The five-second rule — the idea that food dropped on the floor is safe to eat if picked up within five seconds — is one of the most universally shared pieces of informal food safety wisdom. It is also, like most universally shared food wisdom, only partially connected to reality.
The rule was properly tested. Here's what happened.
Quick answer
The five-second rule is not reliable. Bacteria can transfer to food almost immediately. Time matters, but moisture, surface type, and how contaminated the surface is matter more than the exact five-second cutoff.
The Rutgers study
The most comprehensive investigation of the five-second rule was conducted by Donald Schaffner and Robyn Miranda at Rutgers University in 2016. It was a well-designed study, published in the American Society for Microbiology's journal Applied and Environmental Microbiology, and it tested conditions that previous informal attempts hadn't.
They tested four surfaces (tile, carpet, wood, and stainless steel), four foods (watermelon, bread, bread with butter, and gummy candy), four contact times (less than one second, five seconds, thirty seconds, and five minutes), and used Enterobacter aerogenes as a bacterial surrogate for Salmonella.
The headline finding: the five-second rule is not supported. Bacteria transfer from contaminated surfaces to food in less than one second. Contact time matters, but not in a way that makes a five-second window safe — it just means less contamination than longer contact.
The nuances that matter
The study's more interesting findings are in the details.
Surface type matters enormously. Carpet transferred bacteria at much lower rates than tile or stainless steel. The intuition that carpet is dirtier than hard floors is apparently backwards in terms of bacterial transfer — the carpet fibers don't make as much direct contact with food.
Food moisture matters more than time. Watermelon picked up substantially more bacteria than gummy candy regardless of contact time, because wet, high-surface-area foods make better contact with surfaces and bacteria transfer more readily to them. The gummy candy, being dry and dense, picked up far fewer bacteria even at five minutes of contact.
Starting contamination is the key variable. The study used a highly contaminated surface. Most floors, most of the time, aren't nearly that contaminated. The relevant question isn't whether bacteria transfer (they do, immediately) but whether the amount transferred is likely to cause harm — which depends on what's on the floor to begin with.
What this means practically
The risk from eating dropped food is real but probabilistic and small in most circumstances. A floor that's been recently contaminated with Salmonella is genuinely dangerous — don't eat anything off it. A dry kitchen floor that hasn't been walked on recently presents a much lower risk, especially for dry foods that don't pick up bacteria as readily.
The FDA's official position is that any food contact with a floor should be considered potentially contaminated. This is a conservative standard appropriate for food safety regulation. It's not the same as saying that eating a dropped cracker off your kitchen floor carries meaningful risk.
The five-second rule as a rule is wrong — transfer happens immediately. But the underlying intuition it encodes — that brief contact is less risky than prolonged contact, and that context matters — is directionally correct even if the mechanism isn't.
The broader lesson
This is a good example of folk wisdom that's imprecise but not entirely wrong. The rule exists because people observed, correctly, that immediately retrieved food seems fine to eat most of the time. The explanation they attached to it (five seconds is safe) is incorrect; the empirical observation behind it (brief contact is usually okay in normal conditions) is mostly right.
Folk wisdom often encodes real patterns with inaccurate explanations. The pattern is worth taking seriously. The mechanism usually needs revision.
For related myth pieces, read sugar and hyperactivity, eight glasses of water, and reading in dim light.