Mental Processes Revealed by Reaction Time Experiments

This experiment is based on the following paper:

Sternberg, S. (1966). High-speed scanning in human memory. Sternberg, S. (1966).Sternberg, S. (1966). High-speed scanning in human memory. Science, 66, 652-654. Reprinted by permission of the author, S. Sternberg.


Mental functions are often conceived as a series of stages through which stimuli are processed. In a simple reaction time (RT) task, the RT is the sum of the time it takes all functions to complete their operations. Sternberg proposed to decompose the RT into its different processing stages, thereby revealing the sequence and processing time of different mental functions.

The results from experiment one, a memory recall task, are described here. In this study, Sternberg explored the functions of recognition and memory recall using an experimental task in which subjects compared a target stimulus to a given stimulus set, and determined whether or not it was a member of that set. He proposed that theoretically this task could be accomplished by two types of a serial searches.

Let serial search (or scanning) be a process in which each of a set of items is compared one at a time, and no more than once, to a target item. …The purpose of the search is to determine whether an agreement (or match) exists between the test item and any of the items in the memorized set. Two types of serial search that might serve this purpose need to be considered. In self- terminating serial search, the test stimulus is compared successively to one item in memory after another, either until a match occurs (leading to a positive response), or until all comparisons have been completed without a match (leading to a negative response). In exhaustive serial search, the test stimulus is compared successively to all the memorized items. Only then is a response made – positive if a match has occurred, and negative otherwise. (p.426-427)

Sternberg proposed that exhaustive search could be distinguished from self- terminating searches by the way in which a target stimulus was positively or negatively recalled.

In an exhaustive search the test stimulus is compared to all items in memory before each positive response as well as before a negative response. Hence, the rate at which RT increases with list length–the slope of the RT-function–is the same for positive and negative responses. In contrast, a self-terminating search stops in the middle of the list, on the average, before positive responses, but continues through the entire list before negatives. The result is that as list length is increased, the latency of positive responses increases at half the rate of the increase for negatives. (p.427-428)


The full text describing the experiment is available here.

The experiments are available below. (18.2 KB) (24 KB)