The serial position effect

¶ … Distractor Task on Serial Position Effect

Serial Position Effect

The serial position effect has been attributed to the labile nature of the short-term memory and the resilience of long-term memory. This theory implies that lengthening the delay between presentation of the last words on the list would impair the recall of these words (recency effect), but not the first words presented (primacy effect). This theory also implies that if test subjects become preoccupied with an unrelated cognitive task in between word presentation and recall, the last words on the list would tend to be forgotten. To test this theory, a 17-item word list was presented to students at a rate of one word per second and half the students were allowed to immediately recall the words, while the other half counted backwards from 300 in threes for 30 seconds before recall. The graphical presentation of the data is consistent with the findings of other researchers and reveals little impact on the primacy effect and a dramatic negative impact on the recency effect. However, when the students t test was used on the item means for the two conditions, only the first four words were significantly different. This latter finding indicates the raw data should have been compared, rather than the item means. Despite this limitation, these results confirm that a short delay impairs the recency effect by interfering with short-term memory.

Impact of Distractor Task on Serial Position Effect

The serial position effect is a well-established phenomenon that reflects human memory processes (reviewed by Murdock, 1962). When a listener is given a word list of about a dozen or more words and is asked to recall them, they will tend to remember the first and last words better than the middle words; thus the position of the words in the list determines how well the human mind is able to recall them. The terms used to describe the differential ability to remember the first and last words is called the primacy and recency effects, respectively.

A number of different variables can be used to influence the shape of the serial position effect curve. For example, short word lists tend to increase the recall rate for middle words, such that the serial position effect curve as a whole moves upward along the vertical axis (Murdock, 1962). The primacy and recency effect is therefore suppressed as the word list lengthens. Both the primacy and recency effect were also independent of the content of the word lists. Such studies reveal the dynamic nature of the serial position effect under different conditions.

Glanzer and Cunitz (1966) tested whether the interval between presentation of the word list and the recall task had an impact on the serial position curve. They proposed that the primacy and recency effects were the products of long-term and short-term memory processes, respectively, and therefore the delay would have an impact only on the recency effect. Their findings supported their hypotheses by revealing a significant difference in recall rates for the last words on the list when there was a 10 or 30 second delay (p < 0.001).

This project will build upon the work of Glanzer and Cunitz (1966) by assessing the effect of a 30 second distractor task inserted between presentation and recall of the word list. We propose that the distractor task will have a no impact on the primacy effect for the first four items in the word list when compared to immediate recall rates. We also propose that the distractor task will have a significant negative impact on the recency effect.

Materials and Methods

Participants

To investigate the impact of a distractor task on the recency effect, a convenience sample of 300 undergraduate psychology students enrolled in a cognitive psychology class (HPS203/773) at Deakin University was utilized. Other demographic information, such as age and gender, were not collected under the assumption they would not have a significant impact on the dependent variable. All subjects enrolled in the study provided informed consent.

Materials

A word list with 17 words, matched for frequency and concreteness, was used to test the serial position effect. The distractor task involved counting backwards from 300 in threes for 30 seconds.

Procedure

The word list was presented to study subjects verbally, at a rate of one word per second in a monotone voice. The recall task was performed after the complete 17-word list was given to the test subjects and under two distinct conditions. The first condition was to recall as many words as possible immediately after the word list was finished. The second condition involved engaging in the distractor task immediately after hearing the word list and then trying to remember as many words as possible. Subjects were not required to recall the words in any specific order and were given an unlimited amount of time to complete the task. The average time for completing the recall task was 2 minutes.

Results

The recall rates for the two conditions, immediate and delayed recall, were noticeably different (Figure 1). The primacy words were recalled at an equivalent rate between the two conditions, as were the middle words, but the recall of the recency words were dramatically impacted by the delayed recall condition. Based on the data presented in Figure 1, the last three words in the 17-word list could not be recalled any better than the middle words when the study subjects were required to count backward from 300 in threes.

The means and standard deviations for the three serial position categories are presented in Table 1. Based on this data the recall rate for primacy words were similar, but the small difference reached statistical significance [t (4) = 0.030, two-tailed] because the students in the immediate recall condition did consistently better at recalling the first four words of the list. In contrast, when the means for the middle words were examined the students in the delayed condition did slightly better on average than the students in the immediate recall condition.

As discussed above, there was a big difference in the ability of students to recall the last few words of the list based on whether they were tasked with counting backward from 300 in threes for 30 seconds. However, When the last seven words in the list were averaged the difference was not statistically significant [t (6) = 0.243, two-tailed]. Even though the means for immediate and delayed recall were 39.73 and 29.52, respectively, the variation in the immediate recall data for recency words was high (SD = 20.64). If the comparison is limited to the last 5 items according to Glanzer and Cunitz (1966), then the difference is significant at the 90% probability level [t (6) = 0.097, two-tailed].

Figure 1: Word Recall Rates Based on Serial Position. The individual item recall rates for the immediate and delayed conditions are depicted with black squares and diamonds, respectively.

Table 1

Recall Means and Standard Deviations Based on Serial Position

Condition

Word Serial Position

Mean

Std. Dev.

Immediate

Primacy

75.18

16.87

Middle

24.57

8.28

Recency

39.73

20.64

Delayed

Primacy

70.43

15.86

Middle

30.66

10.12

Recency

29.52

9.08

Discussion

We proposed that the 30 second distractor task would have no effect on the primacy effect and have a significant impact on the recency effect. The findings presented here, from a statistical standpoint, revealed the opposite. These results are controversial, because it contradicts the findings of Glanzer and Cunitz (1966). However, when viewed graphically in Figure 1 the negative impact of the distractor task was limited to the recency effect. This suggests the use of a student t test on the item means, rather than the raw data, produced a statistical artifact.