So I was teaching some introductory data analysis with Python/Pandas and using the CDC's National Health and Nutrition Examination Survey for 2013-2014 to illustrate some features of Python/Pandas. I did some aggregation across gender, ethnicity, and household to look at differences in BMIs across these groups. The difference in median BMI for adult white females 30 to 50 years age across household income is striking. Here is a plot of median BMIs for above and below 100K in household income (so basically either upper middle class and above or not):

There are now really big differences in BMI between white upper middle class women and below. For lower income female whites, median BMIs are almost 30 but they are about 24 for higher income female whites. On the other hand, there is almost no difference for men.

There are now really big differences in BMI between white upper middle class women and below. For lower income female whites, median BMIs are almost 30 but they are about 24 for higher income female whites. On the other hand, there is almost no difference for men.
Comment