There is one big variable here that is being overlooked - which is the test itself. I remember my mother doing the marshmallow test on me, aged 3. She put two

There is one big variable here that is being overlooked - which is the test itself. I remember my mother doing the marshmallow test on me, aged 3. She put two packs of play-doh moulds on the stairs where I could see them - one with three in it, one with five in it. She told me that if I didn’t touch them, then she would give me the pack with five for my birthday a week later, and the pack with three to a friend. If I did touch them, she would give them both to said friend. After passing them on the stairs for the umpteenth time, after a few days, I caved, and opened the small pack. She made good on her threat. This was 40 years ago, and remains seared into my mind as a learning moment, and I have, since that moment, been absolutely ruled by delayed gratification - I never made that error again. So, in short, merely administering the test likely influences the outcome, as humans have memories, and children learn from experience.