Wednesday, 29 August 2012
The task that Andrew Gilbey and Stephen Hill presented to dozens of pilots (some of whom had 160 hours flight experience) required that they choose which of three geographical features to focus on (e.g. a picnic area, bush covered hills, or an unmade road) as a way of determining their location. They were to imagine that they were lost either on a motorbike, in a plane, or on a yacht and, with time short, they needed to use one of these three geographical features to check whether they really were where they thought they were (marked as a "best guess" circle on a map), or if they were in fact located elsewhere nearby.
Over 82 per cent of the time, the pilots chose to focus on one of the three available geographical features that was present both in the best-guess location and elsewhere nearby. In other words, they sought confirmatory evidence to support where they thought they were located. They almost entirely failed to focus on the one geographical feature that was not present in the best-guess location (but was located elsewhere). That is, by failing to seek disconfirmatory evidence, they fell victim to the confirmation bias and missed the best strategy in this situation.
"It appears that having extensive experience with map reading and flight navigation does not help in and of itself [to prevent confirmation bias in lost procedures]," the researchers said.
In other experiments, a group of psychology undergrads consistently made the same error as the pilots, even though they'd just had a lecture on the confirmation bias. A short presentation on confirmation bias also failed to improve the navigation performance of another group of trainee pilots (average flight experience 55 hours). Gilbey and Hill speculated that maybe the presentation failed because it was passive and didn't require the pilots to practice seeking disconfirmatory evidence.
Intriguingly, a group of 21 orienteers performed much better at the task, choosing the disconfirmatory evidence 67 per cent of the time. Gilbey and Hill said this result could help inform future training programmes for pilots - "Although confirmation bias has been the focus of a great deal of research, the current findings suggest that further applied research, particularly in the area of applied aviation, may further improve understanding of this pervasive phenomenon."
A Gilbey, & S Hill (2012). Confirmation bias in general aviation lost procedures Applied Cognitive Psychology DOI: 10.1002/acp.2860
-Further reading- People don't follow their own directions when walking from A to B.
Post written by Christian Jarrett for the BPS Research Digest.