People make wrong assumptions based on stupidly small and biased samples. You are a good example of this. You don't know all stereotypes, in fact it's likely you only know a small fraction of stereotypes and that fraction is biased by geography and self-selection and generally poor education in statistics and/or logic (i.e. not understanding that even if most avid shoe shoppers are women does not imply that most women are avid shoe shoppers). From this you have generalized that this says something about all stereotypes. Not only that but I'd wager you probably don't know many of the true values (i.e. How many Americans really are X).
This one paragraph wins. It is so full of stereotypes about people that believe in stereotypes
*sigh* No.
Much of that paragraph is refuting one persons stereotype (about stereotypes if we weren't getting "meta" enough). There is an air of refuting the general case but there's really not much of a stereotype there. If you aren't basing your stereotypes around large, well-randomized samples then you are by definition making wrong assumptions and unless you are doing so knowingly then it would be the result of poor education in statistics.
With regard to the OP's argument that "if you look at things rationally 90% of stereotypes hang around because they are true". Given that I doubt anyone knows the total number of stereotypes it seems pretty likely that he/she is part of the "poorly educated in statistics" group.