I'm guilty of not reading the article, but I find pleasure in poking apart statistical analysis. There are so many things to consider, like average sugar intak

I'm guilty of not reading the article, but I find pleasure in poking apart statistical analysis. There are so many things to consider, like average sugar intake of affluent/non-affluent families at the time of the first test and now. If both classes consume candy, perhaps the affluent have access to tastier candy (maybe important because the wealth distribution in the early 90s was more equal), so marshmallows are not tasty to them. (Note: I haven't had a marshmallow in a long time, but they are no where near as tasty as candy bought off the check out shelves). If present affluent families are educated about sugar (which was less so the case in the early 90s), they would probably feed their children less junk food, so the desire of wanting more of it is not as hard wired in their neurons, or it would taste overly sweet to them. Maybe less affluent families give their children sweets all the time because it shuts them up. Maybe the recipe of marshmallows has changed between then and now (see point 1).