[SteveEisenberg]

Originally Posted by AngryD

They only asked 1500 Americans.

Let's do some math. America has 331 million people in it.

They polled 0.000453172205438% of the population. From that you draw your conclusions?

See Slide 5 here:

Interestingly, and surprisingly to most students, the sample size we really care about is n [the sample size]. As long as N [the size of the sample universe] is much larger, that is as long as n<<N, then it doesn’t matter how big N is. The sample size of our sample, n, is what determines the standard errors of our means, and the power of our tests.

This is how I was taught and what I believe. Standard statistical tests assume a near-infinte N to n ratio because anything else would add complexity without real-world significance.

However, it's undeniable that a locality sample size of 1,500 gives more accuracy in a town of 1,505 people than it does in a town of 5,000 people. So there must be some statistical power relationship as the N to n ratio approaches 1. I just looked for it on the internet and did not find it. Can anyone who went further than I did in statistics quantify this?

It could make a good term paper topic in a liberal artsy statistical methods class. No need to credit me