Nobody was saying that you must eat eggs to survive - the point is to show the flaws in the hypothesis of the study when related to the sample group.
If you are sampling 3000 mothers in New Hampshire and looking for those who eat less poor people food and more rich people food you should expect to see a correlation that can be equally described by socioeconomic status as it can by diet.