Cognitive Human Factors & Mental Models (9/27)
On the psychological and cognitive processes that influence how we perceive, move through, and act on the world.
Cognitive Human Factors involve an examination and understanding of cognitive ability to design optimal interactions that take human cognition into consideration.
What exactly is cognition, you may ask?
For the purposes of this course, we’ll loosely define it as any mental process associated with gaining knowledge and comprehension.
Knowledge
Per Don Norman, there are two types of knowledge that we need to be aware of as designers since both play a part in any type of interaction:
- Knowledge of (factual/conceptual: perceived facts and rules)
- Knowledge how (procedural: skills)
In practice, we know that there is a difference between knowing of something and knowing how to do something.
In addition, The combination of knowledge in our head (mind) and in the world (environment) helps us navigate complexity and ambiguity in complementary ways. Considering our mental limitations (we cannot know or remember everything), the knowledge required for new interactions, decisions or behaviors can exist partially in the world. One of the ways in which we rely on knowledge in the world is through environmental constraints (or possibilities/affordances). As shown in the example below, the affordances found in these trashcans enable and limit some actions, aiding our quick processing and decision-making :
The openings of each recycling container act as signifiers for the types of materials that should be dispensed. These signifiers and affordances provide the information that helps us “know” what to do and how to do it.
Natural Mappings
Natural mapping refers to when the relationships between an element and its representation (including location) match. When controls map accurately to their resulting actions and positions, they create an intuitive relationship that makes systems faster to understand, adopt and remember.
In Practice
- Putting sufficient cues in a design (knowledge in the world) can result in good performance, even without previous knowledge on behalf of the user
- When the knowledge needed for a task is available in the world, the need for acquisition and learning decreases (reduces cognitive load, speeds decision-making)
Decision-Making
In practice, approximate answers are often good enough, even if they’re not the optimal strategy. According to Don Norman,
- Using simplified models and approximations of the underlying technical “truth” simplifies thought and complexity
- Practitioners don’t need truth: they need quick results that are “good enough” for the purpose and context to which they will be applied
We can give form to these notions through the concept of bounded rationality, put forth by Herbert Simon. This acknowledges human’s limitations in making decisions, where rationality is limited and not the only mental process that informs our actions (we are not fully rational machines or driven by optimisation objectives alone).
- Our cognitive capacity, available information, and time are limited in the real world
- To respond to these limitations, we opt for the option that’s “good-enough” (satisficing), which is not necessarily the best choice overall (and is not the result of a cost-benefit analysis).
One of the ways we make such decisions is through heuristics: shortcuts or rules of thumb we create for ourselves that help us make quick decisions based on what we have learnt from past experiences.
What determines our behavior?
Our mind? Environment? Both? We briefly discussed different approaches to understanding behavior, using Herbert Simon’s scissor model (adapted by Dan Lockton as shown below) to illustrate the interaction of these different dimensions shaping behavior. The graphic depicts how this is not an either/or neat process, but a product of the two.
This leads us to consider two Contrasting Views:
As Designers…
- We can learn from the strategies that people adopt on their own
- We should design to facilitate or support human needs and capabilities by learning from their mental models
In summation: Are the problems we identify as “human errors” in the use of different artifacts really people’s fault? Can we solve human problems instead of forcing a “correct” action or behavior based on the constraints imposed by technology?
Mental Models
Mental models can be described as internal representations of the world that are shaped based on an individual’s personal experiences, perception, and past understanding.
In design, we can specify this further as “what the user believes about the system at hand”. It’s a model of the world that helps people understand and address novelty.
But what if there’s a gap between a designer and user’s mental models? We have two options:
- We can adjust the system to conform to our users; or
- We can improve user’s mental models such that they more accurately reflect the system - help users learn
Luckily, we all adjust our mental models through experience and repeated practice. Therefore, feedback is crucial to help people adjust their behaviors- it dictates the kind of learning that can occur.
Single & Double Loop Learning
We briefly discussed varying ways feedback is integrated in different learning models.
Single-loop learning acquires 1 channel of information and takes corrective actions based on that, by making a new attempt based on the same information. Double-loop learning on the other hand, uses feedback from past actions to reevaluate assumptions that underlie the current model. This leads to trying something new (learning has occurred).
Altering Mental Models
We often need to reexamine and alter mental models to reflect the changing realities in which our problem space lives. Identifying a mental model gap offers hints on how to address it. This can be achieved through various research methods such as:
Think-aloud protocol: by asking people to verbalize their actions and decisions, various aspects of a system are revealed
Mental model diagrams: help visualize and articulate the foundational causes that drive behaviors. This is powerful becuase mental models are otherwise hidden, inaccessible structures of thought.
- This allows development of solutions that deeply resonate with people’s consistent beliefs
In-Class Activity
We then split into groups to examine our mental models of Amazon’s Alexa voice assistant or Amazon Prime grocery delivery service.
Just from this short activity, it was evident that everyone holds vastly different mental models based on their personal experiences, knowledge, and interactions with the object/service. By creating a group mental model of such interactions and systems, each member learnt something new about the system. This already started hinting at where we identify pain points in our everyday uses of such products and services or where some of those processes are highly hidden or obscure (black boxes). What was evident is that initially, no-one had a clear and accurate representation of the system as a whole.