The Difference of Main Recall Between Pictures and Words When Presented on a Background or No Background.
Tara M. O’Donnell
State University of New York at StonyBrook
The picture superiority effect has been studied for many years and many theories have been proposed to explain it. Craik and Loackhart (1972) have found that pictures are always remembered better than words. Swets and Birdsall (1967) on the other hand have used the signal detection theory to try to explain how the picture superiority effect could be reversed. The signal detection theory states that it takes longer to find critical items when there is other extra information in the way. This is what we hope to prove. By showing participants pictures and words on a background or no background we hope to show that words will be remembered better when on a background. This hypothesis is supported by the signal detection theory because we believe that it will take longer to find the picture on a background, giving the participants less time to encode and store it, thus making retrieval of the item harder.
The Difference of Main Recall Between Pictures and Words When Presented on a Background or No Background
An experiment was performed to test the finding that pictures are remembered better than words. This finding has been termed the “picture superiority effect” and has been demonstrated in numerous experiments (Craik and Loackhart, 1972, and Paivio, 1986). In a typical experiment the participants were asked to study a mixed list of words and pictures and were then asked to recall the studied items. According to Paivio, Rogers, and Smythe (1968) it was found that the pictures of objects were recalled significantly better than their names. Over the years, many theories have been proposed to explain the picture superiority effect.
One possible explanation that has been proposed concerns levels of processing as given by Craik and Loackhart (1972). Based on levels of processing, it has been proposed that there is a deeper level of encoding for pictures than for words. There are three levels of encoding; graphemic, phametic, and semantic (Craik and Loackhart, 1972). Words are remembered more in the first two levels, where as pictures are remembered better at the third and deepest level (Craik and Loackhart, 1972). The deeper the level, the better something is encoded, and thus, remembered and retrieved quicker and better.
Another explanation for the picture superiority effect is the dual coding hypothesis. Paivio (1986) states that the verbal and pictorial information are stored independently, but are interconnected. In other words, independence means that one system may be active while the other is not, or that both can be active at the same time. Interconnected-
ness, on the other hand, suggests that one system can be activated by the other through related pathways that connect verbal and pictorial units. Pictures are more likely to elicit both verbal and pictorial codes than are words, which could be a reason why they are remembered better.
Although several studies have shown the picture superiority effect may be explained by the theories above, other theories such as signal detection suggest the picture superiority effect may be reversed (words remembered better than pictures). Since a problem with encoding might have to do with the interference of encoding in which extraneous information interferes with the critical items, we proposed that words would be remembered better than pictures on a background. According to signal detection theory, it takes longer to find the critical items when there is a background (Swets and Birdsall, 1967).
To test this hypothesis we used a mixed design in which all participants viewed both pictures and words, but one group viewed the critical pictures and words on a background and the other group viewed the same items on no background
Based on the theories discussed above, we predict that pictures will be better remembered than words when they are both presented on a plain white background. However, we predicted that words would be remembered better than pictures if they were presented on a background. This outcome would support the signal detection theory. Two other hypotheses were also derived for the picture superiority effect, they were that regardless of the fact if the stimuli was a word or a picture there would be no difference when presented on no background. The second alternative hypothesis is that pictures will always be remembered better than words whether or not if they are on a background or no background.
There were a total of 60 participants who took place in this experiment. Some were volunteers and others were recruited from the psychology subject pool. All of the participants were students at the State University of New York at StonyBrook. The participants were both male and female of all different ethnical backgrounds. All participants were treated in accordance with the American Psychological Association ethical standards.
A two by two mixed factorial design was used for this experiment, with the type of background varying between subjects (no background or scenic background) and type of target stimulus (pictures or words) manipulated within subjects. We chose 20 common objects (e.g., ambulance, house, tree, etc.) as the critical items. Counterbalancing was achieved by rotating the 20 items through two lists, as either a picture or a word, in both the background and no background conditions.
The scenic backgrounds and picture files that we used were downloaded from the Web. The pictures and words were both superimposed on the scenic backgrounds. All the stimuli were centered on the screen with a border around them. Both the pictures and words were saved on the computer as BMP files in PAINT. The words were typed in Arial, with a font size of 26 and were in bold.
The computer used for this experiment was a Dell Dimension XPSM 200S with a monitor size of 35.56 cm. The program that was used to create the experiment and present the stimuli to the participants was Superlab. The stimuli remained on the screen for a duration of 2000 ms each, with an inter-stimulus interval of 500 ms. The recall test consisted of a blank piece of paper and something for the participants to write with.
When the participants first arrived they were asked to read and sign a consent form. Having that done they were sat in front of the computer, a normal distance from the monitor. The first thing that was presented on the computer to the participants were the instructions. The instructions read as follows:
You will be presented with a series of items. Please pay attention to them.
Hit the space bar to begin.
They were not told that they would be getting a memory test later. They were then asked if they had any questions before starting. After pressing the space bar the stimuli was presented to them at a fixed interval until all 20 items had been seen.
When all items were seen, the experimenter noted the time. For the first seven and a half minutes after they saw the stimuli they were given a piece of blank paper and told to write down all the television shows that they could think of. After the time was up the participants were then asked to play solitaire on the computer for another seven and a half minutes. If they did not know how the experimenter briefly taught them how. After that time was up they were given a second piece of blank paper with their designated subject number on it and asked to recall as many of the bordered items as they could. When five minutes were up we stopped them and then debriefed them on the experiment. The participants were told that this was a memory experiment designed to see whether or not pictures are remembered better than words and if there was a difference when they were presented on either a scenic background or not. They were also told that the two additional tasks that they were asked to do were distracter tasks so they would not go over the items in their heads.
A mixed design was used in this experiment. There were two levels of the independent variable, pictures and words presented either on a background or no background. We analyzed the means and standard deviations for all four conditions. The mean was highest for pictures presented on no background (M = 4.67), followed by pictures on a background (M = 4.3), followed by words on no background
(M = 3.63), followed by words on a background (M = 3.4). The standard deviation was highest for words presented on a background (SD = 2.28), followed by words on no background
(SD = 2.01), followed by pictures on no background (SD = 1.71), followed by pictures on a background (SD = 1.67). We conducted a two by two mixed factorial ANOVA with context as a between subjects factor (background vs. no background) and stimulus type as a within subjects factor (words vs. pictures). No significant difference (F