I can't make sense of it mathematically. A statistical distribution fitting these characteristics does not exist.

If non-weeders have an average seat belt wearing, and if weeders also have an average seat belt wearing, then the proportion of weeders inside of the seat belt non-wearing class is just equal to the proportion of weeders inside the whole population.