Running head: MNEMONICS AND SERIAL POSITION

 

 

 

 

 

 

 

 

 

 

The Effect of Imagery and Narrative Mnemonics on Serial Position Recall

Katelyn R. Boguski

Saint Bonaventure University

Abstract

This study was designed to see if the serial-position effect could be differentially influenced by varying mnemonics. Forty-five participants were each given three word lists to account for serial position. The participants were divided into three conditions, which required the use of either the narrative technique, imagery, or no mnemonic to study the words. After studying each word list they were asked to recall as many words as they could. A 3 x 3 ANOVA was conducted. Results indicated a main effect for serial position but no main effect for type of mnemonic, nor an interaction. These results challenge the notion that mnemonics can influence the serial-position curve, which had been shown in previous research by Persensky and Senter (1969) and Boneau (1990).

The Effect of Imagery and Narrative Mnemonics

on Serial Position Recall

Mnemonics are defined as a specific reconstruction of target content that is meant to link new information more closely to the learner’s existing knowledge and thus facilitate retrieval (Scruggs & Mastropieri, 1990, as cited in Alper, Eggers, & Raschke, 1999). They are used to help learn new information. Using a mnemonic allows a connection between what one already knows and what they are trying to learn. One type of mnemonic is the narrative technique. This technique instructs people to develop stories of any kind that link the series of words that they are trying to learn together. As cited in Matlin (2002), Bower and Clark (1969) conducted an experiment in which they told a group of people to make up narrative stories using sets of words. There were 12 lists for this narrative group as well as for a control group. The control group was told to study and learn each list. The results showed that the people in the narrative technique group recalled about six times as many words as the control group did. However, the narrative technique is only effective if the person can generate the narrative easily and reliably during learning and recall.

            Imagery as a mnemonic device involves developing mental representations of objects or actions that are not physically present (Matlin, 2002). Previous research has shown that visual imagery is an effective means of improving memory. A study was done with pairs of nouns to test participants in different conditions. In the repetition condition participants repeated the pairs of words silently to themselves and in the imagery condition participants tried to construct a mental picture of each word-pair. The results were that participants in the imagery condition recalled over twice as many words as did those in the repetition condition.

            According to research by Persensky and Senter (1969) mnemonics reduce or distort the serial-position effect. The serial-position of a word is where it is placed in a list of words in relation to the other words. The serial-position effect is the U-shaped relationship between a word’s position on a list and its probability of recall (Matlin, 2002). This curve illustrates strong primacy and recency effects. A primacy effect is better recall for items at the beginning of the list; a recency effect is better recall for items at the end of a list (Matlin, 2002). Atkinson and Shiffrin (1971) did research which showed that the primacy region of the curve is evidence for retrieval from long-term storage, whereas the recency region of the curve is evidence for retrieval from short-term storage. This says that the information that is recalled from the beginning of the serial-position curve shows which material has made it into long-term memory (LTM), whereas the information that is recalled from the end of the serial-position curve has only made it into short term memory (STM). The information in the middle has been lost from STM but not much of it has made it into LTM.

            After learning that mnemonics distort the serial-position curve I was curious to see which mnemonics do this, it not all of them. Boneau (1990) did research which found that imagery distorts the curve by producing a smaller recency effect. However, each mnemonic is different and so I wondered if different mnemonics would have different effects on the serial-position curve. Looking at the curve resulting from using different mnemonic techniques can show how the mnemonic affects memory storage. The curve will show what influences LTM by looking at the primacy effect and what influences STM by looking at the recency effect.

I am looking at where a recalled word is on a list, to see its serial-position, for the two different types of mnemonics and a control. The variable of serial-position has conditions which are the beginning of the list, middle of the list or end of the list. The variable of mnemonics has conditions of either narrative technique, imagery, or no mnemonic. The variable being measured is the number of words recalled.

The mnemonic conditions are using elaborative rehearsal which gets in the way of maintenance rehearsal. Elaborative rehearsal is a form of learning information that consists of doing more than just repeating the words silently to yourself, which is what is done in maintenance rehearsal. However, the narrative technique is a different type of mnemonic than the imagery mnemonic and thus should produce a different effect than the imagery mnemonic. The narrative technique connects the entire list of words into a story. This connection helps in recall if the participant can remember the basis of their story. The story should provide cues to the words that they were studying as they think of it.  Imagery uses elaborative rehearsal, but it is not forming the same connection that the narrative technique does. Therefore, the curve for the narrative technique should be flatter than imagery condition’s curve because it is forming a connection based on what is in LTM and the other words on the list whereas imagery is only forming a connection based on what is in LTM. For imagery there should be less of a recency effect based on what was found by Boneau (1990). The no mnemonics condition will be using maintenance rehearsal to study the words and thus should produce the normal U-shaped curve for serial-position. My hypothesis is that due to the nature of each mnemonic, there will be different serial-position curves in each condition.  

 

Method

Participants

            There were 45 participants in this experiment, 15 in each condition. All of them were students (at St. Bonaventure University). 

Materials 

There were 45 packets. Each packet had three lists of words with a blank page after each list. Each list had 30 words so that on each page there were 10 words in the beginning, 10 words in the middle, and 10 words at the end. This was to account for serial position. The words on each list were chosen randomly from Battig and Montague (1969). This was done for each word by randomly picking a page in the book and then using the most common word on that page. Common words were used so that they would not be difficult to imagine or to make up a narrative from. For the narrative technique condition there were 15 sheets with instructions that said to make up a narrative with the words on the sheet to help study the words that would be given. For the imagery condition there were 15 sheets with instructions that said to use imagery by forming a visual image to help study the words that would be given. For the no mnemonic condition there were 15 sheets that said to study the words that would be given. Each participant received the same packet of word lists, and the instructions they got depended on which condition they were in. A watch was used for timing.

Design and Procedure

            This was a 3 x 3 design which was within subjects for the serial position and between subjects for the mnemonic variable. The instruction sheets were attached to the word packets and given out at random as people came in. Everyone was told to read the sheet. Once everyone read the sheet I told them to follow the instructions and start by turning to the first sheet and that I would tell them when to stop. After the time was up I told them to turn the page and write down as many words as they could remember. This was repeated for the other two lists.

Results

The nine scores for each person were averaged so that there was one number for each of the three serial position conditions, therefore three scores for each participant. This resulted in 15 scores in each serial-position for each type of mnemonic. They were grouped in a 3 x 3 matrix where each condition contained 15 scores. The means and standard deviations for each condition are shown in Table I and graphed in Figure I. An ANOVA was performed, which can be seen in Table II. The alpha level of .05 was used for all statistical tests. The effect of type of mnemonic, F(2, 44) = .21,  p = .81, was not statistically significant. The effect of serial position, F(2, 90) = 37.47,  p < .0001, was statistically significant. The interaction between the two variables, F(4, 84) = 1.38, p = .25, was not statistically significant. 

Discussion

The results did not support my hypothesis. My hypothesis was that there would be an interaction between the variable, type of mnemonic, and the variable, serial-position.  There was no interaction between the variables; therefore the type of mnemonic and serial position did not have an effect on the amount of words recalled. There was a main effect for serial position which says that the words were recalled either more or less depending on where they were on the list. There was not a main effect for type of mnemonic which means that the scores were not different depending on what the instructions said.     

The main effect for serial-position shows that there was a relationship between where the words were on the list and how well they were recalled, but it was not consistent with the normal U-shaped serial-position curve for any of the conditions. The means for each serial-position decrease as serial-position goes from beginning to middle to end.

It was surprising that there was not any interaction between the variables or a main effect for the type of mnemonic. Persensky and Senter (1969) had said that mnemonics distort the serial-position curve. My results show a distortion of the U-shaped curve for both mnemonics conditions, but also for the no mnemonic condition which served as a control.  Therefore, this does not support my hypothesis because the control did not produce the desired effect.  Previous research by Boneau (1990) has shown that imagery mnemonics can produce less of a recency effect therefore distorting the curve. This curve is not a normal U-shaped curve, and in this experiment that was shown for all types of mnemonics. It seems like there is less of a recency effect for all of the conditions. Maybe if the participants were given less time more of the words would stay in STM, thereby showing a recency effect. Previous research (Scruggs & Mastropieri, 1990, as cited in Alper, Eggers, & Raschke, 1999) has also shown that mnemonics facilitate retrieval.  Therefore, there should have been more words recalled in each mnemonic condition compared to the no mnemonic condition but there was no significant difference which does not support previous research. 

Participants were tested together and thus there were the normal distractions that always occur. Some people would talk and maybe this disturbed others. Perhaps the setting effected how serious participants were when they did the experiment. If they did not take it seriously they may not have done exactly what the instructions said to do. Maybe if participants were tested separately in a room with no distractions the results would be more consistent with the hypothesis, or at least with previous research.

While scoring the lists of recalled words I noticed that most participants recalled the same words. I got so accustomed to seeing the same words that when a different word was on the list I was not even sure if that word even appeared on the word lists that they studied from. Although the word lists were made by randomly selecting words,  some participants said that some words were just easier to remember than others. I thought that since they were all common words they would all be equally likely to be remembered. Maybe if some of the lists had hard words and other had all easy words it would be more effective. 

           Unfortunately the results were not consistent with previous research or the hypothesis.  This could have been due to the methods used and therefore it might be advantageous to do a similar experiment while changing some of the methods that may have caused problems. 

 

References

Alper, S., Eggers, E., & Raschke, D. (1999). Recalling alphabet letter names: A

mnemonic system to facilitate learning. Preventing School Failure, 43, 80-84.  

Atkinson, R. C., & Shiffrin, R. M. (1971). The control of short-term memory. Scientific

American, 225, 82-90.

Battig, W. F., & Montague, W. E. (1969). Category norms for verbal items in 56

categories: A replication and extention of the Connecticut category norms.

Journal of Experimental Psychology Monograph, 80, 25-29.

Boneau, C.A. (1990). Short-term recognition memory under rehearsal instructions and

imaging instructions. Bulletin of the Psychonomic Society, 28 (4), 297-299.

Matlin, M.W. (2002). Cognition (5th ed.). New York: Wadsworth.

Persensky, J. J., & Senter, R. J. (1969). An experimental investigation of a mnemonic

system in recall. Psychological Record, 19(3), 491-499.

 

 

Table I  

 

Words Recalled at Each Serial Position

 

 

 

 

 

 

Standard

 

 

 

 

 

 

 

 

Type of

 

Mean

 

 

deviation

 

 

 

 

 

 

 

 

mnemonic

SP 1

SP2

SP3

SP 1

SP2

SP3

 

 

 

 

 

 

 

No mnemonic

5.813

4.261

4.21

2.259

2.016

2.367

 

 

 

 

 

 

 

Imagery

6.568

4.085

3.203

2.158

1.745

2.033

 

 

 

 

 

 

 

Narrative

6.533

4.103

3.857

1.961

2.289

2.149

 

 

 

 

 

 

 

 

Table II

Analysis of Variance for Type of Mnemonic and Serial Position

Source

SS

df

MS

F

P

Between Subjects

104.96

44

 

 

 

Type of Mnemonic (A)

1.06

2

.53

.21

.811

Subjects within A

103.9

42

2.47

 

 

Within Subjects

372.03

90

 

 

 

Serial Position (B)

169.36

2

84.68

37.47

<.0001

A x B

12.52

4

3.13

1.38

.248

B x Subjects within A

190.15

84

2.26

 

 

TOTAL

476.99

134

 

 

 

 

 

Figure Caption

Figure I. Mean words recalled as a function of type of mnemonic and serial-position.