Picture three popcorn sizes at the cinema: small, a medium that costs nearly as much as the large, and the large. Plenty of people buy the large and feel clever about it. Did the medium exist to be bought, or to make the large look like a bargain? That small question opens a big research area on how the structure of a choice shapes the outcome.
This page walks through four well-known effects. For each, we separate what controlled studies found from the confident version that circulates online, and we flag where the evidence has weakened.
The option nobody buys but everybody needs
laboratory findingIn 1982, Joel Huber, John Payne and Christopher Puto published what became known as the decoy or attraction effect. When people choose between two options and a third is added that is clearly worse than one of them, the option that beats the decoy tends to gain share. Dan Ariely popularised a striking example in Predictably Irrational (2008): a magazine offer where a print-only subscription costing the same as print-plus-web made the bundle look irresistible.
The effect is real in lab settings with numerical attributes. But a 2014 paper by Shane Frederick, Leonard Lee and Ernest Baskin, titled "The Limits of Attraction," found it was much weaker or absent when options were shown as pictures or described in more realistic, less tidy ways.
Study limit: works best with clean numerical comparisons; fragile with realistic products
Defaults: the decision someone already made
field evidenceEric Johnson and Daniel Goldstein's 2003 paper in Science, "Do Defaults Save Lives?", compared European countries that registered people as organ donors unless they opted out with countries that required people to opt in. Registered consent was dramatically higher in the opt-out countries. In the UK, automatic enrolment into workplace pensions, phased in from 2012, saw participation rise sharply among eligible employees.
"Defaults work because people are lazy."
A neat story: changing the form costs effort, so people don't bother.
At least three forces at once
Effort plays a part, but defaults can also signal what the designer recommends and set a reference point that makes leaving feel like a loss. Which force dominates depends on the decision.
One important caution: registered consent is not the same as the number of organs actually donated. Spain, often cited for high donation rates, attributes much of its success to hospital coordination through its national transplant organisation, not just a form design.
Anchors: numbers that should not matter
laboratory finding · replicatedIn their 1974 Science paper, Amos Tversky and Daniel Kahneman described participants who watched a wheel of fortune land on 10 or 65, then estimated the percentage of African countries in the United Nations. The median estimates were 25% and 45%. A meaningless number had pulled the answers.
Anchoring is one of the sturdier findings in the field. In the Many Labs 1 project (2014), dozens of labs across several countries repeated classic studies, and anchoring effects replicated strongly. You see the everyday version in "was £120, now £60" labels and in opening offers during negotiations.
When the famous finding shrinks
overstretched claimSheena Iyengar and Mark Lepper's 2000 jam study found that a tasting table with 6 jams converted far more tasters into buyers than one with 24. It became the backbone of the "paradox of choice" message. Yet a 2010 meta-analysis by Benjamin Scheibehenne, Rainer Greifeneder and Peter Todd found an average effect close to zero across dozens of studies.
A later meta-analysis by Alexander Chernev, Ulf Böckenholt and Joseph Goodman (2015) offered a more useful picture: overload appears when choices are complex, decisions are difficult, people have no clear preference, or they want to buy quickly. In other words, context decides.
Observation prompt
Next time you're at a café with three cup sizes, notice which one most people order. Then visit one with two sizes. Don't treat it as proof of anything. Treat it as a better question about menus, prices and habits.
What to take away
Choice architecture is real, measurable and useful, but it isn't magic. The strongest effects, like defaults and anchors, have survived large replications. Others, like decoys and choice overload, depend heavily on how options are presented. When someone promises that one design trick will transform behaviour, ask which of these categories it belongs to.