An article from the New York Times, so you know it's legit...
[A] thorough microbiological study of the five-second rule: the idea that if you pick up a dropped piece of food before you can count to five, it’s O.K. to eat it.
We’re reminded about germs on food whenever there’s an outbreak of E. coli or salmonella, and whenever we read the labels on packages of uncooked meat. But we don’t have much occasion to think about the everyday practice of retrieving and eating dropped pieces of food.
[T]the true pioneer of five-second research was Jillian Clarke, a high-school intern at the University of Illinois in 2003. Ms. Clarke conducted a survey and found that slightly more than half of the men and 70 percent of the women knew of the five-second rule, and many said they followed it. She did an experiment by contaminating ceramic tiles with E. coli, placing gummy bears and cookies on the tiles for the statutory five seconds, and then analyzing the foods. They had become contaminated with bacteria. For performing this first test of the five-second rule, Ms. Clarke was recognized by the Annals of Improbable Research with the 2004 Ig Nobel Prize in public health.
It’s not surprising that food dropped onto bacteria would collect some bacteria. But how many? Does it collect more as the seconds tick by? Enough to make you sick?
Clemson ha[s] now put some numbers on floor-to-food contamination. Their bacterium of choice was salmonella; the test surfaces were tile, wood flooring and nylon carpet; and the test foods were slices of bread and bologna. First the researchers measured how long bacteria could survive on the surfaces. They applied salmonella broth in doses of several million bacteria per square centimeter, a number typical of badly contaminated food.
[A]fter 24 hours of exposure to the air, thousands of bacteria per square centimeter had survived on the tile and wood, and tens of thousands on the carpet. Hundreds of salmonella were still alive after 28 days. [They] placed test food slices onto salmonella-painted surfaces for varying lengths of time, and counted how many live bacteria were transferred to the food. On surfaces that had been contaminated eight hours earlier, slices of bologna and bread left for five seconds took up from 150 to 8,000 bacteria. Left for a full minute, slices collected about 10 times more than that from the tile and carpet, though a lower number from the wood.
What do these numbers tell us about the five-second rule? Quick retrieval does mean fewer bacteria, but it’s no guarantee of safety. The infectious dose, the smallest number of bacteria that can actually cause illness, is as few as 10 for some salmonellas, fewer than 100 for the deadly strain of E. coli.