musikkkalprostitute wrote:Elessar wrote:

I’m not going to spend my time talking to you about the difference between a self-selected group and a randomly selexted group, because you’ve got no desire to learn anything about scientific methodology, so we’ll leave it at that.

A very, very tiny group study does not give a fair picture of an entire group of people. People are far too diverse. But seeing as you have no desire to learn anything about diversity and how prejudices are formed, etc. we'll definitely leave it at that.

Oh go on then, I’ll bite.

If you take a random sample of 300+ people and find a statistically significant correlation between a desire for muscularity and sexism, you have a statistically significant finding with reasonable power. You can therefore extrapolate that finding to the wider population with a fair degree of confidence. A p-value of <0.05 is usually the accepted standard; in this case the p-value was <0.001. Statistician purists would argue that a p-value of much less than the pre-determined accepted upper threshold doesn’t actually mean the finding is ‘more’ statistically significant, but I’d argue that it certainly makes it a bit more compelling. In any case, it’s <0.05 which is enough to reach the significance threshold.

So actually the study didn’t find 300+ sexist bodybuilders. It looked at a cross-section of male society and looked at correlations. They were able to do so with a fairly small sample size because both bodybuilding and sexism are quite common.

You used an example of terrorism. A sample size of 300 would be wholly inadequate to try to draw conclusions about how widespread terrorism is, because in any given random sample of 300 people, the chances are there would be zero terrorists. If you had a rough idea in advance of the prevalence of terrorism you could do a power calculation in advance to determine how large the sample size would need to be - I suspect it would be in the hundreds of thousands.

Statistics are hard and most people have no need to know anything about them. But it would be wise to learn how to science before you criticise science. That would eliminate so many unnecessary misunderstandings, and you’ve had your fingers burnt in that department a few times in the last few weeks.