the task(s) do require more resources than are available, then we are in a task overload situation and our performance degrades. When our performance degrades. So, the filter does what it can to keep us out of an overload situation.
Expertise, Resources, & Cognitive Workload
It is important to remember that the filter passes along patterns and these patterns are unique to each human, the amount of resources needed to complete a task is unique to each person. In other words the number of resources required to complete a task varies from person to person. Therefore the cognitive workload experienced in any given task varies from person to person.
One of the most studied phenomena in psychology and decision making is the differences between experts and novices on cognitive tasks. Through this research, scientists that the key distinction between experts and novices is the patterns in long term memory . This research has shown that experts’ patterns typically are more complex than those of novices (Grier, 2015). That is experts’ patterns contain more information than novices’ patterns. This means that in experts the filter passes on more information or at least the information is organized in a way that requires less resources to obtain situation awareness, which is a key input to the decision making process.
As described in the second article in this series, situation awareness is most commonly defined as “the perception of environmental elements and events with respect to time or space, the comprehension of their meaning, and the projection of their future status” (Endsley, 1988). More succinctly there are three levels of Situation Awareness: (1) perception, (2) comprehension, and (3) projection. The more situation awareness a person has, the better quality of his/her decision making.
As experts have more complex patterns than novices, they need to spend less resources to develop situation awareness. So experts experience less workload. These more complex patterns typically result in better performance with less workload for experts. That is the expert does not need to use resources to create meaning and projection as the filter has recognized the pattern and passed it on as 1 chunk to Working Memory. Malcolm Gladwell described in “Blink: The Power of Thinking Without Thinking,” they may instantaneously and unconsciously make a decision.
However, when this occurs experts rarely can describe why they came to that decision. And in fact, asking an expert to describe how they know something can lead to worse performance as they try to reconstruct the patterns in their long term memory and begin to question their decision making ability. Novices are far more capable of explaining why or how they do something than doing it. Thus supporting the adage, “those who can — do; those who cannot — teach.”
Another disadvantage of the expert is that when the pattern is passed on, it may pass along information that is not there. I am certain that everyone has written text and not realized there was a typo. This is an example of the filter passing along a pattern and not what was actually sensed. It is also how magicians do their illusions. They take advantage of the universal patterns so that we sense what they want us to sense and not what is happening.
To be clear, the cognitive workload does not correlate to the demands of the task. The same task can lead to different levels of cognitive workload depending on the people who are performing the task and their expertise. Physiological factors such as intoxication, illness/injury, hunger, and fatigue can impact the workload experienced. All of these and other non-physiological factors can cause us to experience stress, which can also impact the number of resources available.
Cognitive Workload, Stress, & Performance
Recall that the definition of workload is the number of resources required by working memory to complete a task in relation to the number of resources available. Resources are physiological; so if we are intoxicated, ill, injured, hungry, or tired, we will have fewer resources available. So even if a task requires the same number of resources on Monday and Tuesday, if on Tuesday we are tired/ill/hungry, we will experience higher workload due to fewer resources being available.
In addition, the number of resources available can change based on the experience of stress. We have heard the stories of people doing extraordinary things in moments of extreme stress. This is because in those instances additional resources (attention and physical) are made available for the fight or flight response. However, these additional attentional resources are only available for short periods of time. After which the number of resources available drops precipitously. The precipitous drop allows our bodies a chance to recharge the pool of resources available. It should be noted this recharge is necessary. If we do not recharge after extreme stress or we experience stress for an extended period of time, there will be permanent changes to the structure of the brain.
With the exception of the extreme situations, it is well documented that performance does not improve as stress increases. In 1908, the Yerkes-Dodson law established that there is an optimal level of stress where our performance increases. Part of this is due to the number of resources available, but part of it is due to task unrelated thoughts.
If the number of resources required by a task is far below the number of resources available, attention resources may be diverted away from the task to the base motivations associated with the 3 S’s (i.e., sleep, sustenance, and love). That is our mind wanders to one of these topics instead of focusing on the task at hand.
When we experience these low demand situations and our thoughts start to wander, one of two things can happen: vigilance decrement or complacency. The vigilance decrement occurs with high stress tasks and complacency occurs with low stress tasks.
The vigilance decrement is one of the most robust, but seemingly paradoxical findings in human factors. It was first studied in WII; when RADAR was invented a new position in aircraft was created where operators would monitor the RADAR displays for a clear signal that there was a German U-boat in the waters. They would then alert other entities, who would attack the U-boat.
It seems like a simple task; look at this display for a clear blip. In fact, detection was near 1005 during the first 15 min of watch. However, after 15min of watch, the chance of detection decreased precipitously. Furthermore, the operators would return to base as tired as the pilots.
To the casual observer it looks like the person is doing very little. However, 50 years of research has shown that the person is expending a lot of attention resources on preventing their mind from wandering or staying awake, because they understand the risks of not detecting every signal. So they are expending a great deal of resources to keep the task unrelated thoughts out of their working memory. This high level of resource usage leads to high cognitive workload. The vigilance decrement is the reason life guards and transportation security guards switch locations every 15min.
Conversely, complacency is the failure to adequately monitor a situation (i.e., maintain vigilance) due to self-satisfaction with performance. In other words, if we believe that the task does not demand all of our attention, then we will allow unrelated thoughts to use resources in our working memory. Complacency has resulted in numerous accidents during “routine” medical transport calls, long duration space flight, war, among other situations. Though it is instinctual to think of complacency as a human failing it is indeed a normal and adaptive behaviour to a situation in which fewer resources are available. The filter is “aware” that fewer resources are available, so it sends fewer things on for processing. Complacency also commonly happens when a task has become routine.
Complacency occurs in low stress tasks with moderate and low resource requirements. The vigilance decrement occurs in high stress tasks with low demands. Cognitive tunnelling is a phenomenon that occurs in high stress tasks with moderate demands. First described in combat pilots, cognitive tunnelling is when an individual has trouble making more strategic decisions. That is they are very much focused on the here and now and not on the wider context. It seems that stress reduces how much information working memory can process. Examples of cognitive tunnelling exist outside of the military domain. For example, medical doctors have been documented to make poor decisions after being awake for numerous hours. It has also been seen in daily life, when people are exposed to longer term stressors such as poverty.
Thus stress and resources required have an interactive effect on an individuals performance. There is an optimal level of stress and resource requirements where an individual will have the highest performance. In future articles, I will return to this concept and describe how leadership and other organizational factors can impact individual performance.
Multiple Types of Resources
Thus far, I have described resources as if they are one type. That isn’t quite accurate. Chris Wickens posited in 1984 that there are multiple pools of resources and all evidence supports this assertion. Each pool of resources is available for different types of information at different stages of processing. The three stages of processing are (1) perception, (2) analysis, and (3) action. Information can come in via different sensory systems, which is one dimension of type. A second dimension of type is whether it is linguistic or graphical/tonal. Each combination of information type and stage of processing has a dedicated pool of finite resources.
So one can be in an overload situation and still have available resources. Those resources are just in another pool. Multitasking is completely possible when using multiple pools of resources. There are limits though. Each of the three stages of processing have a finite limit that cannot be exceeded and each dimension of type of information have a finite number of resources. So, we are able to do more multi-tasking when the tasks use very different resources than when there is overlap in the tasking.
The careful reader will notice that the 3 stages of processing do not align perfectly with the stages of processing in working memory. Perceptual resources are dedicated to attaining the first level of situation awareness, perception. Action resources are dedicated to taking action. Analysis resources for everything else that occurs in working memory. This includes attaining the second and third levels of situation awareness (i.e., comprehension & projection), decision making, sensemaking, and metacognition. Thus the patterns that are used by the filter are even more critical to reducing workload. As the analysis resources are used for most of working memory.
Summary
In summary, all the cognition that occurs in working memory requires attention resources. Cognitive workload is the number of attention resources required for a task compared to the number of attention resources available. The number of attention resources available can change based on our health and our experience of stress. When we are doing a task that does not have a lot of stimulation, our working memory will allow our mind to wander to our base motivations. If the task is important, then we will expend more resources to prevent mind wandering. Thus tasks that appear easy can be just as challenging as tasks that appear overwhelming in terms of information to be processed. Furthermore, there are multiple pools of resources. When tasks are spread across these pools, we can do more than when relying on a single pool. So when adding automation, we must consider how it changes the cognitive processes. This is best done through a task analysis and function allocation.
The next article will dive deeper into how to utilize this knowledge of the dual system model, resources, workload, & stress in the design of systems. This will talk in more detail about task analysis and function allocation. Future articles will also discuss how the user interface helps with creation of patterns known as mental models and how that impacts trust with automation, which is critical to successful human-system integration. Another topic for future articles in about how leadership and teammates can aid each other in staying in the zone of optimal performance.
