How much information will an individual store in his or her iconic memory? In a given time, individuals are able to perceive more information than they can verbally encode. This is the process by which a visual stimulus is transformed to neurons to enable the brain to store information in the immediate memory. The rate of transfer is how fast an individual can encode something in a given time, which is stored in the immediate or short-term memory. It is stated that visual input can be stored in some medium, that later will be recalled. When the duration of the stimulus is limited, information is not properly encoded from a stimulus to a verbal code and it is lost from immediate memory. This is a cognitive process. The term “icon” was introduced by Neisser (1967) to refer to the brief persistence of information from a visual display after the “display” is no longer present. Early experimenters, such as Erdman and Dodge (1898), had been concerned with this phenomenon and asked how much information could be acquired at a single fixation reading.
The typical finding from briefly presenting a set of letters and having the subjects report as many letters as possible (full-report) was that the perceptual span was 4 to 5 letters. In replicating Sperling’s experiment we hope to see why iconic memory as well as duration recall is limited. He has shown that the duration of an icon has to do with the ability of an individual to encode the visual information. One limitation to the study of the icon is, its very brief duration. As individuals begin reporting the contents of the icon, it is already disappearing. Sperling invented the partial-report technique to overcome this difficulty. His third experiment is being replicated to understand the problem at hand. He used manipulations to control the rate at which information is verbally coded as well as the span of apprehension (memory). Consequentlyб increasing the ability of subjects to report more of the information available to them.
Sperling’s test presented a 3 x 3 matrix of letters, flashed for 50 milliseconds (msec.). Instead of using the full-report method of previous studies (report as much as you can of the whole display), he used a new partial-report method. In his within subjects study, participants saw the display of letters, and then heard a tone that was high-pitched, medium-pitched or low-pitched. The high-pitched tone indicated to report the top line of the display. The medium tone indicated to report the middle line of the display. The low tone indicated to report the bottom line of the display.
According to Sperling’s research, individuals were able to report 80% of the letters on a line, if the tone was sounded just before the letters were presented. That drops only slightly, to about 75% when the tone immediately follows the letters. Sperling’s experiment showed that the limit in report of letters from a brief visual display was due to a rapid decay of the icon. The longer the icon or stimulus is present the more information one can encode and therefore recall. Limited duration recall is due to the rate of decay of the icon. The longer an individual waits to recall a given stimulus the less he or she can recall. When knowing the duration and the capacity of the visual sensory register one can, to a point, determine if the subject have a better chance of remembering certain in formation of any visual stimulus.
Sixteen Experimental Psychology students from Queens College participated in the experiment to meet a course requirement. The mean age was 22.61 years with a standard deviation of 2.89. The mean years of education were 15.44 with a standard deviation of 0.98.
The experiment was conducted with 16 MEL Lab manuals as well as formatted 3.5” disks MEL Lab. The experiments were run in MS-DOS mode; on 16 different IBM compatible computers with Microsoft Windows 95 program that are kept in the Psychology laboratory at Queens College.
The independent variable was the delay of instruction tone. The delay occurred in three intervals or levels of 0, 150, and 500 msec. between the presentation and the instruction tone. The dependent variable is the percent of accuracy of letters recalled.
To control for attention as well as perception subjects were instructed to perform the experiment in an isolated room. Subjects were told, if possible, to perform the experiment in a dimly lighted room to control for the brightness of the icons on the screen. There were 120 practice trials before the actual experiment, to allow the subjects vision to adapt to the light and get used to the tones. All practice trails involved partial report with no delay between presentation of the letters and the tone to indicate which line to report. The actual experiment will test the subject in four blocks of 30 trials each. There was a 100msec presentation of a matrix of letters that consisted of three rows of four letters each in each trial. The 12 letters presented were randomly chosen from the 21 consonants in the alphabet.
One of the block of trials will require full report, where the subject must report as many letters as they can from anywhere in the display; this trial would not have instruction tone. The three remaining trial will have an instruction tone, but differ in the delay of the instruction tone (0, 150, and 500 msec). Over subjects, presentation of conditions was counter balanced in a Latin Square design. This was done to control for sequence effect in the within-subject design. A high-pitched, medium-pitched or low-pitched tone indicated which line to report. The high-pitched tone indicated to report the top line of the display. The medium tone indicated to report the middle line of the display. The low tone indicated to report the bottom line of the display. The rows of letters to be report were selected at random. However, after the cue they had to enter as many letters they could from the row specified.
The instructor of Experimental Psychology 213 first instructed the students to read section 1.2 in their MEL LAB manual, to become familiar with the experiment to be conducted, and to bring their MEL LAB disk. Later each subject was told to go to a cubical, where they would find an IBM compatible computer to run their experiment in MS-DOS mode. They were instructed to close the door of their perspective cubicles before they start experiment. Subjects were also informed that displays used in this experiment were only visible for 1/10th of a second. So subjects had to be sure they were looking right at the focus cross when they start each trial.
To run the experiment each subject inserted the formatted 3.5” disk into the computer. Then right clicked on the start icon and selected shut down. Afterwards a screen appeared where they had to select Restart in MS-DOS mode and enter A: EXPERIMENT. With the arrow keys, they selected 1.0 Experiments in Perception from the topic menu. In an experiment selection screen, again with the arrow keys, they selected 1.2 The Duration of the Icon. Afterwards they selected the mode for the experiment, NORMAL. This mode allowed the individuals to serve as subjects in the entire experiment. From this point on the subjects followed the instructions on the computer screen; that explained the procedures for each trial and what keys they had to use for their responses.
The 120 practice trials started and were followed by the actual experiment. The experiment tested the subjects in four blocks of 30 trials each. The order for the four blocks were counter balanced across subjects with a Latin Square design. One of the block of trials required full report. The three trials remained required partial report. They differed in delay of the instruction tone (0, 150, and 500 msec). Tones were randomized. A high, medium or low-pitched tone indicated which line had to be reported. The high-pitched tone indicated to report the top line of the display. The medium tone indicated to report the middle line of the display. The low tone indicated to report the bottom line of the display.
There was a significant effect in the delay of instruction tone between conditions found. Table 1 reports the mean, the standard deviation, and the sample size for each condition. The actual trials were used to analysis the difference for the descriptive statistics and inferential statistic. The partial report of 0 msec. had a mean of 57.09 and a standard deviation 32.65 for the delay of instruction tone. For the 150 msec. condition there was a mean of 50.08 and a standard deviation 33.90 for the delay of instruction tone. For the 500 msec. condition there was a mean of 41.40 and a standard deviation 31.51 for the delay of instruction tone. The full report had a mean of 46.37 and a standard deviation 29.90 for the delay of instruction tone. Table 2 reports the repeated-measures ANOVA (N=16) since there was more than one condition. Please refer to table 2 for the Df, Ms, F, and P. An independent F-test yielded a F (2,128) = 31.46, p * .05. Table 3 illustrates how the percent of recall varies as a function over the recall condition.
According to our findings, individuals are able to perceive more information then they can verbally encode. This is illustrated by the fact there was a significant difference found in the F-statistic. The ANOVA done on the partial report helps to illustrate that short-term memory is limited as well as duration recall. The experiment showed that the limit in report of letters from a brief visual display was due to a rapid decay of the icon. The longer the icon or stimulus is present the more information one can encode and therefore recall. Limited duration recall is due to the rate of decay of the icon. The longer an individual waits to recall a given stimulus the less they are able to recall. When knowing the duration and the capacity of the visual sensory register one can, to a point, determine if the subject have a better chance of remembering certain in formation of any visual stimulus. This information is very useful in everyday accounts such as eyewitness incidents, where we are able to estimate how much the witness really saw and how much is made up. We are also able to use this when it comes to exams, how much time is really need to study a given amount of work.
- Erdman, B., & Dodge, R. (Psychological studies on reading.) Sited in R. Woodworth & H. Schlossberg Experimental Psychology. New York: Holt, 1958.
- James, J., Schneider, W., and Rogers, K. MEL LAB: Experiments in Perception, Cognition, Social Psychology & Human Factors. Pittsburgh: Psychology Software Tools Inc., 1994.
- Neisser, U. (1967). Cognitive Psychology. New York: Appleton-Century-Crofts.
- Sperling, G. (1960). The information available in brief visual presentations. Psychological Monographs, 74 (Whole number 11).