The first group was exposed to a series of unreliable experiences.For example, the researcher gave the child a small box of crayons and promised to bring a bigger one, but never did. Then the researcher gave the child a small sticker and promised to bring a better selection of stickers, but never did.
Meanwhile, the second group had very reliable experiences. They were promised better crayons and got them. They were told about the better stickers and then they received them.
You can imagine the impact these experiences had on the marshmallow test. The children in the unreliable group had no reason to trust that the researchers would bring a second marshmallow and thus they didn’t wait very long to eat the first one.