Turns out, they can — and psychologists have the research to prove it.
Back in 1976, a team of four psychologists decided to investigate the idea. Their big takeaway, published in the Journal of Personality and Social Psychology: Left to their own devices, it's mostly kids who trick or treat in groups who steal additional treats.
The psychologists ran the experiment on more than 1,000 trick-or-treaters in Seattle between the hours of 5 and 9 p.m. The kids would show up at a house in a neighborhood, where a woman greeted them. She would then excuse herself, saying she had work to do, and that each child should take one piece of candy. There was also a bowl left out with nickels and dimes in it, also left unsupervised.
Most children actually did follow the instructions: 69 percent took a piece of candy and went on their way. Of those who stole, 65 percent took an extra candy (on average, between 1.6 and 2.3 additional pieces). Thirteen percent pocketed some of the change. One in five kids took both money and candy.
A few other interesting findings about trick-or-treating behavior: Kids who went out in groups were about twice as likely to steal additional candy as those who went candy hunting solo. Some of that was due to a sort of modeling effect: In 85 percent of the cases observed, kids would mimic the behavior of whoever dug into the candy first. If he took an extra piece, the friends routinely did the same.
Rates of stealing also went way, way down when the experimenter asked each kid his or her name and where he or she lived prior to leaving the room. The researchers thought of this as the "non-anonymous" condition: Although kids didn't write down fully identifying information, there would be some details for the experimenter to work on should she want to track down the thieves.
The takeaway, for those looking to ward off candy thieves, seems to be: Get a kid's name before you let them dive into the goods.