In one sentence: Dan Ariely’s 2008 bestseller argues that human error is patterned rather than random, and shows through a decade of his own experiments that we judge value only by comparison, treat free as a category rather than a price, and let expectation and ownership rewrite what we want and what we feel.
At a Glance
Author: Dan Ariely
First published: 2008, in this Revised and Expanded edition 2009 (Harper Perennial)
Category: Economics / Psychology
Length: about 384 pages, about 101,000 words (Harper Perennial Revised and Expanded paperback)
ISBN-13: 978-0-06-135324-6 (Harper Perennial Revised and Expanded paperback)
Summary reading time: about 12 minutes
Book reading time: about 7 hours
Notable adaptations: no film or series, though the book helped launch the behavioral economics genre and Ariely followed it with The Upside of Irrationality in 2010 and The Honest Truth About Dishonesty in 2012
Content note: the evidence behind three of this book’s chapters has since been retracted or formally questioned. Psychological Science retracted Ariely and Wertenbroch’s 2002 procrastination study in September 2026, the experiment chapter 7 is built on, after a failed replication and an analysis concluding the data had been tampered with or fabricated. No one has been publicly identified as responsible. The 2008 paper cited for the honesty chapters carries a journal expression of concern dated December 2024. Separately, a 2012 honesty-pledge paper Ariely co-authored was retracted in 2021 over falsified data, and Duke University investigated for three years, concluding, by Ariely’s own account of it, that data had been falsified but that there was no evidence he knowingly used fake data. The analysis section sets out which chapters are affected. Figures and prices are as of 2009.
Ariely came to decision making through injury. At eighteen a magnesium flare burned over 70 percent of his body, and through years of treatment he watched nurses tear bandages off fast, sure that speed was kinder. His own experiments later suggested the reverse, that lower intensity over a longer duration hurts less, and that experienced professionals could be wrong for decades seeded a career. He trained in psychology, then spent a decade at MIT and Duke running small, inventive studies in dorms and cafeterias. This book gathered them in 2008 and gave the genre its shape.
Read it if you want the genre’s founding popular text, written with real wit, and are curious how experiments on choice get designed. Skip it if you want evidence you can lean on without checking, because knowing which chapters to discount matters more than the chapters do.
The Big Idea
Standard economics begins from a flattering premise, which Ariely paraphrases as the idea that we are capable of making the right decisions for ourselves. His counterclaim is not that people are stupid. It is that we go wrong in patterned ways, in the same direction, over and over, so the errors can be measured, exploited by sellers, and designed around by anyone trying to help us. Error as structure, not noise, is the title’s claim.
Underneath most chapters sits one mechanism, relativity. People carry no absolute scale of worth, so we judge each option against whatever comparison is nearby. Arrange the comparison and you arrange the choice. Each chapter then opens with a cheap, clever laboratory manipulation, reports a surprising gap between group averages, and widens into an argument about pricing, dating, medicine or policy. The ingenuity is real, and the difficulty is how far the extrapolation travels on how little data.
Key Ideas
1. Value is relative, and a decoy can move a decision
Ariely’s demonstration is a real subscription offer from The Economist: web for 59 dollars, print for 125 dollars, print plus web also for 125 dollars. Of 100 MIT Sloan students shown all three, 16 took web only, nobody took print only, and 84 took the bundle. Drop the print-only option and the pattern inverted, with 68 choosing web only. The dominated option sells nothing and earns its place, because it makes the bundle look like a bargain.
2. Anchoring and arbitrary coherence
A meaningless number can become the reference point for everything after it. In an auction run with Drazen Prelec, 55 MIT students wrote down the last two digits of their social security number, then bid on wine, computer accessories and chocolates. Those whose digits fell in the upper range bid 216 to 346 percent more than the lower range, while the ordering of their bids stayed sensible. That is arbitrary coherence: random at the root, consistent above it.
3. Free is not a price but a different category
Ariely ran a stand selling Lindt truffles at 15 cents and Hershey’s Kisses at one cent, where 73 percent of buyers took the truffle. Cut both prices by a cent, so the truffle costs 14 cents and the Kiss nothing, and the truffle’s share fell to 31 percent. The gap between them had not changed. Zero had appeared on one side. Amazon’s French site, charging about 20 cents for shipping rather than waiving it, saw none of the sales jump free delivery produced elsewhere. Every transaction has a downside we weigh, and zero stops the weighing.
4. Social norms and market norms are separate systems
We live in two worlds at once, one run on warmth and vague obligation, the other on prices and explicit exchange, and the second drives out the first. Asked by AARP to advise members for 30 dollars an hour, few lawyers agreed. Asked to do it for nothing, most said yes. In a dull dragging task run with James Heyman, people paid five dollars moved 159 circles and those paid 50 cents moved 101, while people asked as a favor moved 168. Once money enters a relationship, the social norm is slow to return.
5. Ownership and the compulsion to keep options open
Duke allocates basketball tickets by lottery after a long campout. Ariely and Ziv Carmon called more than a hundred students: losers would pay about 175 dollars for a ticket, winners would not sell below about 2,400 dollars, roughly fourteen times as much for the same seat. Owning something, even by luck, rewrites its worth. A second study used a laptop game of three paying doors, each vanishing if left unvisited for twelve clicks. Players spent real clicks keeping doors alive they had no use for, earning about 15 percent less than those who settled on one.
6. Expectation and price change the experience itself
In a pub study with Leonard Lee and Shane Frederick, students tasted beer spiked with balsamic vinegar. Told beforehand, they disliked it. Told afterward, they rated it as highly as drinkers never told at all. The warning did not change the report of the taste, it changed the taste. Price does the same to medicine. Around a hundred Bostonians took a placebo painkiller before rated shocks, and nearly all reported relief at 2.50 dollars a capsule against about half at 10 cents.
7. Most people cheat a little, and context sets how much
These are the book’s most quoted studies and the ones to treat with most caution, because the only empirical paper cited for the two honesty chapters, by Mazar, Amir and Ariely in 2008, carries an expression of concern from the Journal of Marketing Research dated December 2024. Take the figures as the book’s claims. Students marking their own 50-question test claimed about 36.2 against a supervised baseline of 32.6. In the matrix version, claimed scores ran at 3.5 under supervision, 6.2 when participants shredded their sheets, and 9.4 when paid in tokens exchanged for cash a step later. Recalling the Ten Commandments first, or signing an honor statement, returned claims to baseline.
Context and Analysis
The integrity problems come first. Chapter 7 rests on a 2002 study by Ariely and Wertenbroch in which students who set their own spaced deadlines for three papers outperformed students given a single final deadline. Psychological Science retracted that paper in September 2026, after a failed replication by Hyndman and Bisin and an analysis by Data Colada, which reported an effect implausibly large for the sample size, eighteen of twenty participants in one condition each having an exact twin with identical scores and a participant number exactly ten apart, correlations absent from the original data but present in the replication, and only about 12 percent of self-reported times falling on round numbers against about 85 percent in the replication. It concluded that the data had been severely tampered with or fabricated, and declined to say who was responsible. Wertenbroch said he never had access to the data and asked the editor to retract. Ariely acknowledged serious anomalies and said he would cooperate. The 2008 paper behind chapters 13 and 14 carries the journal’s expression of concern. Further back, a 2012 honesty-pledge paper co-authored by Shu, Mazar, Gino, Ariely and Bazerman was retracted in 2021 after its field data were found to be falsified, and Duke investigated for three years. Ariely said in 2024 that the inquiry concluded data had been falsified but found no evidence he knowingly used fake data, and because Duke declined to comment, that account is his own. No public investigation has named who falsified anything. Three chapters of fifteen are implicated. The others rest on different experiments.
Set the retractions aside and ordinary criticisms remain. Anchoring and arbitrary coherence have a contested replication record, and the original effect sizes look large beside later work. Samples are small and convenient: about a hundred Sloan students, 55 in the auction, 25 Berkeley men in the arousal chapter, nearly all from MIT, Berkeley or Duke, with class sizes sometimes unreported. Much of the evidence is hypothetical or self-reported, and famous illustrations such as the Starbucks story are anecdote rather than test. In the cheating chapters group averages are read as individual conduct, so a majority is described as having cheated and one condition as not cheating at all, with no per-person detection and no reported variance. Mechanisms arrive after the result and sometimes contradict each other, free producing over-reach in chapter 3 and restraint in chapter 5. The policy conclusions, on health-care pricing and the end of cash, travel far from a cafeteria and a shock machine.
Plenty survives. The zero price effect has been replicated widely and is now standard pricing practice. The split between social and market norms is well supported and useful to anyone managing volunteers. The endowment effect is among the better-established findings in the field and owes nothing to Ariely’s data. Placebo and expectation effects sit in a large independent literature. The framing claim, that our errors are systematic rather than random, has been vindicated where particular demonstrations have not. For the same territory handled with far more care about what is established, read Thinking, Fast and Slow, slower and drier and openly flagging the findings Kahneman would now weaken. On the honesty chapters’ ground, approached through theory rather than experiment, The Elephant in the Brain by Simler and Hanson is the better companion here.
What It Teaches
Five things worth keeping from the parts that hold up.
1. You carry no internal price list, so whoever sets the comparison controls your decision. 2. Treat any first number, a sticker price or an opening bid, as contamination rather than information. 3. Free moves behavior far out of proportion to the cent it saves, worth exploiting when you offer it and distrusting when you take it. 4. Paying people for what they did willingly can cost more than it buys, and the goodwill does not come back. 5. Expectation, ownership and price shape the experience itself, so a cheap remedy can work worse than an identical costly one.
Memorable Lines
“We don’t have an internal value meter that tells us how much things are worth.” (Ariely)
“Zero is not just another discount. Zero is a different place.” (Ariely)
“Social norms are not only cheaper, but often more effective as well.” (Ariely)
“We have an irrational compulsion to keep doors open. It’s just the way we’re wired.” (Ariely)
“The truth is that placebos run on the power of suggestion. They are effective because people believe in them.” (Ariely)
“A second main lesson is that although irrationality is commonplace, it does not necessarily mean that we are helpless.” (Ariely)
Should You Read the Full Book?
Verdict: Summary is enough
The ideas are worth having and several are solid, but a reader can no longer take this book’s evidence at face value, and the most useful thing to know is which chapters to discard. A summary delivers that in a paragraph and the book cannot deliver it at all. Read Kahneman for the careful version. Sample Ariely for the experimental imagination and the prose, better than almost anything in the genre, but read him on how to think up an experiment, not on what it proved.
The Predictably Irrational book page has the full details and where to get a copy.