What is the serial position effect?
How the first and last items of a list come to be remembered best, what moves the two ends of the curve, and what retests have added.
Answer to the question
What is the serial position effect?
The serial position effect is the tendency to remember the first and last items of a list better than the middle ones. Bennet Murdock described the curve in 1962. Later work shows the two ends behave differently: a 30-second distraction removes the recency advantage, and a final test can reverse it.
Key takeaways
- Bennet Murdock's 1962 paper described a steep primacy effect over the first three or four words, an S-shaped recency effect over the last eight, and a flat stretch between them.
- In published accounts of Glanzer and Cunitz's 1966 work, counting aloud for 30 seconds after a list removed the recency advantage, while slower presentation helped early and middle items.
- In final free recall, Kuhn, Lohnas and Kahana (2018) replicated negative recency within lists: end-of-list items were recalled less well than middle items.
- Kelly and Risko (2019) found that offloading to notes weakened early items and removed the primacy effect; their 2022 repeat judged any such cost small and less robust than first reported.
- Daniel Levitin's The Organized Mind draws on a word-list demonstration to argue that memory works best when the environment holds part of the load.
The words that win the first test can come last on the final one. Joel Kuhn and colleagues found it by testing listeners on every list again at the end of a whole session: items from the end of each list were then recalled less well than items from the middle. The turn sits against one of the oldest and most repeated findings in memory research, the serial position effect, in which the first few and last few words of a list come back more often than the ones between. What sixty years of work have added is a clearer view of what bends the curve, and the two ends of it respond to different things.
How was the serial position effect first measured?
Bennet Murdock gave the curve its standard description in 1962. Words at the start of a list showed a steep primacy effect over roughly the first three or four items. The last eight or so words showed an S-shaped recency effect, and a flat stretch of lower recall ran between them. His paper called it the serial position effect of free recall, free recall meaning that listeners may report the words in any order they like.
A free recall test has three parts:
- A list of unrelated words is read or shown one at a time.
- An optional task, such as counting aloud, fills the gap before recall.
- Listeners report as many words as they can, in any order.
Four years later Murray Glanzer and Anita Cunitz asked whether the two ends of the curve come from one process or two. In accounts of their work, slowing the pace at which words were shown improved recall of early and middle items and left the final words alone. Having listeners count aloud for 30 seconds after the list did the reverse: the recency advantage disappeared and the early items were barely touched. Two manipulations moved two different ends of the curve.
The authors read this as evidence for two storage mechanisms, a durable one that early words reach through attention and repetition, and a short-lived one that holds the latest words until something displaces them.
Daniel Levitin runs a word-list demonstration of his own in The Organized Mind, and readers tend to recall the opening word more often than the closing one. His larger point is that memory works best when the environment carries part of the load, which makes the serial position effect a reason to write things down and not a verdict on anyone's ability. A list that stays on the page does not have to survive the middle of someone's memory.
What did later tests do to the curve?
The serial position effect looks fixed on a single test, yet the same end-of-list words that win a first test can lose a later one. Joel Kuhn, Lynn Lohnas and Michael Kahana studied final free recall, where listeners are tested on every list again at the end of a whole session. Within each list, items from the end were recalled less well than items from the middle, a result called negative recency. Across lists the pattern ran the usual way, with more words coming back from recent lists than from distant ones.
Their proposal was a spacing account. Words recalled early in the first test are repeated close together, and close-together repetition pays less than repetition spread out. The reversal was largest for words recalled early and remained when a task between words limited rehearsal.
Offloading adds another bend to the serial position effect. Megan Kelly and Evan Risko had undergraduates hear lists of 20 words and write them down, with their notes available at recall on the first three trials and withheld on the last. In two experiments of 64 participants each, memory for the early items suffered and the usual primacy effect vanished, while the recency effect stayed. Their 2022 repeat matched the earlier work in most respects. It found primacy in both conditions and concluded that any cost of offloading to the early items is small and less robust than first reported. Participants who leaned harder on their notes did worse once the notes were gone.
When does the serial position effect matter?
The serial position effect describes one test at one moment, and the moment decides which end wins. Where recall is immediate, the closing items are easy to retrieve, and the evidence above says that advantage fades quickly once attention moves on. The opening items are the steadier asset. Anyone who saves a favourite for last, a habit taken up in Why We Save the Best for Last, is betting on the end of the list, which is the part of the serial position effect least likely to survive a delay.
Would the curve show up on a shopping list, a slide deck or a menu? Nobody in these studies tried, so what follows is our inference and not a result: when something must be remembered later, put the item that matters most early, and give the middle a cue or a written place to live. Writing a list down hands the job to paper, the mechanism covered in What is cognitive offloading, and is it making us forget.
Order also steers choices, not only recall. The pull of the first position on what people pick is set out in Why the First Option Gets Picked, and it sits beside the memory result as a separate finding. In memory, the serial position effect shows that where an item stands in a sequence is a feature of the item, as real as its meaning. The position that helps first can cost later.
Common questions
What is the serial position effect?
It is the tendency to remember items from the start and end of a list better than items from the middle. Bennet Murdock described the curve for free recall in 1962.
What is the difference between primacy and recency?
Primacy is the advantage of the first few items, and recency is the advantage of the last few. Glanzer and Cunitz found that a 30-second counting task removed recency and left the early items alone.
Has the serial position effect been replicated?
The basic curve has been reproduced many times. Newer work shows that its ends respond differently: a repeat of an offloading study found a smaller cost to early items than first reported, and a final-recall study found end-of-list words recalled less well than middle ones.
Why do the last items sometimes drop out?
After a delay filled with another task, the closing items lose their advantage. In a final test at the end of a session, they can be recalled worse than middle items, which one team explained through spacing.
Sources
Levitin, D. J. The Organized Mind. Seed book; vault hub Books/The Organized Mind.
Murdock, B. B. (1962). The serial position effect of free recall. Journal of Experimental Psychology. doi:10.1037/h0045106 https://doi.org/10.1037/h0045106
Glanzer, M., & Cunitz, A. R. (1966). Two storage mechanisms in free recall. Journal of Verbal Learning and Verbal Behavior. doi:10.1016/S0022-5371(66)80044-0 https://doi.org/10.1016/S0022-5371(66)80044-0
Kuhn, J. R., Lohnas, L. J., & Kahana, M. J. (2018). A spacing account of negative recency in final free recall. Journal of Experimental Psychology: Learning, Memory, and Cognition, 44(8), 1180-1185. doi:10.1037/xlm0000491 https://doi.org/10.1037/xlm0000491
Kelly, M. O., & Risko, E. F. (2019). Offloading memory: Serial position effects. Psychonomic Bulletin & Review, 26(4), 1347-1353. doi:10.3758/s13423-019-01615-8 https://doi.org/10.3758/s13423-019-01615-8
Kelly, M. O., & Risko, E. F. (2022). Revisiting the influence of offloading memory on free recall. Memory & Cognition, 50, 710-721. doi:10.3758/s13421-021-01237-3 https://doi.org/10.3758/s13421-021-01237-3