The Stanford Marshmallow Experiment, AKA Mischel’s delayed gratification experiments, investigated attentional mechanisms in the delay of gratification. The basic design was to show children a snack (such as a marshmallow) and tell them that they would get a greater number of snacks later if they didn’t eat the one they could see while the experimenter was out of the room.

While the age of the children had an effect on their ability to wait until the experimenter reappeared, there were also effects of whether they could see the immediate or delayed reward, and follow up studies have shown correlations with later SAT scores, reaction times in a response inhibition task, and differential activation in the prefrontal cortex.

But we haven’t found any mention of participants stealing all the rewards.