Obviously in the Stanford graduate culture this is a lot less likely.
But this makes me think. Let's figure out a few dozen (or hundred) various properties that are important in distinguishing cultures. Then let's let everyone in the world determine their stats on these properties, throw it all into a computer, and break it into clusters.
What kind of clusters would we get?
What are the most important distinct groups?
I have the feeling the results would surprise people.