Reaction Time

History

Some of the earliest recorded attempts to evaluate human performance with RT were made by 17th-century astronomers. They worried about the personal equation, which is simply the fact that different observers vary in their estimates of the transit times of stars as the stars moved across the visual field. These astronomers were not as interested in why observers had different personal equations as they were in how much they needed to recalibrate their equipment so that transit times were as accurate as possible. The first serious attempt to explain why RTs varied under different conditions—the first use of RT to determine how people's brains perform mental tasks—was made by F. C. Donders (1818–1889).

Donders's idea, called the method of subtraction, was to estimate the time taken by different components of a mental task. The tasks he used are now called simple reactions, go/no go reactions, and choice reactions. These tasks, he reasoned, could be broken down into smaller stages of processing: perceptual encoding, stimulus identification, response selection, and response execution.

[Page 631]Consider, for example, a task where an observer must respond to the presentation of red and green lights by pressing a button. For a simple reaction, an observer presses a button as soon as he sees any light, no matter what color it is. Donders reasoned that this task could be performed only with perceptual encoding and response execution. For a go/no go reaction, however, the observer presses the button only when the light is green. This requires not only the perceptual encoding and response execution stages but also stimulus identification. For a choice reaction, the observer presses one button for a green light and a different button for a red light. The task now requires response selection, by requiring an observer to determine which of the two possible buttons are appropriate for a presented stimulus.

Donders measured his observers’ RTs in the simple, go/no go, and choice reaction tasks. To estimate the duration of the stimulus identification stage, he subtracted the simple RT from the go/no go RT. To estimate the duration of the response selection stage, he subtracted the go/no go RT from the choice RT. This method of subtraction assumed that the task stages were arranged in serial order (so no two stages could be operating at the same time) and that the stages were independent from each other (so if one stage took a very long time to complete, that would not affect how long it took any other stage to complete).

...