That's the idea of generalizations. They don't need to be 100% fair, just statistically useful.

I'd wish people would get that, and only complain when a generalization is statistically wrong (e.g. doesn't really hold for the majority of the people it lumps along), instead of everytime they find a counter-example...

Simpsons paradox though. There are more conservative people in California than there are in Georgia (e.g. more people in California voted for Mitt Romney than people in Georgia). Slicing people up into states to make nice sound bite statistics is useless and leads to stupid stereotypes.

Well, that more people in California "voted for Mitt Romney than people in Georgia" might not tell the whole story.

Are the people in California traditionalist, religious, backwards, or even racist, etc, as the people in Georgia? They might not be, regardless of who they chose to vote for -- which takes into account many factors and policies.

That said, California is also traditionally regarded as conservative (at least by us in Europe) -- it's places like Massachusetts or Vermont that we'd count as progressive. Especially since California is regarded as hardcore pro-money/big-business, which is large part of being a "conservative" here.

Same percentage voted for Romney in Massachusetts as did in California. Again, grouping people by states is ridiculous to represent the views of the whole population of the state. It usually comes down to urban vs rural and that's about it.

Atlanta is not all that different from LA and Boston politically.

Thank you. Thus my using "51%" as a hypothetical proportion.