Every time I get my eyes checked I’m just waiting for bad news. It doesn’t help that every time I visit the optometrist, they seem to have installed a new piece of equipment to test another aspect of my visual abilities. Some of them are really hard to cheat.
The most recent one (which added an extra $10 to my visit), involved sitting with my head pressed against a pair of goggle-frames which looked into an apparently blank white screen. I had a small button in front of me, and any time I saw some activity I was supposed to press it. What I ended up seeing was like a faint fluttering – almost like a large cursor blinking on and off – that would appear in various spaces around the screen, presumably to measure the effectiveness of my eyesight in each quadrant. At first it seems straightforward, but the fluttering would get fainter in some regions and more pronounced in others. The point was to track user error, and what I was doing isn’t far removed from the flanker test.
An article in the Economist last week discussed some recent research that has been conducted into the links between human error and brain activity. We still don’t understand a lot about the loss of concentration that tends to happen when we do routine tasks, even though we know it causes major problems. “In a factory it might mean that a component has to be thrown into the scrap bin. But in some occupations, like operating a giant crane or piloting an aircraft, the consequences can be devastating,” the Economist said. In IT it’s more about information being entered (or not entered, or entered incorrectly) into a database or other system, but it still leads to duplication of effort, botched processes or (in health care) lives put in jeopardy.
The researchers use the flanker test, which is available via a Java applet online, to gauge how distracted we really get. Here’s how it works: you see arrows and boxes flanked by other arrows and boxes. You have to choose (with your keyboard’s arrow keys) which direction you think they’re moving. There tends to be a 10 per cent error rate. According to the Economist article, tracking brain activity suggests there might a way to predict those errors about six seconds in advance, which would obviously be huge.
While the cognitive scientists pursue this area of study, it might be worthwhile to use the flanker test as in the early stages of application development. Look at how well users perform and examine the workflow which constitutes the technology-driven process. If the application requires even more response time, or throws something more distracting at them than the flanker test’s arrows, prepare for a higher level of failures once the system is up and running.
We tend to think of software projects as a way of ridding repetitive, mundane tasks but the inputting and retrieval processes alone leave a lot of room for missed steps. The successful use of technology demands, at a minimum, lightly focused attention. Human error may simply be explained by what we unconsciously, occasionally, ignore.
