Iyengar's Jam Study Wasn't About Jam. Here's the Data on What It Was Really About

In 2000, two psychologists set up a folding table in an upscale grocery store and started selling jam. Some shoppers saw 6 varieties. Others saw 24. The study that came out of that table has been cited in TED talks, business books, and — more recently — every argument about why dating apps feel exhausting. But almost nobody who cites it has read past the abstract.

So this is a deep dive into slow dating's favorite piece of evidence: what Sheena Iyengar and Mark Lepper's 2000 study actually measured, what the numbers actually showed, what the study does not prove, and what it genuinely means for anyone deciding how to spend their time meeting people in 2026.

What the Study Actually Set Up

Iyengar and Lepper ran the experiment at Draeger's, a gourmet market in Menlo Park known for absurd selection. Every few hours, researchers alternated a tasting table between two conditions: a limited assortment of 6 jams, or an extensive assortment of 24. Shoppers could taste as many as they liked and got a coupon for $1 off any jar in the store.

The researchers tracked two separate things, and this is the part that gets flattened in every recap: how many people stopped to look, and how many people who stopped actually bought.

• More people stopped at the 24-jam table — 60% of passersby, versus 40% at the 6-jam table.
• But conversion ran the other way entirely: 30% of the 6-jam group bought a jar, versus just 3% of the 24-jam group.

Do the multiplication and the limited table outsold the extensive one by roughly a factor of six, despite drawing fewer lookers. The paper, "When Choice Is Demotivating: Can One Desire Too Much of a Good Thing?", was published in the Journal of Personality and Social Psychology — a real peer-reviewed venue, not a pop-science paraphrase.

Why This Isn't Just a Metaphor for Swiping

The jam study gets invoked constantly in dating-app criticism, usually loosely — "too many choices, just like Tinder." That's fair as an intuition, but the mechanism matters more than the analogy, and the study is specific about it.

The extensive assortment didn't just fail to help. It actively suppressed action. People facing 24 options weren't paralyzed forever — most eventually walked away without buying, having tasted a jam or two and lost the thread of what they were even comparing. The researchers' interpretation: past a certain point, more options don't add information, they add cognitive load, and that load gets misread as "I guess none of these are quite right."

That's a mechanically different claim than "swiping is tiring." It's a claim about judgment quality degrading under volume — which is exactly what a deck of profiles asks you to do, repeatedly, for hours.

The Follow-Up Studies That Actually Matter More

Here's what most citations of this study skip entirely: the original jam finding didn't fully replicate on its own. A 2010 meta-analysis by Benjamin Scheibehenne, Rainer Greifeneder, and Peter Todd, published in the Journal of Consumer Research, pooled 63 studies attempting to reproduce the choice-overload effect and found the average effect size was close to zero.

That matters, and it would be dishonest to leave it out. But the same meta-analysis identified the conditions under which the effect reliably shows up — and those conditions describe a dating app almost exactly:

• High complexity per option. Choice overload appears when each item takes real effort to evaluate — not identical jam flavors, but items that differ on many dimensions at once (a person's values, humor, appearance, politics, life plans).
• No clear preexisting preference. The effect is strongest when people don't already know what they want before they start browsing — which describes most dating app users more than most jam shoppers.
• Time pressure or decision fatigue already present. Overload compounds when the chooser is already tired or choosing repeatedly, not once.

So the honest version of this evidence isn't "the jam study proves swiping is broken." It's narrower and, frankly, more useful: choice overload is a real, replicated phenomenon under specific conditions, and a swipe deck of dozens of complex, high-stakes profiles satisfies nearly all of those conditions simultaneously. That's a stronger claim than the pop-science version, not a weaker one — it's just more precise about when it applies.

What This Predicts About Swiping Specifically

Put the choice-overload research next to Pew Research Center's finding that roughly half of U.S. adults under 30 have used a dating app, and a similar share report the experience as at least somewhat negative — and a pattern emerges that lines up with the mechanism, not just the mood. People aren't reporting exhaustion because dating is inherently exhausting. They're reporting it in a format specifically engineered to maximize the number of comparisons per session.

This also connects to a separate, related finding worth naming precisely: Eli Finkel and colleagues' review in Psychological Science in the Public Interest found that compatibility algorithms have little demonstrated power to predict real-world relationship success from profile data alone. That's a different claim from choice overload — one is about volume degrading judgment, the other is about the algorithm's predictive ceiling — but they compound. You're being asked to make more comparisons than you can meaningfully process, using a ranking system that wasn't shown to predict outcomes in the first place.

What the Study Does Not Prove

Precision cuts both ways. Here's what this research does not establish:

• It doesn't prove fewer options always wins. The meta-analysis is clear that overload is conditional, not universal. Six jams beat 24 in that store, that day — not in every choice context.
• It doesn't prove people who date successfully were "just less picky." The mechanism is about processing load, not about lowering standards.
• It doesn't measure dating apps directly. No version of this research put subjects in front of an actual swipe deck. The connection is a reasonable, well-supported inference from adjacent findings, not a direct measurement — and it should be described that way rather than oversold as a one-to-one proof.

What Actually Reduces the Load, According to the Research

If overload is conditional on complexity, volume, and absent preexisting preference, the fix follows from the mechanism rather than from willpower:

1. Shrink the number of live comparisons at once. Not "browse less out of discipline" — structurally reduce the deck size you're evaluating in a sitting.
2. Front-load your preferences before you look, not while looking. The meta-analysis found overload is worse when people don't know what they want yet. Deciding your actual constraints before opening an app changes which condition you're in.
3. Let someone or something else do the first filter. Reducing the field before you personally have to evaluate it is the direct antidote to the mechanism the studies describe — which is functionally what a matchmaker, human or AI, is doing.

This is also where slow dating, as a practice, and Ofcom's separate work on choice architecture line up with the same lab finding from two different directions: fewer, better-curated options aren't a consolation prize for people tired of swiping. They're the condition under which the original 2000 finding — and the narrower version that survived replication — actually holds.

What This Means If You're Deciding How to Date Right Now

None of this says dating apps are worthless, or that everyone experiences overload identically. It says something more specific: if you're evaluating dozens of complex profiles per week without having fixed your own preferences first, you're running the exact experiment Iyengar and Lepper ran — and the replicated portion of that literature says the outcome tilts toward fewer purchases, not more, however many jars are on the table.

The practical move isn't abstinence from choice. It's changing who — or what — narrows the field before you have to. That's the entire premise behind agent-mediated matchmaking: an agent briefed on your actual constraints does the first filtering pass, so what reaches you looks more like six well-chosen jars than twenty-four.

Frequently Asked Questions

Was the jam study ever actually retracted or discredited?

No. It's a real, peer-reviewed finding. What happened is more nuanced: a large 2010 meta-analysis found the effect doesn't replicate universally, but does replicate reliably under specific conditions — high-complexity options, no preexisting preference, and decision fatigue. Those conditions describe dating apps closely.

Does this mean dating apps definitely cause choice overload?

The research doesn't measure dating apps directly, so that's an inference, not a direct finding. But the conditions the meta-analysis identifies as necessary for the effect are unusually well matched to how swipe decks are designed, which makes the inference stronger than most analogies drawn from this study.

Is more choice ever actually better in dating?

Yes, in the sense that some initial breadth helps you learn your own preferences. The issue the research flags is volume without filtering — comparing far more complex options per session than you can meaningfully process, repeatedly, which is a different thing from having options at all.

How is this different from the compatibility-score problem?

They're related but distinct. Choice overload is about volume degrading your own judgment. The compatibility-score question is about whether the ranking system itself predicts outcomes — a separate, earlier concern documented in the Finkel review. Both point toward the same practical fix: narrow the field before it reaches you.

Does slow dating just mean using fewer apps?

Not exactly — it's a set of mechanics around cadence and filtering, not app count. The full explanation of how slow dating works covers the mechanics in more depth than this piece does.