3 ms·

One thing that’s commonly true in statistics: You don’t need a very big sample size to get an accurate representation of a population. I’m willing to bet you co

One thing that’s commonly true in statistics: You don’t need a very big sample size to get an accurate representation of a population. I’m willing to bet you could get away with meeting 1,000, well chosen people, and get an accurate sample of society. If you think about it, people aren’t all that unique. Most people share the same personality, same hobbies, same interests and problems in life. Most of the things in life we think apply to us (for example navigating romantic relationships) are just common patterns that are repeated over and over again with millions of people. In fact we’re all a lot less special than we think we are.

codesushi42 7y ago
I'm willing to bet there are more than 1000 configurations, and sampling a mere 0.00001% of the population isn't going to get you much. In other words, that's sampling 1 out of every 7.3 million people. An utterly useless exercise. Statistics is fun, isn't it?
dredmorbius 7y ago
Correct, though the exercise may not be statistical. If you're looking at a system in which elements require some amount of mutual interaction on some regular basis, this sets your maximum grouping size. To scale up you need clusters of groups (or, sufficiently large, supergroups, etc.) to be able to function. This acts similarly to Amdahl's law. Or, if you are dealing with a system of some size and want it to spontaneously decompose to smaller units, change its ability (or requirements for) 1:1 exchange. Mass media accomplished this in reverse, enabling mass-scale coordination and culture.
PavlovsCat 7y ago
The map is not the territory. You could also say "all of Earth is just terrain", would that make every place on Earth actually the same? We are all unique. There are no degrees to that, there's just a lack of resolution, and abstraction and simplification.

codesushi42 7y ago

I'm willing to bet there are more than 1000 configurations, and sampling a mere 0.00001% of the population isn't going to get you much. In other words, that's sampling 1 out of every 7.3 million people. An utterly useless exercise. Statistics is fun, isn't it?

dredmorbius 7y ago

Correct, though the exercise may not be statistical. If you're looking at a system in which elements require some amount of mutual interaction on some regular basis, this sets your maximum grouping size. To scale up you need clusters of groups (or, sufficiently large, supergroups, etc.) to be able to function. This acts similarly to Amdahl's law. Or, if you are dealing with a system of some size and want it to spontaneously decompose to smaller units, change its ability (or requirements for) 1:1 exchange. Mass media accomplished this in reverse, enabling mass-scale coordination and culture.

PavlovsCat 7y ago

The map is not the territory. You could also say "all of Earth is just terrain", would that make every place on Earth actually the same? We are all unique. There are no degrees to that, there's just a lack of resolution, and abstraction and simplification.