The Serial Position Effect: Why You Remember Some Parts of Books More Than Others

Quick Answer: The serial position effect — first documented by Hermann Ebbinghaus in 1885 and refined by Bennet Murdock in 1962 — describes the tendency to remember items at the beginning and end of a sequence better than items in the middle. The strong memory for early items is called the primacy effect; the strong memory for late items is the recency effect. For readers, this means you naturally remember the opening pages of a chapter and the final paragraphs but lose much of what happens in between. Understanding the mechanism lets you structure your reading sessions, note-taking, and reviews to push more of the middle into long-term memory.

You finish a 350-page book. A week later, someone asks what it was about. You can describe the premise vividly. You can recall the ending. The middle 200 pages are a haze. You read them, you remember the experience of reading them, but the specific arguments and examples are gone.

This is not a personal failure of memory. It is one of the most replicated findings in cognitive psychology, and it has a name.

The Original Experiments

In 1885, the German psychologist Hermann Ebbinghaus published Über das Gedächtnis (On Memory) — the founding text of experimental memory research. Ebbinghaus used himself as the only subject. He memorized lists of nonsense syllables (zoq, dax, fim) and measured how many he could recall after various delays. Among his findings: when recalling a list, syllables at the beginning and end were remembered far more reliably than syllables in the middle. Plotted on a graph, recall accuracy formed a U-shape across position in the list.

Almost 80 years later, Bennet Murdock at the University of Vermont ran a more rigorous version of the experiment. He gave subjects lists of words of varying lengths (10, 15, 20, 30, 40 items) and asked them to recall as many as possible immediately afterward. The same U-shape appeared in every condition. The first few items had high recall (the primacy effect). The last few items had even higher recall (the recency effect). The middle items collapsed to a low, flat baseline.

Murdock's 1962 paper, "The Serial Position Effect of Free Recall," became the canonical demonstration of the phenomenon. Decades of follow-up work have confirmed it across word lists, faces, music, locations, narrative content, and — most relevantly — extended prose like book chapters.

Why the Middle Disappears

The serial position effect is not a single mechanism. It is the result of two distinct memory systems doing two different jobs.

The Primacy Effect

When you read the opening of a chapter, your working memory has nothing else to compete with. You can rehearse the early ideas — turn them over, connect them to prior knowledge, form initial impressions of the argument. This rehearsal helps consolidate the information into long-term memory. The first idea in a chapter gets, on average, more cognitive processing than any subsequent idea.

This explains why chapter introductions tend to stick. You arrive at the chapter fresh, the author's framing is clear, and your brain has bandwidth to encode it deeply.

The Recency Effect

The end of a sequence sits in a different memory store. Short-term or working memory holds roughly four to seven items at a time, and the last things you read are still in that buffer when you stop. If you are tested immediately, you can recall them with near-perfect accuracy — not because they were processed deeply, but because they have not yet been displaced.

This is why you can summarize the final pages of a chapter the moment you close the book. The information is fresh in working memory. It also explains why the recency effect disappears if you wait. After about 30 seconds of doing something else — reading the next chapter, checking your phone, talking to someone — the working memory buffer clears, and the recency advantage is gone. The primacy effect, which is rooted in long-term memory, persists much longer.

The Middle Falls in a Gap

The middle of a chapter loses on both fronts. By the time you reach the middle, your working memory is already partially loaded with earlier content, so each new idea gets less rehearsal. And there is plenty of subsequent content still to come, which will displace the middle from working memory before any test happens. The middle is too late for primacy and too early for recency. It falls into the dip in the U-shape.

Cognitive psychologists Atkinson and Shiffrin formalized this in their 1968 multi-store model of memory. Items at the beginning have time to move from short-term to long-term memory; items at the end are held in short-term memory; items in the middle never make it into either store with strong encoding.

What This Means for Reading

For a reader, the serial position effect operates at multiple scales simultaneously.

At the sentence level, you remember the first and last words of a sentence better than the middle. This is why writers are taught to put emphasis at the end of a sentence — the recency effect makes those words land harder.

At the paragraph level, the topic sentence (usually first) and the conclusion (usually last) are encoded better than the supporting evidence in between.

At the chapter level, the introduction and the closing summary stick. The arguments, examples, and case studies in the body of the chapter are vulnerable.

At the book level, you remember the opening chapters (you were excited and engaged, with no prior content competing) and the final chapters (still in recent memory). The middle of a long book — the long stretch where the author develops the argument in detail — is where retention falls off most sharply.

This pattern holds across genres. It is true of novels (you remember the inciting incident and the resolution; the middle plot twists fade). It is true of nonfiction (you remember the thesis and the conclusion; the supporting research becomes a blur). It is true of textbooks. It is true of essays and articles.

The Connection to the Forgetting Curve

The serial position effect interacts with another well-documented memory phenomenon: the forgetting curve, also discovered by Ebbinghaus. The forgetting curve shows that retention decays exponentially over time without review. Most of what you read is lost within a week.

When you combine the two effects, the picture becomes sharper: the middle of every chapter starts at a lower retention level than the beginning or end, and then decays from there. The primacy and recency portions decay too, but they have further to fall before they hit zero. The middle is already near zero on day one. By day seven, it is functionally gone.

This is why "I read it but I don't remember anything" is such a common complaint among readers of nonfiction. They are not imagining it. The serial position effect, compounded by the forgetting curve, predicts exactly this outcome.

Practical Strategies to Beat the Effect

Knowing the mechanism gives you specific levers to pull. The goal is not to fight your memory's defaults — it is to structure your reading so that the parts you most want to retain land in the high-retention zones.

Break Long Chapters Into Shorter Sessions

If a chapter is 30 pages long and you read it in one sitting, the middle 20 pages are in the dip. If you split it into three 10-page sessions, you create three new beginnings and three new endings. The same pages that would have been "middle" in one session become "primacy" or "recency" in another.

This is one of the highest-leverage adjustments you can make. Short, focused reading sessions of 15 to 25 minutes naturally create more boundaries — and more high-retention zones — than marathon sessions. Distributed practice does not just help with stamina; it changes which parts of the book your brain is biased to remember.

Pause and Summarize at the Midpoint

If you cannot break a chapter into multiple sessions, you can still create an artificial boundary. At the midpoint, stop reading. Spend 30 seconds summarizing what you have read so far in your own words — out loud or in writing. This does two things. First, it gives the early-middle a chance to be rehearsed (boosting primacy-like encoding for material that would otherwise be lost). Second, it creates a recall event, which is a form of retrieval practice — and retrieval practice is one of the most effective techniques for moving information into durable long-term memory.

The brief pause feels like an interruption to the flow of reading. It is. That is the point. The interruption is what creates the encoding boundary.

Highlight Strategically in the Middle

Most readers highlight more in the opening and conclusion of a chapter than in the middle, partly because authors tend to put their cleanest formulations in those places. But the middle is exactly where you need extra encoding help. Be intentional about highlighting one or two key claims from each middle section, even if the writing is denser or less quotable. The act of selecting the highlight forces processing that passive reading does not.

If you use a digital highlighting tool, those middle highlights are also what you most need in your spaced repetition queue, because they are the most vulnerable to being forgotten.

Use Headings as Boundaries

In a book with frequent subheadings, every subheading is a boundary that resets the serial position effect. The last paragraph before a subheading benefits from recency; the first paragraph after benefits from primacy. Authors who use frequent subheadings are inadvertently making their books more memorable.

When reading a book with sparse subheadings, you can add your own. Mark a horizontal line in the margin every few pages where you sense a topic shift. Treat each marked section as its own mini-chapter. Your brain will encode the boundaries the same way.

Review With Spacing

The serial position effect describes encoding. The forgetting curve describes decay. Both can be partially countered by spaced repetition — periodic reviews of material at expanding intervals. The first review, ideally within 24 hours of reading, catches information before it has fully decayed. Subsequent reviews extend the durability further with diminishing effort.

Spaced repetition is particularly powerful against the middle-of-chapter problem. The first time you read a passage in the middle of a chapter, encoding is weak. The second time you encounter it (in a review the next day), it is now in a position of "primacy" relative to the rest of that review session. By the third or fourth review, the middle material has been encoded as deeply as if it had been the opening line of the original chapter.

Discuss the Middle, Not Just the Ending

In book clubs and casual conversations about books, people gravitate toward the beginning and the ending — the parts that are easiest to recall. If you want to remember the substance of a book, force yourself to discuss the middle. Pick a specific chapter from the middle of the book and try to explain its main argument to someone else. The act of articulating a memory strengthens it dramatically. This is the Feynman technique applied to the most fragile part of the book.

Read the Conclusion First (Sometimes)

For nonfiction, reading the conclusion before the body of the book serves two purposes. First, it tells you where the argument is going, which provides scaffolding for the supporting material. Second, it shifts the recency effect: the body of the book, normally a long stretch where retention degrades, is now contextualized by an ending you have already absorbed.

This is the Mortimer Adler "inspectional reading" approach. It can feel like spoilers in a novel, but for nonfiction, it usually improves both comprehension and retention.

What the Research Doesn't Cover

The serial position effect was originally established with simple lists. Reading is more complex. There are interaction effects with attention, prior knowledge, narrative interest, emotional engagement, and the inherent memorability of specific content. A particularly vivid scene from the middle of a novel can stick because of picture superiority or emotional intensity, not because it occupied a primacy or recency position.

Some research also suggests a "serial position effect of long-term memory" that operates over longer time scales — the first books you read on a topic and the most recent ones tend to have stronger influence on your thinking than the books in between. This is harder to study rigorously, but it appears in subjective reports often enough to take seriously. If you read 50 books on a subject across a decade, you probably remember the first few (when you knew nothing) and the last few (recent) more clearly than the dozens in the middle.

A closely related effect governs how you judge a book rather than how much of it you retain. The peak-end rule holds that your verdict on an experience is set largely by its most intense moment and its final stretch, while the duration and the average quality of everything in between count for much less. Serial position explains why the middle of a book fades; the peak-end rule explains why an uneven book with a brilliant closing chapter can still feel like a great one.

The point is not that the serial position effect is the only thing happening. It is that it is one consistent force pulling on your memory, and you can either work with it or against it.

The Reader's Takeaway

Your brain is not equally good at remembering everything you read. The opening and closing of every reading session, every chapter, every paragraph get a memory advantage that the middle does not. This is not a bug — it is how memory has always worked, going back to Ebbinghaus's nonsense syllables in 1885.

You cannot turn the effect off. But you can multiply the number of openings and closings in your reading life by reading in shorter sessions, pausing at midpoints, marking your own boundaries within long chapters, and reviewing with spacing. Every new boundary creates two new high-retention zones.

The middles of books are where most of the substance lives. Most authors put their main arguments in chapters three through seven of a ten-chapter book, not in the introduction or the conclusion. If you only remember the openings and endings, you have a vague memory of every book and a deep memory of none. Counter the serial position effect, and you start remembering what is actually inside the books you read.

Frequently Asked Questions

What is the serial position effect?

The serial position effect is the tendency to remember items at the beginning and end of a sequence better than items in the middle. First documented by Hermann Ebbinghaus in 1885 and confirmed by Bennet Murdock in 1962, it produces a U-shaped recall curve across any list or extended sequence. In reading, this means chapter openings and closings encode more deeply than the content in between.

Why do I remember the beginning and end of a book but forget the middle?

Two separate memory mechanisms are responsible. Material at the beginning benefits from the primacy effect: your working memory is fresh, so you can rehearse early ideas more thoroughly, helping them consolidate into long-term memory. Material at the end benefits from the recency effect: those pages are still held in short-term working memory when you stop reading. The middle falls in a gap — too late for primacy and too early for recency — so it encodes weakly from both directions.

What is the primacy effect in reading?

The primacy effect is the memory advantage given to the first items in a sequence. When you begin a chapter, your working memory is unloaded, giving early ideas more processing bandwidth. You can rehearse and connect them to prior knowledge before subsequent content crowds them out. This is why the introduction of a chapter or the opening argument of a book tends to stick better than equivalent material encountered later.

What is the recency effect in reading?

The recency effect is the memory advantage given to the most recently encountered items. The last passages you read before stopping a session remain in short-term working memory and are highly accessible immediately after. However, the recency advantage is fragile — it disappears within 30 to 60 seconds of switching to a different activity. The primacy effect, rooted in long-term memory consolidation, persists much longer.

How can I use the serial position effect to improve reading retention?

Work with the effect rather than against it: break long chapters into shorter sessions (each creates new primacy and recency zones), pause at the midpoint of a chapter to summarize what you have read (turns middle material into a recency event), highlight one or two key claims from each middle section deliberately, and schedule spaced reviews of your highlights so middle content receives the review sessions it was denied during initial reading.

Does the serial position effect apply to entire books, not just individual chapters?

Yes. At the book level, you are more likely to remember the opening chapters (high engagement, no competing prior content) and the most recent chapters you read (recency advantage) than the chapters in the middle of the book. This compounds with the forgetting curve: middle-of-book content starts at a lower encoding level and decays from there. Readers who finish a long book and remember "the premise and the ending but nothing in between" are experiencing the serial position effect operating at book scale.

How This Interacts With Multi-Book Reading

The serial position effect gets sharper, not weaker, when you read multiple books at once. Each switch between books creates two new boundaries — the close of session A and the open of session B — and each boundary gets the primacy/recency boost. A reader running a deliberate two- or three-book rotation, as described in interleaving for reading, is essentially manufacturing extra primacy/recency zones across the week.

The catch is that the boost only operates when each session is genuinely "loaded" — in other words, when the situation model has been reactivated. The retrieval pass on re-entry discussed in the Pomodoro article is the move that converts a switch into a real opening rather than a fake one. Without the retrieval pass, the start of the new session is muddled by reload friction and the primacy boost never triggers. With the pass, you get a clean opening every time you return to a book, and the serial position effect starts working in your favor across the whole week instead of just at the bookends.

A 2026 Note on Audio and the Recency Effect

One adaptation worth flagging for audiobook readers: the recency effect is weaker for audio than for text. Visual recognition reinforces the last paragraphs read in a way that fading audio does not, so the end-of-session retention bump is smaller for listening. The practical fix is to pause an audiobook one or two minutes before stopping and recall, in your own words, what just happened. That recall pass restores most of the recency effect that the audio modality otherwise loses, and it doubles as a retrieval-practice event that strengthens the fading middle of the session at the same time.

How the Serial Position Effect Interacts With Other Memory Phenomena

The serial position effect is not a standalone curiosity — it composes with the other major memory effects in ways that change what the practical fix looks like.

With the generation effect. Generated content is remembered better than received content at every position in a sequence, but the boost is larger in the middle than at the edges, because the edges already had primacy or recency working for them. The implication for reading: the highest-leverage place to insert a generation step (a paraphrase, a prediction, a self-quiz) is in the middle of a chapter, where the serial position effect is leaving the most retention value on the table. Concentrating your active-reading effort at chapter openings and closings double-counts retention you would have gotten anyway.

With dual coding. Visual encoding helps most for the items that get the weakest verbal encoding — which, again, means the middle. A diagram, sketch, or mental image of a mid-chapter concept produces a larger retention gain than the same image attached to an opening or closing concept, because the verbal channel was already doing most of the work at the edges. Treat sketches and mental images as middle-of-chapter tools first, and add them at the edges only when the material genuinely warrants both channels.

With spaced repetition. A spaced-review schedule effectively flattens the U-shape because every card, regardless of its original position in the book, gets the same review treatment. This is one reason flashcard systems built from highlights produce more even retention across a book than naturalistic re-reading does — the system does not know or care which chunk came from chapter two versus chapter seven, and it reviews all of them on the same forgetting-curve-aware schedule. For readers who want their long-term memory of a book to match the book's actual structure rather than the artifacts of how human encoding fails in the middle, spaced repetition is the most reliable intervention.

With the encoding-specificity principle. Items at primacy positions are encoded under "session-just-started" conditions (fresh attention, no working-memory load); items at recency positions are encoded under "session-just-ended" conditions (full working memory, often tired). Items in the middle are encoded under the most variable conditions, which means they have less consistent retrieval cues. Deliberately varying the contexts in which you review middle material — different times of day, different physical settings, different modalities — gives that material the retrieval-cue diversity the edges acquired automatically.

The general principle behind all of these compositions: the serial position effect is not a bug in your brain, but the uneven distribution of retention it produces means the practical leverage is in the middle of any sequence — chapters, books, study sessions, presentations. Optimizing the middles is what separates readers who remember everything they finish from readers who remember what they happened to encounter at the right position in the curve.

Chapterly turns the middle of any chapter into spaced repetition reviews — so the parts you would normally forget become the parts you remember best. Try it free.