The interviewed people had been experiencing weight discrimination, but they were _still overweight at the time of the survey_. If there are some people that you can shame into changing their weight, many of them would have already changed in response, and either no longer be overweight (by BMI) or have changed something so that they no longer get discriminated against.

I also am very disappointed by the use of BMI for any study. Obviously this accounts not at all for muscle mass, which is actually hard and expensive to deal with so understandable, but it also has a significant _height_ bias. It was designed as a very coarse measure which was easy to calculate before the prominence of calculators. Human weight fits a curve that is closer to height^(2.5) rather than height^2. BMI will make tall people register as overweight, and short people register as more underweight. It's only accurate (and not precise at all) for people within a couple inches of average for their sex.