Your critical thinking gym
Home
The Commandments
Join the Club
Your critical thinking gym
● Cognitive bias
← From Latin primus ("first") and recens ("recent"). Formalized as the serial position effect by psychologist Bennet Murdock, 1962, building on psychologist Hermann Ebbinghaus's 1885 memory experiments.
Give your brain a list of anything, in order, and it remembers the start and the end far better than everything sandwiched in between, then quietly treats "remembered best" as "was best."
Also called: serial position effect (umbrella term for both halves), first impression bias (primacy half), recency bias (recency half). Not the same as anchoring bias (Section 09).
The tendency, when encountering an ordered sequence of items (words, arguments, candidates, applicants, impressions), to more easily recall and more favorably judge whatever came first (primacy) and whatever came last (recency), while items in the middle receive disproportionately little weight, regardless of their actual merit.
Plain version
Read someone a list, in order, and ask them later what they remember: they'll nail the beginning and the end, and the middle turns to mush. That's just how memory works. The problem is that "what I remember best" quietly gets treated as "what was actually best," even when the two have nothing to do with each other.
Aliases
Serial position effect is the umbrella academic term covering both halves together (the name for the whole U-shaped curve, not just one end of it). First impression bias and recency bias are looser, colloquial names for the primacy and recency halves individually.
Not the same as anchoring bias
This is the pairing worth getting right: anchoring bias is about a single reference point, often a number, pulling a later estimate toward it, even when that reference point is arbitrary or irrelevant. Primacy/recency is about position within a whole sequence of separate items affecting memory and judgment across all of them, with no single "anchor" value involved. A ballot candidate listed first isn't a numeric anchor pulling your vote toward a figure; they're just first, and first is memorable. Full distinction in Section 09.
Scope
Gets confused with
Anchoring bias: precise distinction in Section 09.
Hard to catch, for a specific reason: the first or last item genuinely does feel more vivid and more memorable in the moment. That's not a false feeling, it's a real one, just caused by timing rather than merit. And "they made the strongest impression on me" is a completely reasonable-sounding thing to say afterward, which is exactly why this rarely gets questioned.
Why your brain does this
Two ordinary memory habits combine to produce this. Early items in a sequence get more "airtime": when something is first, there's nothing before it competing for attention, so your brain has spare capacity to rehearse it, which pushes it into longer-term storage. That's primacy. The most recent items, meanwhile, are still sitting in short-term memory at the exact moment you're asked to judge or recall anything, no retrieval effort required, they're just there. That's recency. Neither is a malfunction: leaning on "what got the most rehearsal" and "what's cheapest to retrieve right now" is a genuinely efficient way to compress a long sequence instead of treating every item as equally demanding to remember.
Where it misfires
It misfires the moment position in a sequence gets mistaken for information about quality. A candidate interviewed last isn't better at the job because they're more memorable; they're just closer in time to your decision. A name printed first on a ballot isn't the stronger candidate; a printer or a clerk put them there, often by an arbitrary or administrative rule that has nothing to do with merit.
The live academic debate, briefly
Researchers in the Tversky and Kahneman heuristics-and-biases tradition tend to frame this as a systematic error against how a rational evaluator should weigh evidence (position shouldn't matter, but it does). Gerd Gigerenzer's ecological-rationality tradition would frame the underlying memory mechanism differently: rehearsal-based storage and short-term availability are sensible, resource-efficient strategies in a world of limited attention, and the "error" only shows up when you then use a memory shortcut as if it were a quality judgment, a mismatch between the tool and the job, not a broken tool. This page doesn't take a side.
How it activates, step by step
1. You're presented with an ordered sequence: words, candidates, applicants, arguments, one after another.
2. Early items get extra rehearsal because nothing preceded them to compete for attention, pushing them toward longer-term memory (primacy).
3. The final items are still active in short-term memory at the moment you're asked to judge or recall (recency).
4. When asked to choose, rank, or remember, you lean on whichever items are easiest to bring to mind (first and last), and the middle gets underweighted or dropped entirely.
Anyone who controls the order something is presented in can lean on this bias without saying a single false word.
Cherry picking / card stacking
A communicator who has already selected which true facts to include still has to decide what order to put them in. The classic move is to open with the single strongest point (claiming primacy) and close with the second-strongest (claiming recency), burying the weakest or most damaging material in the middle, the exact position this bias makes least memorable.
When an audience is flooded with more claims than they can process, their attention naturally collapses down to just the first thing they saw and the last thing they saw. A communicator relying on primacy/recency doesn't need anyone to absorb the whole flood, just to be overwhelmed enough that they fall back on positional shortcuts instead of genuine evaluation.
Often front-loaded, placed early in a speech, campaign, or broadcast, before the audience has built up any skepticism or fatigue, specifically to claim the outsized memory weight that comes with going first.
Outside politics
Restaurant menu design and retail product listings exploit this directly: items placed first or last on a list get disproportionately more orders or clicks regardless of actual quality, which is why menu and shelf-placement consultants fight over top and bottom slots. Sales and pitch coaching gives the same advice from the other side ("save your strongest point for last") specifically to claim the recency advantage in a listener's memory.
What increases it
Strong: low information about the individual items. Ballot-order research finds the effect strongest specifically in nonpartisan, low-profile races where voters have little else to go on, and weaker or statistically undetectable in high-information, high-profile races between well-known major-party candidates.
Strong: the overall quality of the option set. Research on sequential judgment (option sets ranging from job candidates to faces) finds that when everything in the set is generally strong, evaluators lean toward a recency effect (the last one wins); when everything in the set is generally weak, they lean toward a primacy effect instead (the first one, being marginally less bad, wins). Position effects flip direction depending on overall quality, not just position.
Strong: delay or distraction between the last item and the judgment. The recency half is specifically fragile: insert almost any gap or distracting task between the final item and the moment of judgment, and recency collapses while primacy holds steady, because recency depends on short-term memory in a way primacy doesn't.
What does NOT predict it
Expertise and formal training don't reliably eliminate it. Psychologist Wändi Bruine de Bruin's 2005 study of real competitive judging (the Eurovision Song Contest and international figure skating) found trained, experienced judges working from formal scoring criteria still showed serial-position effects in their official scores. Knowing the rules and being good at the domain isn't the same as being immune to where in the lineup you're judging from.
Open question, stated honestly
Time pressure and fatigue are commonly assumed to worsen this bias, but the research specifically isolating those factors for primacy/recency (as opposed to related but distinct phenomena like general decision fatigue) is thinner than for the factors above; treated here as unconfirmed rather than asserted.
The pattern across all three: whether it's a ballot, a televised singing contest, or a hiring queue, being positioned first or last is worth something measurable, and it has nothing to do with the actual quality of what's being judged.
First documented
Psychologist Hermann Ebbinghaus, in his 1885 self-experimentation on memorizing lists of nonsense syllables (published as Über das Gedächtnis, later translated as Memory: A Contribution to Experimental Psychology), observed that items at the start and end of a list were retained better than items in the middle, before the effect had a formal name.
The landmark study
Psychologist Bennet Murdock's 1962 free-recall experiments empirically established and named the serial position curve. Participants were shown lists of 10 to 40 words, one at a time, then asked to recall as many as possible in any order within 90 seconds. No matter the list length or presentation speed, recall traced a consistent U-shape: strong for the first several words (primacy), strong for the last several (recency), and weak for everything in between.
Since then
Psychologists Murray Glanzer and Anita Cunitz (1966) showed that inserting a distractor task between the last word and the recall test eliminates the recency effect but leaves primacy intact, evidence used to support the idea of separate short-term and long-term memory stores. The effect was later extended well beyond word-list memory into judgment and social evaluation: psychologist Solomon Asch's 1946 impression-formation experiments found that reversing the order in which personality traits were presented changed the overall impression participants formed of a person; Miller and Krosnick's 1998 ballot-order study and its replications established the effect operating in real elections; and Bruine de Bruin's 2005 study found it operating inside expert competitive judging.
A contested claim worth separating out
Asch's 1946 paper is also widely cited as evidence for a specific "primacy-of-warmth effect," the claim that warmth-related traits carry outsized weight in impression formation specifically, more than competence-related traits. A 2014 replication and re-analysis (Nauts, Langner, Huijsmans, Vonk & Wigboldus, published in Social Psychology) found no clear evidence for that particular warmth-primacy claim in Asch's own data, concluding warmth's role is much more context-dependent than textbooks often suggest. This is a narrower, separate claim from the general order-effect finding: several of Asch's own experiments that directly reversed list order did find real primacy effects on overall impressions. The two shouldn't be conflated. The order effect itself is well-supported; the specific "warmth always wins" version of it is contested.
Cross-tradition note
See Section 02 for Gerd Gigerenzer's ecological-rationality framing, the relevant non-English research tradition for this page's adaptive-origin section.
Score: 2 out of 5 (see the breakdown in the card after Section 01).
The in-the-moment question
Before you settle on a ranking or a choice among several options: "If I'd encountered these in a different order, would I still pick this one?" If you genuinely don't know, that uncertainty is more honest than the confidence you started with.
Why this is hard to catch
The first or last item really does feel more vivid, more memorable, more "right" in the moment. That's not a false signal your brain is faking, it's a real one, just caused by timing rather than merit. And there's always a completely reasonable-sounding story available afterward ("they just made the strongest impression on me"), so the judgment never gets flagged as suspicious. Ties to both the rationalization and, to a lesser degree, the payoff sub-scores: a vivid, confident-feeling conclusion is more comfortable than an uncertain one, even when the vividness is coming from the wrong source.
Honest note on debiasing
Rotation is the one intervention with genuinely strong evidence behind it, but the way it works is worth understanding precisely: the New York ballot-rotation research showed that rotating position across precincts cancels out the average first-position advantage across the whole electorate, even though any single voter looking at any single ballot is just as susceptible as ever. It fixes the population-level outcome, not the individual's psychology. Blind or structured review for interview panels helps reduce the effect, but the evidence there is more mixed than for ballot rotation; be wary of any claim that a single trick eliminates it entirely.
Commonly co-occurs with
Often mistaken for
The one this gets confused with most. Anchoring is about a single reference point, usually a number, pulling a later estimate toward it, even when that number is arbitrary or stated to be irrelevant (the classic demonstration uses a spun wheel of fortune to anchor unrelated numeric guesses). Primacy/recency doesn't require a number, an anchor, or even a single reference point; it's about position within a whole sequence of separate items, and it's driven by memory mechanics (rehearsal and short-term storage), not by a value silently distorting arithmetic. A ballot candidate listed first isn't anchoring your vote to a figure; they're just occupying the most memorable slot in a list.
Compounding effect
A primacy-formed first impression tends to get entrenched by confirmation bias afterward: once an early impression sets, later information about the same person or claim tends to get filtered to fit it, which is part of why first impressions are so hard to dislodge even after a lot of contrary evidence follows. The halo-horns effect can compound with primacy in a similar way: an early positive trait can cast a positive glow that colors how everything presented afterward gets interpreted, reinforcing the very first impression that triggered it.
Primary sources
Accessible reading