The Invisible Gorilla

Today's newspaper report about distracted driving will not be surprising to social and cognitive psychologists, most of whom are familiar with the “invisible gorilla” demonstration by Christopher Chabris and Daniel Simons. Before going any further, please take a look at this video.

The video shows two teams, wearing white and black shirts, passing basketballs. The viewers were told to count the number of passes made by the white team, a doable task, but one which requires their full attention. Midway into the video, a person in a gorilla suit walks into the court, turns to face the camera, thumps his chest, and leaves after 9 seconds on screen. About half the participants don't even notice the gorilla. A striking finding is the confidence with which some people deny that a gorilla was present in the scene. As Daniel Kahnemann notes, “We can be blind to the obvious, and we are also blind to our blindness.”

This is a classic demonstration of inattention blindness. (Several other demonstrations can be found at Chabris and Simons' website.) It has obvious implications for the dangers of using electronic devices while driving. 

David Strayer and his colleagues at the University of Utah have released a study, supported by the AAA Foundation for Traffic Safety, comparing the amount of distraction caused by six different secondary tasks:

• listening to the radio
• listening to a book on tape
• conversation with a passenger
• conversation on a hand-held telephone
• conversation on a hands-free telephone
• interaction with a speech-to-text email system

To provide a basis for comparison, there were two control conditions:

• a baseline condition in which the driver performed no secondary task
• a cognitively demanding task (OSPAN) requiring participants to memorize words and solve math problems

Three experiments were performed: in a laboratory, in a driving simulator, and driving an instrumented vehicle in a residential neighborhood. Several measures were used to assess distraction: brainwave activity, eye and head movements, brake reaction time, reaction time and accuracy to a peripheral light detection task, and self-report measures of difficulty. (The peripheral light detection task was included to measure “tunnel vision,” the tendency to concentrate only on the center of the visual field.)

In addition to looking at these measures separately in each experiment, Strayer standardized the measures and combined them to create an overall measure of distraction. The baseline condition was arbitrarily set at “1,” the OSPAN task at “5,” and the six tasks were compared to these anchors.

Task	Workload Rating
Baseline	1.00
Radio	1.21
Book on tape	1.75
Passenger	2.33
Hands-free cell phone	2.27
Hand-held cell phone	2.45
Speech-to-text system	3.06
OSPAN	5.00

The tasks fall into three clusters. The two listening tasks are the least distracting; the three tasks requiring conversation fell in the middle; and the task requiring speech-to-text interaction was the most distracting. For comparison, a 2007 study by Strayer and Drews found the distraction produced by hand-held and hands-free cell phones to be equivalent to a blood alcohol level of .08, which is legally intoxicated in most American states.

Unfortunately, auto companies are installing voice-activated infotainment systems such as Chrysler's UConnect, which they describe as a “solution to distracted driving.” Both the auto companies and the public equate hand-held with dangerous and hands-free with safe. For example, in one survey, 66% of drivers said hand-held devices were unacceptable and 56% said hands-free devices were acceptable. However, Strayer found only a trivial difference in distraction between hand-held and hands-free cell phones. Relying on public opinion to decide which devices are safe is a bit like asking people whether they saw the gorilla.

Not surprisingly, the Alliance of Auto Manufacturers dismissed the study as “misleading,” since, they said, it focused only on the cognitive aspect of driving and ignored the visual and manual aspects. The statement is false, since eye movement and reaction time measures were included. It is also inaccurate in its implication that the cognitive, visual and manual aspects of driving are separable.

You may also be interested in reading:

Book Review:  Failed Evidence, by David Harris

New Year, New Blog