Information Processing, Human Error and the Learning Process — Page 164, Lesson 258

I want to walk you through the final part of the information processing model — the output side. We've covered how we take in information and how we make decisions; now let's look at what happens when we actually act on those decisions. Actions can be taken as a result of information coming from three different sources. First, actions can come from Motor Programmes via the Attention Mechanism — this is the skilled pilot flying straight and level, where the action is almost automatic and the attention mechanism is just monitoring. Second, actions can come from Motor Programmes via the Central Decision Maker — for example, when carrying out a fire drill, you have a stored programme for the drill, but the central decision maker has to consciously select and initiate it. Third, actions can come directly from the Central Decision Maker — this is when you're solving a non-routine or unfamiliar problem, and you have to think through each step before acting. These three sources correspond to what are called Skill-Based, Rule-Based, and Knowledge-Based Behaviour respectively. I'll just note that those are dealt with in the next chapter — we'll cover them in detail when we get there. Now, Feedback. The feedback mechanism is continually in use. Even when motor programmes are being used to fly the aircraft straight and level in visual conditions, your senses are constantly scanning the environment to maintain the desired configuration. You're not consciously thinking about every tiny control input, but your senses are feeding back information to keep you on track. Response is the actual action you take, and any action undertaken will cause a detectable change. That change is noted by your senses, and this feedback may alter the action you're taking. Let me give you the example from the book: a pilot banking the aircraft receives feedback from the visual horizon or the artificial horizon. From the rate of bank detected, the pilot may increase or decrease the control forces. When the desired angle of bank is reached, the feedback causes the pilot to return the controls to the neutral position. So it's a continuous loop — act, sense, adjust, act again. When there is pressure to make a rapid response — perhaps in an emergency — there are several factors to bear in mind. First, there will frequently be a trade-off between speed and accuracy. A delay in some situations could be dangerous — engine failure after take-off is the classic example. But there will also be pressure to make a response before sufficient information has been processed. Second, high arousal level leads to faster but less accurate responses — you react quickly, but you're more likely to make mistakes. Third, auditory stimuli — noises — are more likely to attract attention than visual stimuli, but they are also more likely to be responded to in error. Fourth, an increase in age between 20 and 60 years tends to lead to slower but more accurate responses. So as you get older, you may take a bit longer, but you're less likely to make a wrong move. Finally, let's look at Response Error, specifically an Error of Commission. If we expect a stimulus and prepare a response, we will respond more quickly if the expected stimulus actually occurs. But if an unexpected stimulus occurs, we are more likely, under pressure, to make the prepared response anyway — and that's an error of commission. You commit the action you had ready, even though the trigger was wrong. Here's the example from the book: a pilot notices engine instrument variations showing parameters approaching out of limits. The pilot mentally prepares the engine shut-down drills in case the limits are exceeded. Now, any stimulus — perhaps as simple as the noise of a tray falling — may be sufficient for the pilot to shut down the engine. The noise wasn't the engine exceeding limits, but the pilot had the response ready, and under pressure, committed the error. That's the core of how actions, feedback, and response errors work in the information processing model.