Where You Put Your Best Point Changes Whether It’s Remembered
By Clara Ferreira · Speeches That Lasted
In the 1880s, the psychologist Hermann Ebbinghaus ran what may be the most patient self-experiment in the history of the field: he memorized long lists of meaningless syllables and tested his own recall of them at intervals, alone, over years.1 His work started the experimental study of memory, and one of the patterns that study went on to measure is now among its best known: items at the beginning and end of a sequence are recalled far better than items in the middle. It's called the serial position effect, and it describes something close to a structural fact about how sequences get remembered, independent of the content.
The shape of the curve
When Bennet Murdock measured the effect carefully in 1962, plotting recall accuracy against an item's position in lists of words, the result was a distinctive U-shape: high accuracy for the first few items, high accuracy for the last few, and a long, low middle stretch where recall drops substantially.2 Researchers distinguish two mechanisms behind the two ends of the U: a primacy effect, attributed to early items getting more rehearsal and a better chance of transferring into long-term memory before later items start competing for attention, and a recency effect, attributed to the last few items still sitting in short-term working memory at the moment of recall.
Two ends, two mechanisms
The evidence that these really are two separate things came from Murray Glanzer and Anita Cunitz in 1966.3 They changed the conditions in two ways. When people had to count backwards for half a minute before recalling the list, the recency advantage disappeared and the primacy advantage stayed. When the words were presented more slowly, giving more time to rehearse, recall of the early items improved and the end of the curve barely moved.
That matters for a speaker. Recency is fragile: it lasts while the last items are still held in mind and fades once anything else takes their place. Primacy is sturdier, because the early items have had time to settle. A point at the very end of a talk is vivid as people leave the room. Whether it is still there the next morning is a different question.
A bias of position
The uncomfortable implication for a speaker is that this is a bias in sequence position, and says nothing about which points matter. An important point placed third in an unweighted list of five is structurally disadvantaged for recall compared to a less important point placed first or last, for reasons that have nothing to do with the content and everything to do with where it sits.
Order changes judgement too
Position affects more than memory. In 1946 Solomon Asch read people a list of traits describing a stranger.4 Some heard "intelligent, industrious, impulsive, critical, stubborn, envious." Others heard the same six words in reverse order. The first group formed a noticeably more favorable impression. The early words set a frame, and the later words were read through it.
The closest test to an actual speech is a 1959 study by Norman Miller and Donald Campbell.5 Participants heard two opposing sides of a simulated court case and gave their judgement, with the timing varied. When the two sides were heard back to back and the judgement came a week later, the side heard first had the edge. When a week separated the two sides and the judgement followed the second one immediately, the side heard last had the edge. Neither position wins in general. Which end wins depends on when the decision is made.
How this differs from the peak-end rule
This is a close cousin of the peak-end rule, and the two are worth distinguishing clearly. The peak-end rule is about the intensity of a moment and the ending specifically. The serial position effect is older, more general, and purely about order: first and last positions in any sequence, regardless of how emotionally intense any given item is. A talk's best point benefits from both effects if it's placed at the start or end and is also the most memorable moment in the room, but they're two separate findings that happen to point the same direction.
Classical rhetoric got there independently
Roman handbooks of rhetoric advised putting the strongest arguments at the beginning and the end of a case and the weaker ones in the middle, centuries before anyone measured a recall curve. The research doesn't validate the classical advice by coincidence; it explains why speakers who never read a psychology paper kept converging on the same structural choice through trial and error.
Lists are not talks
Murdock's and Ebbinghaus's findings come from controlled recall of lists (syllables and words) over short delays, rather than from twenty-minute talks with real content and days between hearing and recall. Applying the finding to speech structure is a reasonable, widely adopted extrapolation rather than a direct replication. Miller and Campbell's study is closer, and it adds a condition rather than a simple rule. We name this as an extrapolation rather than presenting it as a one-to-one transfer.
Where to put your best point
Open with it, if the decision comes later. A pitch judged after a week of other pitches, or a proposal discussed at next month's meeting, is the delayed case. The side heard first had the advantage there.
Close with it, if the decision comes now. A vote at the end of the meeting, a question asked as soon as you sit down: the last thing said is the freshest in the room.
When in doubt, use both ends. State the point early, develop it, and return to it at the close. That is the old advice to tell them what you'll tell them, and it now has a mechanism behind it.
Don't bury anything in the middle that you can't afford to lose. The middle is where supporting detail belongs. If a talk sags there, see why your talk falls apart in the middle. And because recency fades, a point you need people to keep needs repeating, which is what the forgetting curve is about.
References
- Ebbinghaus, H. (1885). Über das Gedächtnis (translated as Memory: A Contribution to Experimental Psychology, 1913). https://openlibrary.org/works/OL28477195W
- Murdock, B. B. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482–488. 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, 5(4), 351–360. https://doi.org/10.1016/S0022-5371(66)80044-0
- Asch, S. E. (1946). Forming impressions of personality. The Journal of Abnormal and Social Psychology, 41(3), 258–290. https://doi.org/10.1037/h0055756
- Miller, N., & Campbell, D. T. (1959). Recency and primacy in persuasion as a function of the timing of speeches and measurements. The Journal of Abnormal and Social Psychology, 59(1), 1–9. https://doi.org/10.1037/h0049330