Maslow's Hierarchy of Needs

I maintained that HF exists at the absolute top of the hierarchy. After all human needs are met, after 99% of the usefulness of a system is harvested, we look for the remaining 1% by optimising that human system interface. Her husband apparently struggles with his contribution to society through teaching, and her contribution through art.

My response to their crises was more charitable than theirs: without teaching even some of the most basic human needs (warmth, food, water) would not be met, so teaching makes a huge contribution to society. Likewise art somehow appeals to some basic and fundamental human need. I firmly believe that the human mind is on a constant quest to learn, to absorb, to problem solve and to explore, and art provides for many of these needs. So while many would say art is at the ’self actualization’ level, I believe it is more visceral than that.

You could argue that HF also addresses some basic human needs, but I would argue without HF, those basic human needs would still be met. It is less likely that some basic human needs would be met without teaching and art.

With so many people in the world lacking the basic human needs of food, water, shelter, etc. is HF an endangered species (note the nod to Darwin – happy birthday)?  Maybe not in the first world, but is this constant existence at the pinnacle of our hierarchy of needs leading to good things?  Is it helping to make the world a better place?

This thinking keeps me grounded and puts things into perspective when doing HF.

Model Human Information Processor

The model accommodates the entry of a stimulus from the environment (although stimuli internal to the processor, triggered by previous external stimuli, may also enter in this manner) via a sensory processing ‘short term sensory store.’ This allows the brain to ‘buffer’ incoming information before stimuli are perceived, a process that is mediated by the processor’s mental models, resident in long-term memory, which are acting at that point in time. The mental models will frame the processor’s expectations about a situation, which will affect the search strategy adopted, and the interpretation given to stimuli.

After perception, the processor can either immediately choose and make a response (in the case of a reflex or an extremely well-learned stimulus-response relationship) or the perception can enter working memory for inclusion into a more complex consideration. A perception in working memory can be ‘chunked’ with other information into more effective and meaningful units. A chunk in working memory can also be ‘the tip of the iceberg’ and give the processor effective access to an entire mental model (thus overcoming the limits of working memory). Specifically, a mental model represents characteristics, behaviours, relationships, expectations, etc. about a system, concept, etc. If the chunk is a sufficiently strong representative of some aspect of that mental model, it may, for the purposes of cognitive performance, effectively draw the entire mental model into working memory, without any penalties that might be associated with exceeding working memory capacity.
The processor can then use the information in working memory (and possibly in long-term memory) to engage in judgement, decision making and problem solving, before choosing an appropriate response and making that response.

The processor will also monitor the response and receive feedback about its success or failure.
The interaction between working memory and long-term memory, specifically the new information being passed between and the feedback information being considered, result in construction and modification of mental models, which is effectively learning.

All processes can be rendered more or less effective according to the attentional resources provided to them. An undifferentiated pool of attentional resources is divided between the various processes. Some processes are more demanding of these resources than others. Some stimuli make more demands on a processing stage, requiring more attentional resources. When multiple stimuli are being considered, translation between different modes of presentation may also place large demands on attentional resources, as would secondary or multiple task performance.

What if I do Human Factors or cognitive engineering (both of which I use in preference to  ‘cognitive ergonomics’)?  Where’s my legitimacy?

As alluded to above, hard ‘facts’ mean that things are reproducable.  That is, people are applying investigative and remedial techniques and are independently coming to the same answers.  I don’t think I can say the same for cognitive engineering.  Give the same technique to 5 practitioners and you will get 20 different applications of the technique (depending upon the day, the hour, the coffee, the colleague…) and 100 different outcomes.  This reduces the legitimacy of the discipline and makes teaching the discipline nigh on impossible.  Often, the success or failure of HF and cognitive engineering relies wholly on the individual and how well they practice the black art.  The HF ‘Artist’ is the one who manages to create a successful result out of a whole lot of noise.

So what’s the noise?  In a nutshell, cognition is noise.  It is a noisy process, prone to distraction, delay, dead ends, and logical leaps, performed in a world that is also noisy, where any one bit of noise may be the bit to which the person is attending.  It takes an artist to figure out what is important in this noisy world and use it to create the successful result.  If we attempt to exhaustively analyse all the noise, the analysis itself is unusable because we end up with too much noise there.

I prefer to focus on things as a means to an end.  The end I am searching for is whatever it takes to make my customer, but more importantly the user population, happy.  Publications for the customer, but job enhancements for my user population.  This point is made in Chapter 5 of this book that we shouldn’t care about the means, only the end.  If, from the outset, we concentrated on achieving measurable improvements in job performance, we could get away from rigid adherence to methods and get to our successful result by any means necessary.  Indeed, if a project as a whole set out to achieve some specific measurable improvement in job performance, a percentage of that could be down to software, a percentage to hardware, a percentage to chemistry, a percentage to whatever, and a percentage to Human Factors.  In this way, HF would cease to be a bolt on, but would be perceived as an important, integral, and integrated member of the team.

Maybe university and college courses should go this way too.  Don’t teach the methods beyond the basics.  Let the student figure them out.  Concentrate on setting performance improvement goals as class projects, then mark the students on how close they come to achieving them.  Sure, we’re still going to get different approaches and different results, but we will see the artist in all the students.  And we’ll be able to compare their results against their peers.

We will get as close as we can to hard ‘facts’ in our noisy cognitive world, and start to realise the legitimacy that we, as HF and cognitive engineers, know we deserve.

I took the car out the next day before the snow plow came round and tried to spin the car.  I got on the road and floored it; I couldn’t fishtail or spin the tires.  I hit the brakes; I didn’t lock it up.  I turned hard up a side street and accelerated as I did it; the car went right where I pointed it.  In short, the slide shouldn’t have happened.

I asked my wife how she thought the ABS worked.  When she slid the car I felt her go on the brakes and immediately felt the ABS.  I felt her come off the brakes, and come back on the brakes, and I felt the ABS go again.  Turns out she believed the ABS was an assistive device, rather than an informational device.  She believed that the ABS would bring the car to a controlled stop if she stood hard on the brakes.  The car doesn’t agree.  When the ABS goes off she needs to ease off the brakes (rather than come totally off and brake hard again) until the car finds grip. (Edit: apparently ABS IS supposed to help you brake safely when you stand on the pedal hard.  It turns out neither my wife nor I understand the system, and neither use it correctly.  Thanks Scott!)

When we are given a new system, there is always a learning curve (Windows 3.1 to Windows 95 was pretty steep, even XP to Vista was!).  In this case the user needs to know how the system is designed to work, and then what it feels like when it works.  I’d never really driven a car with AWD or traction control, so when we got this one and I felt like it was fishtailing, I looked up how the AWD worked so I understood what the car thought it was doing and could work with it.  After all, AWD and traction control are supposed to make the car safer.

There are some famous examples of such conflicts between the operator and the system.  The best one I know of is the crash of an Airbus in France.  Here, the pilot brought the aircraft in as if landing, then when he tried to put the power on to climb out, the plane didn’t respond and ended up crashing into the trees.  The sophisticated computer thought the pilot was trying to land and went into that mode and didn’t exit it when the pilot tried to take off again.  This led to redesigns of the computer logic and a change in the training given to pilots, so they could understand the logic the computer was bringing to their activities.

I don’t do it, but always take the time to understand not only the interface (in my recent case, the brake pedal is the interface) but also how the system is designed to work.  I’ll be taking my wife out to a snow parking lot so she can get comfortable with the ABS, AWD and traction control so this will never happen again.